Dan Shipper is the CEO and cofounder of Every. Every week he explores the frontiers of AI in his column, Chain of Thought, and on his podcast, ‘AI & I.’
Artificial Unintelligence
Returning risky results is a dimension of performance for AI models
Dec 10, 2022 · 7 min readUpdated Jul 7, 2026
Sponsored By: Reflect
This article is brought to you by Reflect, a beautifully designed note-taking app that helps you to keep track of everything, from meeting notes to Kindle highlights.
If you want to build a sustaining advantage in AI the conventional wisdom says you have to build the technology required for a powerful model. But a powerful model is not just a function of technology. It’s also a function of your willingness to get sued.
We’re already at a point in the development of AI where its limitations are not always about what the technology is capable of. Instead, limits are self-imposed as a way to mitigate business (and societal) risk.
We should be talking more about that when we think about where sustainable advantages will accrue and to whom.
. . .
ChatGPT is a great example. It is awesomely powerful, but it’s also profoundly limited. It’s limitations though, are not mostly technological. They’re intentional.
On the awesome side, it saved me ten hours of programming for a project I used it on this weekend. But for other use cases it completely fails:
Thanks to our Sponsor: Reflect
Think faster and clearer with Reflect, a beautifully designed note-taking app. Reflect is built so that you never lose track of all notes, meetings, kindle highlights, calendar, web snippets and more.
Create a free account, or log in.
Every members live and work at the edge of AI. Join now.
By continuing, you agree to the Terms of Sale, Terms of Service, and Privacy Policy.
Enjoy unlimited access to all of Every.
See subscription options