zlacker

The problem is that once you have proof of these things, it may be too late. That's an even larger problem when you consider the speed at which our government moves on things.

I recognize why many folks see it as Altman bringing up the ladder, and I'm certainly cynical enough not to discount that. Still, I don't think the facts that you're citing here are evidence that he's doing that.

> We do not even have a consensus around a working definition of "intelligence"

This doesn't seem like a good reason not to regulate - intelligence has been around for a very long time, and you're correct that we don't have a consensus around how to define it. There's no reason to believe we'll get to a consensus on that soon, so if you're saying that we should wait to get there before we regulate, you're effectively saying we should never regulate.

>>idopms+(OP)
> The problem is that once you have proof of these things, it may be too late.

If the thing you're worried about is Skynet, regulation isn't going to do anything. Once someone builds it, you lose regardless of whatever the law says, and the person building it may not even be in your jurisdiction.

If the thing you're worried about is AI being used for copyright infringement or something, in what sense will it be "too late" if that sort of thing happens for a period of time before any laws are passed?