The Creator of ChatGPT Thinks AI Should Be Regulated

This is like encryption: the cat is out of the bag and there's no getting it back in. As computing power increases, control of LLMs will be democratized / made widely available, and no amount of regulation will stop that.

More to the point, it will be incredibly difficult to reset the gen pop's impression of AGI. More structure to LLMs may give more reliability, but in the meantime, I have no doubt LLMs will wreak havoc in the world. We're in a mess it's not clear can be cleaned up.