Anthropic IPO filing will show AI backlash as a risk factor, sources say

34 points by newsomix9xl 6 hours ago on hackernews | 76 comments

1123581321 | 5 hours ago

This is a non-story; for those who haven’t read them, prospectuses include every potential risk under the sun. There will be risks that never make the news alongside the exciting backlash.
“Everyone hates us and the very fundamentals of what we are promising, will fight to prevent us establishing infrastructure, may pass laws to limit our activities in their states, and will extend that hatred to our major investors, potentially materially damaging their businesses” is not a common thing in an IPO prospectus, I imagine, for quite obvious reasons.

But that is the gist of the situation is it not? If it’s an honest statement, it is going to say something like this.

Which prospectus got close to anything like this? People hate SpaceX products less.

montagg | 5 hours ago

The non-story part is that they would include this risk even if it were 5 people.

sumeno | 5 hours ago

I could easily find 5 people who hate SpaceX and yet that wasn’t mentioned as a risk

I think you’re going to have to provide evidence that this is a commonly documented risk

jmalicki | 5 hours ago

"Our AI products and X platform are subject to complex and evolving U.S. and foreign laws and regulations regarding privacy, cybersecurity, data use, data combination, data protection, content, AI, competition, youth protection, safety, consumer protection and notification, advertising, e-commerce, sanctions, export controls, and other matters. Many of these laws and regulations are subject to change and uncertain interpretation, and we could be required to make changes to our products and business practices, and be exposed to monetary penalties, increased cost of operations, declines in user growth or engagement, or loss of customers, or other harm to our AI products and X platform. "

"Certain of our AI products, including Grok, offer features or modes designed to generate more candid, direct, or less reserved or irreverent outputs, such as “Spicy” Imagine Mode and “Unhinged” Voice Mode. These features are intended to provide users with greater flexibility and control in how they use our tools. Because these modes may be more irreverent and harsher than our standard offerings, they present heightened risks, including reputational harm, the generation of potentially explicit content and misinformation or deceptive outputs, potential nonconsensual or exploitative imagery, intellectual property infringement, or content that could be viewed as exploitative, harmful, harassing, abusive, or discriminatory. The availability of such features may also increase the risk of regulatory scrutiny, enforcement actions, litigation, or claims of harm, as well as reputational damage, user or advertiser backlash, or limitations on our ability to distribute or monetize our products in certain jurisdictions or through certain partners."

Or the wonderful:

"We, Mr. Musk, and other companies Mr. Musk is affiliated with frequently receive an immense amount of media attention. The actions and statements of Mr. Musk and his affiliated ventures, whether or not directly relating to us, may draw significant public attention and scrutiny to us and could potentially have a positive or negative impact on our business, relationships with customers and regulators, or stock price."

1123581321 | 4 hours ago

Thank you. Hopefully a good lesson for the GP to actually read prospectuses before making claims about them!

sumeno | 4 hours ago

I should have specified the space part of SpaceX, of course the AI part has the same risk.

1123581321 | 3 hours ago

Market acceptance risk, geopolitical funding risk, reputational harm risk, for space flight and starlink specifically. The Musk risk section isn’t specific to AI.

jmalicki | 5 hours ago

Uber and Airbnb both had similar risks in their S1s, for obvious reasons.

cavem0nkey | 5 hours ago

And both have been damaging to society. And are getting fined and chased out of certain countries.

Why would you throw investment in this unless you’re just leveraging the growth in the time it takes for regulatory issues to appear? Which is terribly unethical.

Wall to wall bastards. I’d rather die poor than contribute to the further decline of society.

They have but this reply was a fully fair answer to my point; perhaps there are more.

[OP] newsomix9xl | an hour ago

Maybe those who seek and profit from the decline in our society also want to promote divisiveness?

Maybe divisiveness itself, is part of the problem?

cavem0nkey | 52 minutes ago

Yep. We’re tribal apes. This is just a foot on the accelerator.
I was thinking maybe Uber but I was not sure it could have been that bad. They would have had states pushing back but against a very popular product with users.

But Airbnb, good point: they would have talked about the risks of states curtailing their businesses.

It’s still a very unusual level of hatred, I would say.

jmalicki | 5 hours ago

Uber called out that they were aggressively doing an end run around both taxi licensing AND minimum wage laws and facing significant pushback on both fronts, among many other risks.

cavem0nkey | 4 hours ago

I think there’s a genuinely huge divisive opinion on AI companies. We don’t see this due to our particular industry echo chambers but really the majority of society has bigger problems and this offers them no solutions and threatens to take their lifelines away.

Divisive technology leads to hate. Flock as well which is an us and them. Same mindset.

We are not as popular as we think we are. I know which side of the fence I want to be on.

[OP] newsomix9xl | an hour ago

Well said.

The divisiveness is not inherent in the technology. Other forces are driving divisiveness around issues: AI seems to simply be caught up in that trend.

I say "AI IS OK!"

cavem0nkey | 53 minutes ago

Well it’s a little more complicated than that. It doesn’t just manifest itself. People create it. And people do that for a reason. So there is intent. And that means it’s not naturally inert and consequence free and “ok” in a social construct.

jmalicki | 51 minutes ago

AI is pretty universally hated, it might be the one thing that large swaths of society have unity about! It's right up there with thinking babies and puppies are cute.

[OP] newsomix9xl | 24 minutes ago

I am willing to concede that AI has a negative reputation, but what I'm saying is that it shouldn't.

AI is a neutral technology.

The telephone (whose automation caused massive layoffs of telephone operators btw) can be used to order a pizza, talk to your mom, or facilitate drug trafficking and organized crime.

No one complains about telephones. It's a utility. AI also is a utility.

csallen | an hour ago

I think phones were similarly hated, but diluted over a ~15 year period as they were gradually introduced to society. AI is happening much faster, practically all at once.

And I think the extent of the hatred in part is also due to how widespread it is, not just how fast it's spread. It's everywhere. If you watch YouTube, you're hearing people narrating scripts read by AI. Social media posts are written with AI. Students are cheating with AI. AI is threatening to take jobs from knowledge workers. AI is copying the work of artists and creators. Etc. There's almost nothing it doesn't touch.

(I'm not talking here about why it's hated or whether it should be hated. More just commenting on the remarkability of how widely it's disliked.)

[OP] newsomix9xl | an hour ago

Its not just pervasiveness, but how it's characterized. "AI data centers are raising electricity rates." or "AI pivot is the reason for 10% layoffs". It's rare that you read "AI cures cancer" or "AI reducing inflation" or "AI raising standard of living for blue collar workers".

The relentless drum beat on AI news is negative even when there is zero - literally zero - evidence around the speculation. To wit: AI will destroy all of humanity, or the related "AI will become sentient and lead to threat to human civ" etc. This is pure negative science fiction (related to Musk's AI will create a utopia in which everything will be free).

> To wit: AI will destroy all of humanity, or the related "AI will become sentient and lead to threat to human civ" etc. This is pure negative science fiction

All of this is true, but it is the two main AI companies themselves who created and marketed this narrative.

It's their negative science fiction. It is part of the marketing campaign they are running that is the foundation of arguments that the government should be their funding backstop (OpenAI have made this explicit), that open weights AI is more dangerous, that China is a threat, that only they can solve the problems they themselves are creating.

(After all, if China can only do what they are doing by distilling American models, which is what they claim, then the big American AI labs could shut down the risk for decades by downing tools out of principle.)

This whole fear scenario is more or less the entire basis for Anthropic's original elevator pitch, after all: that AI is dangerous and that they will specialise in making it safe.

Churlish to blame the media for accurately reporting what Musk, Amodei and Altman have said and the sort of blasé, detached, sociopathic way they say it.

jryle70 | 5 hours ago

"Everyone hates us but they can't have enough of us. Our revenue can't stop rising"

Anthropic's credit line is now $10B - [0], and the IPO is expected to beat or match SpaceX - [1]

[0] https://finance.yahoo.com/technology/ai/articles/anthropic-p...

[1] https://www.bloomberg.com/news/articles/2026-08-20/anthropic...

Not what we are discussing.

nozzlegear | 3 hours ago

This doesn't really refute what they said. The fact that they make a lot of money doesn't mean the public doesn't hate them or that they won't find themselves facing anti-AI backlash in the near or far future.

jryle70 | 2 hours ago

Huh? "Everyone hates us and the very fundamentals of what we are promising"

First of all everyone? Based on what?

Second of all "they hate our guts and the core of what we're doing (fundamentals), but they can't have enough of us". Are you sure the society (everyone) is that self loathe or HN is in a giant bubble?

AI clearly causes disruption, anxiety, uncertainty, and will be negative for certain groups, just like anything with major impacts in the history. But hate by everyone? people who consider themselves smart and well informed should really be more objective and level headed.

bragr | 5 hours ago

I'm sure you can find similar stuff in the prospectuses of many controversial companies spanning oil, cigarettes, alcohol, cannabis, adult entertainment, guns, gambling, defense, etc... Anything someone is trying hard to ban.

[OP] newsomix9xl | 5 hours ago

Why is AI in the category of oil, cigs, and beer? Its software ....

bragr | 4 hours ago

Social media and gaming companies probably have it too

dragontamer | 4 hours ago

The electricity usage is off the charts. Every place that gets a data center has their electricity prices go up.

Consumers are now fighting to stop data centers from coming into their hometowns. It's not like Data centers have lots of jobs that benefit a community anyway. It's basically a building that eats a local city's resources and gives barely anything back in return.

At best, an AI data center is a year job for the local blue collar workers. A temporary benefit at best, while the price of electricity rises will negate any local benefits by a long shot.

Plus water and noise, and other such issues.

I think that was maybe a rhetorical question, but that indeed sums it up.

ToValueFunfetti | 4 hours ago

Just read this interview[1] the other day that contends that the blue collar workers, or at least their union leaders, are thrilled about the data center boom. There are loads of these projects underway and they show no sign of stopping, so the fact that it's only work for a few years per building hardly matters. Plus, the pay is great because the projects have to fight each other for labor. It also says that localities competed for these projects because of the substantial tax revenue they'll bring in for the communities.

Separately, while electrity usage is a valid concern with respect to pollution, data centers are not really driving electric costs up for residents, with (maybe?) a couple exceptions. Noise is likewise a non-issue by-and-large with (definitely existent) rare exceptions. And the water usage issue is totally invented; data centers use a tiny fraction of the water other industries that would otherwise occupy industrially-zoned land use.

You didn't mention it, but the interview also contends that the "this will take our jobs" angle that gets passed around here as a reason isn't on anybody's mind.

Instead, it seems that the AI companies used NDAs to prevent localities from efficiently pricing their deals. That's obviously sketchy on its own and worthy of attack, plus as a bonus it prevents governments from making statements to clarify that the populist narratives that surround the things and are, bluntly, false.

[1] https://www.natesilver.net/p/why-does-everyone-hate-data-cen...

nozzlegear | 3 hours ago

This is a very long interview, and I really don't think it's as rosy for AI as you're painting it. Jasmine makes it clear throughout that people are very concerned about data centers in their areas, and AI in general.

From the comments on that post, Rachel says:

> With all due respect to Nate and Jasmine I think they are overthinking why people hate data centers - it’s a fairly direct answer. People hate data centers because they are a tangible manifestation of a technology that is expected to take away their jobs and lifestyle, mess with their societal place and sense of self, and uglify their neighborhoods and the natural world, all while providing little obvious benefit to the regular person and the tantalizing prospects of all the weird elites they hate the most becoming brains in a jar who lord over them forever.

Bingo.

ToValueFunfetti | 2 hours ago

I read that comment as well at the time and was taken aback. After such a lengthy discussion where Jasmine went and talked to the people doing the disagreeing over multiple weeks and reported her findings, Rachel replies that no, this is wrong, her presumably pre-existing belief is correct, and she cites nothing at all in support of this. Due respect seemed lacking despite claims to the contrary.

Likewise, I am taken aback that you, having apparently read this lengthy interview to find it in disagreement with my summary, elected instead to quote Rachel's unsupported comment dismissing the whole thing. Why not just quote the journalist disagreeing with me instead of some rando disagreeing with her? Does Rachel have a reputation that I'm unaware of?

nozzlegear | 27 minutes ago

I feel like we read two different articles. Both Nate and Jasmine think the outlook for AI looks grim. They both think that the locals are right to protest, they think that the concern about AI and its environmental impact is real, they both think that AI is politically toxic and SV is out of touch.

Meanwhile, your comment downplays all of the legitimate concerns that Jasmine discusses and worse, even invents concerns to downplay as if Jasmine had reported on them. In particular, you said that noise is a non-issue – that's not discussed at all in the interview (CMD+F "noise" – zero results).

You said that data centers "are not really driving electric costs up for residents," but Jasmine doesn't say that at all. Instead, she says that power and electricity are the most serious of everyone's environmental concerns, and that electricity rates going up is just one of the concerns that motivates opposition.

You said that the "they will take our jobs" argument is not on anybody's mind, which is literally false. Jasmine says job displacement wasn't the dominant explanation that people gave for opposing local data centers, but they both repeatedly discuss job displacement. Jasmine even quotes people she interviewed as saying "my boss keeps making me use AI, I have to say that I’m using AI or people won’t resource my project."

You said that NDAs prevented governments from clarifying "bluntly false" narratives, but that's, at best, cherry-picking what Jasmine was saying. You're right that NDAs were used and officials couldn't correct rumors about deals or water usage, but the interview does not say that the concerns people had were just straight up false. She notes that they had real concerns like grid strain, gas generation (from the power plants they're putting on data center property), ugly facilities, temporary jobs, subsidies, secrecy, and an unequal distribution of the benefits.

> Likewise, I am taken aback that you, having apparently read this lengthy interview to find it in disagreement with my summary, elected instead to quote Rachel's unsupported comment dismissing the whole thing. Why not just quote the journalist disagreeing with me instead of some rando disagreeing with her? Does Rachel have a reputation that I'm unaware of?

I find it strange that you're putting more stock in a substack journalist from the Bay area who went to Wisconsin to figure out why people are doing something, than the people who live there and are doing that thing. Do the people who live in the places where data centers are going up have no value to their opinions? Does Rachel need her own Substack before we can take her bonafides seriously?

I'm taken aback.

[OP] newsomix9xl | 2 hours ago

The electrical price argument I find unconvincing.

If this is a shortage of electrical capacity the data center didn't arrange that. I'm pretty sure they want plentiful cheap electricity.

And the data center doesn't set the price: that's the utility and regulators.

Yet the consumer is blamed for poor regulation and low supply?

jmalicki | 4 hours ago

Because oil, cigs, and beer don't create the existential threat to the human race that Anthropic is trying to claim it creates.

Yes, it is software, something more powerful and thus potentially more harmful than any of the above physical products.

[OP] newsomix9xl | an hour ago

I don't buy the "existential threat" argument for a second.

Nuclear weapons are an existential threat. AI is not. If AI can control nuclear weapons, then maybe. I don't believe any nuclear power has added an AI interface to their nuclear controls.

Until that time, it's general purpose software and is mostly useful and mostly benign. That's my take.

seattle_spring | 4 hours ago

Is your comment meant to be making fun of the people who think AI is doing no harm?

The world is considerably worse because of AI than the effects of beer and cigarettes. No, I'm not joking.

[OP] newsomix9xl | 56 minutes ago

Last time I checked alcoholism is heavily linked to cancer, crime and violence, birth defects, and poverty. Cigarettes cause cardiovascular issues and preventable deaths.

I am not aware of a single death due to AI, or any cancer, or physical harms. Maybe I missed it?

I believe AI has "facilitated fraud" by allowing already existing bad actors to make more plausible scams, but that is not an AI harm, any more than the phone company being used by the Mafia to order murders being a "telephone harm" (both are tragic, as should go without saying).

There are claims that AI has harmed people by encouraging their delusions. I am not convinced that these claims have merit (maybe they do, who knows?).

AI has apparently enabled cheating on a vast scale and test scores have gone down. I dont think that's an AI problem: that's an ethics and school administration problem (some teachers create AI traps to punish students, or have live tests, etc to stymie AI cheating - which is to say they actually adapt to the technology landscape).

I am not making fun of anyone. AI is not harming society, threatening civilization, raising electricity prices, or threatening massive job loss. Seriously.

Betelbuddy | 4 hours ago

Well there is a story, and the story is we are going to have an Anthropic IPO? Believe it when I see it...

mihaelm | 4 hours ago

We've known Anthropic has filed for an IPO since early June.

Betelbuddy | 4 hours ago

One risk they should include is: Our vision is working on a product and business model, that if we succeed, will make our business model and product obsolete.

JacobAsmuth | 4 hours ago

Not a risk.

1123581321 | 4 hours ago

I expect this will effectively be stated, that their success will depend on developing a succession of products and evolving the company in response to the market climate.

jqpabc123 | 5 hours ago

negative sentiment towards AI and data centers will be listed as a risk factor

The fact that the technology can "hallucinate" and is unreliable is not just a "sentiment" --- it is a well established fact. And it is more than a risk, it is virtually guaranteed to happen if you use it repeatedly.

Calling it "sentiment" is like an "alternative fact".

uejfiweun | 5 hours ago

...what? I don't think they're talking about hallucinations here, they're talking about public opinion.

jqpabc123 | 5 hours ago

Is it just an "opinion" if it is based on proven facts?

LPisGood | 5 hours ago

Yes. Let’s say I saw a person kill someone. That’s a fact, but my perception of them would be low; that is an opinion.

I think it’s really swell how the last 10 years have made it so the basic difference between fact and opinion is not totally obvious.

jqpabc123 | 5 hours ago

We're talking stated "risk" to the public.

Using your example --- as far as "risk" to the public is concerned, which would be more pertinent info --- the fact that this person is a murderer or your low opinion?

bluefirebrand | 5 hours ago

"Public Opinion" -

Public opinion, or popular opinion, is the collective opinion on a specific topic or voting intention relevant to society. It is the people's views on matters affecting them (From Wikipedia)

It has nothing to do with fact. It broadly represents peoples views, factual or otherwise.

paul7986 | 5 hours ago

Tell that to graphic designers as do you think that's a career a student going into college should pursue? Same with web design and web development (AI companies now are trying / working on shifting us away to less screens/more voice).... like would you tell a freshman in college to pursue those degrees vs. medical stuff and or skip college altogether, rather pick up tools to help build those data centers?

simianwords | 5 hours ago

its a bit sad that the public can't grasp tools that have high variance but still end up being useful. lots of things in real life follow the same pattern yet when it comes to AI..

seattle_spring | 3 hours ago

AI backlash is not because of "high variance"...

simianwords | 2 hours ago

i didn't say it was

cm2187 | 5 hours ago

Not sure what you are talking about. The negative sentiment isn't about hallucinations, it is about job replacement and high electricity cost.

watwut | 5 hours ago

And water. And oligarchs relations to Trump administration. And it way AI was forced everywhere. And how they get massive investments while rest of economu suffers. And doomtrolling. And how unappealing their culty vision of the world is. And plan to create monopoly. And how they claim to be creating doom machine.

Like ... basically everyone can pick their reason to hate.

azan_ | 5 hours ago

Isn’t data center water usage actually non-story?

watwut | 4 hours ago

It is a story. It is not global story, water usage is highly local.

But yes, they are building or want to build water thirsty datacenters at locales with water issues. So, yes, it is a story.

iLoveOncall | 4 hours ago

I said it [1][2] when OpenAI was supposed to go public "next week" and I'll say it again today: neither Antropic nor OpenAI are going public in the coming few years.

The moment they go public is the moment the music stops and everyone who still has their head in the sand realizes that they now can't deny there's no money to make in selling AI models, only money to lose.

[1] https://news.ycombinator.com/item?id=48211272

[2] https://news.ycombinator.com/item?id=48358945

I rather think OpenAI will never get there.

I think the AI industry desperately needs one of the two to get to IPO for its own sense of identity, though; the morale boost and the presumption of validation are crucial at this point.

I think it’s possible that the remaining power of the delusion gets Anthropic to IPO.

4fffaa | 4 hours ago

Lol nobody's buying the BS.

It is all about greater fool theory now. The substance OAI and Anthropic had is gone.

danilocesar | 4 hours ago

++

I don't know if they will never go public,but their business is clearly not profitable as it is today.

I'm a heavy user but the moment they stop subsidizing I will pick something else because I know how damn expensive it actually is (my usage is ~4k usd/month, but paying 200 in their max plan).

I guess lot of people will do the same.

4fffaa | 4 hours ago

The funny thing is the govt would prob prefer all the resources were concentrated (e.g. across MSFT and Alphabet) rather than diversified across many firms in a head to head fight vs China.

So therefore theres no argument to keep anthropic and OAI afloat if they cannot do so on their own.

nswizzle31 | 4 hours ago

Anthropic is absolutely going public now that Leopold got taken for a ride and the market is back at its very tip top. There is no other option at this point. It is imminent.

What you get wrong is who the music stops for. It stops for us (retail). They dump their bags and consolidate into a more manageable business in the long run. S&P drops 10% and life moves on in a few years.

I think you’re right about OpenAI though. They got outplayed at every step and wont go public.

mimischi | 4 hours ago

Care to explain how this relates to Leopold? I know they hold some stake in Anthropic, but I’m not proficient enough in the topic to jump to the conclusion you’re alluding to.

nswizzle31 | 4 hours ago

Sure, a month ago the market was roaring and Anthropic first floated going public. Then suddenly all AI-related stocks drop 30%+ over a week and people claimed the jig was up. No chance for Anthropic IPO.

Turns out it was mainly just caused by one hedge fund (Citadel) targeting another hedge fund (Leopold). He got liquidated and Citadel picked up his portfolio for pennies. Market subsequently soared to even higher highs in the last few weeks and, poof, Anthropic is going public again.

That’s the narrative anyways - the big drop was just Leopold getting liquidated and not a larger market downturn.

downrightmike | 3 hours ago

It all implodes once they can't afford to lock up all the RAM and SSD, and/or when the reduced resources environment(private/China) models get good enough to negate frontier lead / become equivalent (almost there)

danilocesar | 4 hours ago

Interesting how that whole conversation about "AGI tomorrow" completely disappeared and it's all about making money now.

I'm a fan of claude and I have been using their top tier plans because it's so damn cheap comparing to my real usage. They will have to stop subsidizing this at some point, and people (include me) will flee. What they are gonna do?

Interesting times we live in.

VCFundedGenYer | 4 hours ago

Sam Altman being like "AGI in <6 months/8 months/2023/2024/2025/2026>" is just sad to listen to now. They don't have it, and never will.

joebuckwilliams | 4 hours ago

The bigger risk is going to be the fundamental inability to meet their buildout commitments. There’s not enough money in the world.

jiaosdjf | 4 hours ago

The risk is that we rise up and violently lynch the elites when they try to implement autonomous AI mass surveillance backed by an army of robots to quell the growing unrest of unemployed "useless eaters".

It's a big risk and the only ways to mitigate it are either:

a) build the dystopian system quickly and violently killing all dissenters on sight

b) build a utopian system where robots build our houses, grow our food and serve us 24/7

Just a hunch they'll try option A.

vrganj | 3 hours ago

Socialism or Barbarism! - Rosa Luxemburg

[OP] newsomix9xl | 51 minutes ago

How about

scenario c) Mass surveillance is used to reduce crime in high crime areas usually ignored by police, and public safety is provided to all citizens, even the poor and working class?

How about the army of robots helps to clean the streets, provide services to the poor and working classes, and provide high quality low cost housing, clean water, and livable cities?

bigtex | 4 hours ago

I would think profitability would be the biggest risk factor