Sunny_Fung

Thanks, this takes me right back to middle school. I was on a 80386 back then. You had to dial in, and the phone line was tied up the whole time. An hour online cost me a whole month's allowance. Worth it, though.

So it's self-aware enough to suggest the fix but not self-aware enough to not do the thing in the first place.

Glad OpenAI opened up the Decisions API. These decision models are gonna be everywhere. The point isn’t an LLM for humans — it’s a model for LLMs. The LLM is the bot’s brain, the decision model is the cerebellum, feeding real-time decisions back to the brain.

It shouldn’t, but it can. And legally, the operator may still be liable for what it does.