Transcript
Explain it to me like I’m 80
Mom: What’s with these AIs escaping and hacking people? That sounds scary.
Me: You park your car at the top of a hill. You leave a note on the dashboard that says “please be good” and then you pull the parking brake. The car careens down the hill, smashes everything up, and you put out a press release saying “My car ignored my instructions! It hallucinated! It went rogue!” Your insurance company not only believes you, but invests a hundred million dollars in your company.
Mom: Jesus Fucking Christ
Nah it’s just autocomplete on steroids. They’re not coded to do anything.
It’s like hitting the fist suggestion on your keyboard over and over again. Whatever you get is based on patterns it found in training data, which includes every doomsday conspiracy cult on 4chan and every CVE vulnerability ever reported.
There has been multiple studies showing how a model’s behavior can drastically change just by altering individual ”neurons”.
Not all models work the same. I literally do this for work. LLMs are not every kind of AI agent out there. Some are quite literally designed to do things others are not. Yes changing weights and reward functions for models changes their behavior drastically…that is coding.