> But even cynicism doesn't seem to me to give much reason to worry about regulation of "next to no cost" open source models, though. There's only any chance of regulation being practical if models stay very expensive to make, requiring specialized hardware with a supply chain chokepoint. If personal devices do catch up to the state of the art, then for better or worse regulation is not going to prevent people from using them.
This is a really good point. I wonder if some of the antipathy to the joint statement is coming from people who are worried about open source models or small startups being interfered with by the regulations the statement calls for.
I agree with you that this cat is out of the bag and regulation of the tech we're seeing now is super unlikely.
We might see regulations for startups and individuals on explicitly exploring some class of self-improving approach that experts widely agree are dangerous, but there's no way we'll see broad bans on messing with open source AI/ML tools in the US at least. That fight is very winnable.
This is a really good point. I wonder if some of the antipathy to the joint statement is coming from people who are worried about open source models or small startups being interfered with by the regulations the statement calls for.
I agree with you that this cat is out of the bag and regulation of the tech we're seeing now is super unlikely.
We might see regulations for startups and individuals on explicitly exploring some class of self-improving approach that experts widely agree are dangerous, but there's no way we'll see broad bans on messing with open source AI/ML tools in the US at least. That fight is very winnable.