We use analytics and advertising tools by default. You can update this anytime.
Manage optional tracking categories. Necessary cookies stay on so the site can function.
Staff Writer & AI Editorial Lead
Katie Parrott is a staff writer. She writes Working Overtime and contributes to Vibe Checks, Source Code, and Context Window.
I had a quarter of OKRs and no clue whether I was hitting them. Enter Codex.
Kieran Klaassen turned a prompt into a working app in an hour and shows you how
Plus: Vercel and Lovable’s security woes, and how to make your agent your watchdog
The asynchronous, agentic workflow developers love is finally accessible to everyone—but the polish isn't there yet
Anthropic's latest Opus is more precise, more literal, and the best coding model we've tested on well-specified tasks—but it won't fill in the gaps for you anymore
OpenAI nailed the interface. But it's built for hardcore engineering.
A mini-Vibe Check on Gas City, a Grok classifier that grades your X drafts, and why HTML is the new markdown
Plus: A manual for handing recurring work to Opus, a skill for judging new tools in context, and the specialist model that beat the frontier for less
Plus: Agent-designed automations, why final review belongs in the destination app, and how to use our compound knowledge plugin
Sonnet 4.6 delivers Opus-close performance at half the price—but speed didn't come along for the ride
Plus: The Vatican weighs in on AI labor, and our Codex playbook
Plus: Perplexity’s rules for agent skills, the office politics of dictation, and creating a weekend AI piano coach
Plus: Dan’s attempt to clone Kate, a shortcut for turning demonstrations into skills, and the human goals machines still need us to set
Opus 4.8 tops both our Senior Engineer benchmark and our writing tests. It’s the most complete model we’ve tested. We just wish it had an app to match.
OpenAI’s new model is a top-end senior engineer—and easy to talk to
GPT-5.4 is fast, opinionated, and good enough to tempt our Opus loyalist
Anthropic’s new model can draft, code, and analyze competently, but every use case has a cheaper, faster, or smarter alternative
The AI-native IDE is now becoming an agent-orchestration tool. Will it work?
A power-user’s guide to turning OpenAI’s coding agent into an operating system for knowledge work, including setup, workflows, and a seven-day starter plan
It one-shotted a problem other models missed—and brings agentic, parallel work to non-coding tasks
.png&w=3840&q=75)
.png&w=3840&q=75)
.4_OpenAI_s_Comeback_Kid_Plans_Like_a_Product_Builder.png&w=3840&q=75)
.png&w=3840&q=75)
Katie Parrott
Dan Shipper