CharlieKey takeaways from your favorite podcasts
← 12 October 2026The Cognitive Revolution

AI Code Needs Specs: Formal Verification & Safer AI Agents

AI agents are becoming the layer between human intent and real-world control, so we need verified building blocks, written specs, retained manual control and fallback plans to stay safe as we delegate more.

Watch on YouTube ↗

Word for word

The model is not your bottleneck. These models are about as smart as most business work really needs. The art now is giving them context and exercising judgment and oversight. And context is something only you can provide.
Every specification that you write is executable. You can test it right away. You can play with it. Maybe it's not as fast as the program will be eventually, but it's good enough to play with. It's like having a bridge model that you can play with for a while before you build the real thing.

Do more with Charlie

Free — sign in with your email, no password.