# How Codex Happened — all six interview transcripts

Captured interviews from Dan Shipper’s reporting. These captures are not certified verbatim transcripts; preserve their provenance and transcription notes.

# Alexander Embiricos

Product

Final interview transcript supplied October 15, 2026. Source file F0C9K0V5ZGS; text, speaker labels and timestamps preserved from the supplied final transcript.

Passage IDs identify this edition's captured paragraphs. Bracketed transcription notes are not speech.

## final-20261015-0001 — Dan Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0001)

So, well, I think the place that I'm interested to start, just looking at the trajectory, the big memory I have in my head is being there, the night before the GPT-5 launch.

## final-20261015-0002 — Alexander Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0002)

Mm.

## final-20261015-0003 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0003)

And I feel like I was trying to politely say to you guys, it's cool, but it feels like it's built for pair programming, and—

## final-20261015-0004 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0004)

This is Codex or GPT-5?

## final-20261015-0005 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0005)

Both. It's like the Codex version of GPT-5 and the way that people were using Codex. And the way that it was being even marketed. It was very pair programmer-y, and I was like, there's this thing called Claude Code and compound engineering and blah, blah, blah. I'm curious, what your memory of that conversation was, how you were feeling then, and yeah, just starting there.

## final-20261015-0006 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0006)

Yeah. So I would say for me, GPT-5 was actually a really big deal and a really positive moment because I would say that before GPT-5, we were incredibly far behind on coding models.

## final-20261015-0007 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0007)

Mm-hmm.

## final-20261015-0008 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0008)

A little bit of the history of models here is, when I was working on coding... Well, you heard a little bit of my backstory sometime in the past. I was working on, how do we make the models useful?

## final-20261015-0009 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0009)

Yeah.

## final-20261015-0010 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0010)

So let's build a desktop assistant, let's build a browser. And basically my bet was that multimodal progress would be faster than I anticipated. And maybe model progress generally, but especially for multimodal inputs. And that wasn't the case, and even as they got better, there was latency issues that maybe we could have predicted, but we just hadn't.

## final-20261015-0011 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0011)

And why is that relevant to the making models useful coding stuff?

## final-20261015-0012 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0012)

Okay. Because for me, a lot of what we need to do to make a model useful is we need to let it work with your con... This is so obvious today, right?

## final-20261015-0013 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0013)

Yeah.

## final-20261015-0014 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0014)

But at the time it was like it needs to work in your context.

## final-20261015-0015 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0015)

Yeah, yeah.

## final-20261015-0016 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0016)

It needs to do stuff for you.

## final-20261015-0017 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0017)

Yeah.

## final-20261015-0018 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0018)

Which means it needs to use your applications.

## final-20261015-0019 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0019)

Yeah.

## final-20261015-0020 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0020)

So this is more just my personal journey.

## final-20261015-0021 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0021)

I get it.

## final-20261015-0022 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0022)

And so then we get to, we kinda realize, hey, the rate of progress for coding is really dramatic. Let's lean on that. And so I start working on coding products, and—

## final-20261015-0023 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0023)

And this is pre-GPT-5.

## final-20261015-0024 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0024)

At this point—

## final-20261015-0025 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0025)

This is way before... where are you at OpenAI?

## final-20261015-0026 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0026)

Yeah.

## final-20261015-0027 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0027)

Yeah, yeah, yeah. Okay.

## final-20261015-0028 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0028)

So I worked on ChatGPT Desktop, and then I helped start Atlas as projects, like, two years ago.

## final-20261015-0029 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0029)

Okay. Got it.

## final-20261015-0030 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0030)

And then I think this is late 2024.

## final-20261015-0031 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0031)

Mm-hmm.

## final-20261015-0032 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0032)

We have our reasoning models coming out. There's a ton of progress. Super exciting. My first—

## final-20261015-0033 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0033)

Like an O series?

## final-20261015-0034 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0034)

Yeah.

The first feel the AGI moment for me, real massive feel the AGI moment was I saw someone running a reasoning model that was connected to a terminal. I know, again, not exciting now, but it was like— I mean it, yeah... it was fucking cool.

## final-20261015-0035 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0035)

Yeah.

## final-20261015-0036 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0036)

Right?

## final-20261015-0037 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0037)

Yeah.

## final-20261015-0038 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0038)

So then we were starting to work on products to ship.

## final-20261015-0039 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0039)

Yeah. Yeah.

## final-20261015-0040 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0040)

But I think the really salient part that's interesting, even this, we'd have to frame carefully, but basically, we didn't, in my opinion, have really excellent coding models compared to the competition at the time. And so we really predicated our product—

## final-20261015-0041 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0041)

This is pre-GPT-5?

## final-20261015-0042 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0042)

Yes.

## final-20261015-0043 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0043)

Yeah.

Yeah.

## final-20261015-0044 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0044)

Around long-running models.

## final-20261015-0045 — Shipper, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0045)

Yeah.

## final-20261015-0046 — Embiricos, 8:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0046)

And so we shipped Codex Cloud.

## final-20261015-0047 — Embiricos, 12:25

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0047)

[Codex Cloud] was really just like a product built around a model, and a model research direction of asynchronicity.

So we've kind of been oriented around this idea that like you need to start with like a slightly more interactive experience, but didn't really have the models to build it with, so we went with this asynchronous product. Thus, that was Codex Cloud.

And Codex Cloud was actually, it was this very cool spike when we launched. Everyone thought it was a great idea. I still think it's a great idea, although I would build it... We're gonna get back to that, and it's gonna look completely different.

## final-20261015-0048 — Shipper, 13:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0048)

It seems like it was like a too early thing rather than—

## final-20261015-0049 — Embiricos, 13:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0049)

Yeah.

## final-20261015-0050 — Embiricos, 13:20

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0050)

Well, it's too early and incorrectly done.

## final-20261015-0051 — Shipper, 13:22

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0051)

Yeah. Oh, interesting.

## final-20261015-0052 — Embiricos, 13:22

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0052)

So the TLDR is that. It was too early because the models weren't smart enough. I would say the models today are smart enough to build Codex Cloud.

## final-20261015-0053 — Shipper, 13:34

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0053)

You said they are?

## final-20261015-0054 — Embiricos, 13:35

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0054)

Yeah, yeah, yeah. But it's—

## final-20261015-0055 — Shipper, 13:35

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0055)

To build Codex Cloud?

## final-20261015-0056 — Embiricos, 13:36

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0056)

Yeah, yeah.

## final-20261015-0057 — Shipper, 13:36

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0057)

Or to make it useful in the—

## final-20261015-0058 — Embiricos, 13:38

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0058)

To make it useful enough. Yeah. It's super clearly there already. Then there's actually two product points that need to be considered. The first is that Codex Cloud was rigidly asynchronous, and it was very annoying to get the work down from the cloud. What we are working on now is making it so that there is no rigid distinction between cloud and local tasks.

## final-20261015-0059 — Shipper, 13:58

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0059)

Mm.

## final-20261015-0060 — Embiricos, 13:58

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0060)

All chats are just like chats. There's no like... There's not a cloud or a local Linear issue, right? There's just issues.

## final-20261015-0061 — Shipper, 14:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0061)

Yeah, yeah.

## final-20261015-0062 — Embiricos, 14:05

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0062)

So all chats are just gonna be chats. The reasoning loop will actually be running in the cloud. This is already partially true today. It's already partially true today.

## final-20261015-0063 — Shipper, 14:16

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0063)

This is so fun. I'm like, I love my life.

## final-20261015-0064 — Embiricos, 14:21

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0064)

Yeah, me too. I have those moments like weekly.

So you already have the engine. Right? And then you have like some server component, right? And then maybe you have some connectors with like, say Slack as an example. Right? And then you have your computer, and your computer has like Bash or whatever, right? So like today, when you use the product, you're already kind of having... Like you spawn this thing, and then at some point we decide to hit the engine. Right? It comes back to us. We're like, “Oh, should I run... What tool should I run next?” And then it might decide, okay, let's tell the computer to run a Bash command, which might actually call Slack.

## final-20261015-0065 — Shipper, 15:05

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0065)

Yeah. Yeah, yeah.

## final-20261015-0066 — Embiricos, 15:06

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0066)

So and then again, we get this back, and then da, da, da, da. Right? And so we actually run the... This is running every time. Now obviously if we wanna run a Bash command, then okay, we need to get all the way down here. But actually, it would make a lot of sense to skip hitting the computer for a lot of these loops, but the loop is here right now. And it actually—

## final-20261015-0067 — Shipper, 15:32

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0067)

And it seems like the problem though is that it's much easier... All my stuff's on my computer already. So I don't have to go like install Slack.

## final-20261015-0068 — Embiricos, 15:43

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0068)

Yeah. Well, so today this is where things run, right? And for simplicity, the loop is running on your computer. Which means that at minimum, even just for thinking, we're coming all the way down to here and up. Right?

And then when we have to run a tool, we're happy we made it all the way here, 'cause this is like zero latency, right? But another way that you could do this that would make a lot of sense is you could say, let's actually run the loop here. And so when you're just thinking, you're just doing this, right? That's it. When you want to hit a tool that is not on your computer, such as a connector, why should that live on your computer?

## final-20261015-0069 — Shipper, 16:21

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0069)

Yeah, yeah, yeah.

## final-20261015-0070 — Embiricos, 16:21

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0070)

You could actually give... So we're—

## final-20261015-0071 — Shipper, 16:23

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0071)

This is gonna be really good for my CPU usage.

## final-20261015-0072 — Embiricos, 16:25

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0072)

Yeah. So we're gonna have a node REPL running here, and that's just so that we can compose things. You're not doing all the work. It's not mental math. It's like you actually can write code. We'll just go here, right?

And increasingly, a lot of stuff will be hosted to the cloud, and if we need to run a Bash command, then we can go hit your computer and do this, right? So the latency doesn't get any worse if you're doing work on your computer. The latency gets better whenever we're not running a Bash command.

## final-20261015-0073 — Shipper, 16:54

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0073)

Mm.

## final-20261015-0074 — Embiricos, 16:54

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0074)

And now we have this computer, which is your laptop. You can close your laptop. The thing can keep going. It won't be able to run commands, but models are smart. It'll figure it out.

## final-20261015-0075 — Shipper, 17:03

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0075)

Yeah, yeah, yeah.

## final-20261015-0076 — Embiricos, 17:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0076)

So you could do a coding task and then later you have a question about it, and you could like, “Hey, why did you do this?” It doesn't need to run a Bash command to ask that.

## final-20261015-0077 — Shipper, 17:09

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0077)

Yeah, yeah, yeah.

## final-20261015-0078 — Embiricos, 17:09

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0078)

Or it could spawn... this is computer as a service. We could spawn a container for whatever we want. And which brings us to the second thing.

So the first product problem was the rigidity of cloud [versus] local. So actually with this green diagram I just drew for you here, there's no longer a distinction between a cloud task and a local task. Like all tasks are just running the loop. There's also a database now that lives in the cloud.

## final-20261015-0079 — Shipper, 17:35

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0079)

And either have access to your computer or they don't.

## final-20261015-0080 — Embiricos, 17:37

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0080)

It has access to the computer or not. That's all.

## final-20261015-0081 — Shipper, 17:38

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0081)

Yeah, yeah, yeah. That's very—

## final-20261015-0082 — Embiricos, 17:39

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0082)

It's like if you run Linear on desktop, maybe the desktop have some features that are not.

## final-20261015-0083 — Shipper, 17:44

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0083)

Yeah, yeah.

## final-20261015-0084 — Embiricos, 17:45

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0084)

So it's obviously in this case, you're gonna run the app most of the time. But you can also close it and it doesn't matter.

## final-20261015-0085 — Shipper, 17:50

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0085)

Yeah, yeah, yeah.

## final-20261015-0086 — Embiricos, 17:50

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0086)

And this is gonna be great for automations, right? Like, oh, check my email, whatever. You check the email. You can continue the same conversation with your computer open. And then this means that we can support bring your own box. So if you're a big enterprise and you can only access dev dependencies on VPC or whatever, then we can also hit this thing. And so these things here are called executors now.

So a lot of this work is actually happening on open source.

## final-20261015-0087 — Shipper, 18:26

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0087)

Yeah, yeah.

## final-20261015-0088 — Embiricos, 18:27

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0088)

But this is where we're going, and a lot of it's actually in open source, and it's gonna be a really cool blog post.

## final-20261015-0089 — Shipper, 18:30

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0089)

That's sick.

## final-20261015-0090 — Embiricos, 18:31

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0090)

And we're hoisting the whole system. So that was—

## final-20261015-0091 — Shipper, 18:33

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0091)

You're hoisting the whole system?

## final-20261015-0092 — Embiricos, 18:34

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0092)

Yeah, so we call this... we view this as like today the loop is here and we're hoisting the system to the cloud.

## final-20261015-0093 — Shipper, 18:39

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0093)

That's a good word. I haven't heard people talking about that. Is that like a thing now?

## final-20261015-0094 — Embiricos, 18:43

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0094)

I don't know.

## final-20261015-0095 — Shipper, 18:43

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0095)

Or you just coined it just now?

## final-20261015-0096 — Embiricos, 18:45

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0096)

No, at OpenAI, we've been using this term. I don't know who coined it. Hoisting? Hoisting.

## final-20261015-0097 — Shipper, 18:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0097)

Hoisting.

## final-20261015-0098 — Shipper, 18:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0098)

I haven't heard that.

## final-20261015-0099 — Embiricos, 18:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0099)

Hoisting is just pulling something up, right?

## final-20261015-0100 — Shipper, 18:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0100)

Yeah.

## final-20261015-0101 — Embiricos, 18:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0101)

So if you go and look at open source today, you'll already notice that the loop is being separated from the executor. We call this the orchestrator, actually. The orchestrator is being separated from the executor in open source today, but they still both run locally on your computer.

## final-20261015-0102 — Shipper, 19:03

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0102)

Yeah, yeah.

## final-20261015-0103 — Embiricos, 19:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0103)

And we'll still always support that 'cause you might not want stuff in our cloud, right? But then we'll put that up. This is a very hard infrastructure project. It's taking us a lot of time.

## final-20261015-0104 — Shipper, 19:12

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0104)

I bet.

## final-20261015-0105 — Embiricos, 19:13

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0105)

But I think in this space,  it's really hard to decide when to go do something like this. Because we wanna run automations in the cloud today, right? How do we do that?

We could either do this, which would really unlock things and I think be a giant leapfrog. Because this takes months to execute. Or we could just go build a new standalone product, with its own page.

## final-20261015-0106 — Shipper, 19:46

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0106)

Yeah, yeah, yeah.

## final-20261015-0107 — Shipper, 19:56

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0107)

That is a really interesting question...

## final-20261015-0108 — Embiricos, 20:03

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0108)

And if you wanna look at our growth chart, this was Codex Cloud launch. You know, obviously at the time, the scale looked like this on the chart, right? But so it was like this, maybe, yeah, kinda started out with zero. So we had some real usage, but not by OpenAI standards. And then GPT-5 happened, and it just started doing this [points up and to the right], you know? And at the time it looked like that.

## final-20261015-0109 — Shipper, 21:12

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0109)

That's usage of Codex Cloud specifically?

## final-20261015-0110 — Embiricos, 21:14

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0110)

This is cloud. This is the cloud era. This is the GPT-5 era.

## final-20261015-0111 — Shipper, 21:18

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0111)

Yeah. GPT-5, but when you say usage of what in particular?

## final-20261015-0112 — Embiricos, 21:22

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0112)

Oh, sorry. Of the CLI.

## final-20261015-0113 — Shipper, 21:23

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0113)

Okay. Got it.

## final-20261015-0114 — Embiricos, 21:24

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0114)

That's the product we ship. It just looked like that, right? And then we shipped the IDE extension, which looked like that. Like, again, the chain. And then we had 5.3 Codex, was like this very significant model leap for us.

## final-20261015-0115 — Shipper, 21:37

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0115)

Yeah, big moment.

## final-20261015-0116 — Embiricos, 21:37

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0116)

5.3 Codex, and then an app. And so then now all these lines look very low now, and it's kind of up here. But that's kind of the journey for me. So yeah, we can unpack it later.

## final-20261015-0117 — Shipper, 21:48

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0117)

That's really interesting. Well, I wanna get into... Okay, at least from my perspective being outside of the company, what I felt was, let's say the GPT-5 launch, or around then, it felt like there was this split between: professional engineering happens in Codex, and then if you're vibe coding stuff, you should do it in ChatGPT, and you should use the canvas and the artifact stuff.

And part of the problem with that is, the bet that Anthropic made with Claude Code was like, if it just has access to your whole computer and you can just delegate tasks to it, that creates all this power and flexibility that compounds because it's on your machine. You're not spinning up a new VM every time, like whatever.

## final-20261015-0118 — Embiricos, 22:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0118)

Right, right.

## final-20261015-0119 — Shipper, 22:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0119)

And I was like, that the GPT-5 as a model felt like I was able to use it less well because it wasn't on my computer in that way. So I'm curious how you were thinking about that, and how you were sort of seeing the Claude Code stuff and your strategy relative to that.

And then I wanna trace it from there, 'cause 5.3 was the point at which I was like, “Oh, I think OpenAI gets it.” I think you guys, you've figured it out. And it sounds like part of it was the model felt like it wasn't quite right, but I also felt like the harness and the harness strategy was not quite right.

## final-20261015-0120 — Embiricos, 23:30

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0120)

So yeah. So there's a few layers to this. I would say that we took a hard pivot after this.

## final-20261015-0121 — Shipper, 23:35

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0121)

Mm-hmm. After Codex Cloud?

## final-20261015-0122 — Embiricos, 23:38

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0122)

Yeah, after Codex Cloud. And I would say that this idea of, hey, we need to build a very flexible primitive that runs on the user's computer and that people can use as interactively as they want, and in fact we think people will start using it very interactively and then kind of onboard a teammate, they will learn how to use it less interactively. That became our strong product thesis and the strategy we started executing against in August.

## final-20261015-0123 — Shipper, 24:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0123)

In August, okay.

## final-20261015-0124 — Embiricos, 24:05

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0124)

Yeah. So look, obviously it took us some time to realize the strategy, right?

## final-20261015-0125 — Shipper, 24:10

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0125)

Yeah, yeah, yeah.

## final-20261015-0126 — Embiricos, 24:10

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0126)

But that's why we were building CLI. We had this idea of, it's open source. So funnily, by the way, we actually shipped the CLI even before Codex Cloud, but that was a very small side project, done heroically by two people.

But around this time after Codex Cloud, like August.

## final-20261015-0127 — Shipper, 24:37

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0127)

Like it'll get it right more than Claude will.

## final-20261015-0128 — Embiricos, 24:47

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0128)

You couldn't even use the Codex model in the CLI.

## final-20261015-0129 — Shipper, 24:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0129)

That's interesting.

## final-20261015-0130 — Embiricos, 24:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0130)

As a company, we have a very conservative safety stance. And we actually had to do a lot of work to figure out how to get comfortable deploying these, that level of capability.

## final-20261015-0131 — Shipper, 25:03

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0131)

Mm.

## final-20261015-0132 — Embiricos, 25:07

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0132)

So initially you couldn't even use this really capable model in the CLI. You had to use, I forget what, maybe o3. I just don't remember. But you had to use some other model that was just not designed for that purpose.

But around the time of August is when we were like, "No, we're gonna bet really hard on the CLI and the interactive coding form factor." And basically our journey is gonna be—and this has been the product strategy that basically we've been executing against—okay, so I don't know, the title's probably Agents. And we're gonna start with ... Here's a two by two.

## final-20261015-0133 — Embiricos, 25:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0133)

But it's good. I'm gonna ... When things are good, I give them credit. So it's like dev, non-dev. And then it's sync, async. And it's kinda like, all right.

Shipper: [laughs]

## final-20261015-0134 — Shipper, 26:02

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0134)

That's a good breakdown, yeah.

## final-20261015-0135 — Embiricos, 26:03

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0135)

We're just gonna start here [the dev, sync quadrant]. Right? And then we're gonna kind of move this way and also move this way. But you have to land here. Like both—

## final-20261015-0136 — Shipper, 26:14

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0136)

Why?

## final-20261015-0137 — Embiricos, 26:14

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0137)

So—

## final-20261015-0138 — Shipper, 26:15

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0138)

'Cause you would imagine one way I could say is well, to be really good for devs, you have to be so smart and so good that we should start with consumers.

## final-20261015-0139 — Embiricos, 26:26

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0139)

Yeah. So there's a lot of reasons that I think developers are the right place to start. And I mentioned this a bit during the talk, but I don't know if you arrived yet.

Basically, I think that they as an audience are very interested in trying new tools. So you're just gonna get a lot more adoption. They are interested in trying open-ended tools, and they will figure out how to use it. They provide... They're, in my opinion, they are the best audience for providing actionable feedback for something that's still early.

And then you can figure out how to productize from there. You can put a feature deep in config.

## final-20261015-0140 — Shipper, 26:58

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0140)

Yeah, yeah.

## final-20261015-0141 — Embiricos, 27:24

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0141)

I think the second thing is actually that the tasks are way easier to build for. So coding tasks, first of all, okay, so this is a fun thing, right? Developers—

## final-20261015-0142 — Shipper, 27:33

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0142)

Coding tasks are verifiable.

## final-20261015-0143 — Embiricos, 27:34

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0143)

Yeah. Coding. So first of all, developers, there's this old stat, which I don't have a citation for, but I'm sure we could find it, they only spend 30% of the time coding, and then they 70% of the time spent not coding. And I would say that actually it's now that we delegate to agents, maybe it's like 90% of the time is spent not coding. And so actually, developer software engineering work looks very similar to knowledge work.

But let's start with the coding tasks. The interesting thing about them is they're verifiable. There is an existing place where the human is in the loop, right? I talk to a lot of enterprise customers today, and they're like, "Hey, we have like a regulated workflow. How do we make sure we trust the agent?" The truth is that a lot of coding is regulated too. It's just we already have established human practices. It's called code review. Right? So we already know how to do this—

## final-20261015-0144 — Shipper, 28:15

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0144)

Yeah, yeah.

## final-20261015-0145 — Embiricos, 28:15

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0145)

for developers. Also, coding tasks are hermetic, right? You can do a coding task in a sandbox — and you might actually disagree with some of this, but you can do a coding task in a sandbox without access to the internet, and it'll work pretty well, right? Maybe, okay, some amount of testing and things will need sandbox access, but as you're getting started, you can literally limit it to a folder on your computer.

## final-20261015-0146 — Shipper, 28:34

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0146)

Yeah, yeah.

## final-20261015-0147 — Embiricos, 28:34

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0147)

It can do its job. As soon as you get outside of coding to knowledge work, every task has a side effect on the internet. And okay, maybe only read actions. It's actually, there's no clean way to know if something is a read action. And actually, there's also no such thing as a read action on the internet. Everything has a side effect even if you get the task.

## final-20261015-0148 — Shipper, 28:50

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0148)

There's a very quantum physics thing there, where it's like you can't read something without modifying the world, you know? Yeah.

## final-20261015-0149 — Embiricos, 28:59

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0149)

Exactly. Yeah, no, 100%. Right? So you have to think about the audience, the sort of the safety question of performing the tasks, the level of side effects, which is also a security question. And I'm missing one third point. Let me see if I can find it. Task specificity. Right?

Basically, my view is that we are gonna progress. Actually, there's non-dev, and then there's straight up consumer. And so non-dev here includes there's work productivity, and there's also life productivity. I would put those both in here. Straight up consumer is like I don't even know what I want, right?

## final-20261015-0150 — Shipper, 29:29

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0150)

Yeah, yeah.

## final-20261015-0151 — Embiricos, 29:29

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0151)

So also the specificity of the task you get is quite high here, and a developer knows how to change how they're specifying to give you what they want. Here [non-dev], the task specificity goes down. We actually see this. It's like less clear what I want, and then eventually we're gonna get here where they don't even know what they want, and we have to be useful. So that's why we want to start there, and kind of productize our way through.

## final-20261015-0152 — Shipper, 29:51

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0152)

That makes sense.

## final-20261015-0153 — Embiricos, 29:52

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0153)

So you can start with this very natural strategy. We're gonna start with CLI, super open-ended, and see what people do with it. Very quickly we're like, we should build a great IDE extension. And then that became our most used surface, so we disrupted ourselves kind of, which is kind of crazy. And then the app actually now is the most used thing.

And so I talked about this a bit earlier as well, but for us, moving to the app was a big part of how we got to ... It was kind of two things actually. The app helps us transition from sync to async. And it also helps us transition from dev to non-dev.

Eventually when this app is running and we have this new architecture here and we're running the task in the cloud, the app will feel very ergonomic for that. You won't even care or know that anything changed.

## final-20261015-0154 — Shipper, 30:42

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0154)

I don't know how much you can tell me, but what did you figure out or what changed in the model that let you be like, “Okay, we can get to this point where it's gonna be okay for synchronous work, and it's gonna be okay for non-devs”?

## final-20261015-0155 — Embiricos, 32:30

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0155)

Now it's working really well because there's a flywheel, right? Once we ship the CLI and GPT-5, well, mostly GPT-5 with the CLI as a light wrapper around it, it was just like it created this self-fulfilling flywheel—

## final-20261015-0156 — Shipper, 33:07

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0156)

Yeah. Yeah.

## final-20261015-0157 — Embiricos, 33:07

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0157)

Of just people being really excited to work on this stuff, and obviously now we've climbed way further in intelligence.

So the change, from my perspective, was that there was a researcher, Michelle Pokrass, who was really interested in pushing on this, obviously working with us, but mostly she was working with other customers we had that were asking us to improve our interactive coding capability.

And so we were able to leverage that momentum and then ship our thing, which very quickly actually became a really significant user, and then the biggest user of these models. And then grew it from there.

## final-20261015-0158 — Shipper, 33:52

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0158)

So there's this really interesting thing that I'm observing that I think you're one of the most interesting or best examples of, which is, it's about how to do product strategy in AI as the models get better, and you use the word, like, reconstruct it ourselves, where it feels like every time there's a jump in models, there's a new way of bringing it to users that becomes important.

## final-20261015-0159 — Embiricos, 34:21

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0159)

Mm-hmm. Mm-hmm.

## final-20261015-0160 — Shipper, 34:21

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0160)

And then your old surface, you have to figure out what to do with it, and there's different ways to figure that out. What it seems like you do is you have maybe parallel bets going on. You tell me if I'm wrong. You have parallel bets going on. When something works, you pour a bunch of energy into it, but then what happens is the models advance, and it is better to have, for example, with Codex, you had this new surface that you could play around on. Both a model line and a product surface that you could then move faster on to explore these new capabilities 'cause you're not held back by supporting the big thing that's already working.

But then you launch it, and then you're like, “Well, what do I do with this old thing?” And so there's this interesting weirdness with you're building these products. As soon as they start to scale, they're also at risk of, you kind of have to disrupt yourself with something new. But then you have to figure out, does that new thing... For Anthropic, what they do is they just put it in a tab. So it's like, is the new thing a new product, or is the new thing a tab on the existing product?

Talk to me about, you've touched on this a little bit, but I'm curious, where you started in thinking about how to do this and where you are now, and what do you think the best practice is for thinking about those decisions?

## final-20261015-0161 — Embiricos, 35:41

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0161)

Yeah. I have to think about this a little bit. It's a great question. It's a very difficult one. I think to do a good job, you really do have to be trying to think in the future and sort of think ahead about what's going to work. And to the extent you can, it is far better to extend something that you have, if it can take advantage of the new capabilities.

And then sometimes, through luck or lack of foresight, you will have a thing that just doesn't quite scale to what you need next, and then you have to go out, make a new thing. But that always introduces challenges, mostly for, I think for our users.

## final-20261015-0162 — Shipper, 36:26

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0162)

Mm-hmm. Yeah.

## final-20261015-0163 — Embiricos, 36:26

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0163)

'Cause then they have to understand more stuff. So for instance, here's an interesting example. Like the Project Atlas, the browser. We had a lot of conversations about whether we should build this as an in-app browser for ChatGPT, or whether we should build it as a standalone browser. And there, the guidance, it's quite funny. I remember thinking we should build an in-app browser. But I could get behind the standalone browser too. But they're kind of different products.

We had a pretty big steer on this, which is like, “Hey, I would love for you guys to at least explore building a standalone browser. Just at least explore it so that we can see just what happens if we're completely disconnected and open-ended. And then let's ship whatever's best.”

So that was at the time that I was working on it. And then I stopped... I moved to Codex, so I wasn't there when we decided that, okay, that was actually the form factor to ship. But it was a very explicit guidance, like, “Let's give the idea as much room as it can. And then, if you decide to ship an in-app browser at the end of the day you can go for it.” So that was quite interesting there.

I think in the case of Codex, we felt like when we were shipping Codex Cloud, we actually built it into ChatGPT as a tab.

## final-20261015-0164 — Shipper, 37:43

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0164)

Mm. Yeah.

## final-20261015-0165 — Embiricos, 37:44

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0165)

'Cause it kinda made sense—

## final-20261015-0166 — Shipper, 37:45

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0166)

Yeah, yeah.

## final-20261015-0167 — Embiricos, 37:45

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0167)

In there. But the moment we realized that we wanted to ship this, we adjusted our strategy. We're like, "We're gonna ship a very flexible tool on your computer." We realized that bundling that into ChatGPT desktop would just be kinda crazy. It would just be weird to understand how to use the product. It's a very different audience.

So even though now we plan to bring all these capabilities to ChatGPT, I feel like for that idea to have room, it needed breathing room and it needed to be a standalone thing that you install with npm install or whatever, right? So there we had to go separate.

And I think looking back, there are certainly some decisions that I would have loved to have made, to make it easier to bring them back together. For example, we built a really strong brand around Codex with a lot of investment.

## final-20261015-0168 — Embiricos, 39:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0168)

But I guess my point is there's certainly some decisions we could've made better there, but I think we needed to create room for separate products.

Now, if you look at the Codex planning, there we've been very oriented... An overriding principle we've had the whole time, that for me is a key thing in designing for AI, is we are building one agent, and actually we spend a lot of time talking about the agent and not the product.

And often when I'm working with someone who hasn't been thinking about Codex as long as me, and they wanna talk about product and they start talking about all these features they wanna build, oftentimes the feedback I give is like, "I mean, these are very cool, but to me it feels like you're building product, but we shouldn't be building product. That's not our job."

## final-20261015-0169 — Shipper, 39:44

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0169)

Mm.

## final-20261015-0170 — Embiricos, 39:45

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0170)

"Our job is to build an agent."

## final-20261015-0171 — Shipper, 39:46

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0171)

Mm.

## final-20261015-0172 — Embiricos, 39:46

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0172)

We have to productize it. Yes, it is ultimately a product for users. Yes, ultimately it is a business that earns revenue for our company, but the way we should think about it is actually we're designing an agent.

## final-20261015-0173 — Shipper, 39:55

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0173)

Can you give me a concrete example?

## final-20261015-0174 — Embiricos, 39:56

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0174)

Yeah. So I think in product land... These are nuanced, right? And maybe there's better words for this, but in product land, you use UI affordances. You put a lot of scaffolding around the model. You just do whatever you need to solve the user's problem.

In agent land, you think much more about primitives, and you leave a lot more up to the model, and then your job is just how to make it ergonomic for the user and the models collaborate. Right? So for instance, let's see. Here's an interesting example: how to decide when to open the in-app browser, right?

## final-20261015-0175 — Embiricos, 40:19

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0175)

I'm not saying these were super deeply thought about for a long time, but briefly some people were thinking that we should have some rules for when we open in-app browser, and you would write those rules in TypeScript.

## final-20261015-0176 — Shipper, 40:51

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0176)

Oh, interesting.

## final-20261015-0177 — Embiricos, 40:51

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0177)

You know, like, we would write those rules.

## final-20261015-0178 — Shipper, 40:51

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0178)

Like code rules, not like prompt rules. Okay.

## final-20261015-0179 — Embiricos, 40:54

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0179)

And I think to you and me, that is obviously the right way to do it, right?

## final-20261015-0180 — Shipper, 40:59

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0180)

Yeah, yeah.

## final-20261015-0181 — Embiricos, 40:59

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0181)

We should just let the model decide.

## final-20261015-0182 — Shipper, 41:00

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0182)

I think the model also wants to do that by default. The model wants to write code for itself 'cause it thinks that that's what engineers would do.

## final-20261015-0183 — Embiricos, 41:07

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0183)

Yeah. That's interesting. That's really interesting feedback. I need to think about that more, but you're probably right. So we're like, "No, we need to leave that up to the model," right? How do automations work? There's loads of ways you could try to do that. But at the end of the day, if you can keep it super simple and it's just like, "Hey, here's a way to schedule a task. It can be repeating or not repeating," and you can just let this be a primitive that the model can use, you do better.

Another example would be, currently I think the UI in the app around Git workflows, it does its job for some people, but it does the job that we designed it for. So there was a lot of onus on the human designer, and that human designer was designing for some people, and that's not all people, right?

So I actually view that as an area of improvement for the app. If that was more just model orchestrated, then you could create all these different workflows that people end up doing anyway, so just tell the model to use Git in the command line.

## final-20261015-0184 — Shipper, 42:00

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0184)

Mm-hmm.

## final-20261015-0185 — Embiricos, 42:00

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0185)

So broadly, the way we think about the agent is you have to have very clear layers. So okay, model coming through the Responses API, open source harness. This harness we need to do our best to keep as simple as possible, and then have the right primitives for extensibility, and you let the model orchestrate as much as possible.

So things like computer use, core capability people talk about all the time, built as a plugin. The browser use, that's also a plugin. Both of these plugins use the node runtime that we have to be able to execute code, local and computer. So that's a core primitive that you can take advantage of. A lot of the new things that we're building now we've forced to go through the plugin path so that they can maintain arm's length from the—

## final-20261015-0186 — Shipper, 42:40

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0186)

Mm.

## final-20261015-0187 — Embiricos, 42:40

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0187)

From the core agent loop. And then if we wanna render UI, we also try to avoid, we try to be... I think we could do an even better job of this, but we try to avoid... You wouldn't believe how many teams want to add UI to the app. How many people, how many PRs we get for different things.

## final-20261015-0188 — Shipper, 42:51

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0188)

Sure. I see.

## final-20261015-0189 — Shipper, 42:56

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0189)

Yeah.

## final-20261015-0190 — Embiricos, 42:56

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0190)

And we were kinda like, "No, we need to really avoid adding UI to the app." But we need to think about how to make sure that the app is extensible so that if you want to render UI as part of your thing, a plugin can be the way.

Now the running joke is that Alexander goes around telling people that they're building plugins. But my point is more we need to really keep the core action loop simple, and then most things that you would normally go build should just be done as plugins. And generally speaking, we should try to not add too much between the model and the user. Sorry, a bit of a ramble.

Shipper: [laughs]

## final-20261015-0191 — Embiricos, 46:10

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0191)

So I think the Codex team and in particular, Tibo is the lead for the overall Codex team, we've done a really good job of just creating a team that has really clear conviction on what we're gonna do. Is willing to say no to a lot of side quests. And just move with intensity towards this goal.

## final-20261015-0192 — Shipper, 46:29

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0192)

Is there anything else that you wanted to get in that we haven't talked about?

## final-20261015-0193 — Embiricos, 46:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0193)

So no, I think I just... We spent a lot of time talking about how we got here, and I think there's a lot of interesting stuff.

'Cause now I basically feel like this year, since Feb when we finally got the app out, I think the world can kind of see where we're going, and it's like finally things are starting to make sense. Like, oh, why is the team... I mean, you could critique this decision, by the way, but we have explicitly not invested as much in the CLI, and you can feel that, you know?

And I still can easily counter-argue that decision. Because a lot of people who compare us to Claude Code try the CLI, and they're like, "This is worse." And I'm like, "Oh man, we need you to compare the app."

Shipper: [laughs]

## final-20261015-0194 — Embiricos, 47:34

[Read this passage](https://every.to/p/how-codex-happened/interviews/alex/?edition=gift#final-20261015-0194)

And then people are like, "This is better." So we need to reverse that narrative and stuff. But anyways, I don't know. So I don't have anything specific to say other than that we should definitely make sure we talk about how does that journey kind of ultimately get us back to some of the founding principles of the team, which were like, when we built cloud, we were like, "We're gonna build you a teammate that lives in the cloud and constantly does work for you."

And now I feel like the pieces are all coming together, and it's gonna happen.

---

# Greg Brockman

Company direction

Final interview transcript supplied October 15, 2026. Source file F0C9K0V5ZGS; text, speaker labels and timestamps preserved from the supplied final transcript.

Passage IDs identify this edition's captured paragraphs. Bracketed transcription notes are not speech.

## final-20261015-0001 — Dan Shipper, 1:01

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0001)

First of all, anything in there that maybe we can go back to the beginning of? I think my narrative seeing it from the outside was, on the GPT-5 launch, there was something happening in Claude Code and the way I was using Claude Code that wasn't reflected in the model and product strategy, or at least publicly.

So what were you thinking? What was going on internally then?

## final-20261015-0002 — Greg Brockman, 1:56

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0002)

Yeah. So I'd say in beginning of 2025, we actually did set coding software engineer as a top-level company goal, right? So we were actually very focused on saying, “Hey, coding is a thing that we want to really deliver on.”

And the thing I think we missed or were slower to was, you know, we'd always had the best numbers on coding competition benchmarks, things like that. But actually connecting it to real world software engineering, that was the thing that we had under-invested in and under-appreciated relative to its importance. And so I think we had kind of the smartest models, but that doesn't mean that they're the easiest to use.

They never had seen a messy code base or how a programmer might interrupt it halfway through. All of that, thinking about the personality, what it is like to talk to the thing. Those were not as keenly in our sights and focus.

And I think that that was a real lesson for us, is to realize that how do we actually make this thing useful actually is a lot about those things. It's not just about the raw intelligence. And I think we became really aware of this maybe midway through the year, and then there was a real sprint to close that gap. And Michelle Pokrass led a team there.

## final-20261015-0003 — Shipper, 3:12

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0003)

Wait, what's her name?

## final-20261015-0004 — Brockman, 3:13

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0004)

Michelle—

## final-20261015-0005 — Shipper, 3:14

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0005)

Michelle.

## final-20261015-0006 — Brockman, 3:14

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0006)

Pokrass. And yeah, so she and her team did a ton of the grind of just talk to customers, see what the gaps are, these little things that people are annoyed about. Let's go fix them. Make sure we add that change into the post-training, that kind of thing.

I think it's actually kind of a beautiful lesson in many ways, right? That it's not just this raw intelligence that we're trying to deliver. We're trying to make something that's useful for people, right? And actually people need to be at the center of that, and that was really the way we started to operationalize this.

And on the tooling side, that we'd experimented with a few different form factors, and I think that the place we'd started was with a... And we actually had several different efforts. We ended up merging two of them and then merging another two of them. So we ended up with a little bit of exploration and then collapse, but I think that a lot of the energy was about what is the right form factor?

And if you project into the future, that the right form factor is an AI in the cloud, right? That has its own laptop, that's able to execute things on its own. And so we actually built that, and that was basically Codex Cloud. But it didn't get PMF, and I think it was too early, right?

And I think that my overall realization, too, is if you look at the evolution of AI coding, I mean, we led with basically working together with Microsoft to deliver GitHub Copilot back in, was that '22? Something like that. And you start out with auto-complete, then you move to more of the Cursor style. Like, okay, it's gonna be a sidebar but most of the time you're still doing the work in your editor. Then it's this agentic, you know, sort of Codex app or CLI. I think eventually it's going to end up being that cloud thing. I think that's coming soon.

## final-20261015-0007 — Shipper, 4:48

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0007)

Yeah.

## final-20261015-0008 — Brockman, 4:50

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0008)

And it makes you really realize you also need to calibrate on where the technology is now and build for the moment.

## final-20261015-0009 — Shipper, 4:56

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0009)

So one of the other things I think is really interesting, and it's a big challenge that a lot of AI companies face, even beyond the model companies, which is the models are moving so fast that any product strategy you have has to change every three to six months as the products change.

And so there's this really interesting thing with the Codex model line, but also the Codex app, where what you're able to do with the models now is fundamentally different than what you were able to do when ChatGPT was first designed. But you have this gigantic app and gigantic user base, and model tuned for those people that have a certain set of expectations.

One of the things I think is super interesting about the journey of Codex is it was this separate model line and separate product that now it seems like you're trying to put back in. But in order to get to where it is, it sort of feels like it had to be separate, and there's a lot of different ways to solve it.

So it seems like one thing that you did is split it out, and now you're gonna put it back in. What Anthropic has done, for example, is they have tabs. It's like Code and then Co-work in the same app, and I think that's kind of messy.

So tell me about that from a product strategy, research strategy perspective, that kind of move and why you have to do it and what's the best way to do it?

## final-20261015-0010 — Brockman, 6:21

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0010)

Well, the challenge here is fundamentally an innovator's dilemma. Right? You have a very successful product. Lots of people are using it, relying on it, loving it. Lots of diversity of different use cases, all of those things. But then you have a new leapfrog coming, and that is rapidly taking off, and you just see the value of if you could bring that to all your existing users, think about the value there. And by the way, no one has seen the consumer agentic moment yet. That just hasn't happened.

And I think that the way we approach this has been very deliberate, and we're taking a very step-by-step strategy. And so a lot of it actually comes down to technological unification, right? We realize that having a single agentic implementation, the harness, the connectors, all of those things, each of those are pretty general purpose regardless of what your use case is. So you really want to build them once in a unified way. And I think the goal ultimately is zero tabs, right? It's like the actual thing you want.

## final-20261015-0011 — Shipper, 7:13

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0011)

Thank God.

## final-20261015-0012 — Brockman, 7:13

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0012)

I know, right?

## final-20261015-0013 — Shipper, 7:14

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0013)

I saw the tab version today, and I was like, "No, don't do the tab."

## final-20261015-0014 — Brockman, 7:18

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0014)

I know. But it's not to say you don't take tabs as a stepping stone, right? And you want as few tabs as possible. And I think that is actually a really key thing.

But again, back to where are we going? You want an AI that you can talk to. Right? And if you think about a voice interface, it immediately suggests to you, you can't have tabs in a voice interface. It just doesn't even make sense.

ChatGPT right now, you have model pickers, and you have different features on the left and things like that. What you want is all that to just blur away, right? You just want an AI, like a agent in the cloud that can solve problems for you, right? That's like constantly out there thinking about how it can help you and what value it can add.

And so sometimes that's you proactively putting things in. Sometimes that's it trying to find news articles of interest, or you wake up and your calendar has been managed well and your inbox has been triaged, and it's like, "I need these three things from you." That isn't really a sci-fi vision anymore, right? We have the technology for it. We just need to do a little bit of engineering, a lot of bit of engineering. But we need to do that to scale and deliver it to everyone.

And so there will be a step-by-step here of how we deliver tabs, and part of the reason you need that is because there still is so much depth to what chat is, right? That it's able to scale to a billion users, which means that there's a lot of diversity and complexity under the hood that isn't obvious if you're just like, "Hey, it's a text box. How hard could that be?" It's like the simplest products are the ones that have the most depth of implementation and thought behind them.

But ultimately, our ambition and where we want to go is to deliver, we call it the personal AGI. We want this to be one unified experience, and you can chat with it. You can get trustworthy answers to questions and advice, but you can also ask it to go and accomplish work tasks for you. And all of that, I think, should be in a single seamless fashion.

## final-20261015-0015 — Shipper, 9:08

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0015)

But let's talk about the way to get there. It's really interesting about what you said is the first thing you said about the innovator's dilemma, because usually the innovator's dilemma, it's like you have got an incumbent, and then you've got a set of startups that are gonna beat the incumbent because the incumbent can't see the valuable thing that the startups can see 'cause it looks smaller.

And normally that's not happening within one company all the time. At least I don't, from disk drives in the '80s, I don't think they were moving that fast in the '80s to see that. And so what does that change about how to run a company if, because the frontier is moving so quickly, you have to expect to disrupt yourself?

## final-20261015-0016 — Brockman, 9:54

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0016)

That is the fundamental structure behind OpenAI in many ways, right? We've always set a goal, at least at a qualitative level, of every single year doing something that just totally obsoletes what we did before. Right? That just feels like a step function better than anything we've seen. And we basically have done that pretty consistently. And so that means you have to be set up for constant self-disruption. And a lot of that is around where you put pressures, right?

One of the things that used to cause me a lot of pain is we'd produce a new model that was no one had ever seen anything like it, and it would sit on a shelf because we're like, "Well, we're busy with the shipping the current things. We're fixing bugs," whatever. And it means, no, you actually really need to have a prioritization engine that says getting a new model out well is a priority, and that actually comes over some of the needs of the day.

Now, you can't take that too far, right? Because you also really need to make sure that you're doing a good job of constantly upleveling based on what you're learning from the world. That's part of why we deploy. And so I think that there's a constant management challenge of how to balance these priorities, and there's no one general purpose answer to it. That is fundamentally the job of management is to internalize these trade-offs and make decisions.

But I'd say that the way that we structure our teams is towards thinking about, [we] want smaller teams that can move faster, can try out new things. The way that we structure our code bases, all of these things are towards the vision of saying we're not just ossifying on a thing. It's like really about the machine that makes the machine.

## final-20261015-0017 — Shipper, 11:20

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0017)

How do you then keep people motivated and excited to do quality work if you're like, "I spent so much time making ChatGPT amazing, and now there's this new model out, which means that the Codex team is getting all this love, and there's so much thrash in our roadmap," and all that kind of stuff. How do you do that well?

## final-20261015-0018 — Brockman, 11:35

[Read this passage](https://every.to/p/how-codex-happened/interviews/greg/?edition=gift#final-20261015-0018)

I think that there's several pieces to it. Number one is I think that really helping people see the value in what they are doing is important, right? It's very easy to just give all the oxygen to the new shiny thing. And that is a mistake, right? And I think this is true across many companies, but you see it especially here, that there's so much core value that's delivered by people doing the behind the scenes unglamorous work, right? So improving reliability and latency, all of these, improving how different regions are handled, things like that. Those actually transfer, right?

And I think that understanding what is one-time work that is specific to the current moment and the current system versus what's gonna transfer to the future, I think the more that you can highlight these investments, we're gonna have to make them for the new system versus the old system anyway. And so just actually internalizing what we're learning from reality and making sure we do a good job of meeting the moment. Sometimes the specific implementation won't transfer, but the lessons will.

And so I think really helping people understand that, honestly, a lot of it is just calling it out, just saying, "This team that is behind the scenes has been quietly just making things better. That's exactly the kind of thing that we wanna see." 'Cause you kind of don't need to put as much highlight on the people who are doing the thing that's going to get all the attention by itself. So I think a lot of it is what management highlights tends to reflect on what the values are within, that people perceive within the org.

---

# Tibo Sottiaux

Research & engineering

Final interview transcript supplied October 15, 2026. Source file F0C9K0V5ZGS; text, speaker labels and timestamps preserved from the supplied final transcript.

Passage IDs identify this edition's captured paragraphs. Bracketed transcription notes are not speech.

## final-20261015-0001 — Thibault "Tibo" Sottiaux, 0:00

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0001)

Cool. No, we're working things through. We have some more creative solutions than others. But this is the shape that we think we'll go out with roughly in two weeks. But it will feel significantly better. Two weeks is eternity in software development these days. So yeah.

## final-20261015-0002 — Dan Shipper, 0:25

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0002)

Well, tell me about why you think that's the right move.

## final-20261015-0003 — Sottiaux, 0:29

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0003)

So it's really about making it work across all devices. And for me, just really what you want is, you want a highly, highly personal entity that you can reach from anywhere at any time. And that's not yet— what you've seen is not yet that thing.

But you want it to work on your phone. You want it to work on web. You want it to work on desktop. You want everything to be in sync. You want your agent to have— It should feel like this one entity.

Most of our users are on web, mobile. We've got two million businesses on ChatGPT. So we're kind of very eager to bring everything that has made Codex successful to them in a way that's also lightweight and appropriate.

And we did an analysis as well of how many actually of the hard tasks that people give Codex around general work require local file system and computer access — [the answer] is actually surprisingly little.

## final-20261015-0004 — Shipper, 1:39

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0004)

Yeah. I bet.

## final-20261015-0005 — Sottiaux, 1:40

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0005)

So what we've done is, we're now capable of running all of that, but outside of your computer environment.

## final-20261015-0006 — Shipper, 1:48

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0006)

You did the hoist. Will it be able to call down to my computer?

## final-20261015-0007 — Sottiaux, 1:54

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0007)

Eventually. Yeah.

## final-20261015-0008 — Sottiaux, 1:57

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0008)

And then you start to really see, you have the agent, it's able to act on any device, depending on whether it's connected to it or not, with a variety of permissions.

Maybe it can not just access your computer, it can also access a Dev Box or multiple Dev Boxes.

## final-20261015-0009 — Shipper, 2:16

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0009)

Yeah. Well, I guess I'm curious about— I wanna back up, actually. Tell me about when did you start working? Were you working on Codex from the very beginning? 'Cause when Codex Cloud launched and in the GPT-5 days, I was mostly talking to Alexander. So I'm kind of curious about your story with Codex and how you came to it.

## final-20261015-0010 — Sottiaux, 2:43

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0010)

There's many beginnings to Codex.

## final-20261015-0011 — Shipper, 2:44

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0011)

Yeah. And 'cause it's the same name for a lot of different products too.

## final-20261015-0012 — Sottiaux, 2:48

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0012)

Yeah. I got involved with coding agents around December 2024. And this was in the pursuit of writing better infrastructure and accelerating the research at OpenAI.

And so what we started to work on was a series of models and way to interact with them to write code really fast. And we trained agents specifically on our monorepo, Python, to write good infrastructure.

## final-20261015-0013 — Shipper, 3:24

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0013)

Oh, you're training it on your monorepo?

## final-20261015-0014 — Sottiaux, 3:26

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0014)

Yeah. That was in December of 2024. Python only really, specializing on OpenAI infrastructure libraries.

## final-20261015-0015 — Shipper, 3:38

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0015)

Oh, for internal use?

## final-20261015-0016 — Sottiaux, 3:40

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0016)

Yeah, for internal use only.

## final-20261015-0017 — Shipper, 3:42

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0017)

'Cause I thought that your monorepo was your benchmark for your new coding agent. Okay. Yeah.

## final-20261015-0018 — Sottiaux, 3:48

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0018)

It is as well, but the origin of how I got involved in Codex was just really accelerating research within OpenAI. Accelerating specifically writing infrastructure for the system.

## final-20261015-0019 — Shipper, 3:58

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0019)

Mm-hmm.

## final-20261015-0020 — Sottiaux, 3:59

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0020)

That was my background. And so did this for a bit and then as things started to work, we got increasing adoption, increased our ambitions step by step. Eventually to the point where we thought we should combine forces with another effort, which was called Scientist Assistant, and there was a merger of teams, and that's when I got to know Alexander.

So as I'm working with Alexander, we were working on what's referred as the cloud sprint, which was roughly a seven-week sprint. And then we came out with Codex, the cloud version, which was really exciting.

It was a little bit too early. Eventually, it's what we're going back to. What everyone will go back to. But it did not have PMF at the time, and it was too frictionful.

## final-20261015-0021 — Shipper, 4:53

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0021)

Tell me about that 'cause I talked to Alexander about it, I talked to Greg about it. I think Greg's framing was interesting, which is something like we went from Copilot to you have a chat inside of your IDE to the next step was, in my mind, Claude Code being like, "We don't need the IDE at all. We just have an agent on your computer." You guys skipped that step initially.

## final-20261015-0022 — Sottiaux, 5:26

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0022)

We didn't skip it. We had it internally. We did not release it. We went through that phase very quickly. We built a version of that and then discarded it after a month to go for the full cloud version.

## final-20261015-0023 — Shipper, 5:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0023)

Why?

## final-20261015-0024 — Sottiaux, 5:50

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0024)

Because we were optimizing also for what we thought we would benefit from the most if we were willing to invest a lot in the setup. And so it was very successful internally. 'Cause everyone having perfectly reproducible environments, provisioned dev boxes at scale, seemed like the way to go. And it was making people a lot more productive.

But then externally that was way too high friction. And then with models getting better and better, also realized you wanted to pull in a lot more context from the system and you kind of felt trapped inside of that little VM. So we went back to local. Then again, it was too early.

## final-20261015-0025 — Shipper, 6:42

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0025)

Yeah, I feel like something really shifted. Oh, by the way, I should tell you the framing of the article.

## final-20261015-0026 — Sottiaux, 6:53

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0026)

I'm happy to talk about it and check—

## final-20261015-0027 — Shipper, 6:54

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0027)

Yeah, totally. The framing of the article is how Codex became king. I feel like there was this... Well, the last year or so has been very Claude Code Anthropic focused in terms of the public narrative. And we were very early to Claude Code, and the team now is almost everyone is using Codex as their daily driver.

And so I think there's going to be a shift in the meta over the next month or two. When that happens, wanna have an article that's like, “Here's how that happened from the inside,” an oral history of that. So that's kind of what I'm putting together, and the sort of framing all the interviews.

I do feel like I was very around Codex... Well, actually, before we get into that, tell me about the decision to split off the model line. 'Cause that feels like one of the things that it was one of the ingredients to success here. Tell me about that.

## final-20261015-0028 — Sottiaux, 8:02

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0028)

So there was no decision to split off the model line.

## final-20261015-0029 — Shipper, 8:07

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0029)

Mm-hmm.

## final-20261015-0030 — Sottiaux, 8:08

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0030)

It was rooted in the very beginning of this research acceleration effort, which I was part of, where we were training agents specifically on coding, to help us write infrastructure and do research faster. And we were doing so off of things that were so far from what we were doing for external models. This is just really a research exploration and an optimization of our velocity.

We were seeing rapid gains on coding. And so we continued doing that, and what we developed is a series of research methodologies that were very different and were hard to reconcile for a while.

One of the relevant things here is, after the Cloud sprint, we did not immediately find PMF. I took over the team, and then I changed the direction quite a bit. We pivoted to local. We had a bit of a lull with the Rust rewrite of the CLI,

## final-20261015-0031 — Shipper, 9:29

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0031)

Oh, interesting. I didn't realize you rewrote it.

## final-20261015-0032 — Sottiaux, 9:32

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0032)

Yeah. So the Codex CLI initially was written in TypeScript. Then there was a Rust rewrite, where we decided... that it wasn't the CLI was the way to interact with the agent, but just really the underlying infrastructure was our agent, which was our most precious infrastructure.

And we needed to take great care in it and think about the correctness, the security, sandboxing, and all of that was low-level system design. And we wanted to invest in it. And for the long term, there was absolutely no lack of conviction. It's like 100% conviction that this is very, very core to our success.

And so we decoupled the textual — the TUI implementation — from the agent implementation very early on. But that also led to, we were kind of slow to have really good CLI features, because we were so busy developing the agent.

## final-20261015-0033 — Shipper, 10:44

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0033)

That's interesting.

## final-20261015-0034 — Sottiaux, 10:45

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0034)

And then we kind of went back to it.

## final-20261015-0035 — Shipper, 10:47

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0035)

Tell me about the open source decision. Does that connect to the open sourcing of the TUI decision?

## final-20261015-0036 — Sottiaux, 10:55

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0036)

Yeah. So the open source decision was a decision I made with the team. One is, when I grew up, I loved open source. It's how I learned everything.

Two, it felt that if we were right and this was going to change how code is written, and very quickly so, open source would struggle, and we kind of also wanted to be part of the ecosystem instead of just causing trouble and not having solutions. Just being part of it.

And then the third thing was, it feels like it was a very — still very early days — and there was a whole bunch of explorations necessary, and nobody really knew exactly what a good agent would look like, what a good implementation would look like. And experimenting with that in the open, and showing that there's not really magic to it, and everyone can fork it and contribute back, and then we can benefit from the ideas — seemed very appealing to me. So that's why we decided to do it open source.

## final-20261015-0037 — Shipper, 12:02

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0037)

Got it. Okay. So it sounds like... 'Cause my memory of some of this was, there was a point before 5.2. There was maybe there was like five Codex, 5.1 Codex, where it was very slow, powerful, hard to use, but often right.

And it sounds like maybe during that time — tell me if the timeline lines up — during that time, part of what you were doing was, yes, making better coding, but also building the underlying fundamentals of the agent itself, like the agent harness, that made it so that the visible progress to users was a bit slower than you would've liked. Is that a good summary?

## final-20261015-0038 — Sottiaux, 13:01

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0038)

Yeah. We took the general cultures, like, move fast on things  where you can move fast. And so you want to experiment and learn, and then for some of the fundamental things, take your time and invest in the long term success. And so for sure, we were really doing a lot of underlying investments during that time.

And then also this was our specific strategy on the model as well, which was to focus on improvements that would benefit — helping solve hard engineering problems. Not necessarily being a model that would do things that are cool, or kind of perceived as cool.

I think around the same time, Gemini was just investing a lot in front end. I was discussing with people over there, like, “Oh, what are you focusing on?” And these kind of things. And it was really like optimizing the model for this wow factor. We made the opposite decision, and we were like, “Okay, how can we develop an accelerating flywheel as quickly as possible? What is the best model as a tool look like, the best agent as a tool look like?” And then build a lot of that underlying infrastructure.

And at some point, we had this sort of realization that we should continue maintaining the CLI, but the biggest next investment should be a native desktop application. And just really optimizing for what the future of agents were gonna be. Which is a lot of multitasking, probably also you want visual output, you want to interact through voice. It has to be this very rich interface if you're gonna spend all your time in it. And just doing all that in your terminal felt very silly when you played it all out. And so decided to invest a lot in that.

## final-20261015-0039 — Shipper, 14:59

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0039)

Tell me about that.

## final-20261015-0040 — Sottiaux, 15:00

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0040)

That's when Andrew got involved. He was like, I found him. He was just the best person for this.

## final-20261015-0041 — Shipper, 15:08

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0041)

Yeah. He seems amazing at it. Well, tell me about that because there was that time where TUIs were so hot. And even people internally who are non-technical were starting to use TUIs for everything. So it was a pretty, and still is to some extent, a pretty counterintuitive bet to be like, no, no, actually GUIs are still better strictly.

Tell me about that decision. 'Cause you can imagine being like, “Holy shit, everyone's talking about Claude Code, and it's all about the terminal and blah, blah, blah. We have to catch up to that or whatever.”

## final-20261015-0042 — Sottiaux, 15:53

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0042)

Mm. I don't believe in looking at the competition to make good decisions. I think it's important to  reach conclusions from first principles. And based on the information that you have.

And the information that I had at the time was that our models were rapidly getting better. And we were seeing the models do longer and longer horizon tasks successfully. We're seeing models capable of ingesting a lot of data outside of code bases as well in order to inform their work.

And then we were early at OpenAI as well on voice and multimodal. We had state-of-the-art image generation, state-of-the-art advanced voice mode that shipped two years ago.

And so this idea of playing it out where is this very rich interaction, it just felt completely silly to expect that that would happen in a terminal. But obviously requires a lot of conviction to then go and do it.

## final-20261015-0043 — Shipper, 17:08

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0043)

Yeah. And tell me about that because you have to convince other people that you're right. Maybe if you're not looking at the competition, someone is.

I guess I'm curious about the culture. It feels like there's a culture of a lot of parallel bets here, and seems like you're able to take some risks, but I'm curious how that works.

## final-20261015-0044 — Sottiaux, 17:36

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0044)

Mm. It was clear and controversial to them within the team that desktop was our main bet. The desktop application native application was our main bet. And so it was not a small parallel side thing.

Obviously with anything controversial, people don't necessarily see it until it happens. But I would say the conviction of the team after talking it through and explaining why we were making that bet, was people were excited about it.

## final-20261015-0045 — Shipper, 18:18

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0045)

Yeah, yeah.

## final-20261015-0046 — Sottiaux, 18:22

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0046)

Especially 'cause it took us less than a month for a lot of the team to have switched it and feel that this was actually going to work. I think pre agents, maybe this would have taken us four to five months, and it would've been very hard to convince people because we were kind of able to make rapid progress.

## final-20261015-0047 — Shipper, 18:47

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0047)

That does make a big difference. Tell me about that.

## final-20261015-0048 — Sottiaux, 18:50

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0048)

Mm. Yeah, I think it takes way less talking now to convince people. And you can just go and do it in like two days. And then show the thing.

And that's a lot more convincing than any document you could write before. And so that was kind of also playing out in real time, which was very cool.

## final-20261015-0049 — Shipper, 19:15

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0049)

And at what point did you decide to do the desktop app as the main bet?

## final-20261015-0050 — Sottiaux, 19:20

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0050)

This was in December of last year.

## final-20261015-0051 — Shipper, 19:22

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0051)

Okay. And that was around when 5.2 came out? Okay.

What about, if you were ever gonna split out a model line again, and maybe a product too, in the same way where it's like you had a ChatGPT desktop app, you didn't necessarily have to do a separate Codex desktop app. When is it the right move to do that kind of split?

## final-20261015-0052 — Sottiaux, 19:50

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0052)

Right. So we were also optimizing for velocity first and foremost. Just deep sense that all of this was going to change again and again in the course of a couple of months, and you need to optimize for how quickly you can ship and move. And that we needed to start with a very forgiving audience as well.

## final-20261015-0053 — Shipper, 20:19

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0053)

That's interesting.

## final-20261015-0054 — Sottiaux, 20:20

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0054)

And doing that with the ChatGPT desktop app felt that was not fulfilling either of those.

## final-20261015-0055 — Shipper, 20:28

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0055)

And by forgiving audience, usually people don't think of developers as being particularly forgiving.

## final-20261015-0056 — Sottiaux, 20:35

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0056)

I think they're amazingly forgiving. And they are very good with feedback as well, which makes it delightful to work with. If you can sort of navigate the fact that you get like a thousand different feelings thrown at you.

Shipper: [laughs]

## final-20261015-0057 — Shipper, 20:48

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0057)

Yeah.

## final-20261015-0058 — Katie Parrott, 20:50

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0058)

Well, they understand what it means to build a product, whereas Joe in Montana who uses ChatGPT to get answers doesn't understand what's going on under the hood.

## final-20261015-0059 — Sottiaux, 20:59

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0059)

Yeah. And if it doesn't work, that person will probably stop using your product, whereas someone who's an early adopter and a developer will come and bother you and be like, “Hey, you broke me. And I wanna know about that, and I wanna fix it. And I wanna learn from you.”

## final-20261015-0060 — Shipper, 21:15

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0060)

That feels like another good generalizable strategy is serving builder type people with new AI stuff 'cause they're gonna figure out all the different corners and all the ways that you might not even realize, and then you can translate that to everybody else. Is that consciously something that you guys are doing?

## final-20261015-0061 — Sottiaux, 21:31

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0061)

Yeah, I think there's a lot that we learned from that strategy. And now we're reproducing in other corners.

## final-20261015-0062 — Shipper, 21:38

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0062)

What are the subtleties of—

## final-20261015-0063 — Sottiaux, 21:41

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0063)

On different models, we will train other specialized models when we feel like it allows us to learn something more quickly. Or when there are research reasons to do it.

## final-20261015-0064 — Shipper, 21:53

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0064)

What are the—

## final-20261015-0065 — Sottiaux, 21:54

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0065)

We did one for life sciences. Rosalind. Yeah.

## final-20261015-0066 — Shipper, 21:59

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0066)

Interesting. And what are the subtleties of doing that sort of strategy well, the serving builder type strategy?

## final-20261015-0067 — Sottiaux, 22:07

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0067)

The subtlety is, and this is something that we are still figuring out for Codex, is that you can build amazing technology that is then overfit, right, to this type. And so it requires then quickly understanding how you can pivot it and position it to be broadly applicable. Or to generalize it, and this is exactly what we're doing with Codex, right?

So we're... Okay, the technology is amazing. The agent can do all these things. It's clearly much more useful than just for coders and builders. We're seeing very quick adoption from people who are doing all sorts of general, day-to-day work. But it has been quite optimized for technical people. And so that's the trade-off.

## final-20261015-0068 — Shipper, 23:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0068)

Yeah. How are you thinking about how to push it back well?

## final-20261015-0069 — Sottiaux, 23:13

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0069)

So a lot of experimentation. We have a broad network with whom we push early ideas, you included.

## final-20261015-0070 — Shipper, 23:26

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0070)

Yeah.

## final-20261015-0071 — Sottiaux, 23:27

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0071)

And then also we have a lot of diversity at OpenAI itself in ways of working. And so we're able to ship a lot and iterate quickly within OpenAI internally.

## final-20261015-0072 — Parrott, 23:45

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0072)

When did Codex sort of escape containment in the technical builder side into some of those other functions of OpenAI, internally?

## final-20261015-0073 — Sottiaux, 23:54

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0073)

I'd say it really started to pick up, what was it? Six weeks ago roughly. Something like this.

## final-20261015-0074 — OpenAI, 24:01

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0074)

That was when we hit 85%. Yeah.

## final-20261015-0075 — Sottiaux, 24:05

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0075)

Yeah. That's when it was clear, when Sarah Friar did an all-hands and explained how the last fundraise was in great part orchestrated by Codex. It's like, okay. Seems like we've got something quite general here.

And then we were seeing a lot of stories across comms, marketing, and sales. So we had this desire of just really generalizing it and solving some of the hard problems around security, safety, when you don't... And having the power of an agent that uses code under the hood, but where code is the tool and not the main output.

Shipper: [laughs]

## final-20261015-0076 — OpenAI, 24:53

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0076)

I feel like auto review was a big moment where people could really just start feeling confident.

## final-20261015-0077 — Sottiaux, 24:57

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0077)

Yes. So we developed internally. Externally we shipped it as a review for me or auto review. And you know what it does, right? It's this second agent that verifies all the actions of the primary agent, but against the original intent. And it reduces approvals by like an order or two of magnitude.

## final-20261015-0078 — Shipper, 25:20

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0078)

I love it.

## final-20261015-0079 — Sottiaux, 25:30

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0079)

That's a reason as well, why I think the team cares deeply about the experience, but also uses it every day. And so we're all using the product that we ship. There's not a different version of Codex internally than there is externally. That was like an important decision. We could have optimized it all kinds of different ways internally than the external version, but the majority of OpenAI is on the Codex desktop app all day, every day.

We have a very, very active feedback channel with over 250 internal messages of feedback every day, which is now in great part triaged and actioned on by Codex itself. And so it's kind of like this, yeah, building itself.

## final-20261015-0080 — OpenAI, 26:24

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0080)

That triage is so crazy 'cause you'll share something and it'll be like, “Oh, someone's already shared this. Here's where it's being worked on.” Or it will just correct you and be like, “That's not right.” Or like, “Yes, this is the right use case.” It will give you education, which is kind of crazy.

## final-20261015-0081 — Shipper, 26:39

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0081)

That's so cool.

## final-20261015-0082 — Sottiaux, 27:07

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0082)

Yeah. The Codex we spin up on feedback has access to the roadmap, has access to all of the documentation that we wrote and has access to the code base, and is able to sort of tell whether something is a regression or not. And then if it's a regression, writes a PR, assigns it to the person, based on the blame layer.

And so it's just a highly efficient sort of magical experience. We wanna bring this to external too.

## final-20261015-0083 — Shipper, 27:38

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0083)

I guess what's it been like for you going from... Codex for a while was mostly under the radar unless you're a technical person. You ran a Super Bowl ad. It's become a thing both internally and externally. What's that been like for you, that ride?

## final-20261015-0084 — Sottiaux, 27:58

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0084)

It was awesome to see. For almost 15 years, I've always built tools and products for people. I love to see the broad use and utility that people get out of it. I always want to put something out there that is the best version of itself,

and seeing the wider and wider adoption in something that we initially built with a high degree of ambition, but was just internal only, going to then external and then continuing to build it with the community, then going from there. And it's the very beginning still.

## final-20261015-0085 — Shipper, 28:42

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0085)

Yeah. Do you have any specific moments where you're like, "Oh shit, this is a real thing"?

## final-20261015-0086 — Sottiaux, 28:48

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0086)

I think there were a lot of small moments. Like the Super Bowl ad was definitely like, oh wow, we're actually doing this.

## final-20261015-0087 — Shipper, 28:58

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0087)

Yeah. Hard to put it back in the box when it's in front of that many million people.

## final-20261015-0088 — Sottiaux, 29:01

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0088)

Yeah, yeah, yeah.

## final-20261015-0089 — Shipper, 29:02

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0089)

Were you watching—I assume you were watching the download numbers when it went live?

## final-20261015-0090 — Sottiaux, 29:06

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0090)

Well, we had a Codex party where we were watching it live—

## final-20261015-0091 — Shipper, 29:09

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0091)

Yeah. Okay.

## final-20261015-0092 — Sottiaux, 29:09

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0092)

and then we were sharing the numbers.

## final-20261015-0093 — Shipper, 29:11

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0093)

What was that like?

## final-20261015-0094 — Sottiaux, 29:12

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0094)

It was exciting. It was exhilarating. The vibes have been really good. I think once we got past the initial period, which is like pre-GPT-5, I would say—GPT-5 made a dent and then 5.2 just really started to accelerate.

I feel like the vibes online, the community, it's been really good since the start. And I've always wanted to sort of like be as nurturing of that as I possibly can.

And sometimes it's challenging, with scale and capacity and all these things. But always try to be very transparent with what we're dealing with.

## final-20261015-0095 — Shipper, 30:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0095)

Yeah. Mm-hmm.

## final-20261015-0096 — Sottiaux, 30:06

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0096)

So no, I mean, it's been awesome. It's been magical.

## final-20261015-0097 — Shipper, 30:10

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0097)

Do you actually have a big red button that resets rate limits?

## final-20261015-0098 — Sottiaux, 30:13

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0098)

I have a big red virtual button. There's a better version coming out today, actually. So up until now, when I press it it actually resets everyone's rate limit. People have been complaining about that, so we changed it where we now give everyone their own button.

## final-20261015-0099 — Shipper, 30:33

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0099)

Hmm. Oh. Cool.

## final-20261015-0100 — Sottiaux, 30:35

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0100)

So you can press it yourself—

## final-20261015-0101 — Shipper, 30:37

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0101)

Whenever you want. Yeah. That's great. Very cool.

Sottiaux: [laughs]

## final-20261015-0102 — Sottiaux, 30:40

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0102)

I think people will prefer that.

## final-20261015-0103 — Shipper, 30:41

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0103)

Yeah, I love that. Katie, any questions on your end?

## final-20261015-0104 — Parrott, 30:45

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0104)

I just wanna know, this is kind of my guiding question as a non-technical person who didn't touch Codex because our Naveen used Codex, and so I was like, "Oh, it can't be for me too." Why am I, as a writer and non-technical person, why is Codex so good for me?

## final-20261015-0105 — Shipper, 31:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0105)

Yeah, she uses it every day.

## final-20261015-0106 — Parrott, 31:06

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0106)

I live in Codex.

## final-20261015-0107 — Shipper, 31:07

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0107)

[laughs]

## final-20261015-0108 — Sottiaux, 31:08

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0108)

Thank you.

## final-20261015-0109 — Parrott, 31:10

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0109)

And then—

## final-20261015-0110 — Sottiaux, 31:12

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0110)

So why?

## final-20261015-0111 — Parrott, 31:13

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0111)

Yeah, just what is it, what changed or what has always been true about it—I don't know which it is—that now makes it so generalizable?

## final-20261015-0112 — Sottiaux, 31:22

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0112)

Yes. So this is just, I think, the natural conclusion of doing a very powerful coding agent, is that over time, the agent needs access to broader and broader context. Ultimately, as much context from the entire world as we can give it. And it needs to be able to perform a broad set of actions to be useful for the builder, right?

And so it tends towards not a coding agent, it tends towards a very general, useful assistant that can help you with everything. And when you realize that, which we did, it's like Codex is not for technical people, it's not for builders. Codex is for everyone. We're building a very, very general kind of assistant.

And so we started then to take steps to make the Codex app more friendly. And if you're not technical, we don't show you code, we don't show you bash commands to approve that you wouldn't be able to possibly understand. We added extra security controls. We added the auto review.

And so like evolved to a place so that it could actually provide that utility for everyone. You still have the name Codex—

Parrott: [laughs]

## final-20261015-0113 — Sottiaux, 32:44

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0113)

which a lot of people like. Some people are a little bit intimidated by it. Yet to find out if it will be sort of accepted by society.

I hope that it will take on a life beyond coding. And Codex will become synonym with a very powerful kind of agent or assistant that can help you with pretty much anything you throw at it.

## final-20261015-0114 — Parrott, 33:10

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0114)

Well, 'cause it's interesting 'cause when I see the word Codex, I don't think of code. I think of a codex, like a book, like the Bible.

## final-20261015-0115 — Shipper, 33:16

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0115)

The scroll.

## final-20261015-0116 — Parrott, 33:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0116)

Again. Because—

## final-20261015-0117 — OpenAI, 33:20

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0117)

Its definition.

## final-20261015-0118 — Parrott, 33:21

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0118)

Yeah. So brand wise, there's this merge happening, and what will be the name?

## final-20261015-0119 — Sottiaux, 33:32

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0119)

Mm-hmm. So Codex will move to ChatGPT, and so the overall experience will be ChatGPT. But you will be able to use Codex within ChatGPT.

## final-20261015-0120 — Parrott, 33:43

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0120)

Okay. Interesting.

## final-20261015-0121 — Shipper, 33:46

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0121)

But the version of ChatGPT that I would be using would be basically the Codex desktop app with ChatGPT added, but it's sort of positioned as ChatGPT with Codex inside it.

## final-20261015-0122 — Sottiaux, 33:57

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0122)

Yes. So we're not taking any of the functionality away. It is exactly the same app, right? But we're actually replacing the ChatGPT desktop app with it. So the Codex app becomes the ChatGPT desktop app, and through that becomes ChatGPT on desktop, right? But it's all the same functionality that you use.

## final-20261015-0123 — Shipper, 34:29

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0123)

Why not just keep it separate?

## final-20261015-0124 — Sottiaux, 34:33

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0124)

We're considering everything. We're learning fast and adapting fast. So one of the things that's for sure is people who love to just use the Codex app as a separate thing will be able to continue to do that. You'll not be forced, not for a while, to move. And then we'll continue to adapt and learn.

## final-20261015-0125 — Shipper, 35:02

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0125)

Does it go to being like, "I have to use the ChatGPT desktop app."

## final-20261015-0126 — Parrott, 35:05

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0126)

Would you rather have a toggle or a separate app?

## final-20261015-0127 — Shipper, 35:08

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0127)

I don't want any toggles. No toggle.

## final-20261015-0128 — Sottiaux, 35:11

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0128)

Yeah, just a very pure a pure experience. And yes.

## final-20261015-0129 — Shipper, 35:16

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0129)

Yeah.

## final-20261015-0130 — Parrott, 35:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0130)

Behavior change is just so hard. That's almost harder than the capability of the technology it seems—

## final-20261015-0131 — Shipper, 35:22

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0131)

Mm-hmm, mm-hmm.

## final-20261015-0132 — Parrott, 35:23

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0132)

[Behavior change] is getting however many millions or billions of people you have to accept this new thing. I imagine that has to be a very complicated process to help people understand this new thing that they're gonna be seeing that is different, so much more powerful, but helping them see how it can help them.

I was really impressed with the rollout of the communications around the plugins, for example. How do you think about handling human behavior change and educating people to use the app?

## final-20261015-0133 — Sottiaux, 35:58

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0133)

So there's two populations we have to bring along, right? So there's the Codex set of users today who are like—

## final-20261015-0134 — Parrott, 36:06

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0134)

Us.

## final-20261015-0135 — Sottiaux, 36:07

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0135)

Yeah. Used to everything that it does. And I think that one we'll manage fairly easily, figure out some details to not upset the, the—

## final-20261015-0136 — Parrott, 36:20

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0136)

The fans.

## final-20261015-0137 — Sottiaux, 36:21

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0137)

The fans of this world. The other thing which I'm spending more of my time on is how do we bring the one billion, almost one billion active ChatGPT users along and have them discover that now our models can do so much more. And it's a really exciting opportunity.

And so we're working on how do we introduce it to them. And this is also why we're doing the move of bringing Codex to ChatGPT, to just really decrease the friction, and make it super easy. Again, we have over two million businesses using ChatGPT. To date are not benefiting from Codex.

## final-20261015-0138 — Parrott, 37:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0138)

Mm. Mm-hmm.

## final-20261015-0139 — Shipper, 37:06

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0139)

Why not just... And I'm honestly asking this question. I could imagine, for example, it just still for them looks like ChatGPT, but when you give it a query that verges on something that Codex can do, it's like, "Hey, I have all these new powers. Do you wanna allow me to access your computer," and whatever and there's no tabs. It's just like... Yeah.

## final-20261015-0140 — Sottiaux, 37:28

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0140)

Yeah. We've played with routers in the past—

## final-20261015-0141 — Shipper, 37:31

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0141)

Yeah. Yeah.

## final-20261015-0142 — Sottiaux, 37:31

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0142)

or using intent detection and it's very hard to do well.

## final-20261015-0143 — Parrott, 37:37

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0143)

You can't really read the user's mind, right? Until we have chips in people's brains, you won't really know what they want.

## final-20261015-0144 — Shipper, 37:45

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0144)

Even then.

## final-20261015-0145 — Sottiaux, 37:46

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0145)

Yeah. They're like, "Sign me up." And so the future is one unified mode that does everything, where ChatGPT can just do things for you. We're not there yet. And—

Shipper: [laughs]

## final-20261015-0146 — Shipper, 38:03

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0146)

It's like a change management.

## final-20261015-0147 — Sottiaux, 38:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0147)

a lot of people just rely on how fast and delightful ChatGPT is, for all sorts of things in their lives. And we don't wanna disrupt that. We wanna respect that. At the same time, we want to show you what's now capable, what you now can do with a Codex-like agent. Especially, I think, I use it on my phone all the time now. Once it's synced across device, it's really good.

So we're trying to do it thoughtfully and gradually. And then eventually enough people will kind of be used to these kinds of capabilities and then we'll be able to kinda unify everything.

## final-20261015-0148 — Shipper, 38:44

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0148)

You feel like you need another playground with less pressure to do more different things once you do this?

## final-20261015-0149 — Sottiaux, 38:50

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0149)

I'm sure that will emerge. Yeah.

## final-20261015-0150 — Shipper, 38:54

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0150)

Do you have a sense for along what dimension or what you need, why you need a playground?

## final-20261015-0151 — Sottiaux, 39:00

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0151)

Yeah. We have a couple of things that are brewing right now, which I'm really, really excited about, which feel qualitatively different from anything that we've put out. And so we're exploring a little bit about how we're gonna bring that to the world. And it might be separate. Yeah.

## final-20261015-0152 — Parrott, 39:21

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0152)

Awesome.

## final-20261015-0153 — Shipper, 39:22

[Read this passage](https://every.to/p/how-codex-happened/interviews/tibo/?edition=gift#final-20261015-0153)

I'm excited for it.

---

# Amelia “Mia” Glaese

Model research

Final interview transcript supplied October 15, 2026. Source file F0C9K0V5ZGS; text, speaker labels and timestamps preserved from the supplied final transcript.

Passage IDs identify this edition's captured paragraphs. Bracketed transcription notes are not speech.

## final-20261015-0001 — Dan Shipper, 0:14

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0001)

Talk about everything, please. Maybe I'll give you a framing, and then would love as much honest detail as you can give.

So my general framing for the article is how Codex became king. I want it to be an oral history of how you went from ...

'Cause for me, I was in the room with Alexander, Alexander, the night before you launched GPT-5. And it was the first time where I remember ... 'Cause I was at that point really psyched about Claude Code and not looking at the code anymore.

And on the GPT-5 launch, it was very pair programmery, and there was this split between Codex for professional engineers and vibe coding inside of ChatGPT. And I was like, “This doesn't work for me,” you know?

And then in like November, December, it feels like you guys have figured something out. And first with 5.3, and then the Codex desktop app. I and most of the rest of the team who are all Claude people started using Codex for pretty much everything.

And so I wanna tell that story a little bit, and both tell it from the perspective of, just help me understand how things happened and why things happened the way they did and what you sort of figured out and what changed.

And to the extent we can, I would love to be able to sprinkle in a little bit of actual artifacts from the time of, like, how do you make decisions, like Slack or strategy documents or whatever. Whatever you feel comfortable sharing, just send Codex off on it, and obviously get it approved. But I'd really love to tell the story in detail. So does that make sense? Okay, cool.

## final-20261015-0002 — Amelia "Mia" Glaese, 2:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0002)

Sounds great.

## final-20261015-0003 — Shipper, 2:05

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0003)

Well, let's talk first about GPT-5 and that launch and where you guys were and what you were thinking and what you were feeling and as a way to set up maybe what changed. And maybe my perception of it is totally different from yours. I'm sort of curious about that launch and that model.

## final-20261015-0004 — Glaese, 2:27

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0004)

This is already a tough question because I was not responsible for the GPT-5 launch.

## final-20261015-0005 — Shipper, 2:35

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0005)

You're not. Okay. Well, where were you? Where were you then?

Glaese: [laughs]

## final-20261015-0006 — Glaese, 2:38

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0006)

So yeah, maybe that's helpful context.

## final-20261015-0007 — Shipper, 2:42

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0007)

Yeah. Yeah.

## final-20261015-0008 — Glaese, 2:42

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0008)

So I've been working at OpenAI for, like, more than two years, but I've been doing a bunch of ... And for that time, I've been responsible for both the data that we use for our model training and alignment, which is more really about can we get models to do the right thing and what humans want them to do.

And I started getting into Codex very early, basically at the beginning of Codex when it was a project of, like, four people and—

## final-20261015-0009 — Shipper, 3:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0009)

Like the separate model line. Yeah.

## final-20261015-0010 — Glaese, 3:19

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0010)

And it was this project of four people where we're like, “We need to have a coding agent kind of thing.” And so it was this somewhat, like, this bet where we were just like, “Okay, we're just gonna do this right now. If we don't do it, that would be a mistake, so let's just do it.”

## final-20261015-0011 — Shipper, 3:36

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0011)

Do you remember what triggered that, especially the idea of having it be a separate coding agent? Because I assume the idea of coding agents has been around OpenAI for a very long time, right?

## final-20261015-0012 — Glaese, 3:47

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0012)

I don't think it was ever intended to be a separate coding agent, right? It was intended to be our coding agent. But it was this very focused effort on just, can we get the best coding agent in the world—

## final-20261015-0013 — Shipper, 4:01

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0013)

Yeah, I see.

## final-20261015-0014 — Glaese, 4:02

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0014)

There, as quickly as possible.

## final-20261015-0015 — Shipper, 4:03

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0015)

Yeah, yeah. Okay, got it.

## final-20261015-0016 — Glaese, 4:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0016)

And so yeah, that's kind of that. But I think—

## final-20261015-0017 — Shipper, 4:10

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0017)

And when was that specifically, when you started?

## final-20261015-0018 — Glaese, 4:13

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0018)

That was ... I'm not sure how much we talk about this publicly, but it was a lot more than a year ago.

## final-20261015-0019 — Shipper, 4:22

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0019)

More than a year ago, yeah. So it was pre-GPT-5, 'cause I think GPT-5 was around a year ago. Yeah. And one of the interesting architectural choices there is a separate model line. Tell me about that decision.

## final-20261015-0020 — Glaese, 4:37

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0020)

I think it was very much focus on the coding agent that we wanted to build. And it was never meant to be a forever thing, right? But oftentimes it can help to just have a focus, grouped sprint on something.

And that was really ... Codex was maybe the most intense thing we've ever done. Well, I've ever done, with a very small group of people.

I think having everything in one model was the start of this plan to create more than a coding agent, right? It was like we were like, “Oh, okay, now our coding agent is very good. It's amazing, and what would make it better?” It would be better if it could do more things than coding.

## final-20261015-0021 — Shipper, 5:35

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0021)

You mean right now?

## final-20261015-0022 — Glaese, 5:37

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0022)

That was, like—

## final-20261015-0023 — Shipper, 5:37

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0023)

Oh. That, that—

## final-20261015-0024 — Glaese, 5:38

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0024)

Basically with 5.3, right? It was in the 5.3 blog where we started to make the model really good at other types of things.

And we did that very deliberately because we were like, “How do we make Codex better?” Codex would be better if it could do more things, right? Just making it better at coding right now is not going to—

The ROI on that is we want to be better, but the biggest blocker or the biggest thing that would make Codex better now is if it could do more things.

## final-20261015-0025 — Shipper, 6:11

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0025)

Mm-hmm. But I still wanna push on the separate model line thing. You said sometimes it's nice to just be able to focus on one particular thing. But one of the AI truisms is the bitter lesson of just the bigger, the more general the model, always beats a more specialized one. So that seems like it would be a different— Help me understand that.

## final-20261015-0026 — Glaese, 6:35

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0026)

Yeah. I think actually the truth is somewhere in the middle, right? A specialized model is probably always going to be better at specialized thing. Right? If you only care about one thing.

But I agree, having everything together was always the ultimate goal. We were never planning to have this be a separate thing for forever.

I've been doing this type of getting models good at working with humans for a very long time, and it is a very complex process. Especially if there's many demands on the models that we have, and so having some focus is helpful.

## final-20261015-0027 — Shipper, 7:33

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0027)

Mm-hmm. Yeah. One of the things that Alexander said, and I'm curious how it fed into—

## final-20261015-0028 — Glaese, 7:40

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0028)

And now we have everything together. And it's been great.

## final-20261015-0029 — Shipper, 7:43

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0029)

We're getting there.

## final-20261015-0030 — Glaese, 7:45

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0030)

Yeah. 5.4 and 5.5 have been some of my most favorite launches that we've done.

## final-20261015-0031 — Shipper, 7:53

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0031)

They're great. I mean, I'm using the current alpha right now all the time, and it's fantastic. It's really good.

## final-20261015-0032 — Glaese, 7:58

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0032)

Oh, I'm so glad. Yeah. We are especially proud of the front-end work because that was a pain for so long.

## final-20261015-0033 — Shipper, 8:07

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0033)

Mm-hmm. Yeah. Okay, so, one of the things— 'Cause I was talking to Alexander about Codex Cloud. Which was a big launch at the time. There was a big—

## final-20261015-0034 — Glaese, 8:19

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0034)

Big launch, yeah.

## final-20261015-0035 — Shipper, 8:19

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0035)

Big bump, and then it didn't really work as a product. And one of the things he was saying is, what the design considerations were for Codex Cloud was talking to the research team, which I assumed was you or someone that you know.

## final-20261015-0036 — Glaese, 8:35

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0036)

Mm-hmm. Yeah.

## final-20261015-0037 — Shipper, 8:37

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0037)

And thinking about, okay, he said, “We're a little bit behind on coding right now, but what we know is that test time compute scaling is gonna work.” And so building a sort of asynchronous style experience was the right product shape for what we thought we could be good at, on the research side. Can you tell me about your perspective on that?

## final-20261015-0038 — Glaese, 9:02

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0038)

I think it's a bit different. But I actually agree with everything that Alexander is saying. But I think in the future, we will have agents that do work for long stretches of time for us, right?

We are already seeing that. I mean, I think you're in the channel where people are showing, “Oh, my model has been working for like 48 hours” or something like that, right?

## final-20261015-0039 — Shipper, 9:31

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0039)

Yeah.

## final-20261015-0040 — Glaese, 9:31

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0040)

And that's awesome. And that's the future I think that we were seeing at the time. And clearly, doing that on your laptop is not the most ideal thing.

And so I think we were kind of approaching it with this more future vision where we are like, yeah, in the future this is clearly how things are going to be. And we didn't do this, let's go to the people where they are. Which is on their laptop. And so that was something that we then had to adjust quickly.

## final-20261015-0041 — Shipper, 10:08

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0041)

Mm-hmm. Who's we?

## final-20261015-0042 — Glaese, 10:10

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0042)

We as a Codex team.

## final-20261015-0043 — Shipper, 10:12

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0043)

Okay. And then tell me what changed in the direction and maybe from the research side on the research direction, because my memory of, for example, 5.1 and 5.2 was at the time I would use Claude Code all the time, every day, and then if I had a really hard problem, I would ask 5.1 or 5.2, and it would be very slow, very powerful, like a little bit just very literal. So tell me about the shift—

## final-20261015-0044 — Glaese, 10:43

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0044)

Which 5.2 Codex or 5.2?

## final-20261015-0045 — Shipper, 10:45

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0045)

5.2 Codex. Yeah.

## final-20261015-0046 — Glaese, 10:47

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0046)

I think this is where maybe your bitter lesson comes in, right? We were very focused on sort of coding. And that's what you—

## final-20261015-0047 — Shipper, 10:59

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0047)

That's what we got. Yeah. You successfully replicated a senior engineer.

## final-20261015-0048 — Glaese, 11:05

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0048)

So yeah. To me actually, I think 5.2 was a turning point where I was like, “Wow, I'm very excited about what this can do.” And what it can do for me.

And it was this— It felt like it started to really accelerate a lot of people. And then 5.3 I think even more, but to me it was sort of like the curve started to be—

## final-20261015-0049 — Shipper, 11:37

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0049)

That's where you started. Okay.

## final-20261015-0050 — Shipper, 11:38

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0050)

And that was— 5.2 was when you were like, “We actually have to make this better, not just for engineering.” And—

## final-20261015-0051 — Glaese, 11:44

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0051)

Yeah. And so then after that we did 5.3 Codex, which was the first model where we put everything together and we're like, “Let's do”—

## final-20261015-0052 — Shipper, 11:52

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0052)

Oh, wait, tell me about that. Because when you say put everything together, what does that mean?

## final-20261015-0053 — Glaese, 11:58

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0053)

So 5.3 Codex was actually the first model where it wasn't just a coding— It's named Codex, but it was—

## final-20261015-0054 — Shipper, 12:06

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0054)

But you put other stuff in it.

## final-20261015-0055 — Glaese, 12:07

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0055)

We put all the other—

## final-20261015-0056 — Shipper, 12:10

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0056)

Yeah, yeah.

## final-20261015-0057 — Glaese, 12:10

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0057)

Knowledge work and that kind of— we started building the model for more than just coding agents.

## final-20261015-0058 — Shipper, 12:18

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0058)

But still called it Codex.

## final-20261015-0059 — Glaese, 12:20

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0059)

Yeah, we still called it Codex. I think we weren't ready. It was sort of like a step. We weren't ready for it to be like okay, is this a thing that is better at all things?

## final-20261015-0060 — Shipper, 12:33

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0060)

Yeah, yeah.

## final-20261015-0061 — Glaese, 12:34

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0061)

Than the models that we had.

## final-20261015-0062 — Shipper, 12:36

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0062)

Yeah. Well, okay, so one of the other truisms is whenever you train models on code, they get better at everything else. But what it sounds like you're saying is the model just got a lot better at coding or just being useful once you trained it on things other than coding. So is there something there that people miss?

## final-20261015-0063 — Glaese, 13:03

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0063)

I would say there's this thing where coding itself, code itself is so useful for everything, right? Code controls your computer. If you can write code, you can do all sorts of things with a computer, almost like all the things with a computer that a human can do with a computer.

And people don't just do software engineering with their computers, right? People use their computers for all sorts of work or even entertainment, and so you can use code to make music. That's very cool.

## final-20261015-0064 — Shipper, 13:43

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0064)

Yeah, yeah.

## final-20261015-0065 — Glaese, 13:44

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0065)

And so coding is really this foundation that we believe is going to be helpful, not just for software engineers but for all sorts of work that happens at a computer.

But it is a little bit separate from, say, knowing what a nice slide looks like. Right? So I can use code to actually create slides, but there's this extra step of, do I know what a good financial analysis is, and how can I present that for humans to consume.

## final-20261015-0066 — Shipper, 14:27

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0066)

What guides the research direction for the model? Is the product telling you, “Okay, here's our product strategy. Make a model for this”? Are you saying, “Here's what models can get good at. Make a product for this”? Are you talking about to users? How do you decide your research program?

## final-20261015-0067 — Glaese, 14:45

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0067)

My favorite thing about the way that we are working, the way that we also worked from the very beginning when we started working on Codex, is that it was very close together.

## final-20261015-0068 — Shipper, 15:01

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0068)

Research, product.

## final-20261015-0069 — Glaese, 15:02

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0069)

Research and product and product engineering. It's a small team.

## final-20261015-0070 — Shipper, 15:06

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0070)

How many people?

## final-20261015-0071 — Glaese, 15:09

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0071)

I mean, the very early versions were, like, handful. And then it got a bit bigger when we started to actually do sort of the external, when the cloud Codex launched, for example.

## final-20261015-0072 — Shipper, 15:24

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0072)

I see. Mm-hmm.

## final-20261015-0073 — Glaese, 15:25

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0073)

But at that time also, it was product engineers and researchers, they're just sitting next to each other.

## final-20261015-0074 — Shipper, 15:35

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0074)

As opposed to maybe culturally at AI labs previously, it's like researchers are in their own computer.

## final-20261015-0075 — Glaese, 15:40

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0075)

Yeah, you could be like here's your function. Research does their thing.

## final-20261015-0076 — Shipper, 15:44

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0076)

Yeah, yeah.

## final-20261015-0077 — Glaese, 15:44

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0077)

And then product looks at what there is and they build a product around it or something like that. And it's like, from the options that you gave me, I don't think either of them is true, right?

I care a lot about product and maybe I'm not a product  professional, right? But I can see what's possible with the models. I can be like, “Here's all the amazing things that we could build.” And also product people are like, “Wouldn't it be amazing if the model could do this thing?” And so we can sit together and we can share our vision and build something together.

And we have a lot of strength in that where we can sort of parallelize that work where we can think about what do we want Codex or our models to be able to do. And then we can build the model for that and we can build the product and the harness and everything that we need for that together. Because we are thinking about this as a joint thing that we build together.

## final-20261015-0078 — Shipper, 16:52

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0078)

That makes total sense. I think that's really interesting and unique, and the reason I care about it is, I think that kind of a team structure is really important but is under-discussed.

But what we talked about so far is when things work well together. But tell me about how you resolve conflicts or a specific thing where the incentives or the desires of someone who cares about research might conflict with either product or users or whatever and how you think about that.

## final-20261015-0079 — Glaese, 17:25

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0079)

I wouldn't say there was a lot of conflict. I think the team had this magical thing where people have a lot of mutual respect for each other's expertise. And so they're like, “Okay, I think we should be doing X, but you're the expert. What do you think?” You make the decision.

And we didn't have necessarily one person who told everyone what to do, but as a team, people were really good at sort of being like, “You're the person who knows the best how to do a thing. You decide how we are going to do that.” So I think that's one thing.

That doesn't mean that there isn't weird things that are quite different between model training and building a product. Where model training is a lot less deterministic, right? Every time we train a model, we go on this magical adventure. Where we're like, “How are we gonna do that?” And then we think about it. And then it's more like science, right? You have to do experiments. You try out a couple of things and then some of them work and some of them may not work. And then you kind of build on top of that.

Whereas on the product side, I think you can think about what do you want to design, what do you want to build, and maybe sometimes there are some surprises. But you can sort of deterministically predict like, okay, this is how we are going to do it. It's probably, with extremely high likelihood, that's possible, like we can do it. Maybe it takes two weeks longer if we are missing something, but, you know, we build this.

Whereas on the research side, we are just like, “Sorry.” Like, we tried, you know?

Shipper: [laughs]

## final-20261015-0080 — Shipper, 19:08

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0080)

Not possible.

## final-20261015-0081 — Glaese, 19:08

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0081)

Or like we got 50%, but it's not—

## final-20261015-0082 — Shipper, 19:13

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0082)

Yeah, yeah.

## final-20261015-0083 — Glaese, 19:13

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0083)

Quite there yet. And so the mutual empathy on these differences, and how both engineers and researchers are really building a thing. But the way that we build and the complexities of building these are different.

And this can be the other way around, too, right? Like researchers have never built like a giant, production-level piece of infrastructure. So it could be they're just like, "Can't you just do that?" And the engineers are like, "No."

And so that's where the trust and the respect kind of comes in, right? Like we have to have that and be like, "That makes sense. You're the expert," you know?

Shipper: [laughs]

## final-20261015-0084 — Shipper, 20:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0084)

Yeah, yeah.

## final-20261015-0085 — Glaese, 20:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0085)

Like, if you're like, "This is a hard engineering problem," I trust that. And if you are like, "Hey, from a research perspective, this is not going to work," or like, "We tried and something went wrong and we have to do another revision on this," that is not necessarily like a failure.

It's just like, that's part of the process that you have to experiment and some things work and some things don't quite work the way that you hoped.

## final-20261015-0086 — Shipper, 20:33

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0086)

What are the qualities of a good researcher?

## final-20261015-0087 — Glaese, 20:39

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0087)

There's many different kinds of researchers.

## final-20261015-0088 — Shipper, 20:41

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0088)

Maybe your kind of researcher.

## final-20261015-0089 — Glaese, 20:43

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0089)

It's a very good question. I think a lot of what we do is very evidence-based, and you have to approach a problem with this curiosity and understanding that this is an experiment, like things might go wrong. You kind of have to take risks also, right? Because that's part of the process.

If you're working closer to products and they are sort of this in sync with product engineering work, then being really well-calibrated on the risk, like what is the worst outcome here, what is the best outcome here, what should we bet on, making sure that we have a variety of, here's some things that we should bet on, and they are a bit more exploratory, and here's some things that we are certain they are going to make the experience of working with the model a lot better, and we should do those, too.

Sort of having that balance is really important. Everyone at OpenAI works really hard, and so that's part of it, too.

## final-20261015-0090 — Shipper, 22:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0090)

We were on your launch schedule, so I know.

## final-20261015-0091 — Glaese, 22:10

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0091)

Yeah. And then for people on my team, it's also really important that they really care about the end product of what the user experiences once they are using the model. Instead of just being on the side and thinking about numbers — it doesn't matter if the users are like, "I don't like this," right?

And so really being caring and being obsessed with the product and the users that we have is super important.

## final-20261015-0092 — Shipper, 22:44

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0092)

How do you think about the challenge of the things that you can quantify about the model that indicate that it's good versus just the vibe of it?

## final-20261015-0093 — Glaese, 22:55

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0093)

Yeah. It's very difficult. I think benchmarks matter.

## final-20261015-0094 — Shipper, 23:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0094)

Which ones?

## final-20261015-0095 — Glaese, 23:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0095)

Some of them. I think benchmarks matter. Not all benchmarks matter. Basically benchmarks, often when they are created for the first time and they measure something, they measure something very valuable, most of the time it's like, okay, models get like 5% or 30%, right? And at that point it's like, okay, there's like a very clear hill to climb.

And if you improve on that benchmark it's probably really, really, it measures something meaningful. Especially if it's corroborated with vibes, right? Like, I think just trusting vibes is not that useful in research. But, like, if you—

## final-20261015-0096 — Shipper, 23:47

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0096)

Why not?

## final-20261015-0097 — Glaese, 23:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0097)

It's hard to measure, right? Like, we have to do many repetitions, and sort of just me trying the model and be like, "I think the vibe is good" — I'm missing a lot of what other people might be doing with the model, right? Like, that's kind of like I'm not the only user for the model, right? And so we would be missing a lot of use cases and preferences and what people need if we just did that.

I think if benchmarks stop being correlated to vibes, it's probably there's something wrong. You can usually see it, like, okay, your benchmark hits like 80%. It's very hard to actually make benchmarks that have less than like 20% noise, right?

## final-20261015-0098 — Shipper, 24:38

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0098)

Yeah, I see.

## final-20261015-0099 — Glaese, 24:38

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0099)

So these kinds of things. So at that moment where it's like you stop really climbing, you're like, "Oh, it's like 81%."

## final-20261015-0100 — Shipper, 24:45

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0100)

It's not 100, yeah, yeah.

## final-20261015-0101 — Glaese, 24:45

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0101)

This is 81.5%. It's all within noise, statistically. But you also know like 20% of that benchmark problem is probably broken already. Like, at that point, the benchmark stops becoming... that doesn't mean it wasn't a good one. It actually served its purpose.

And we have to continue to build new benchmarks that capture what people, that capture the value that people might get. But then having evaluations and having benchmarks is very important for research because we do need to be able to measure our progress in not just vibes.

## final-20261015-0102 — Shipper, 25:25

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0102)

Yeah. Well, let's talk about the new model. I mean, I have one that I'm using now. And I assume we're projecting that'll be out for the embargo. And I'm happy to hold this until whenever, so we can talk about it. Is that okay?

## final-20261015-0103 — OpenAI, 25:44

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0103)

Yeah, we can chat about it. I think we'll still just determine what is the model still and share updates.

## final-20261015-0104 — Shipper, 25:50

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0104)

Yeah, yeah.

## final-20261015-0105 — OpenAI, 25:50

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0105)

But yeah.

## final-20261015-0106 — Shipper, 25:50

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0106)

But there will be a new model.

## final-20261015-0107 — OpenAI, 25:53

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0107)

We're testing, yes.

## final-20261015-0108 — Shipper, 25:53

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0108)

We've tested a couple different release candidates.

## final-20261015-0109 — Shipper, 25:56

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0109)

Yeah, I think we have a fairly good sense of—

## final-20261015-0110 — Glaese, 25:58

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0110)

Which was your favorite?

## final-20261015-0111 — Shipper, 26:00

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0111)

I think [redacted]'s the best. Yeah.

## final-20261015-0112 — Shipper, 26:24

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0112)

What were you trying to achieve with it? And how do you know that you've achieved it or whether you have?

## final-20261015-0113 — Glaese, 26:31

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0113)

Whenever we set out to build a new model, as a team we kind of decide what are the things that we really want this model to be better at. There's like unlimited things that you could kind of optimize for, and so we pick a couple of directions that we really want to focus on.

And some of them might be like things that— I talked about this before, right? Like a bit more like bets. That could work out, could not work out. And some of them are like things that we think as just like, it's like this is super important. It's table stakes. I think like front end for example. Right?

We're like very happy about this because we did set out for this model to make it significantly better at front end. And I'm very happy we succeeded. So there's like a number of these types of improvements that we pick, and then sometimes they continue, right? Like you never want to not be good at front end, right? That's our sort of—

## final-20261015-0114 — Shipper, 27:41

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0114)

Yeah, yeah.

## final-20261015-0115 — Glaese, 27:42

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0115)

Yeah. But we think about it as a team.

## final-20261015-0116 — Shipper, 27:47

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0116)

Are there any other dimensions that you care about for this model other than front end?

## final-20261015-0117 — Glaese, 27:52

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0117)

Yes. But did you find anything different other than front end?

## final-20261015-0118 — Shipper, 28:00

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0118)

Yeah. I think for me specifically, it has this combination of being really fast. And generally just like doing what I want for knowledge work tasks. So it writes all my emails. It's a very good writer actually.

In a way that I think like Claude models— what we talk about a lot is Claude models are a bit denser and more literary. And this is actually like one-shotting like marketing emails and stuff for us. So it's—

## final-20261015-0119 — Glaese, 28:34

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0119)

I think, yeah, some things that— I mean, you often see that in the blog post, right? So things like token efficiency is something that we care about because that is like how fast do you get your answer? Like a lot of users care about that.

We also care about the other side of things, right? Like how long can the model go? I've been having fun seeing people send me screenshot of like, "Ooh, it's been 147 hours."

## final-20261015-0120 — Shipper, 29:01

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0120)

Yeah, yeah, yeah.

## final-20261015-0121 — Glaese, 29:01

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0121)

And still going.

## final-20261015-0122 — Shipper, 29:02

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0122)

Yeah, yeah.

## final-20261015-0123 — Glaese, 29:02

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0123)

And I'm like, "Oh, that's great." Like it's working. And then we have this big moment of like merging— the bringing like the Codex sort of capabilities—

## final-20261015-0124 — Shipper, 29:16

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0124)

Oh, yeah.

## final-20261015-0125 — Glaese, 29:16

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0125)

to ChatGPT. And so generally like knowledge work and things that are important for other, like writing emails but also say like slide decks. Or making good sheets with correct formulas and these kinds of things are things that we paid a lot of attention to.

## final-20261015-0126 — Shipper, 29:34

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0126)

Is there something in particular that you're particularly proud of that people might underestimate how hard it was other than— I mean, the design one we talked about, but—

## final-20261015-0127 — Glaese, 29:44

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0127)

I think that one was hard.

## final-20261015-0128 — Shipper, 29:46

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0128)

Yeah.

## final-20261015-0129 — Glaese, 29:48

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0129)

That one is hard. It's like a lot of it is like taste, right? Taste is like we talked about benchmarks. Taste, super hard to measure. And so that makes it actually like a complex—

## final-20261015-0130 — Shipper, 30:02

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0130)

What is taste?

## final-20261015-0131 — Glaese, 30:03

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0131)

But yes. I know. I don't know.

## final-20261015-0132 — Shipper, 30:05

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0132)

Well, I assume you have to define it a little bit in order to hill climb it, right?

## final-20261015-0133 — Glaese, 30:11

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0133)

Yeah. We often work with experts, and that's why it's like amazing, right? Like the Codex team has Ed, he's our designer. He's great. And we can be like, "What do you think?"

## final-20261015-0134 — Shipper, 30:28

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0134)

Yeah, right.

## final-20261015-0135 — Glaese, 30:29

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0135)

And so I actually try to not be like this is what good taste is like, because I think— I think maybe I'm a little bit better than maybe the average researcher on taste. Maybe not. I don't know. It's hard to tell. But clearly I'm not Ed, right?

And I want the model to be like Ed. And so I have to talk with Ed to figure out what does good look like.

## final-20261015-0136 — Shipper, 30:57

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0136)

Yeah. Anything on your mind that we haven't covered?

## final-20261015-0137 — Katie Parrott, 31:01

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0137)

What is something about this model that surprised even you?

## final-20261015-0138 — Glaese, 31:07

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0138)

Hmm.

## final-20261015-0139 — Shipper, 31:09

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0139)

That's a great question.

## final-20261015-0140 — Glaese, 31:28

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0140)

I think this is such a hard question to answer because we're so close to it. Right? Like, it's not like we just produce... I mean, you've tried three models, right? But it's not like we just produce the models, like this black box thing, and then it appears and then I explore it. It does happen sometimes.

Something that I thought was magical, not necessarily with this model before, but the first time that we did steering, where you could talk to the model and it responds. And it was able to do a lot of things that we didn't actually train it to do. We thought like, "Oh, yeah, it should give updates and stuff, and you should be able to correct it." But you can actually ask it a question and it will respond. And the answer— and we're like, "Oh, that is, that is like quite cool."

But we are in general, we are so close to the model and we obsessively, every day, we look at how our model is doing today. So there's rarely things where I'm like, I'm—

## final-20261015-0141 — Shipper, 32:38

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0141)

Shocked.

## final-20261015-0142 — Glaese, 32:39

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0142)

Shocked. Yeah. I have had delightful experiences. Yeah.

## final-20261015-0143 — Shipper, 32:46

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0143)

So it's no longer a separate part of... So no longer a separate model line. What was it like to merge it? And do you expect to be doing more of these, like, kinda like break it off and improve it on a particular thing type moves?

## final-20261015-0144 — Glaese, 33:03

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0144)

Merging. It was crazy time.

Shipper: [laughs]

## final-20261015-0145 — Shipper, 33:13

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0145)

That's going in the piece for sure.

## final-20261015-0146 — Glaese, 33:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0146)

No. It was actually— I had a great time merging it, because people came together and we really planned out the thing, and then we went into it together and built the thing together, and I was very proud of how we did that. And how...

I mean, like, 5.3 Codex and 5.4 were sort of like the results of that, and I'm still very proud of what we did there.

## final-20261015-0147 — Shipper, 33:48

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0147)

That's great. And then do you think that the move of splitting off a model line is one that you'll continue to do?

## final-20261015-0148 — Glaese, 33:55

[Read this passage](https://every.to/p/how-codex-happened/interviews/mia/?edition=gift#final-20261015-0148)

It's not something that we are sort of... I mean, we have some— there's like GPT Rosalind and these somewhat separate models. We're not planning to do a new Codex separate line anytime soon.

---

# Andrew Ambrosino

Design & engineering

Final interview transcript supplied October 15, 2026. Source file F0C9K0V5ZGS; text, speaker labels and timestamps preserved from the supplied final transcript.

Passage IDs identify this edition's captured paragraphs. Bracketed transcription notes are not speech.

## final-20261015-0001 — Dan Shipper, 0:00

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0001)

I always like to check. Where do I even start? Okay. Katie actually just said something that I think is really interesting, and I would love your take, which is Anthropic just came out with this new model, and I still don't use it every day because it's not in the Codex app. And we have this—

## final-20261015-0002 — Andrew Ambrosino, 0:25

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0002)

Many such cases.

## final-20261015-0003 — Shipper, 0:28

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0003)

I think there's this underlying feeling that models are all that matter. And for sure, I guess, there are cases where a model jump is so big and so important that even if the harness sucks, you'd probably still use the model anyway. But harnesses seem to be becoming more and more important.

Tell me about that and your perspective as someone who's focused on the app.

## final-20261015-0004 — Ambrosino, 0:58

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0004)

Yes. Many such cases. I see this feedback all the time. And I think one of the great successes of Codex as a team was pairing the product team and the research team together very closely, especially when it came to the harness, right? So we do training on the harness now. We use the harness for far more of our products, including upcoming ones. The Workspace Agents products that we have out uses this. And then it's not just a harness, it's also how the app works with that, right? So we don't necessarily have the concept of, like, open the browser tab in the harness, but the harness and the app work together really well to do this.

There is a bit of a meme of model jumps stop mattering to you when they are smarter than you and you don't notice anymore. But I think it's sort of true and kind of funny. But the big deal, the big jumps that I've seen have been when it's easier to get more context to the model, especially automatically, which is app, it's connectors, it's computer use.

When the model can verify more of what it has built or what it's done, right? So when it can click around in a browser or on a computer, right? Like, that to me has been, depending on the task, bigger, equal or larger of a jump than model jumps. And this isn't to say that we're not working on new models. We always are. But I think that's why you're seeing some of the feedback on this.

## final-20261015-0005 — Shipper, 2:36

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0005)

Why do you think people don't understand that?

## final-20261015-0006 — Ambrosino, 2:41

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0006)

I think it's a more boring topic. It's less new. It's like what software has always been, right? It's like the same thing with the terminal on prior eras of this. It's like, oh, yes, you can do more with the terminal. You hear this all the time from developers and such, but there's a reason why operating systems and browsers and applications have been the way things get to mainstream, right?

## final-20261015-0007 — Shipper, 3:09

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0007)

Yeah. I was just talking to Tibo about this, so I'm curious your perspective. When the Codex desktop app came out, it was at a time when TUIs were so hot. Can you just take me into the room of deciding to make a desktop app for Codex?

## final-20261015-0008 — Ambrosino, 3:34

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0008)

Yes. It was a very small room. It was a big room, but small group.

Shipper: [laughs]

## final-20261015-0009 — Shipper, 3:40

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0009)

How many people were on that team?

## final-20261015-0010 — Ambrosino, 3:43

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0010)

We started with, like, three people.

## final-20261015-0011 — Shipper, 3:46

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0011)

Okay. And what was the breakdown of who did what?

## final-20261015-0012 — Ambrosino, 3:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0012)

We all did everything.

## final-20261015-0013 — Shipper, 3:50

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0013)

So—

## final-20261015-0014 — Ambrosino, 3:50

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0014)

I mean, we've heard exaggerated claims of role collapse in the industry. And I think, to extent they are a little bit overblown, but on the Codex team, role collapse has happened faster than the rest of the company. So designers are technical, and our engineers have product mind.

So we had, I think there were four people. If you looked at the original team, it was two engineers, one product manager, and one designer. But we all kind of did very similar work in terms to—

## final-20261015-0015 — Shipper, 4:21

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0015)

How about on the research side?

## final-20261015-0016 — Ambrosino, 4:22

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0016)

On the research side?

## final-20261015-0017 — Shipper, 4:24

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0017)

'Cause I thought there was research together with—

## final-20261015-0018 — Ambrosino, 4:26

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0018)

Yeah, I mean, we were piggybacking off of the same harness that the CLI used, right? And so there was active research, and if you remember, the time of app development was, I think, mid-November to when we released it in February. A very special time in coding models research, right? That's when everything seemed to hit this inflection point.

In fact, we often say that if we had finished the app to the quality it was in November and not in February, it would've been unreleasable because of the quality of the models at that point.

## final-20261015-0019 — Shipper, 5:08

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0019)

Hmm. Yeah.

## final-20261015-0020 — Ambrosino, 5:08

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0020)

It's like the models changed enough during—

## final-20261015-0021 — Shipper, 5:11

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0021)

Yeah, yeah.

## final-20261015-0022 — Ambrosino, 5:11

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0022)

this course of time that you could actually have an app that you weren't editing the code by hand.

So back to the inception, so there's a room with four of us or maybe a little more if you count Tibo and such. And we didn't really know what it was gonna be. There were a few examples out in the wild of lighter weight things that were not IDEs.

## final-20261015-0023 — Shipper, 5:34

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0023)

Like Conductor?

## final-20261015-0024 — Ambrosino, 5:35

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0024)

Like Conductor, yep. We know Charlie — that's a great team over there. They have done great work. We knew that we wanted a surface that could be a daily driver, but that was more lightweight than an IDE, and that we could surface things that were not typically done, like we ended up doing with our plugins marketplace and putting a browser in this thing.

And we kind of had the idea then that we would do this stuff, but we were trying to figure out, what does this look like? We spent a lot of time on the worktree model, we spent a lot of time on really basic things like how the sidebar's organized, how is your work organized?

'Cause it wasn't really — there was ChatGPT, there were the IDEs that had kinda started to dabble in multi-agents, but still, you look around the industry and everything looks like the Codex app now. But at the time, it was like all of these questions were still brand new. And I don't think we get it all right, which is funny to see people do the same thing that we did, even the things that we don't like. But it was interesting.

Along the way, we had a few moments. One of the moments was when Antigravity came out, and we just kinda looked at each other in the office and we were like, “Did we make a huge mistake by not forking VS Code and trying to do our own thing?” And we felt very conflicted on this, and I don't think the team even agreed with each other. And we said, “You know what? Let's sit on it.” And we decided to commit and keep doing it.

## final-20261015-0025 — Shipper, 7:16

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0025)

This is before it had come out?

## final-20261015-0026 — Shipper, 7:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0026)

Before Codex had come out?

## final-20261015-0027 — Ambrosino, 7:19

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0027)

Before Codex had come out.

## final-20261015-0028 — Shipper, 7:20

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0028)

And was it being used internally at that point, or a little bit before?

## final-20261015-0029 — Ambrosino, 7:22

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0029)

Somewhat. Yeah.

## final-20261015-0030 — Shipper, 7:23

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0030)

Okay. Yeah.

## final-20261015-0031 — Ambrosino, 7:24

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0031)

The dogfooding story, I think, is incredibly important. And I think people have pointed out that you can feel the dogfooding of the Codex app in the way that, some of the competition, you can kind of feel that they don't use it.

Our first mandate was that the team needed to get this into a place where it was our main thing day to day, right? And just the four of us.

Once we did that, every day Alexander would come in and be like, "Let's go around the table. Are you maining this app? Are you maining this app?" And the answer, it was a full nos at first. And then as the time went on, it was like, all right, one person is now maining the app. One person—

## final-20261015-0032 — Shipper, 8:07

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0032)

You call maining, like, I use it as my daily driver? Yeah.

## final-20261015-0033 — Ambrosino, 8:09

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0033)

I am not opening VS Code. I am—

## final-20261015-0034 — Ambrosino, 8:12

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0034)

'Cause you build it from nothing. It can't do anything, and then all of a sudden it hits.

And so once we were all yeses on maining the app, we released it to the company. And we said, "All right, if we can get PMF internally, then this is something we can release." And if we can't, then back to the drawing board.

And then we got there, and now I think, if we need numbers, we can set it. But I think now 90% of the company uses Codex. Not 90% of engineers. 90% of everybody.

## final-20261015-0035 — Shipper, 8:43

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0035)

Yeah, yeah.

## final-20261015-0036 — Shipper, 8:45

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0036)

And then when Antigravity came out and you were looking at each other, how many of you approximately on the team were maining the app?

## final-20261015-0037 — Ambrosino, 8:54

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0037)

At that time? It might have been half, but it wasn't all of us yet.

## final-20261015-0038 — Shipper, 9:01

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0038)

And what was the argument for, like, we should just maybe just go fork VS Code versus this? And how did you decide to let the current version win?

## final-20261015-0039 — Ambrosino, 9:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0039)

Anytime you try to stand up an alternative to an IDE or a browser or even an operating system, to an extent, there's always this moment of, wow, there's more stuff than we—

## final-20261015-0040 — Shipper, 9:32

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0040)

I see. Yeah.

## final-20261015-0041 — Ambrosino, 9:33

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0041)

I think the file editing piece—VS Code is very much like a arrange your own panes type of situation, where it's like, yeah, I need all these things open at the same time. And I need to customize the layouts and the workflows and all this.

I think that it wasn't clear at that point yet where the models would get to on what time horizon, and that's kind of when they started coming out monthly. And the thing that needed to happen was that most of the time you don't need to edit lines of code. And that you can steer them. But if the model's getting too much wrong or doing work that isn't up to your bar, then steering ends up being kinda slow, and you actually wanna dive back in and have, like, a tab model.

But we were able to get to this point on the research side where it crossed some threshold that we felt that, yes, I do have to steer this a lot, but I don't have to edit it by hand. That was a big moment. And I don't remember the exact sequence there, but that was kind of the—the tab models have gotten really good.

And so there was some cases early on that it was actually faster to be in the file, start typing and then tab, tab, tab. Right? And notably, we've never had a tab model. So.

## final-20261015-0042 — Shipper, 10:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0042)

Hmm. Interesting. You said that it wouldn't have succeeded if you had launched the same app in November when you first conceptualized it, which means that you sort of had to make a bet about model progress in order to do it. And you also had to convince other people that that bet was good before you could show them something that was working.

How did that work? How did that happen? And why did you make the bet?

## final-20261015-0043 — Ambrosino, 11:24

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0043)

I think Codex has been a very good model at the company for how closely we work with research. And I think that closeness gives us a bit more sense of where things are going and which speed. I think it was still a bet.

I think at this time, you sort of saw two form factors out in the wild. You saw this style application on the web which was what kind of was called vibe coding at the time, which was not production grade stuff. It was like, you wanna make a toy thing that you put on the internet, that's fine. You've got chat over here, you've got files over here. There are several versions of this.

And so there was precedent for this type of vibe coding. It just wasn't quite at production grade. It wasn't even at the grade of a code base that we wanted for this app, and that was the thing that changed.

So it's not like overnight, in this period, we went from not being able to code to being able to code. It was more of a quality bar. That in that time, we got to something where we could basically vibe code at the company and still meet the bar that we needed.

## final-20261015-0044 — Shipper, 12:33

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0044)

The idea I have of the Codex team and how this all happened is there is a small intrepid band of product people and researchers who got together and were able to split off from the big behemoth 800 million user ChatGPT and play around with a separate model and a separate product until you found something that was pretty sick.

And then that sort of is, in some ways—and actually Greg said this earlier—it's like you're almost disrupting ChatGPT yourself. What are the generalizable best practices for when that should happen inside of a company? When do you need to split out like that and have that sort of playground? 'Cause it's a very different kind of company strategy than you would expect, like, 10 or 15 years ago. People didn't really do this that much.

## final-20261015-0045 — Ambrosino, 13:20

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0045)

Yeah. And it's interesting because you do actually hear about startups within the big company—

## final-20261015-0046 — Shipper, 13:27

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0046)

Yeah, but they never work.

## final-20261015-0047 — Ambrosino, 13:28

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0047)

And they don't work. Right? And I think that I hope that we keep doing this, by the way. I hope that at some point this team feels disrupted in a way that's good for progress.

I don't know if I have a generalized hypothesis of when you do it and how you do it to be successful. I know that this one, I don't think that we would be here without having done this. And I assume there will be future versions of it.

## final-20261015-0048 — Shipper, 13:58

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0048)

What makes it different though than the bajillion big companies that are like, "Yeah, we're gonna disrupt ourselves. We're gonna make a product in a little team that does stuff," and it just never works?

## final-20261015-0049 — Ambrosino, 14:10

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0049)

One of the special things about OpenAI that I think is distinct even from the other labs is how bottoms-up the culture is. And so some days it is very frustrating to me because I will be like, "Hey, why are there 16 different efforts that are doing the [redacted] thing," right? Can we focus a little bit? It sometimes is a very unfocused situation we're in.

But the good thing about that is that there's a whole lot of agency to prototype things. And I think that that's where there is complete freedom to prototype. There is complete freedom to put things out into the company to see what sticks. And if you can do that while also collaborating with research I think that it puts these little spin-out processes on steroids.

Codex started before it was a formally different organization too. So it was a little thing that gained traction. It was actually two different teams that joined forces at some point and—

## final-20261015-0050 — Shipper, 15:14

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0050)

What were the two teams?

## final-20261015-0051 — Ambrosino, 15:15

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0051)

I don't know. It predates me. I also don't know if we share that, but I think we will see this happen again, where it's like, "Hey, there's this prototype," and then people start to use it, and then they join up with research to get some weight behind it on the model training. And all of a sudden it's taking off, right?

## final-20261015-0052 — Shipper, 15:38

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0052)

Yeah. It seems like the benefit of this strategy is because the landscape's changing all the time, you've got all these explorations happening, and something will randomly go right, and then it gets more oxygen in a way that if you're at the top of the org, you're not gonna be able to figure that out yourself. You need the sort of bottoms-up motion.

But the downside is it's fucking messy. Especially towards when something starts to work and you're like, "Oh my God, I've got Codex and I've got ChatGPT." Tell me about that.

## final-20261015-0053 — Ambrosino, 16:09

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0053)

It is messy, and it's kind of one of those things too where you love it when you're the one prototyping and then, if you're running a product, you're like, "Why do we do all of this? Everybody should just listen to what I wanna do."

Shipper: [laughs]

## final-20261015-0054 — Ambrosino, 16:25

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0054)

No, but I mean, I think when it's bad, this version at its worst is the version of it where convergence is not a competency, when it's just chaos, and then nobody knows how to bring this to the world. And I think that that's what we want to avoid here, right?

There are all sorts of things that we have added to the Codex app that started as experiments that nobody on my team told anybody else to do. Like, computer use was one of these things where it's like, Ari's team is doing computer use. They're really good at this thing, and then one day they've got a prototype to work with the Codex app. Like, you've got... What else do we got that's like—

## final-20261015-0055 — OpenAI, 17:20

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0055)

Pets is a good example.

## final-20261015-0056 — Ambrosino, 17:21

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0056)

Pets. Yeah, Pets started as this... And I wasn't totally convinced on Pets, and I was wrong about this thing.

## final-20261015-0057 — Shipper, 17:28

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0057)

Is Pets working?

## final-20261015-0058 — Ambrosino, 17:30

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0058)

I mean, people love it.

## final-20261015-0059 — Shipper, 17:31

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0059)

Really?

## final-20261015-0060 — OpenAI, 17:31

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0060)

It was interesting 'cause it started as someone's onboarding project. And then decided, oh, it blew up internally. Everyone started making pets, and then once you could customize to it, it just further exploded. And everyone's like, "Oh." So then we have launched it, and I think we're at over 100,000 in the first... In the first two weeks of launch it was over 100K.

## final-20261015-0061 — Shipper, 17:52

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0061)

Yeah. What are the other Pets, are there new Pets metrics that I should know about? 'Cause I've never made a pet, to be honest.

## final-20261015-0062 — OpenAI, 17:58

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0062)

I mean, it's fun.

## final-20261015-0063 — Ambrosino, 18:00

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0063)

I don't think the company publishes top-line Pets numbers to the shareholders.

## final-20261015-0064 — Shipper, 18:05

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0064)

[laughs]

## final-20261015-0065 — OpenAI, 18:05

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0065)

We have approval for sharing Pets numbers. At least that 100K one I did get 'cause I thought that was just so—

## final-20261015-0066 — Shipper, 18:12

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0066)

Wild.

## final-20261015-0067 — Shipper, 18:25

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0067)

It shows that people still want something that's familiar.

## final-20261015-0068 — Shipper, 18:30

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0068)

It's the same OpenClaw kind of energy too. We loved our claws and now they're all dead, but you know.

## final-20261015-0069 — OpenAI, 18:38

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0069)

Go grab Peter [Steinberger] if you need him.

## final-20261015-0070 — Shipper, 18:40

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0070)

I would love to talk to him about this at some point. That'd be really interesting, yeah. Katie, any questions on your mind?

## final-20261015-0071 — Shipper, 18:50

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0071)

She's a rabid Codex user.

## final-20261015-0072 — Katie Parrott, 18:59

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0072)

I think the question, and I asked this to Tibo as well, is just why do I like Codex? What did you do that made someone like me who's not technical, English, German literature double major over here, why do I like Codex so much?

Ambrosino: [laughs]

## final-20261015-0073 — Ambrosino, 19:20

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0073)

Why do you like Codex so much?

## final-20261015-0074 — Parrott, 19:22

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0074)

I love how it uses my context. I love that it always know... I set up my entire desktop to be like Codex friendly, so it has like my context folder and my voice MD for writing like me and all this stuff, but it did all that for me.

I basically took a transcript of a call that I had with Dan and dropped it into Codex, and it set up my entire environment for me to be really user friendly. And every time somebody asks me, “How do you use Codex?” I'm like, “Tell Codex how you wanna use it and it'll set itself up for you.”

So that's kind of the kernel of, I think, what I respond best to. It's just 'cause I don't... I skipped the whole no code, low code, no code era of Airtable and stuff because I did not wanna click around and—

## final-20261015-0075 — Ambrosino, 20:14

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0075)

Correct.

## final-20261015-0076 — Parrott, 20:14

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0076)

—connect things, and just being able to speak or type—but I mostly Monologue. Just being able to say I want this done and it gets done is just really compelling to me. And I just am so curious about, from a product design experience, how did that come to be?

## final-20261015-0077 — Ambrosino, 20:33

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0077)

Yeah. Well, I think one thing that's central to the current feel of Codex is that you can put together worlds in a way that with the chatbot era was a little bit harder, right? So you can say, “Hey, I want to have this, this—what do they call it? Mind palace of thousands of documents, and I want the memory organized like this, and the guidance for how I speak and write over here, and I'm gonna make these automations that loop over this set and loop over this context and then do this thing,” right?

And I don't think that that's for everybody, but I think what that does in our minds is that is a... It's gamified, it's fun, it gives people control, and it lets us all—not to be too grand about this—but as a society figure out some of the patterns that work that I think then we want to bring to things like the ChatGPT experience for everybody.

I think there was a while where ChatGPT has this library and it's got a certain way that memory works, but it's very much it is baked, right? It's something that we bring to a billion people.

And I think the thing that people really love about Codex even when they're not writing code is you can put it together yourself in ways that are very experimental and kind of speak to something about how we... To me and to many people here, Codex reminds us of the tools we had growing up that we feel like kind of got sanitized as the years went on and—

## final-20261015-0078 — Shipper, 22:12

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0078)

Like what?

## final-20261015-0079 — Ambrosino, 22:12

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0079)

Like Lego Mindstorms, for example. I really liked Legos and I also really liked computers, and this was a combination where it was like, you build something, you put together the blocks.

## final-20261015-0080 — Shipper, 22:23

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0080)

That's sick.

## final-20261015-0081 — Ambrosino, 22:23

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0081)

There's so many stories.

## final-20261015-0082 — Shipper, 22:25

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0082)

What was your favorite Mindstorm?

## final-20261015-0083 — Ambrosino, 22:27

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0083)

I don't remember.

## final-20261015-0084 — Shipper, 22:28

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0084)

What was your favorite Mindstorm build?

## final-20261015-0085 — Ambrosino, 22:29

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0085)

When I was in the first grade I built a monorail for a science project that I programmed to have little... the rail would have electrical tape around it, and then the sensor would read it and stop and go to the next thing.

## final-20261015-0086 — Shipper, 22:44

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0086)

Oh, sick.

## final-20261015-0087 — Ambrosino, 22:45

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0087)

Yeah. But it feels—the Codex feels like that, and it feels like that to so many of us, even the— you know, dev—

## final-20261015-0088 — Shipper, 22:53

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0088)

You're the first person to say that. That's really... But it's so true, but it's weird because I wouldn't have... It's why I love it and I don't necessarily put it in that category because you're very good at making it very simple to start. Tell me about that.

## final-20261015-0089 — Ambrosino, 23:08

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0089)

I think this is the eternal push and pull about the text box product shape.

## final-20261015-0090 — Shipper, 23:16

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0090)

Yeah. Yeah.

## final-20261015-0091 — Ambrosino, 23:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0091)

And it's one that we are working on differently every week. We knew we wanted something simple. We knew that we didn't wanna lean too far into needing to know technical details, that this should feel like it draws a lot of inspiration from ChatGPT. More inspiration from chat than it does VS Code, which given the audience is a sort of an interesting choice.

But we made it denser. We made it not quite as simple, but we still wanted this shape that felt like you could morph it into anything.

I think that this is especially the case when you start to use the plugins, especially the role specific plugins, right? Where I'm doing financial modeling, or I'm doing sales and wanna keep notes on all my calls. The plugins start to change the way that the software feels for every person.

I think the challenge here is the empty box problem. And I think we need to do a better job here, that the software is infinitely customizable, but not everybody has that time and appetite and, just free hours, to explore what they want it to be. They wanna say, “So and so set up theirs to do this,” and, I just want that.

## final-20261015-0092 — Parrott, 24:40

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0092)

One of the first things that I did, once I saw the light and I wanted to spread the gospel to the rest of the editorial team, was I just had Codex create a Proof doc with a setup guide that I could send to our other writer that would help her go through her process. But not my process—her process. Have it interview you about what your role is and what you need to have, and I just had Codex do it for her.

But it is interesting where there seems like a slight difference between her and me, in terms of I was willing to just go in there and really kind of tinker myself, and she's a little bit more—needed a little bit more guidance, a little bit more possibility. So I'm curious how you think about, as you sort of expand this out to Codex for work for everyone, that customer education piece of the puzzle.

## final-20261015-0093 — Ambrosino, 25:33

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0093)

I've long been somebody who is skeptical about product education as a solution. I had a startup for a number of years. It was in finance and insurance for independent contractors, and people would come in and say, "You guys just need more education about how 1099s work and how the health insurance marketplace works and all this stuff." And our answer was always kind of like, "Actually, these systems are so complicated that we really want to abstract some of that instead." It's not about tool tips, it's not about guides. There's some amount of that that matters, but we really wanna figure out the end product story.

One example that I really like to give is the deep research product on ChatGPT, which has been pretty popular and gone through a few iterations. And it was framed as like a top-level feature or even a product sometimes, right? In many ways, the Codex harness is a generic version of the thing that powered deep research. We don't go out of our way to say there's a deep research setting in Codex that you can do, but if you prompt it the right way and you use /goal or whatever, it'll kick off and it'll do just as thorough, if not more, than deep research.

But what's interesting about this new generic version is that companies may have specific ways that they want artifacts written. They may have certain sources that they rely on, especially if you look at something like investment banks, right? They may say, "Hey, our reports need to look exactly like this in this format, and these are the proprietary connectors that we have in the app," right? How do you basically make that look like a turnkey feature for a certain user group inside of your company?

So it's not a generic deep research feature for everybody. It's also not everybody's gotta know all the skills that they need to tag in the thing and exactly how to use it. It's, "Hey, I'm on this team, and I have this capability that's now just turnkey for me."

## final-20261015-0094 — Parrott, 27:35

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0094)

And that's just baked into the product, and you don't need people like me who are content marketers anymore.

## final-20261015-0095 — Ambrosino, 27:41

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0095)

No. It's that you can do infinitely more. And people don't come to you for incredibly basic things, right? They come to you for things that you are uniquely able to provide value on.

Parrott: [laughs]

## final-20261015-0096 — OpenAI, 27:53

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0096)

Codex feels like the first tool where, as a non-technical person, I'm being acknowledged that my workflow is complex. And that it's personalized, and I'm gonna want different things, and that's okay.

## final-20261015-0097 — Shipper, 28:31

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0097)

Who are the tabs for?

## final-20261015-0098 — Ambrosino, 28:34

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0098)

So in this release, we are trying to bridge a few different worlds, right? We're trying to bring Codex to the cloud in a meaningful way, right? So that you can say, "My laptop doesn't need to be open."

## final-20261015-0099 — Shipper, 28:48

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0099)

You're doing the hoist.

## final-20261015-0100 — Ambrosino, 28:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0100)

Right. Doing the grand hoist. Did somebody call it [that]?

## final-20261015-0101 — OpenAI, 28:52

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0101)

Alexander. Yeah.

## final-20261015-0102 — Ambrosino, 28:54

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0102)

[laughs]

## final-20261015-0103 — Ambrosino, 28:55

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0103)

I think I gave him that name.

## final-20261015-0104 — Shipper, 29:00

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0104)

I'll credit you.

## final-20261015-0105 — Ambrosino, 29:03

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0105)

Yeah. We call it the grand hoist. And we're gonna have a lot more very interesting things here with all of Codex. But that's one world that we're trying to bridge.

Another world that we're trying to join in this vision is all of the ChatGPT context that you've built up over the years. Bring that into the desktop experience that has worked so well, right? We've got a ChatGPT app. It's separate. It hasn't gotten a lot of love. We wanna bring it into here. All of your memory, all of your projects, all of your research reports, and let you go wild in Codex with that.

Then we also have our Codex remote control, which is another world that we're bringing into this. So we're bringing these things together. How these evolve over time together will change, right? First version will look different than version two, three, four. And they will likely converge over time even further than they do now. But we're pretty happy with this.

## final-20261015-0106 — Parrott, 30:00

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0106)

What are you most nervous about with the great merge?

## final-20261015-0107 — Ambrosino, 30:05

[Read this passage](https://every.to/p/how-codex-happened/interviews/andrew/?edition=gift#final-20261015-0107)

One of the things that I've always been nervous about here, just in terms of execution on products, is that ChatGPT serves a billion consumers and also a smaller amount of prosumers, and then also enterprise. And that it is quite difficult to build a cohesive product and even a cohesive brand around that whole spectrum. And that's something that we're gonna have to figure out.

I think Codex has gotten off pretty easy with its focus. Maybe that's a template, maybe it's not. But that it's very difficult to build an everything product.

---

# Ed Bayes

Design

Final interview transcript supplied October 15, 2026. Source file F0C9K0V5ZGS; text, speaker labels and timestamps preserved from the supplied final transcript.

Passage IDs identify this edition's captured paragraphs. Bracketed transcription notes are not speech.

## final-20261015-0001 — Ed Bayes, 7:20

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0001)

I kinda came in, always worked on research projects, and then worked on chat, and then kinda worked on this, and I think you mentioned that it's been like priority since 2025. I mean, the original Codex model, which was a coding model, came out in 2021, so it's always been this, always been there. And I think it was just very clear, early on, kind of anyone that's worked on it, how powerful this stuff is, right? Like if you used the early Codex, even pre-GPT-5 or any of the Claude Code models, you could kind of see where it's going. So I think it's more of first trying some of these new models or seeing some of these new capabilities, you're just like, "Gosh, this is gonna change everything." So it's less about some kind of corporate event and more, for example, with the merge, we kind of announced the merge internally and what we were gonna do and bring these things together. And I just remember coming out of it speaking to a few people just thinking, yeah, it's just crazy that we've gone from this tiny group of a few people in the corner, most of the company obviously paying a lot of attention from the research side but from the product, chat, consumer product, very much focused on different ends, and here we had this kind of coding product which is kind of enterprise-y. So from a product and design perspective, it was very much just kind of like, oh, this team in the corner doing their own thing to kind of coming out these meetings of, okay, this is now the future of ChatGPT, basically. And that was just quite astounding. I've just certainly never been part of a team that has had such a trajectory of users and the kind of the power of the product.

## final-20261015-0002 — Dan Shipper, 9:16

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0002)

Do you have a specific memory of that specific meeting of like how that happened? Like now this is ChatGPT.

## final-20261015-0003 — Bayes, 9:24

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0003)

Yeah. I mean, I don't know if there were like specific—I—the one I'm thinking of is just an all hands basically, where Tibo or someone's basically kinda saying, now we're gonna bring the power of Codex to chat and this is the rough shape that it's gonna take and I had some early designs I kind of set up as—and presented. So it was just a team all hands basically coming away, and it was very much like, congrats to the team. We've done so well that now it's gonna bring it to the kind of billion user space.

## final-20261015-0004 — Shipper, 10:02

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0004)

One of the things I'm trying to understand is like, this doesn't—this just doesn't happen. You don't like start a small team inside of a big company and then end up disrupting the main 800 million or billion user product. People have tried for many years.

## final-20261015-0005 — Bayes, 11:13

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0005)

The way that OpenAI is structured is it's a large company but it's still structured with very small teams building products, and when you're on one of those teams, it feels like you're part of a scrappy startup. You have these pods, you have a PM, a designer and small team of engineers, a research lead, maybe a few researchers. And, these projects really start very small, like five to 15 people working on very ambitious but quite discrete projects, and that's kind of how a lot of the company has worked. It's this kind of like a constellation of these groups. People talk about coding agents being this new thing. I mean ChatGPT is an agent, right? It just has a different harness so it's not two completely separate products. So it's realizing the original promise so it's more of an evolution.

## final-20261015-0006 — Shipper, 14:14

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0006)

When did you join the Codex team?

## final-20261015-0007 — Bayes, 14:18

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0007)

After we launched Codex Cloud. I was asked "Okay, can you build this like CLI thing, and then we're gonna do this IDE extension?" And I was like, "Yeah, cool. Pretty chill." [chuckles] And then we launched the IDE extension with GPT-5. We started to have this steep curve, and you look back now and it's like tiny, but at the time we were like, "Wow, so many users!" Then we decided to do the desktop app and built that out with a very small team and then we launched that earlier this year and then it's now grown into the merge with ChatGPT.

## final-20261015-0008 — Shipper, 15:46

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0008)

What are some of the like design challenges—I think the Merge is a really interesting one that I'd like to get your take on in a second, and I've talked to Andrew so much about tabs and just whatever. But before we get there, what are some of the design challenges for the desktop app that are like maybe unintuitive, like that you ran into?

## final-20261015-0009 — Bayes, 16:07

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0009)

The original desktop app was very opinionated and the thesis that I discussed with Alexander a lot was this is a command center for your agents. So as a result we optimized for being able to run multiple agents at once. This might sound like a small design detail, but at the time other similar products had two lines in the left sidebar. They had a bunch of icons. It was very noisy. I very intentionally flattened it into a single line, very clean, very similar to chat in that sense, so that you can basically work on double the number of tasks at the time. And also orienting it around projects, which is I think another pattern that others have now moved towards. Previously if you're in an IDE, you have to open a new window to work in a different workspace. Here you can work on front end, back end next to each other. So you can kick off 10 tasks and work in parallel. The challenge was, we're building a developer tool and developers want all the bells and whistles, so it's this healthy tension of trying to simplify things and delete things and make it very, very simple. And then just really obsessing over a few core flows for developers, thinking through the entire software development lifecycle from coming up with an idea to creating a PR, to code review, to merging it, and building a bunch of things like worktrees around that.

## final-20261015-0010 — Shipper, 20:00

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0010)

Tell me about the merge.

## final-20261015-0011 — Bayes, 20:04

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0011)

Yeah. What about it? [chuckles]

## final-20261015-0012 — Shipper, 20:07

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0012)

I mean, it just seems like the gnarliest thing ever from a design— I mean, from a lot of perspectives, but in particular from a design perspective.

## final-20261015-0013 — Bayes, 20:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0013)

There are a few really interesting design problems here and ideally you'd abstract all of these away from users, particularly non-developers. So first you have ChatGPT. You have all of these cloud-based tasks in a ChatGPT harness, and then you have Codex, which is this coding agent that works on your computer and it works with local files. So first you have this kind of incongruity between local and cloud. Then you have these different types of tasks which do different things. One of them has access to a computer, either on your computer or in the cloud, and another doesn't have that. So we're working with a lot of these constraints. There's a clear path to reducing the constraints or making it a little simpler to users. But until then, you have this like messy middle. You also have an innovator's dilemma type challenge where you have this existing product that people have a very clear expectation for, and then we have this new, powerful kind of way of using it, and you have to teach users how to use it. So they're the kind of constraints and you have to work around them. But you can expect over the next few months some of those become simpler.

## final-20261015-0014 — Shipper, 23:15

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0014)

That's interesting. Do you have any— Do you have a specific memory of like a specific design challenge that you were working on where you're, you figured something out that you think is unintuitive or underappreciated?

## final-20261015-0015 — Bayes, 23:29

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0015)

I guess it's more of a like high level principle that I generally believe in, which is with ChatGPT we have this universal interface for asking questions and with Codex we have a universal interface for getting work done. Until now we have these bifurcated products where you have a developer tool called Codex, and then you have this general product called ChatGPT. But if you're a PhD scientist studying rocket science, you don't use a different ChatGPT to a grandmother in Michigan who's learning how to knit, right? And I think we'll get there with Codex in ChatGPT too. Maybe that's slightly unintuitive in that we've got to this space and developers are very much like, "this is for developers." But I do think it benefits everyone if you simplify. And we saw that with some of the examples that I gave of like simplifying the left sidebar and moving things around projects. If you simplify products, that generally benefits all users, even power users.

## final-20261015-0016 — Shipper, 25:22

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0016)

What do you think of Tibo and what makes Tibo effective at what he does?

## final-20261015-0017 — Bayes, 25:30

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0017)

He's very good at getting things done. He straddles research and product, and that's really rare. His deep understanding of the research org and the research program, and bringing that in line with product. At a frontier lab, the best thing you can do if you're building product is thinking about where we're going to be in six months and building for that world. And that's basically the bet we made with the Codex app as it only really made sense once you had this generation of models that came out in December. So not only is he just very effective at getting things done but also lives in that world where you kind of look forward a bit and design for that. He's just a great guy as well. I think people really like working with him, and that helps as well when you're a leader and you're trying to bring people along.

## final-20261015-0018 — Shipper, 27:13

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0018)

Anything else that we haven't talked about you think we should before we end?

## final-20261015-0019 — Bayes, 27:19

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0019)

I don't think so. I'm curious your reflections on the journey over the past year.

## final-20261015-0020 — Shipper, 27:49

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0020)

I mean, I think that it's, like I said, it's like super rare to fall behind and then come back by disrupting yourself. It's just like super rare. And right now, I think of you guys as being the one to beat again. And it feels like Anthropic made a big bet on a gigantic model that is, in my view, more powerful than 5.6 but is very hard to use. And you guys have an incredibly well post-trained model that is in the best harness and that is actually better than the super ultra intelligence that can't talk to you. Which is—that's an unintuitive thing. And not only that, but the super ultra intelligence, it seems like is making the rest of their models worse because they're using it to make those models, and those models now have the same personality. So—I wouldn't have guessed that. Yeah. I also think that—you guys had 4.5, which was the big model that didn't really work that well, and I don't think Anthropic had that. I think they just got into that, some of those difficulties.

## final-20261015-0021 — Bayes, 29:17

[Read this passage](https://every.to/p/how-codex-happened/interviews/ed/?edition=gift#final-20261015-0021)

Hmm. Interesting. Yeah. See, yeah, maybe one other reflection I think, and this is actually something that Lindsay once said to me, which is I think one thing that she always is very impressed by with the Codex team was it was very AGI-pilled. And a lot of that came from leadership of Tibo. And I think the very unique thing as well is the team was just like very uniformly very strong, very founder-led, very AGI-pilled. It's quite rare to have a small team who are so aligned and similarly pointing in the same direction and philosophically aligned as well.