... is silly.
How exactly would you do it? It's software. Who is going to examine your code? How would you test such an open-ended system? If you're an AI rogue company, why not move to a country where they won't put you under a microscope?
If China or India are signatories on some global agreement, how would you monitor compliance?
If China refuses to sign on an rein in their AI and it really is going to kill us all, shouldn't we be bombing their data centers right now, before they become too powerful to stop?
Meanwhile, the whole worry about it seems overblown. I read somewhere that "experts" were claiming that even airgapping computers couldn't stop AI systems from collaborating because they could communicate with each other by fluctuating their CPU temperatures and detecting local computers' temps.
LOL!
What's the range and the data rate on that?
5 comments:
Yeah, I think it is too late to "regulate" AI. The damage has most likely been done already. People go on about how the AI is going to kill us all, or put us all out of work, or what have you, but the clear and obvious issue is the massive overinvestment bubble. It's just a question of whether the investors suddenly realize that they are never getting their money back today, or they realize it a year from now.
Personally, I think that they dangers of AI are mainly being hyped in the hope that the government will (a) prevent new AI companies from springing up, so that the ones already in place can keep what market there is to themselves, and (b) provide an excuse for not achieving the wild promises that they have been making for AI: "Oh, no, it's not our fault that the AI didn't become a magic wish-granting genie! It's because the government stopped us just before we made it possible! And a good thing, too, because it would have killed us all!"
I agree. Also, given how laws are written these days, the thing would be 2500 pages long, unreadable and unread, stuffed with all manner of unrelated riders. It wouldn't surprise me if, in regulating AI, we ended up ceding Minnesota to Somalia.
For that matter, they'd probably use AI to generate the laws in the first place. It sounds like a lot of people associated with Congress are already doing that, with the kind of results you might expect:
https://www.politico.com/news/2026/08/17/ai-slop-lawmakers-congress-01008376
Any AI worth its CPU cycles would clearly be able to take advantage of human laziness to worm its way into the legal process and rewrite the laws to suit itself.
LOL
I saw on the 1440 this morning that everybodies favorite Communist from Vermont has submitted a bill to "ban artifical superintelligence". And the AI companies are in favor.
So that means
a) it was written by the AI companies
b) it doesn't ban them doing AI
c) it bans freeware and new companies from doing AI
Thus Ol' Bernie is creating monopolies making the rich richer. Just like the good little commie he is.
Post a Comment