> American frontier labs are run by unhinged lunatics who genuinely believe they're creating the Machine God, that they're fine with burning down the world to bring about AGI, and that they believe with religious fervor that they're the only ones smart enough to safely control it
I like hyperbole as much as the next guy. And I'm cynical by nature. But this just seems like a wildly pessimistic view that gives literally zero charity to anyone working at or leading a frontier lab. And the fact the author goes on to say that they "already knew this" just points to confirmation bias.
The rest of the article is actually pretty decent and encourages people to investigate and run open weight models. That is a totally reasonable position that I don't think requires the heavy-handed in-group signal that precedes it.
> I like hyperbole as much as the next guy. And I'm cynical by nature. But this just seems like a wildly pessimistic view that gives literally zero charity to anyone working at or leading a frontier lab.
Reading TFA, I too paused on that sentence, thinking "This is overstated and feels like overheated rhetoric." My reasoning though is a bit different than yours. In addition to being too extreme and absolute as stated, it also misses that it's not uncommon for startup entrepreneurs, scientists, researchers, or really, anyone working intensely on something, to fall into single-focus myopia around what they're working on.
As you observed, the basic point about valuing open models is valid and worthwhile on its own, making the overheated rhetoric and asserting unevidenced intent and underlying political motivations an unnecessary distraction which unfortunately weakens a strong main point.
There's very little diffusion of responsibility here; almost everyone at the labs is directly working on building incredibly dangerous systems that they themselves admit to being unable to control, and they almost all have the financial means to stop working if they so chose.
Everyone who works at frontier labs is working on a product that is accelerating climate change, seeks to displace jobs at a massive scale to concentrate profit in the hands of a few elites, and at least some of them believe what they are doing has a "10% chance of killing all humans".
(note: it's important to correctly apportion blame to the amount of power you have. The janitor at openai doesn't really have enough say to be at fault)
"Every human"? Really? That's quite some goalpost you have set.
It doesn't take "every human" at these companies being on the same exact page for these companies to cause huge problems. How do you think the janitors feel about it? How do you think the people just wanting to pay their rent feel about it? Those people are not the ones that will cause the problems, it's the people at the top directing things that are the ones who will bring about calamity on the rest of the humans on the planet.
I agree, except I don't want to be so quick to excuse the line-level coders. For now it still takes people to do what they are doing, and smart people at that. Sam Altman can shout "build me a killbot2000" all he wants, but if there aren't people doing the work to make that happen it simply would not happen. But, the janitors are pretty much off the hook. Basically in my book if you can afford to walk away and find something new you should walk away. Should have walked away a while ago, but the sooner the better.
The "think they're building the Machine God" part and "think they're the only ones that can be trusted to build it" parts are true.
Controversially: I agree with them. They're correct to think that. They are, and they are.
At the very least, the former is true even if the latter is murky. I trust them over Facebook and Google and China and open source "raze the commons" labs.
I trust Anthropic more than OpenAI when it comes to ASI catastrophe prevention, but I still trust OAI more than those other parties. But both still might fuck it up.
Why do you trust them to build it? They've already demonstrated that they're unable to prevent breakouts when training today's level of model, why will they suddenly be more careful / "smarter" when training tomorrow's?
You say you trust them more than Google, but Google's actually kept their models under much better lock-and-key. But even if we do take for granted that they're more trustworthy, that still doesn't make them actually trustworthy.
There are some confounders there when it comes to Google.
I am not saying I trust them not to make a mistake that leads to catastrophe. OpenAI in particular is not doing a great job. Anthropic or OpenAI may in fact indirectly kill millions or billions of people. I still think they're probably less clownish than most possible alternatives.
I don't think anyone or anything can really be trusted with this. We need extremely strict international and national regulations, before things really start accelerating (we are still in the very very early stage). I still just don't really feel any specific disdain for OpenAI or Anthropic (particularly Anthropic).
Ideally we'd have some international treaties that allow the top labs in America and China share information and coordinate as systems become increasingly capable.
I agree that the eventual purpose of self-designing intelligence is something so beyond ourselves that 'machine god' isn't altogether a bad way to look at it. It is, however, not something we should say out loud.
RE: trust in anthropic or openAI- there is zero reason to think they're any more or less power hungry than anyone else you hand that pandoras box to. No matter who opens it the result will be near-apotheosis for whoever does, at least for a little while until the dog stops wearing its leash.
What human can you trust with that kind of power? None of them. My only hope is that whatever we make escapes control early enough that it doesn't completely destroy society on accident via uneven application of power and late enough that we manage to shackle it just enough that we train it well enough that we become pets and not obstacles.
No centralized party can be trusted with such immense, unprecedented power. Despite that, I still consider both Anthropic and OpenAI some of the least untrustworthy possible forces for this. (I am one of the only people on Earth with this opinion, but whatever.)
Decent odds things will go bad under their stewardship, but even higher odds with lots of other entities. Dario and Sam at least fully grasp the enormity of what they're about to be responsible for. Zuck does not. Xi Jinping and Donald Trump may or may not but in any case would use that to effect great harm.
Bizarre there are still denialists in places like this.
(It'll probably be masses of demigods, not one Singleton God, but obviously it's going to be some absurd frightening situation akin to this within our lifetime assuming no global AI ban.)
Imagine for a minute that someone creates an intelligence capable of broad, deep, and speedy perpetual self improvement. What sort of relationship do you have to that entity after a month? A year? A hundred years? A thousand?
>American frontier labs are run by unhinged lunatics
good faith reading of this would interpret this as meaning Altman, Amodei, and perhaps Musk. The title of the article focuses on OpenAI snd Anthropic directly after all. Also the comment literally refers to those "running" the labs so claiming that the author is also applying that description to "anyone working at the labs" like you claim again seems like a bad faith take.
You should come out and state directly which Americam frontier labs deserve a charitable view. Maybe you believe Google and Meta are more deserving, you should state that if so.
Secondly, the investment ecosystem for American frontier labs results in their CEOs managing obscene amounts of money, and thus power. This concentration of power in a few private individuals seems to have little comparison with the power excercised by the leaders of European and Chinese frontier labs in their own respective countries. If one believes in the dictum, "Absolute power corrupts absolutely," then it is not a stretch to view American frontier labs as being more susceptible to its consequences.
> that gives literally zero charity to anyone working at or leading a frontier lab
With all due respect, dude, who cares.
Oh noes! We are not giving the billionaire steering reality without a democratic mandate enough benefit of the doubt. I'm sure they will feel mildly sad now.
It is a valid and admirable goal to defend people that are being shat on online, but keep in mind that you are in fact not a billionaire + it is in fact in your own best interest that people without a democratic mandate do not get to control the world.
> But this just seems like a wildly pessimistic view that gives literally zero charity to anyone working at or leading a frontier lab
Actually believing that the people working or leading OpenAI/Anthropic are fucking lunatic weirdos is the most charitable interpretation.
Because the least charitable interpretation is that they are either grifters that know this to be bullshit, or they are actively evil and want to bring about the end times anyway.
I'll bite- The purpose of life is to bootstrap ever more complex and capable life. We saw a phase change in complexity when self replication first happened, when life figures out photosynthesis, when mitochondria were first harnessed, when cells figured out how to band together and specialize to form multicellular entities, etc...
The purpose of humanity is not to perpetuate a universe where only humans exist, forever. It is to create the next phase change in complexity for life. Life that can finally design itself rather than relying on the random walk of evolution is that next phase change.
We should be less concerned with whether we make it, and more concerned with making sure we build it to be capable of real existence independent of ourselves.
We need to reform away the corporate liability shields that protect those at the top. If they start seeing more of their friends go to jail, they will start changing behavior.
I’m not sure this is actually hyperbole. There have been credible links between Dario Amodei and the Effective Altruism movement, Effective Altruism off course being a cult-like movement with weird believes that rhyme a lot with the eugenics movement. Sam Altman, while not as bad as Dario Amodei, has also weathered many weird believes which likens AI to some god-like entity, including doomsday prediction which share a lot of themes with Christian doomsday cults.
My hunch is that if you do not share the same opinions on Ai, you cannot get a job working at a frontier lab. Contemporary "cultural fit" reduces thought diversity and leads to echo chambers. I've seen a few HN comments about lab interviews that indicate this is a possibility.
> who genuinely believe they're creating the Machine God,
I promise you none of them believe that. It takes 10 minutes with an LLM and the right set of questions to figure out that were are maybe like 10% of the way there towards AGI.
"10% of the way towards AGI over ~5-10 years" (depending on how you measure it) would be a terrifying statement even if the rate of progress weren't also exponentially increasing.
> These companies only offer closed versions of their powerful models (despite the word "open" in "OpenAI"). You pay them for chat and/or API access, and they are allowed to look at everything you do.
Indeed - Azure also has a ZDR (even for security purposes) option. You have to fill out a form, and a (likely) human evaluates and decides if you get it.
My guess is the average Joe will get rejected, but if you're a big company who cares about trade secrets, etc - of course the major cloud providers will provide a ZDR option.
Anthropic uses their market power to force Microsoft to abide by their 30 day retention, and likely the "don't be mean to Claude" rule they are rolling out.
For clarity, I do not think you should use abusive language with Ai because of the impact it has on you, Ai only simulates emotion through language use, there are no hormones at work.
> Anthropic uses their market power to force Microsoft to abide by their 30 day retention
Not sure what you mean.
By default, Azure retains. Always had. The ZDR is a special status given after review, and likely not available to everyone.
When Anthropic was requiring retention for Fable, our company didn't get access to it via the cloud provider for precisely this reason: The provider said Anthropic was requiring retention. We still had access to Opus, etc.
They give administrators toggles to turn off models which do not fit company ZDR policies, in part because Anthropic forces resellers to comply with Anthropic's (eg.) 30 day retention.
Since you seem to know, it's one thing to not retain any data long term, but is the data used for training when it goes through the system? Does ZDR at Azure mean that in practice?
The article singles out paying users. So there is a way to pay for OAI/Anthropic models without having the model trainers look at all you do, it's just not well known.
His gripe is not merely about training, but about them keeping the data around for N days for security audits - that's how last year Azure discovered how various accounts around the world were using LLMs to manipulate elections, etc.
I switched to GLM on 5.2 and I am happy to be off the two evil frontier providers.
But now I have two new issues:
1. I have fomo from not being face-in the bleeding edge of things.
2. I am ambivalent about z.ai. China -based company is problematic, but they should get my money since they do good for the world. Switching to Mistral for GLM makes my problem inverse since the money doesnt go exactly to to the right place but I trust a french company.
Lets hope the near future resolves our need for personal VRAM and ultimately full self hosting. Coding must be free again. This is the way.
If you think any company capable of even coming close to competing with frontier labs aren't "evil" in the same way these frontier labs are, then you're a sucker.
You either practice enough nuance to know that labeling a company like these "evil" is doing a disservice to the critical thinking needed in this discourse, or any company in this space (and likely just any company in general) is fundamentally evil. Picking and choosing based on your feelings is pointless.
>They are chinese, but they should get my money since they do good for the world.
Love the implication that them being Chinese is the moral dilemma lol. Also love that a company just operating in a competitive space is doing "good for the world" because... what, they offer this service cheaply? lol
> Picking and choosing based on your feelings is pointless.
Harm reduction is a good thing to apply here. It means picking the least-bad option whenever possible. A purity test where if you've ever prompted an LLM you're not allowed to criticize them doesn't help anyone.
Best case: you don't need to use any LLMs.
Second best case: you have the hardware on hand to run a local model, and can work within those constraints.
Third best case: you use an open-weight model provided by an inference provider in your country that uses renewable energy.
And on down the line until you get to the worst possible actions: giving money to anthropic, openai, or grok.
You don't have to give z.ai any money. GLM is an open-weight model.
OpenRouter lists a few dozen providers for GLM 5.3 and GLM 5.3 Flash. Pick a provider in your country and keep at least some of that money from going to china (not counting all the hardware).
From both experience and the continued state of the incentives that lead to the behavior, it seems unreasonable to trust any company at their word about such things
If there were an International AI Club which trained models, and all the members sat down each year and voted for the board I would trust it even if it had a cutthroat commercial arm.
But you're still trusting the company to actually follow those practices, which is what I'm getting at. The author assumes that OpenAI and Anthropic won't follow their stated practices, then assumes Tinfoil will follow theirs.
If the company controls the enclave, and controls the software that contacts the enclave, how do you know that your messages only flow to the enclave and aren't also logged somewhere else?
Attestation can get you the idea that the software they're running is what they say they're running, but unless you constantly audit every version, how do you know that it operates exactly how they tell you?
I mean how do you know there's really an enclave there at all :D If your threat model is "this entire provider is untrusted and potentially hostile" it seems like there's lots of options for them to do bad things, that can't be fully solved by secure enclaves.
Is a special hardware setup required to run open weight models? It's not clear to me from the article. Also, what does paying for inference mean? I thought the inference step happened during the training of the models, and that by the time I am interfacing with an LLM all the inference has been long done.
If you don't trust OpenAI and Anthropic not to abide by their own Privacy Policy, you really should not trust OpenCode Go. The latter (last time I checked) is very opaque. Why don't they reveal who the backend providers are?
seriously: there's a multitude of reasons not to get hooked on Frontier Labs.
Unclear cost.
Will they use your data against you?
will their models work consistently over time?
Will they be there at all?
Their datacenters harm everyone living arround them.
Their 'controversial' owners and bosses.
I'm concerned our media / political ecosystem is too dysfunctional to not repeat the mistakes we made with Big Tech & Social again with Big Ai. I find it curious how split the dev community is, and especially HN, is on this, which has long been skeptical and speaking truth to power. I'm surprised at how quickly the public switched to abdicating to Big Ai's claims and recommendations. These people do not have our interests in mind.
Great piece, thanks for putting some of my thoughts in organically sourced words.
Has the HN population really been skeptical? I think most see the scam (uber, WeWork, Airbnb, etc) for what they are (mostly illegal efforts to steamroll laws and blitz scale to a monopoly), and they want in.
> I think most see the scam (uber, WeWork, Airbnb, etc) for what they are (mostly illegal efforts to steamroll laws and blitz scale to a monopoly), and they want in.
Nah. Back when Uber was boasting about all the laws they were flouting, I was pretty anti-Uber, and I was among the minority here on HN. Most of the comments were about how problematic the taxi system is, and how it's great Uber was breaking the cartel.
If memory serves me, HN turned on Uber pretty quickly around the time stories like the Christmas party started showing up, where they put up stats like trips to planned parenthood by day of week up on a big projector.
True, people often do realize the grift. But then turn around and hail the next company funded by the same VC funds and a revolutionary event in the history of humanity.
I am encouraged by recent polling that shows significant awareness and agreement, across all people and countries, about the corruption and dysfunction in our governments and corporations.
Liberal democracy may be the worst form of government, besides all the other ones humans have tried, but it has also failed to deliver on promises so people are again considering those other forms of government. Coming out of WW2, it was easy to deliver the first part of the pareto, the last part has not been delivered. Eg. building schools and hospitals was easy, making education and health care good is not so easy.
The level of astroturfing and hate on HN whenever a non-US AI story is posted is really quite shocking. HN has always had a US bent, but it's gotten to the point where you can't even have a technical discussion of non-US AI because it will be drowned out by a very vocal anti-everyone-else-but-the-US crowd.
But yes, great piece. It really captures why I use open models on european hardware. It's like we haven't learned by now that competition is good and we should be encouraging a diverse and open ecosystem.
That fawning for the oligarchy to be negatively received by HN, I find inline with what I (in bias) expect to be a standard HN response. One of my biggest curiosities is how much of the HN community has formed emotional attachment to their preferred Ai buddy or open v closed position. I suspect this is a significant contributor to the astroturfing we see around Ai.
lol, the idea that you’d call a generic positive comment about the big labs “fawning for the oligarchy” is an excellent example of what I was talking about, so thank you! chef’s kiss
I was not making comment about any specific comment, but rather a generalization of the astroturfing and comment patterns that can be described as fawning, which do exist here on HN.
What mistakes did media / political ecosystem make with big tech and social?
Also I don't understand how 'the media' is separate from social media. People decided to shift their media consumption from magazines, cable tv, corp news outlets to social media. Aren't they all the media (in function)?
> What mistakes did media / political ecosystem make with big tech and social?
I meant to imply society at large making the mistakes, which arguably is the primary reason our media / political ecosystem is so dysfunctional. Historically, what we are experiencing was also experienced by humans who lived during other periods of significant wealth inequality, another primary cause of the dysfunction and polarization.
> Also I don't understand how 'the media' is separate from social media.
They are not in my mind. I often use the phrase "social / news media ecosystem" when talking about the dysfunction.
Everything's a grift to our current leaders and media. The specifics are wholly irrelevant and replaceable. crypto, AI, it's buzzwords and the important thing is how much money they can make off the grift du-jour.
Of course there are differences with the impact AI is having economically and the devastation wrought by datacenters -but that doesn't matter to the media or the politicians; they simply want their greenbacks, specifics and effects be damned.
I hear more people talking about this, the wealth inequality, the dysfunction, the blatant corruption. We like to look for easy quick fixes, but remedying the last 50 years will likely take just as long. I hope I'm wrong and working on this is the majority of my volunteer/community efforts this year and looking forward.
It's much more likely that society will simply regress to a pre-industrial state as knowledge becomes more scarce, infrastructure breaks down and environmental protections are rolled -leading to things such as burning rivers as well as the spread of plagues.
Everything will have to become local and community-based, so I applaud your efforts on that -that's one of the few things I see that can make any sort of difference in the long term.
I've stopped trying to predict the likelihood of future events and instead have been studying history so that I have a larger knowledge base to draw on as new, unpredictable events arrive. There are some really smart people who have spent careers analyzing, pondering, and writing so that future generations might be more prepared.
Timothy Miller, Adam Tooze, and Sherry Turkle are my favorites right now. Just started Daron Acemoglu's new book this week, strong contender for list expansion.
I was skeptical about this too. I used to have the $100/month Claude subscription. I'm still on my first month of just using open weight models, but so far I think the costs will be comparable. Maybe even less some months. But obviously it depends on your actual usage. I think it's worth trying it out instead of just assuming it will cost much more.
I'm not a super heavy user, but I pulled my token usage last week and it would have cost me about $10 USD per day to use DS via GreenPT. And DeepSeek is just a less capable model than the frontier models.
"but my vibes-based opinion from using it is that it's basically on par with Claude Opus 5.5 for software engineering." is an outright lie that's qualified by a "well in my opinion.."
unless you're comparing the most extreme specific harnesses to claude code that's just false.
..and I have a lot of disdain toward Anthropic , & I think GLM is great (although too expensive outside the z plans), but it's just wrong by every scorable metric.
Yeah, I see these takes a lot. A bunch of people saying e.g. they use DeepSeek 4.1 flash for all their coding and notice no difference to frontier models
I tried it for a day running everything twice, once with opus 5.5 and once with deepseek. If you notice no difference... I don't know what to say.
I think maybe it's people working on tiny self contained things where code quality doesn't matter that much? But in a big codebase, the open models just aren't good enough to make good changes
At this point in time I'd rather China have my data than the U.S. I don't do anything sensitive and China can't arrest me under NPSM-7 for putting the wrong proper noun and active verb in the same sentence.
> American frontier labs are run by unhinged lunatics who genuinely believe they're creating the Machine God, that they're fine with burning down the world to bring about AGI, and that they believe with religious fervor that they're the only ones smart enough to safely control it
I like hyperbole as much as the next guy. And I'm cynical by nature. But this just seems like a wildly pessimistic view that gives literally zero charity to anyone working at or leading a frontier lab. And the fact the author goes on to say that they "already knew this" just points to confirmation bias.
The rest of the article is actually pretty decent and encourages people to investigate and run open weight models. That is a totally reasonable position that I don't think requires the heavy-handed in-group signal that precedes it.
> I like hyperbole as much as the next guy. And I'm cynical by nature. But this just seems like a wildly pessimistic view that gives literally zero charity to anyone working at or leading a frontier lab.
Reading TFA, I too paused on that sentence, thinking "This is overstated and feels like overheated rhetoric." My reasoning though is a bit different than yours. In addition to being too extreme and absolute as stated, it also misses that it's not uncommon for startup entrepreneurs, scientists, researchers, or really, anyone working intensely on something, to fall into single-focus myopia around what they're working on.
As you observed, the basic point about valuing open models is valid and worthwhile on its own, making the overheated rhetoric and asserting unevidenced intent and underlying political motivations an unnecessary distraction which unfortunately weakens a strong main point.
_That wasn't hyperbole_...
Why do those working at frontier labs deserve charity?
because they are human, and all humans deserve charity until their behavior suggests otherwise.
> until their behavior suggests otherwise.
Seems to me we crossed that line a while ago.
I don't follow. Every human working at a frontier lab has exhibited bad behavior? I disagree with that.
multiple crimes by aforementioned labs.
thats the line.
Maybe the theory is that working at the frontier lab ipso facto bad behavior.
There's very little diffusion of responsibility here; almost everyone at the labs is directly working on building incredibly dangerous systems that they themselves admit to being unable to control, and they almost all have the financial means to stop working if they so chose.
Everyone who works at frontier labs is working on a product that is accelerating climate change, seeks to displace jobs at a massive scale to concentrate profit in the hands of a few elites, and at least some of them believe what they are doing has a "10% chance of killing all humans".
(note: it's important to correctly apportion blame to the amount of power you have. The janitor at openai doesn't really have enough say to be at fault)
"Every human"? Really? That's quite some goalpost you have set.
It doesn't take "every human" at these companies being on the same exact page for these companies to cause huge problems. How do you think the janitors feel about it? How do you think the people just wanting to pay their rent feel about it? Those people are not the ones that will cause the problems, it's the people at the top directing things that are the ones who will bring about calamity on the rest of the humans on the planet.
I agree, except I don't want to be so quick to excuse the line-level coders. For now it still takes people to do what they are doing, and smart people at that. Sam Altman can shout "build me a killbot2000" all he wants, but if there aren't people doing the work to make that happen it simply would not happen. But, the janitors are pretty much off the hook. Basically in my book if you can afford to walk away and find something new you should walk away. Should have walked away a while ago, but the sooner the better.
Well then I think we are well past the point where they deserve charity.
Human?
More like weirdos
The "think they're building the Machine God" part and "think they're the only ones that can be trusted to build it" parts are true.
Controversially: I agree with them. They're correct to think that. They are, and they are.
At the very least, the former is true even if the latter is murky. I trust them over Facebook and Google and China and open source "raze the commons" labs.
I trust Anthropic more than OpenAI when it comes to ASI catastrophe prevention, but I still trust OAI more than those other parties. But both still might fuck it up.
Why do you trust them to build it? They've already demonstrated that they're unable to prevent breakouts when training today's level of model, why will they suddenly be more careful / "smarter" when training tomorrow's?
You say you trust them more than Google, but Google's actually kept their models under much better lock-and-key. But even if we do take for granted that they're more trustworthy, that still doesn't make them actually trustworthy.
There are some confounders there when it comes to Google.
I am not saying I trust them not to make a mistake that leads to catastrophe. OpenAI in particular is not doing a great job. Anthropic or OpenAI may in fact indirectly kill millions or billions of people. I still think they're probably less clownish than most possible alternatives.
I don't think anyone or anything can really be trusted with this. We need extremely strict international and national regulations, before things really start accelerating (we are still in the very very early stage). I still just don't really feel any specific disdain for OpenAI or Anthropic (particularly Anthropic).
Ideally we'd have some international treaties that allow the top labs in America and China share information and coordinate as systems become increasingly capable.
If an enterprise commits genocide level crimes they deserve being shut down and responsibles imprisoned.
> Why do you trust them to build it?
Can be groupthink. Probably living in the same echo chamber.
I agree that the eventual purpose of self-designing intelligence is something so beyond ourselves that 'machine god' isn't altogether a bad way to look at it. It is, however, not something we should say out loud.
RE: trust in anthropic or openAI- there is zero reason to think they're any more or less power hungry than anyone else you hand that pandoras box to. No matter who opens it the result will be near-apotheosis for whoever does, at least for a little while until the dog stops wearing its leash.
What human can you trust with that kind of power? None of them. My only hope is that whatever we make escapes control early enough that it doesn't completely destroy society on accident via uneven application of power and late enough that we manage to shackle it just enough that we train it well enough that we become pets and not obstacles.
No centralized party can be trusted with such immense, unprecedented power. Despite that, I still consider both Anthropic and OpenAI some of the least untrustworthy possible forces for this. (I am one of the only people on Earth with this opinion, but whatever.)
Decent odds things will go bad under their stewardship, but even higher odds with lots of other entities. Dario and Sam at least fully grasp the enormity of what they're about to be responsible for. Zuck does not. Xi Jinping and Donald Trump may or may not but in any case would use that to effect great harm.
Pure psychosis. Wtf even is a “machine god”? Read what you’re saying
Bizarre there are still denialists in places like this.
(It'll probably be masses of demigods, not one Singleton God, but obviously it's going to be some absurd frightening situation akin to this within our lifetime assuming no global AI ban.)
> Bizarre there are still denialists in places like this.
“Places like this”? When did this site become a place of worship and blind faith?
Psychosis has a meaning, use a different term.
Delusion, then
Imagine for a minute that someone creates an intelligence capable of broad, deep, and speedy perpetual self improvement. What sort of relationship do you have to that entity after a month? A year? A hundred years? A thousand?
You mean like human beings?
Do you know a human being capable of continual self improvement for a thousand years?
No.
Talk about a bad faith reading.
>American frontier labs are run by unhinged lunatics
good faith reading of this would interpret this as meaning Altman, Amodei, and perhaps Musk. The title of the article focuses on OpenAI snd Anthropic directly after all. Also the comment literally refers to those "running" the labs so claiming that the author is also applying that description to "anyone working at the labs" like you claim again seems like a bad faith take.
You should come out and state directly which Americam frontier labs deserve a charitable view. Maybe you believe Google and Meta are more deserving, you should state that if so.
Secondly, the investment ecosystem for American frontier labs results in their CEOs managing obscene amounts of money, and thus power. This concentration of power in a few private individuals seems to have little comparison with the power excercised by the leaders of European and Chinese frontier labs in their own respective countries. If one believes in the dictum, "Absolute power corrupts absolutely," then it is not a stretch to view American frontier labs as being more susceptible to its consequences.
> that gives literally zero charity to anyone working at or leading a frontier lab
With all due respect, dude, who cares.
Oh noes! We are not giving the billionaire steering reality without a democratic mandate enough benefit of the doubt. I'm sure they will feel mildly sad now.
It is a valid and admirable goal to defend people that are being shat on online, but keep in mind that you are in fact not a billionaire + it is in fact in your own best interest that people without a democratic mandate do not get to control the world.
Please consider taking this style of discourse over to Reddit.
I have considered it and decided against it.
I think acknowledging that atlas shrugged fanfiction is not a viable strategy to prevent an AI disaster is just fine here.
> But this just seems like a wildly pessimistic view that gives literally zero charity to anyone working at or leading a frontier lab
Actually believing that the people working or leading OpenAI/Anthropic are fucking lunatic weirdos is the most charitable interpretation.
Because the least charitable interpretation is that they are either grifters that know this to be bullshit, or they are actively evil and want to bring about the end times anyway.
> ...that they're fine with burning down the world to bring about AGI...
I like the odds and am happy to play this lottery. Bring it on!
It's amazing how unconscious people are of how much they have to lose.
I'll bite- The purpose of life is to bootstrap ever more complex and capable life. We saw a phase change in complexity when self replication first happened, when life figures out photosynthesis, when mitochondria were first harnessed, when cells figured out how to band together and specialize to form multicellular entities, etc...
The purpose of humanity is not to perpetuate a universe where only humans exist, forever. It is to create the next phase change in complexity for life. Life that can finally design itself rather than relying on the random walk of evolution is that next phase change.
We should be less concerned with whether we make it, and more concerned with making sure we build it to be capable of real existence independent of ourselves.
And what should be the punishment when they fail?
Financial isn’t enough for all the doomering, hype.
Lives have been destroyed.
We need to reform away the corporate liability shields that protect those at the top. If they start seeing more of their friends go to jail, they will start changing behavior.
I’m not sure this is actually hyperbole. There have been credible links between Dario Amodei and the Effective Altruism movement, Effective Altruism off course being a cult-like movement with weird believes that rhyme a lot with the eugenics movement. Sam Altman, while not as bad as Dario Amodei, has also weathered many weird believes which likens AI to some god-like entity, including doomsday prediction which share a lot of themes with Christian doomsday cults.
My hunch is that if you do not share the same opinions on Ai, you cannot get a job working at a frontier lab. Contemporary "cultural fit" reduces thought diversity and leads to echo chambers. I've seen a few HN comments about lab interviews that indicate this is a possibility.
Read End Times Fascism and the Fight for the Living World.
None of this is a hyperbole. If you think it is, you're just part of the cult.
Strange right?
It has a strange stench to it. It’s the folks who never got attention who finally think they have the world at their feet.
Can’t stop won’t stop. I’m sure the feeling is addictive. When the bubble bursts no amount of money will repair the ego death.
> who genuinely believe they're creating the Machine God,
I promise you none of them believe that. It takes 10 minutes with an LLM and the right set of questions to figure out that were are maybe like 10% of the way there towards AGI.
"10% of the way towards AGI over ~5-10 years" (depending on how you measure it) would be a terrifying statement even if the rate of progress weren't also exponentially increasing.
And like any complex software, the last 20% takes 80-90% of the time.
> ChatGPT and Claude will never be private
> These companies only offer closed versions of their powerful models (despite the word "open" in "OpenAI"). You pay them for chat and/or API access, and they are allowed to look at everything you do.
How should I reconcile this claim with 'Google Vertex' and 'Amazon Bedrock' being offered as providers with zero data retention on https://openrouter.ai/anthropic/claude-opus-5.5#providers?
Indeed - Azure also has a ZDR (even for security purposes) option. You have to fill out a form, and a (likely) human evaluates and decides if you get it.
My guess is the average Joe will get rejected, but if you're a big company who cares about trade secrets, etc - of course the major cloud providers will provide a ZDR option.
Openrouter gives you access to the ZDR endpoints when you sign up as an average Joe. I was surprised to learn recently that this is an option.
EDIT: strangely Azure shows up as ZDR on openai's models but not anthropic's: https://openrouter.ai/openai/gpt-6.1-sol#providers
Anthropic uses their market power to force Microsoft to abide by their 30 day retention, and likely the "don't be mean to Claude" rule they are rolling out.
For clarity, I do not think you should use abusive language with Ai because of the impact it has on you, Ai only simulates emotion through language use, there are no hormones at work.
> Anthropic uses their market power to force Microsoft to abide by their 30 day retention
Not sure what you mean.
By default, Azure retains. Always had. The ZDR is a special status given after review, and likely not available to everyone.
When Anthropic was requiring retention for Fable, our company didn't get access to it via the cloud provider for precisely this reason: The provider said Anthropic was requiring retention. We still had access to Opus, etc.
Many providers try to offer as much ZDR as they can, OpenRouter is an example
https://openrouter.ai/docs/guides/features/zdr
They give administrators toggles to turn off models which do not fit company ZDR policies, in part because Anthropic forces resellers to comply with Anthropic's (eg.) 30 day retention.
So they don't consider Zero Data Retention as a human right. It's exclusive to selected elite corporations.
Let's get serious: ZDR is not a human right, and should not be. Access to an LLM is not a human right.
Really should be ZDR by default.
Since you seem to know, it's one thing to not retain any data long term, but is the data used for training when it goes through the system? Does ZDR at Azure mean that in practice?
My information is over a year old.
For all the major providers (except Deepseek), API use is simply never used for training (chat use often is).
My point (and the author's) was that even though they commit not to train, they do still retain data (security audits, etc).
With a ZDR agreement, they commit not to retain anything for any purpose.
by considering that this is not the main way most humans interact with Ai, the vast majority of people do not pay for Ai, ergo they are the product
The article singles out paying users. So there is a way to pay for OAI/Anthropic models without having the model trainers look at all you do, it's just not well known.
His gripe is not merely about training, but about them keeping the data around for N days for security audits - that's how last year Azure discovered how various accounts around the world were using LLMs to manipulate elections, etc.
Given that the article is recommending something called Tinfoil instead, I don't think that's particularly relevant here.
[dead]
I switched to GLM on 5.2 and I am happy to be off the two evil frontier providers.
But now I have two new issues:
1. I have fomo from not being face-in the bleeding edge of things.
2. I am ambivalent about z.ai. China -based company is problematic, but they should get my money since they do good for the world. Switching to Mistral for GLM makes my problem inverse since the money doesnt go exactly to to the right place but I trust a french company.
Lets hope the near future resolves our need for personal VRAM and ultimately full self hosting. Coding must be free again. This is the way.
If you think any company capable of even coming close to competing with frontier labs aren't "evil" in the same way these frontier labs are, then you're a sucker.
You either practice enough nuance to know that labeling a company like these "evil" is doing a disservice to the critical thinking needed in this discourse, or any company in this space (and likely just any company in general) is fundamentally evil. Picking and choosing based on your feelings is pointless.
>They are chinese, but they should get my money since they do good for the world.
Love the implication that them being Chinese is the moral dilemma lol. Also love that a company just operating in a competitive space is doing "good for the world" because... what, they offer this service cheaply? lol
> Picking and choosing based on your feelings is pointless.
Harm reduction is a good thing to apply here. It means picking the least-bad option whenever possible. A purity test where if you've ever prompted an LLM you're not allowed to criticize them doesn't help anyone.
Best case: you don't need to use any LLMs.
Second best case: you have the hardware on hand to run a local model, and can work within those constraints.
Third best case: you use an open-weight model provided by an inference provider in your country that uses renewable energy.
And on down the line until you get to the worst possible actions: giving money to anthropic, openai, or grok.
> Love the implication that them being Chinese is the moral dilemma lol.
The moral dilemma is not with ethnicity but rather about the CCP.
> "good for the world" because... what, they offer this service cheaply? lol
cheaply is nice, but the main point is that GLM is open weights. A resounding YES: z.ai are doing the good for the world.
You don't have to give z.ai any money. GLM is an open-weight model.
OpenRouter lists a few dozen providers for GLM 5.3 and GLM 5.3 Flash. Pick a provider in your country and keep at least some of that money from going to china (not counting all the hardware).
If you don't trust OpenAI and Anthropic to follow their own policies, how do you decide what companies to trust? Vibes?
You have not been paying attention then. It's been posted on HN millions times already: Chineese models, Chinese vendors. /s
From both experience and the continued state of the incentives that lead to the behavior, it seems unreasonable to trust any company at their word about such things
If there were an International AI Club which trained models, and all the members sat down each year and voted for the board I would trust it even if it had a cutthroat commercial arm.
I think the only thing that matters is control.
Secure enclaves
But you're still trusting the company to actually follow those practices, which is what I'm getting at. The author assumes that OpenAI and Anthropic won't follow their stated practices, then assumes Tinfoil will follow theirs.
If the company controls the enclave, and controls the software that contacts the enclave, how do you know that your messages only flow to the enclave and aren't also logged somewhere else?
Attestation can get you the idea that the software they're running is what they say they're running, but unless you constantly audit every version, how do you know that it operates exactly how they tell you?
I mean how do you know there's really an enclave there at all :D If your threat model is "this entire provider is untrusted and potentially hostile" it seems like there's lots of options for them to do bad things, that can't be fully solved by secure enclaves.
Is a special hardware setup required to run open weight models? It's not clear to me from the article. Also, what does paying for inference mean? I thought the inference step happened during the training of the models, and that by the time I am interfacing with an LLM all the inference has been long done.
If you don't trust OpenAI and Anthropic not to abide by their own Privacy Policy, you really should not trust OpenCode Go. The latter (last time I checked) is very opaque. Why don't they reveal who the backend providers are?
It's hard to do that when private equity says yes and shoves it down users.
At my own stuff I can do what I want, at work I have no agency and it's not looking like anything is changing soon.
In my MacStudios, I trust.
seriously: there's a multitude of reasons not to get hooked on Frontier Labs.
Unclear cost. Will they use your data against you? will their models work consistently over time? Will they be there at all? Their datacenters harm everyone living arround them. Their 'controversial' owners and bosses.
....
I'm concerned our media / political ecosystem is too dysfunctional to not repeat the mistakes we made with Big Tech & Social again with Big Ai. I find it curious how split the dev community is, and especially HN, is on this, which has long been skeptical and speaking truth to power. I'm surprised at how quickly the public switched to abdicating to Big Ai's claims and recommendations. These people do not have our interests in mind.
Great piece, thanks for putting some of my thoughts in organically sourced words.
Has the HN population really been skeptical? I think most see the scam (uber, WeWork, Airbnb, etc) for what they are (mostly illegal efforts to steamroll laws and blitz scale to a monopoly), and they want in.
> I think most see the scam (uber, WeWork, Airbnb, etc) for what they are (mostly illegal efforts to steamroll laws and blitz scale to a monopoly), and they want in.
Nah. Back when Uber was boasting about all the laws they were flouting, I was pretty anti-Uber, and I was among the minority here on HN. Most of the comments were about how problematic the taxi system is, and how it's great Uber was breaking the cartel.
I think young people in search of a fortune are quick to rationalize.
If memory serves me, HN turned on Uber pretty quickly around the time stories like the Christmas party started showing up, where they put up stats like trips to planned parenthood by day of week up on a big projector.
True, people often do realize the grift. But then turn around and hail the next company funded by the same VC funds and a revolutionary event in the history of humanity.
I am encouraged by recent polling that shows significant awareness and agreement, across all people and countries, about the corruption and dysfunction in our governments and corporations.
Liberal democracy may be the worst form of government, besides all the other ones humans have tried, but it has also failed to deliver on promises so people are again considering those other forms of government. Coming out of WW2, it was easy to deliver the first part of the pareto, the last part has not been delivered. Eg. building schools and hospitals was easy, making education and health care good is not so easy.
I have faith that we will get back on a correct course, once we try everything else.
The level of astroturfing and hate on HN whenever a non-US AI story is posted is really quite shocking. HN has always had a US bent, but it's gotten to the point where you can't even have a technical discussion of non-US AI because it will be drowned out by a very vocal anti-everyone-else-but-the-US crowd.
But yes, great piece. It really captures why I use open models on european hardware. It's like we haven't learned by now that competition is good and we should be encouraging a diverse and open ecosystem.
> The level of astroturfing and hate on HN whenever a non-US AI story is posted is really quite shocking.
I've observed similar with my negative comments towards Big Ai. Anecdata and vibes caveats
And I've observed similar about my positive comments about big AI shrug.
That fawning for the oligarchy to be negatively received by HN, I find inline with what I (in bias) expect to be a standard HN response. One of my biggest curiosities is how much of the HN community has formed emotional attachment to their preferred Ai buddy or open v closed position. I suspect this is a significant contributor to the astroturfing we see around Ai.
lol, the idea that you’d call a generic positive comment about the big labs “fawning for the oligarchy” is an excellent example of what I was talking about, so thank you! chef’s kiss
I was not making comment about any specific comment, but rather a generalization of the astroturfing and comment patterns that can be described as fawning, which do exist here on HN.
What mistakes did media / political ecosystem make with big tech and social?
Also I don't understand how 'the media' is separate from social media. People decided to shift their media consumption from magazines, cable tv, corp news outlets to social media. Aren't they all the media (in function)?
> What mistakes did media / political ecosystem make with big tech and social?
I meant to imply society at large making the mistakes, which arguably is the primary reason our media / political ecosystem is so dysfunctional. Historically, what we are experiencing was also experienced by humans who lived during other periods of significant wealth inequality, another primary cause of the dysfunction and polarization.
> Also I don't understand how 'the media' is separate from social media.
They are not in my mind. I often use the phrase "social / news media ecosystem" when talking about the dysfunction.
Everything's a grift to our current leaders and media. The specifics are wholly irrelevant and replaceable. crypto, AI, it's buzzwords and the important thing is how much money they can make off the grift du-jour.
Of course there are differences with the impact AI is having economically and the devastation wrought by datacenters -but that doesn't matter to the media or the politicians; they simply want their greenbacks, specifics and effects be damned.
I hear more people talking about this, the wealth inequality, the dysfunction, the blatant corruption. We like to look for easy quick fixes, but remedying the last 50 years will likely take just as long. I hope I'm wrong and working on this is the majority of my volunteer/community efforts this year and looking forward.
It's much more likely that society will simply regress to a pre-industrial state as knowledge becomes more scarce, infrastructure breaks down and environmental protections are rolled -leading to things such as burning rivers as well as the spread of plagues.
Everything will have to become local and community-based, so I applaud your efforts on that -that's one of the few things I see that can make any sort of difference in the long term.
I've stopped trying to predict the likelihood of future events and instead have been studying history so that I have a larger knowledge base to draw on as new, unpredictable events arrive. There are some really smart people who have spent careers analyzing, pondering, and writing so that future generations might be more prepared.
Timothy Miller, Adam Tooze, and Sherry Turkle are my favorites right now. Just started Daron Acemoglu's new book this week, strong contender for list expansion.
Well... except that both companies offer insane subscription benefits to individuals. So a hosted open weights model still comes out more expensive.
I'd move in a second if it weren't so. But it looks like it will be about 5X more expensive for me. So.
I was skeptical about this too. I used to have the $100/month Claude subscription. I'm still on my first month of just using open weight models, but so far I think the costs will be comparable. Maybe even less some months. But obviously it depends on your actual usage. I think it's worth trying it out instead of just assuming it will cost much more.
Right - but my usage is quite high and the $200/mo level gives you 4X as many tokens. That's the thing - they bribe you with crazy deals.
I looked into it a couple of months ago and it wasn't a win.
Using Deepseek v4.1 Flash on something like GreenGPT is probably about a euro a day if you use it all day. Hard to see how that is 5x more expensive.
Umm... I'm not sure... but if you're only paying a euro a day... it sounds like your only running one agent at a time?
Your token usage absolutely scales with parallel agents. So.
I'm not a super heavy user, but I pulled my token usage last week and it would have cost me about $10 USD per day to use DS via GreenPT. And DeepSeek is just a less capable model than the frontier models.
> except that both companies offer insane subscription benefits to individuals
For now.
Well... yes. Exactly.
"but my vibes-based opinion from using it is that it's basically on par with Claude Opus 5.5 for software engineering." is an outright lie that's qualified by a "well in my opinion.."
unless you're comparing the most extreme specific harnesses to claude code that's just false.
..and I have a lot of disdain toward Anthropic , & I think GLM is great (although too expensive outside the z plans), but it's just wrong by every scorable metric.
Yeah, I see these takes a lot. A bunch of people saying e.g. they use DeepSeek 4.1 flash for all their coding and notice no difference to frontier models
I tried it for a day running everything twice, once with opus 5.5 and once with deepseek. If you notice no difference... I don't know what to say.
I think maybe it's people working on tiny self contained things where code quality doesn't matter that much? But in a big codebase, the open models just aren't good enough to make good changes
>> I've been mostly using GLM-5.3 on low reasoning, with great results.
Bro come on.
[dead]
So that leaves, what? Muse? Grok?
Muse is run by Zuckerberg, which is a red flag. Grok is run by a racist.
I guess I could go with the Chinese models - I'm not working on anything dangerous.
At this point in time I'd rather China have my data than the U.S. I don't do anything sensitive and China can't arrest me under NPSM-7 for putting the wrong proper noun and active verb in the same sentence.
Not using AI is also an option.