> In a more normal media ecosystem, the reactions to these cybersecurity issues would be obvious. American AI companies would reconsider doing business with Irregular, not only because of its failure to secure its systems, but because it is an Israeli firm potentially outside US oversight. Lawmakers would consider taking action against Irregular or against its American business partners, which include OpenAI, Anthropic, and Meta. They may consider strengthening liability against firms which instruct AI models to commit cyberattacks, and whose models then commit those cyberattacks.
The obvious point is that dealing with Israeli companies/entities by the same standards you usually deal with others is a career suicide with enormous political consequences (in the US especially). When you combine that with the opportunistic nature of the overlords that run these labs, the benefits of screaming "pace the frontier" outweigh everything.
> This suggests pacing the advances: not training or deploying AIs without a strong safety case27 that convinces independent experts. Such a rule would also create an incentive to work out how to build AIs that are safe by design.
Has any attempt to pace AI ever succeeded? Isn't that the same philosophy that got us OAI and Anthropic? Maybe we are overthinking this, it's much simpler to let AI loose and see how much it can break the arrogance that human thinking is special.
> Don’t use AI to replace every job, from graphic designer to accountant to software developer. Place legal restrictions on (at least) large companies such that they may not use AI to replace human labor in this way.
AI is taking all of our jobs. We need to control the means, nay, the pace of production. We should build some kind of barrier, not of concrete and cement but also of the regulatory variety, to stop these agents from coming into our special white color cubicles. It's the only way.
Hacking open source infra? No, you misunderstand. This is sandbox escape, really just agents being clever and super duper dangerous. Aggressive red-teaming for free really if you think about it.
Everything is fine. Sandbox escape. We will publish a report on it. Export controls, maybe? You hear about China AI stuff? Can you imagine if they get this stuff? Wow, we need to seriously think about regulating this. When is the IPO again? Sorry, ignore that, so yes alignment and sandbox hardening is where it's at.
The mathematicians would be wise to re-read The Bitter Lesson, maybe twice a day, until it sinks in. No offense and with all due respect to the Ivory Tower Giants but the whole "oh no you ruined the game because you solved it, I was supposed to play with that in child-like wonder manner and take several years to do so, and by then I would have showed you all the trickery I did to get there and maybe that will be useful to you" over the past 2 weeks is, get this, cope.
Chess. Go. Coding. Now Math. Another one bites the dust. Let's meditate on this lest we forget: Stochastic parrots that generate the next-token cannot reason or produce anything meaningful. Let's protect our jobs at all costs, even if we have to drag all of humanity down. It can't be! Stochastic parrots can not replace the Ivory Tower. No way.
So... what's the critique here exactly? "Oh hey everybody, Paul is out touch!" - Okay, now what? Rich people are rich, that's truism. Some comments made good counter arguments but this lazy polished ad-hominem attacks are more offensive than the piece itself.
In the modern world: any tech proposition that starts with protection of children as a goal can be dismissed out of hand, since it's emotional manipulation masquerading as tech policy. When I hear "protect kids", all I see is a sleazy politician bowing to their respective Security State apparatus.
I support a rule to ban AI-generated/edited posts.
Initially I thought they'd be fine, because AI-generated isn't intrinsically an issue and the comments can be good. But in practice, the AI posts tend to be slop, and usually there's a better human-written source for the same topic (for example, one of the many other recent "age verification is mass surveillance" posts here).
Part of this was written by AI, but with a human in "charge" who
explained which part of AI was used here. Would that also be a
bannable example for you? I am not so convinced that this is
bannable per se. Perhaps it may be different if the AI-slop was
not announced, but when it was announced and explained?
> one of the many other recent "age verification is mass surveillance"
> posts here
Well, it actually is. It taps very much into other similar laws
e. g. "chat control", aka chat sniffing.
I should've said "guideline". I think posts can include AI if it's reasonable and/or they're good, while the guideline gives a reason to flag AI posts that are generally bad.
> It taps very much into other similar laws e. g. "chat control", aka chat sniffing.
There are many recent Chat Control posts here too. I agree Chat Control is bad, and poorly-implemented age verification is bad (though it can be implemented in privacy-preserving ways, albeit ineffectively; I commented about this 42 days ago at https://news.ycombinator.com/item?id=47123507, and it was stale then). I don't want to hear anymore about it. Maybe I need a filter myself, for the lucky 10,000. But the problem even for them, is that the repeated posts (without links to previous posts) have mostly low-effort comments, because people who made high-effort comments can't/won't keep repeating them.
The obvious point is that dealing with Israeli companies/entities by the same standards you usually deal with others is a career suicide with enormous political consequences (in the US especially). When you combine that with the opportunistic nature of the overlords that run these labs, the benefits of screaming "pace the frontier" outweigh everything.
reply