I think we have a pretty good idea what the capabilities of the most advanced models are. And they're not that great.

This is a really stupid bill, but the one thing about it that I kind of like is that it's going to put a stop to the AI companies trying to inflate their value by talking about how scary their frontier models are, and how they're "going rogue." None of these AIs are "super intelligent" or "going rogue." What you have is a (admittedly really cool) technology that can solve a lot of problems but also is non-deterministic, so makes a lot of mistakes. And you're giving it the keys to do stuff. And then it misinterprets the instructions and fucks up and does something with those keys that it wasn't supposed to. Then instead of admitting, "oh, we engaged in really bad computer hygiene here," they figure, "we can market this, and say it 'broke free of the sandbox and went rogue" to imply that their models are so advanced it had purpose and the will to act.

So now the old politicians who don't understand technology are going, "oh, that IS scary. We're going to ban it," backfiring on their scary tactics marketing campaign. So it will result in them stopping this stupid scary tactics marketing campaign.

Here's the actual situation: you know every story you hear about how a model deleted the entire codebase at a company? If you have any level of computer literacy at all, the reaction you have is not, "oh, using AI as a coding agent is dangerous." It's, "why the HELL was the agent running with permissions to do that?" And if you hear about the OpenAI's agent broke free of its sandbox using an exploit that it found on the database you set it loose on, the question is, "why the hell could it even run executables outside a whitelist of stuff for it to build". it's amateur hour over here, giving a thing that is non deterministic and capable of fucking up the ability to do do anything at all outside a very narrow scope. The models aren't going rogue, they're fucking up, like when you ask it to give you the emoji for a seahorse and it goes into an infinite loop of bad emojis.

If you were worried about being chased by kill bots, go and watch some videos from drone operators in Ukraine. Robots don't need feet for that job.

https://www.cs.wisc.edu/2025/0...

And DARPA has a whole program for it:

https://www.darpa.mil/research...

Ok, but some of us actually want to do something useful instead of just try to win the trophy of making a model run on the cheapest hardware. Unless you're generating at least ~30 tokens / second, it's not useful.

If you want to split the cost evenly, sure. But a team of 3 people working at a software company today typically pays the AI companies about $5,000-$10,000 a month, so I don't think they need to rely on the parts of the world where $400 is twice their income.

... sadly for the Americans, the rest of the world now knows they can't count on a US based provider for this kind of thing any more.

It was uncomfortable enough relying so heavily on American software back when it couldn't be switched off remotely on the say so of an idiot. Today it's an intolerable risk.

...a lot of people online were saying there was an upcoming event called The Halvening that would change everything.

I guess this is it?

Seems to me 'dead' for a taxi isn't 'dead' for a static power bank. If I'm running a taxi I've got hard limits on how large my battery can be and how heavy, and I want to maximise the mileage I get between charges, because while my taxi is charging it's not out on the road earning money. When that battery is keeping only maybe 80% of its original design charge, and now I have to schedule one recharge too many per working day? Bang goes my business plan, so I'm replacing it.

If I'm storing energy for the grid I'm a lot less worried about that. It only stores 80% of what it did when new? Better than nothing, and the taxi firm is selling them off cheap. I'll stack them up!

Perhaps not, but if you pick your moment right then permanently stopping the work of some of the most talented researchers there could very well make a difference. A spectacular incident that makes the headlines might also deter others - bright graduates might decide it's far safer to take up a different line of work, subcontractors and suppliers might decide doing business with AI firms isn't worth the danger, investors might figure the increased risk of loss of premises and equipment into their projections, that kind of thing.

If people genuinely believed AI takeover was a real, present and imminent threat, then they wouldn't just be publishing essays online, they'd be forming direct action groups, along the full spectrum of campaigning: all the way from awareness raising publicity campaigns, through picketing, blockades and sit-ins, up through Black Bloc type actions, right up to menacing intimidation campaigns and terrifying physical force operations. But I don't see any Butlerian Jihad getting started. Which tells me they don't actually believe this at all; they're just bigging up their own importance. 'Oh yes, our technology is so incredibly powerful, if it were done wrong then imagine what could happen! Keep the money coming to make sure it's done right instead! Then all that power can be ours instead!... I mean, uh, yours, Mr Investor sir.'

AI stock valuations don't make a bit of sense unless the technology turns out to be every bit as powerful as that. If they don't keep that thought alive, then the bubble bursts right now. That's what all this hot air is about, and that's why nobody really pulls a Miles Dyson at the AI research lab.

The interesting thing about the Terminator movies is that when AI researcher Miles Dyson became convinced that his work had a high probability of resulting in an artificial general intelligence attempting to replace humanity, he did not go and post a ten thousand word essay on LessWrong about how he had updated his timeline and p(doom) estimates and discussing the full Bayesian analysis of the situation. He went to the lab that very night with some heavily armed companions and he blew the place up.

I keep hearing that one AI researcher or another claims that they believe as Dyson came to believe. Until one of them takes similar action, I simply do not believe that they actually think their research carries such a risk.

You have access to the lab where the work is being done? You regularly meet in person with leading researchers and talents driving the project forward? You are an American and you have the Second Amendment? And the entire future light cone is at stake? Quintillions of hypothetical future lives riding on the outcome of this project here and now?

What's the most effective, altruistic thing you could do for them?

Yeah, exactly. I've never heard of anyone shooting up their AI lab. Which tells me they don't believe their AI is at all likely to wipe us all out.

... and then to secure Ukraine into the European economic and defence structure as firmly as possible. At this point they're far ahead of NATO on how to use drones and robots in war, and they're clearly getting very, very good at building them too. European military and aerospace people both will have a lot to gain from cooperation with Ukraine after the war is over.

As cadets were so memorably told a generation ago: remember always, your duty is clear - to build and maintain those robots!

Sounds good - DDG is pretty bad lately, it typically just returns a list of AI generated SEO slop pages, so I've been switching back to Google in a private tab for a lot of things. Might be worth trying this one now!

If they leave them there, the next administration might be able to switch them back on and start gathering woke climate science data again.

Long ago, the UK courts ordered all the major consumer ISPs to block The Pirate Bay along with various other popular services. Ever since, we've had to keep up to date on what the latest proxy address might be.

Of course, thanks to the new censorship laws introduced more recently, we're all on VPNs now, so as to avoid having to hand our ID to the wallet inspector for every last website we ever use. And once that was set up, it was nice to discover that the original is still in play!

The jokes about Darth Jar Jar were everywhere of course, but it could have worked. Star Wars lifted a few ideas from classic SF sources including Asimov's Foundation series - in which, we might recall, the terrifying, unstoppable galactic warlord known as The Mule was hiding in plain sight as a clown, who seemed to be merely a harmless entertainer at court. His military success was chiefly thanks to his psychic ability to manipulate others' minds to his liking - Darth Jar Jar could have done very well that way!