Hacker News new | past | comments | ask | show | jobs | submit
And you need to factor in Anthropic's growth.
> And you need to factor in Anthropic's growth.

The problem with growth as a metric is that you can more-or-less linearly trade cash for growth. This is the same game that Uber and WeWork played, giving away product below market for years, in a straight trade of investor cash for growth

And it worked out very well for Uber?

For anthropic in first half of 2026 their actual spend on commute is 8.9B$ while revenue is 16.3B$.

Not worthy 1.5T$ valuation still in many people opinion, but these are very good numbers compared to companies that went bust in dotcom period

loading story #50562435
loading story #50563605
That would be true, except Anthropic has a pretty sizeable margin.

edit - Obviously fixed cost of training is a thing, but with more token usage / more customers that gets amortised out.

Yes this is like just how it works..
That’s true; public sentiment about AI is already trending downward, and I suspect we’ll reach a point where these companies start experiencing steady negative growth, once people realize the tech is only ever going to be great for certain specific use cases (e.g. coding). So Anthropic seems equal to or worse than AltaVista from that perspective.
I don’t know, every professional lawyer, doctor, accountant, nurse etc is using AI now in some way. Every person under the age of 60 is using it.

It’s going to be embedded in our professional class and we will have to see who wins. I agree there isn’t as much of a moat for any specific model here, it’s really the last mile who will capture the user/integrations.

> Every person under the age of 60 is using it.

I don't think this is true, or is a reflection of being in tech (assumption on my part about you, correct me if wrong).

Anecdotally, my peers (26ish, so gen z), none of them use AI except me, and that's for work as a SWE. My 68 year old step-dad uses it, but I suspect that's out of boredom more than anything.

I know students are using it for tons to essentially cheat, but I don't know anyone in that cohort using it for personal use of any kind.

People will hate AI and use it daily regardless. I hate the smart phone yet here I am, it’s too useful to avoid.
loading story #50563035
loading story #50562765
the best thing antrhopic/openai could do would be to get rid of their free tier and focus on enterprise/b2b.

it'll cut down the school cheating and the laziest, highest-visiblity slop on social media that is driving so much negative sentiment. AND they dont have to worry about providing capacity for people that ndo not (and never will) sign up for a subscription.

loading story #50561234
> I don’t know, every professional lawyer, doctor, accountant, nurse etc is using AI now in some way. Every person under the age of 60 is using it.

This is obviously false on its face. For example, my household consists of four people all under the age of 60, and only one of them (me) uses AI at all.

And I don't believe my accountant uses AI, at least not for anything of importance. If I found out they were, I would fire them immediately.

Well, as a lawyer, for the past several years, people, especially here, insisted that I'd just be out of a job entirely because of AI. Now AI providers come to my office to beg me to sign up for their products, which apparently cannot replace me. The idea that I need cutting edge AI as opposed to a more specifically well-trained model is also just as silly. Harvey is the company that will benefit from law firms, not Anthropic and not OpenAI
loading story #50566883
In 5 years all the subscription fees your firm pays to LexisNexus or TR will be captured by some AI player. I’m not saying it will replace professionals but it will be a substrate in the professional class. It really is just the new SaaS, just highly targeted.
loading story #50569277
They use it like Google, just kinda there because it's cheap and easy. Not all in omg I need a max plan like some devs.

Can the AI companies be as successful as Google ended if with web? Maybe, just feels like the ad model will not work with these use cases like it did with Google 25 years ago..

Yes, I know a few myself. LLMs are useful but the huge distruption has to be slowed and thought out. These companies are pedal to the metal, they don’t care about anything as if laws don’t apply to them. And we should still keep our people educated, now people are wondering if education is worthless since the LLMs took the limelight. This can’t lead us to a good place. Profit is driving everything.
> I don’t know, every professional lawyer, doctor, accountant, nurse etc is using AI now in some way.

Can you give names so we can sue them for HIPAA and other violations?

loading story #50560019
loading story #50562798
I would have expected on HN some interesting figures on AI use.

Last gallup survey says :

- Frequent workplace AI use rose to 32% of U.S. employees in Q3 2026

- daily users is 16% and total ai use among employee is 53%

The adoption curve is interesting. Total use seems a little bit to plateau but not frequent use which is still linear upward.

https://www.gallup.com/workplace/716585/ai-changing-employee...

Edit : focus stats on fréquent/daily/total users.

Public sentiment against data centers is definitely trending downward, public sentiment against AI is a little trickier. A lot of the anti-slop crusaders are being forced to admit that AI output is getting pretty good, and it's slowly making inroads in more industries and roles. I think there's a begrudging acceptance from haters that generative AI is just how things are going to be from now on.
echo chambers are a helluva drug
loading story #50562861
{"deleted":true,"id":50559952,"parent":50559373,"time":1795095547,"type":"comment"}
...and their ballooning costs.
I heard a news bite this morning, something like

> Anthropic has burned $9 billion in the lifetime of the company, and will have $200b+ in cash post-IPO

loading story #50560931
loading story #50561167
It's incredibly cash efficient, but they've spent compute credits that they've raised in previous funding rounds (that was a great deal for their investors)
loading story #50561438
loading story #50560909