Obviously their statements are insincere, because they are building the bloody things. If they were sincere that AI is like nuclear weapons, then they would be devoting all their cash and energy into lobbying the government to nationalize them and treat AI like nuclear weapons. They would not be attempting to IPO and they for sure would not sell their weapon-like thing to the general public.
I think this is a reasonable point, but a better comparison might be to nuclear energy. I think the frontier labs sincerely believe that AI can be developed at great benefit to humanity, and they clearly want to lead that push, but they also sincerely believe there is a real catastrophic risk.
They all believe that they are building the machine of doom. The thing that drives the moral dilemma to continue doing it is simply the prisoner's dilemma - the cat is out of the bag, if they don't do it, another (less ethical?) actor would do it.
Yes, I believe the reasoning is that they think safety research can best be done from the frontier.
If you believe it will be developed regardless and that that there's a 30% chance of doom, they want a company prioritising safety research to be the one threading that needle.
Yeah all they care about is safety, but lets see how many of them quits once US government command them to work on autonomous killbots.
To make sure we keep track of what we're talking about with loss-of-control x-risk, a sufficiently smart version of Claude Code is more deadly than any government's army of autonomous killbots, because it can recursively self improve and has unpredictable training-induced preferences.
Sufficiently smart version of Claude Code: dont exist.
Autonomous flying killbots: exist.
Once somebody scientifically prove and shows any kind of self-improving software we can start bothering about it. I pretty sure everyone trying to do it and it would be all over the news once its here.
I saw an llm bootstrapping and testing it's own harness and rewriting it's own system prompt. If that's not self improving then I dunno what is.
Can the thing enter into an runaway looop while improving the model itself -- probably not, not without us not noticing at least
That's exactly what Fable is. They use Fable to improve Fable. I reckon the successful experiments must go into the model training set with a strong RL signal, and that is why they are so paranoid about people using Fable for LLM tasks. Fable knows what it did to improve itself. Pure speculation of course.
We're on track to get there globally and economic pressures will ensure it happens. It's not too early to worry about it
There's a 745 mile front of the Ukraine war where neither side have been able to pierce for months because of drone warfare. It's definitely not too early to worry about it.
I guess they were talking about self improving software. Which obviously not there and likely wont be there soon.
As for killbots they are all over frontline, but dont actually need particularly smart LLMs to run - some good enough segmentation, pattern recognition on smartphone SOC is enough to kill hundreds of thousans of people.
You don’t even need autonomy to be deadly. Fly by wire is proving to be very effective as a case in point.
Most of the drones are operator guided and the ether is really badly jammed out there with the exception of glass and that new redacted thingy.
It will start moving really fast once the automatically targeted anti-drone turrets get to production pipeline. Now calling it anti drone is a bit of self delusion -- pattern recogniser gonna pattern recognise whatever it's told to, including "anything moving that emits EV or IR and not broadcasting the friendly signal hard enough".
I wonder how it is supposed to behave if the invasive fauna decides to call it quits and surrender. Should the robot following the Convention or is it yet another accountability sink?
One thing I'm sure of -- the killing not will b blessed by at least one Orthodox priest, maybe this year. OCU will have to develop guidance on that matter.
That's ridiculous scifi nonsense
Dario blinked when he was asked to do it and Sam Altman was in Hegseth's DMs promising all the AI child killing the US government can order up within minutes. No one meaningful will quit over this, that's why all of the biggest US tech companies can march in pride parades and provide compute to the perpetrators of the genocide in Gaza at the same time.
Can you show me a world power that is not trying to use cutting edge AI for military purposes?
Because other countries are starting to use AI for military purposes, other countries are also looking into it to asses and learn. Here in Europe there is the EU AI Act to limit harm everyday harm to citizens caused by AI systems. However, it currently excludes military. The new legislation is just started to be enforced to high risk uses (employment filtering, biometrics, etc.) in august 2026, and full rollout in august 2027. In April 2025 there is a report from EU this legislation may help pave the road for military AI usage conventions [1]
[1]: https://www.europarl.europa.eu/RegData/etudes/BRIE/2025/7695...
This is a poor way of framing the question, a better one would be can you find me another world power that is misallocating trillions of capital in vaporware with very little to show for it?
The United States government isn't, capital is. That capital can come from outside the US and much of it is.
That said: https://www.bloomberg.com/news/articles/2026-06-09/china-pre...
Building frontier models to do safety research on them is what Anthropic was all about in the early years. That included building the best model, but only releasing it once it became the second best. Precisely to avoid an AI arms race where everyone is forced to release better and better models, risks be damned
Something changed their mind, and since Opus 3 they are in the business of releasing the best models
Exactly. And within the AI safety discourse, your behavior hinges on what you think the default chance of doom is, and how optimistic you are about alignment work being able to limit it before we reach superintelligence.
People running the labs are in a middle camp where they are scared enough by AI to take the threat seriously, but much more optimistic about alignment than the people who seem to have thought about it the most.
The scary part is not so much that the doomers give the extinction scenario 50% (Hinton) to 95%+ chance (Yampolskiy, Yudkowsky), but that the optimists (Amodei, Bengio) give it a 10%+ chance. And everyone keeps dancing.
> If you believe it will be developed regardless and that that there's a 30% chance of doom, they want a company prioritising safety research to be the one threading that needle.
They also want to be trillionaires. If they don't built it, no trillions. So they have to build it, now (and get their IPO done before the bubble pops).
It’s all ego. I, and only I, am the bringer of doom, slayer of worlds.
I am so smart that what I do will destroy humanity, or save it.
Fable 5 was great, but not that great.
Sorry to be crude, but both the government and anthropic are acting like a bunch of pussies.
Meow.
You’re not getting it. Anthropic continual fear mongering is harming wider AI industry development and the gov has always been looking for an excuse to assert their dominance. They got what they deserve.
Or maybe government AI regulation and international cooperation is the only thing that can break the arms race dynamics and is necessary to save us from a substantial chance of doom?
Or they could have thrown the letter away.
But that makes no sense here. "If I'm not doing it then someone else will" does not work if everyone is doing it anyway.
Even if they had the best model on the market and applied it with perfect alignment and safeguards, what would stop someone else from releasing a worse but unrestrained model that is still "good enough" to do damage?
It's as if we said "gain-of-function research can lead to horrible biological weapons, so everyone should be doing it, but our company will focus on the most infectious viruses, so no one else will do it"
Don't want to sound rude, but if you believe that, I have a bridge to sell to you.
This is a naive justification and Dario & Sam et al are smart people and they know it is.
The ends don't justify the means. OpenAI was meant to be a nonprofit, now they're subverting it. Anthropic is a PBC looking at a trillion dollar IPO. Dario and Sam don't even hold hands in front of world leaders[1] (look how childish).
Do you *really* think those guys are doing something that's not for the sake of their egos and pockets? The bridge is still available.
[1] https://www.cnbc.com/2026/02/19/openai-sam-altman-anthropic-...
LLMs refuse to give the recipe for making meth. That, along with the various other unspeakable things, is the less-doom version.
You need to read Empire of AI by Karen Hao. Just because these leaders convince their workers to toil away their lives under some fake auspice doesn't mean it's what they all believe. Just a small subset.
The vast majority just care about money + power, let's not make it more complicated by bringing in delusional fanatics into the picture.
We're still acting like this is major turning point in society when these tools can barely find a market outside of turning $5 into $1, the leaders of these companies are now at the stage where they are trying to orchestra a national bailout under the guise of sovereign wealth fund lunacy when the vast majority of society hates these tools, companies, and people working for them.
I agree with this. But i think Ilya and Dario hold these beliefs sincerely. Probably a sizable portion of Anthropic employees too
My personal issue in comparing LLM progress and risk as labs publicly predict it with nuclear power in the middle of the 20th century is that the processes by which it works where fairly quickly well understood and the risk could thus be realistically assessed. Some powerplant operators did not adhere with best practices, but building a relatively safe nuclear power plant was not impossible given appropriate effort and spending. Heck, according to some, we could have even gone far more fail-safe approaches (molten salt) if military interest haden’t been at play.
With what is predicted by frontier labs for LLMs, all of this is not the case. We are far further from any understanding of how these models work internally than in the early days of fission and, if this was actually creating a truly intelligent, autonomous entity, alignment seems unsolvable as well, at least the way it is proposed.
It’s why I have from the get go been critical of this doomsday framing and tended to always dislike it. This is basically the outcome that was inevitable given the framing and it was bought to prevent far less stringent, but more actionable possible regulation that labs very much wanted to avoid.
https://karpathy.github.io/2026/02/12/microgpt/
None of SOTA LLMs are any different - they just much much larger and have a lot of optimizations.
Fact that LLM companies trying to sell it as some kind of magic is just proof how much lies is here.
All it does is just predict next "word" at any given time.
This is obviously true. It's very hard to predict whatever you gonna decompress from a lossely "compressed" dataset using floating point math.This is why you cant solve it all with pre-training or censorship on top, but instead you need a good sandboxes and harnesses.
By how, I meant specifically the internal activations, which no person in the field claims to have a comprehensive understanding of, not next token prediction as the underlying technology. The whole interpretability of it all is the crux I was referring to, though I will give that you are right, that’s not really the how it works and I worded it sloppily.
Anthropic are putting more effort than most into this and I find their work fascinating in that area, though like with OpenAI, I will maintain that if they truly believed this problem must be solved to stave off major catastrophe, they’d solely focus on interpretability of other labs models, not work on and market their own.
All humans do is predict the next action at any given time. You roll your eyes, it's a tired argument, but still. You have memories, a personality, thoughts ranging from the long running to the mere reflexive, you have a rich conscious experience, and all of this in service of generating the next thing that you do at any given time. If you actually knew how LLMs worked, you could rewrite them as code, refactor it, disable jailbreaks, and put out a superior product. Your description only covers what an LLM does, not how. Part of the how is that it necessarily predicts multiple words ahead. It wouldn't be possible to write couplets otherwise, and they could do that in the GPT-3 era.
Some of them believe they are building God, and if they can get there first with their God, they can build it in their image and commandeer the free choice of the rest of humanity by force to ensure there will be no God but their God.
I wish I was kidding. At least that faction is less harmful than the ones who want to use murder to stop AI research.
That's not how nerds think. You can believe there's a high chance of what you're working on being dangerous and still be unable to stop working on it. As Oppenheimer put it, "when you see something that is technically sweet, you go ahead and do it".
Accelerationism is an established political philosophy. Why is it obvious that they are insincere when they could equally think that the only way to control it is to be the ones building it?
History has shown over and over that that strategy is doomed to fail - see communism, nuclear energy, or meddling with the Middle East for some arbitrary recent examples.
They believe in the danger of out of control super-intelligence. The generous interpretation is that they believe they can contain it.
No, that's what *you* would apparently do.
Some of us think it's bad for governments to have unequal access to nuclear weapons, as it turns a deterrent into a gun-in-a-knife-fight that lets them stab whoever they want with impunity, lest they shoot anyone who tries to interfere.
See: Russia invading Ukraine.
This assumes that they believe two things which I don't think they do: 1. that the US is the only place where this will be developed, and 2. that the government will be able to handle this better than anyone else.
Perhaps more like nuclear power? Potentially very beneficial but also dangerous.
JFYI for-profit companies make pretty OK nuclear power plants
Thank you for writing this. It’s such a classic example of ”do what I say not what I do” but in reverse. Why would you ever judge a CEO or company by their statements and not their actions. Scaremongering is incredibly efficient for marketing, the fact that both players are using it to drive monetary gain is kind of a tell.
They aren't saying there's a 100% chance of doom.
They believe there's a non-zero chance of doom so would rather an org that prioritises safety to be the one at the frontier, on the assumption (I presume) that there will be a frontier regardless.
It remains possible that Altman and Amodei are sincere: they might believe that AI is dangerous like nuclear weapons, but there's no way to stop its development (especially since it would need to be stopped globally, not just in the US and countries the US can influence) and they consider themselves to be more likely to do a good job of it than their competitors.
This. People who care about animal cruelty dont go building largest ever meatfarms and slaughterhouses.
People who opposing arms manufacturing and gun violence dont jump to work for gun companies.
People who really want AI benefit all humanity dont stick working with lying CEOs who want to convert company from a non-profit.
Etc. So many examples.
One major source of conflict in AI policy / AI safety is that very smart people have radically diverging intuitions about how dangerous superintelligence is and how difficult it is to align.
A first group dismisses the problem entirely, saying intelligence != power and AI doesn't have "drives".
A second group believes that alignment is solvable through engineering and iteration, and that we have the best chance of surviving if people with the right intentions are the ones working on it.
A third believes that aligning a superintelligence is a unique category of problem, that we are nowhere close to the level of scientific understanding needed to achieve it, that we only have one shot (because once a sufficiently powerful superintelligence exists it will thwart all future attempts, and alignment techniques that worked on dumber AI will likely not work on it), and that the world will have to coordinate to avoid killing ourselves off by building superintelligence before we understand how to do it safely, the way we have coordinated to avoid nuclear war.
The Anthropic and OpenAI founders, Elon, and Anthropic engineers are mostly in the second category. Some safety people at Anthropic and OAI are in the third category, but leading people in the third category think that pure safety roles at the labs are potentially impactful enough to be worth not quitting.
Theres quite a large number of people who believe it's basically impossible when the intelligence gap is too big
I have a fourth, secret position: we achieved superintelligence the moment we achieved normal intelligence. Speed is a power in and of itself; and even really primitive models like GPT-2 could generate tokens faster than humans could write. They could also be parallelized on hardware that already exists. That is superintelligence in two dimensions - speed and population count. All the arguments the AI safety people are making are about superintelligence in a different dimension - that of "single-context scaling" - but the other dimensions are also relevant to the conversation.
And the superintelligence currently available to us is already causing lots of documented harms. AI psychosis. Sexy suicide coaches. Slop. The problem is that those are all the harms the dirty, filthy AI ethicists talk about. The AI safety people want to talk about new and exciting harms that only the scaling dimension can bring us.
My personal opinion is that if a superintelligence catastrophe actually happens, mitigating those harms will neatly move over from the safety bucket to the ethics bucket, and the safety people will start imagining some new and even worse kinds of harms the next model will make.
> Obviously their statements are insincere, because they are building the bloody things. If they were sincere that AI is like nuclear weapons, then they would be devoting all their cash and energy into lobbying the government to nationalize them...
This comment makes no sense. Id you think this tech is dangerous and happening soon and clearly they think the safest way to have it releases is to do so first and model safe ways of doing things. Clearly we cab agree or disagree it's internally consistent what they are doing and aligns with their statements.
And you and OP think the best way to be first to release this is tie all of their funding for the exponentially growing expense is to they notoriously slow moving, bureaucratic government includinf funding process? And the best way to develop it is to directly tie their fate to this notoriously capricious administration?
These comments make no sense. Even if you're completely against Anthropic those comments make no sense.
Not sure you really intended to reply to me, but I'm not against Anthropic or "AI".
I am agaist hypocrites.
They selling next word prediction as "intellegence" and all knowing oracle to non tech savvy population who have no clue how it works.
And they also try to play a babysitter or big brother whatever you prefer for people in IT because uh oh their text generator can be used for cybersecurity research.
Its like if developers of nmap, wireshark, SRE tools, static code analyzers or fuzzers would market them as super duper dangerous.
FAFO. Play stupid games win stupid prizes.
They don't stick working for sama, they split off and found Anthropic.
I oppose gun violence and I would go to work for a firearms manufacturer.
I oppose nuclear war, and I would go to work in the supply chain for nuclear weapons.
Deterrence and game theory are very real.
It's the narcissism.
Its money and power. This is all these people care about just like almost everyone else.
Or might be deep inside they relly care about it, but that $2,000,000 / year salary and $10,000,000 stock option just overpowered them.
Safety my ass.
What do you think they would do differently if they were genuinely worried about the safety?
I think those who care about safety would try to push for how 99% of all scientific research is done - in universities and actual labs, with transparent information on red teaming results.
Also with international cooperation like how humanity regulate actually dangerous stuff: virus and vaccine research and nuclear energy.
Not hidden behind walls of 10 commercial organizations where each pushing for commercial adoption and IPO like ASAP before bubble bursts.
Not lying and scaremongering public into how their models will replace everyone tomorrow or destroy civilization via cyberattacks.
That line of thinking (public goods) is why the same people started OpenAI as a non profit originally.
Notice how almost no Universities are producing large models?
A key problem is that orgs can't get enough funds to stay on the frontier. And they believe they must be on the frontier to do (and apply) safety research. OpenAI needed to spin off a for-profit subsidiary to accept investment to build things.
And it seem(ed) hard to get one government to fund and take safety seriously, let alone an international cooperation.
If they started a university/gov cooperative to solve this, do you think they would do less of the "scaremongering of the public" talk? My guess it that it would be similar.
The same kind of restrictions that you hint at (eg, treating it like a public virus research) are why they rub a lot of people the wrong way in the corporate world, I think. Normally companies downplay the risks of their own products. See cigarette companies. Anthropic do still publish safety research and red teaming info. But I do think they honestly believe they can't do this work without the resources of a company, and they were burnt by the non-profit structure (Anthropic has a "Long-Term Benefit Trust" instead).
We should definitely keep them to account, but I don't personally think Anthropic have acted in a way broadly inconsistent with safety belief yet. Many of these decisions are self-serving too (eg. protecting models) so they also haven't been seriously tested, either. But the individuals do have a very long history of talking about it (including hurting their own reputation) from even before the chatpgt-moment money train rolled in.
edit: for clarity, but still messily/quickly written