Hey, I'm Jeff.
Co-host of What the AI?! podcast. I write about AI, technology, and the choices we're making with both. Previously: Upstart (founding team), Google. Subscribe on Substack to get new posts by email.
Latest Posts
-
Stop Telling AI What You Want — Show It
Most advice about working with AI boils down to some version of “be specific about what you want.” Write a better prompt. Describe the output in detail. Give clear instructions. That’s fine — but I’ve found there’s a much more powerful move that most people skip entirely.
Show it what good looks like.
-
Why I'm an AI Optimist
I spend most of my time these days reading, thinking, writing, and podcasting about AI. And I’m optimistic about where it’s headed — substantively optimistic. Not in the “everything will be fine, don’t worry” sense, but in the “history gives us strong reasons to believe this will be net positive” sense.
-
Regulators Make Bad Product Designers
At long last EU regulators are going to do something about the horrific slate of cookie banners that have descended on the web due to poor EU regulations (that have done nothing to protect privacy, best I can tell).
-
Models Aren't Defensible
There’s a lot of AI news this week — but much of it kept bringing me back to the reality that ultimately models aren’t going to be defensible. That’s not to say that aren’t incredibly valuable and hard to design. But while models may ultimately create a lot of value, I think it will be hard to rely on creating a model to capture that value?
-
AI: Car, Bus, or Road?
The conversation around public AI is messy and often unproductive. Part of the problem is that we don’t have a clear framework for thinking about how to govern AI. We end up with vague calls for government to “do more,” without clarity on what “more” actually means or where it makes sense.