Hacker News Reader

Comment by cmiles8 | Hacker News Reader

Outside a relatively small world of circular investment and FOMO feeding FOMO the general consensus seems to be “let it burn.”

It appears very unlikely we will ever see an IPO of OpenAI. Anthropic appears less doomed, but still iffy at best. Tons of other large, but little discussed, AI startups are just dead-companies-walking at this point.

The likes of AWS are showing good headline numbers but are taking out massive debt to build infrastructure that looks increasingly unneeded. Those with capacity are looking to offload it, quickly. Yes AWS has “committed contracts” for this capacity but if those commitments are with shaky AI startups then it’s mostly just fluff PR and these hyperscalers will get left holding the bag on all this debt.

View on HN

Anthropic is almost purely a model company. They own close to no data centers.

If models are becoming commodities, and the main bottleneck is actually serving them at scale, Anthropic does not appear particularly well positioned.

If AWS had a ~60-80% margin for decades, I see no reason why inference can't have a ~60-80% margin for quite some time.

The problem is, if costs continue to drop ~90% for the same level of quality every 18 months, demand is unlikely to grow 10x to keep the revenue stable.

Who knows. Jevon's paradox. But the cost/quality is dropping too fast that it's hard for me to imagine demand keeps up long term to keep revenues (and profits) GROWING.

Anthropic is perceived as the leading AI company in the world. Even if they started selling inference alone, people would prefer to pay them a premium for it rather than figuring out how to operate opencode or other replacements. Add to that government contracts, enterprise and other markets, and I think Anthropic would be totally fine even if models plateau.

Transistor count has increased exponentially for decades and so did demand for compute. I think we will see a similar phenomenon with AI.

My company has started asking people to convert to open code all the sudden to avoid vendor lock in with anthropic

> I see no reason why inference can't have a ~60-80% margin for quite some time.

Same as anything with low margin: Viable competition and low switching costs.

Many sysadmins lazily do 100% AWS because they don't know any alternatives. CFO's might enforce using more economical inference providers if the cost is even 10% less.

> Same as anything with low margin: Viable competition and low switching costs.

So why didn't this happen to AWS?

High switching cost.

Now do Google

Advantages: Lowest inference cost due to scale, experience, advanced datacenters and custom chips. Huge customer base. AI automation will make it much easier to switch from AWS to GCP. Serious organization that doesn't let it's product unknowingly hack companies.

Disadvantages: Top model is slightly behind the frontier for now. Bias against using their products because it's not cool/trendy.

Their big disadvantage is their terrible, no-good, awful product management (which has always been their issue).

> If models are becoming commodities, and the main bottleneck is actually serving them at scale, Anthropic does not appear particularly well positioned.

That has been my thesis ever since Dario went on about how difficult it is to forecast capacity on the Dwarkesh podcast right about the time Claude started having 9's comparable to GitHub's.

Even as an outsider, seeing all the cloud providers bemoan their lack of capacity and their growing backlogs, I could tell compute capacity was already the biggest constraint on industry growth and that it would be very hard to come by in the future... which also probably made it the only moat.

And now we see Anthropic making very expensive deals with competitors to secure the necessary capacity. Claude likely still has enough momentum to make it worthwhile. But I can't believe Dario, probably the most AI-pilled of them all, would under-estimate demand so severely.

> But I can't believe Dario, probably the most AI-pilled of them all, would under-estimate demand so severely.

To be fair, this mess up actually increases my respect for him, as he focused on making sure that they could continue to run the business under reasonable assumptions rather than YOLOing compute into existence like OpenAI.

My surprise is rooted in the point that he seems to be the biggest believer in AI there is -- to the extent the label "zealot" gets bandied about at times -- and he still underestimated the demand they would get. I would not have classified Altman as a bigger believer than Dario.

My guess is OpenAI, having the first mover advantage and the bulk of AI traffic for a very long time, could gauge the trends much better than anyone else.

> infrastructure that looks increasingly unneeded

This is the crux that needs to be substantiated. Without substantiation none of your other arguments hold up.

Are you saying people haven’t been throwing unneeded GPU capacity on the market? That is happening.

Frankly it’s the opposite scenario (that’s there’s all this demand) which is struggling for any hard evidence.

Which players, specifically, are you talking about? Everything you said is opinion passed off as fact.

SpaceX has put a ton of excess capacity on the market.

Meta tanked chip stocks by saying it was considering the same.

Worries about compute overcapacity, via folks saying they want to offload capacity, is literally the thing that nuked the “situational awareness” fund last week.

It’s long term demand that matters, not just “price.” High prices without true demand is the literal definition of a bubble.

SpaceX is selling compute to a willing buyer. You can't point to one side of the trade as evidence for under/over capacity. The way we usually judge such things is the price. If the price for compute is going up, which it is, then that's the thing to look at.

SpaceX has put a ton of excess capacity on the market that was subsequently absorbed by Anthropic. There's a reason why Grok is no longer free.

Meta said it was considering doing the same because they saw how valuable excess capacity was.

You're misinterpreting what actually happened.

> Are you saying people haven’t been throwing unneeded GPU capacity on the market? That is happening.

That, by itself, doesn't have to mean anything.

AWS is far more compute capacity than Amazon needs, but that's not because Amazon misjudged how much capacity they need. They built it out to sell it to others, leveraging know-how and economies of scale to do so very profitably.

They only need companies to suddenly need their GPU compute if those committed contracts don't make it.

How hard can that be?

It's holding up the entire US stock market and therefore the entire US economy and the dollar value. You probably don't want this particular Atlas to shrug.

It's inevitable though. The funny-money has outpaced actual money by lightyears at this point.

On the plus side, our interest rates aren't 0% right now, so there's some room there.

On the down side, our national debt generation now exceeds 125% of GDP and bond rates are shooting up because nobody wants to buy our debt.

This guy makes a decent argument that the funny money party doesn't really get started until Q3 2027. To which I would respond probably, but also probably not if another coding agent sized market opens up between now and then. But predicting the future is hard and all that.

https://overweightskepticism.substack.com/p/ais-cash-cushion...

I do. The faster it happens, the faster a recovery begins and we can devote resources back to doing more practical things. Get it over with as soon as possible, rip the bandaid off, insert your idiom of choice.

Have you converted all your dollars to other currencies and non-financial assets?

But what is the point of this discussion? That we deny a possible burst because we stand to lose personally? Or is it, we believe it could burst because reasons?

You can’t threaten me with two good times. Long term that would help correct the overemphasis on the U.S. economy and AI.

The “Let it burn” attitude doesn’t mean people think they are immune.

Yes the overall market will take a hit but, like a forest fire we need a healthy burn to just wipe out the weaker players so the older more mature trees can get on with it. Yes the big trees will get burned a bit but they’ll be fine in the long run.

We need a good brush fire to just wipe out all the iffy startups and investors that over-indexed here. Thats what people want with “let it burn.”

I would rather be market-corrected out of a job for a few weeks or months than have this inflate into an economy-wrecking bubble that derails my career and life for years. I'm worried that we're at that state already...

The thing about the past is that it's the past. Things change, big changes do not bring back the past. They change things.

If you are market-corrected out of a job that situation most software engineers will be out of a job for years and wages will not recover for decades, or just never.

The trajectory of feudalism has always bent towards low wages and high rent extraction. There is no possible future where we stay with feudalism but increase wages.

Let's see ... feudalism started with slave labour (which is what muslims are still fighting for, fighting to bring it back), there were many bills of rights. Feudalism more or less started with Christianity forcibly implementing the right to quit your job, any job, including slavery, and ended with equality for all.

So: nope. Quite the opposite. Feudalism was a very slow increase in rights for normal people over time.

TSLA's 400 PE ratio (along with Elon Musk's trillionaire status) was never sustainable. Where I disagree is with companies running market average PE ratios supposedly being doomed as well.

But also, don't count out OpenAI and Anthropic, no matter how shaky the numbers. The market has also demonstrated how to price SPCX so if they IPO, they will see their true FMV and I'm sure it's >0 and much much less than what they believe. For even in the worst case scenarios, they have valuable personnel, experience, and IP deploying AI at scale for what is to come, even if it's based on Chinese open weight models. RIP lightcone of all future value though.

As for the hyperscalers embedded in FAANNG, they've been using profitable divisions to cover the expense of growing but unprofitable divisions for a while now. In fact, that's business as usual. They're all going to land on their feet with PE ratios of ~20 or higher. Now take your favorite tech stock you prefer to scapegoat and figure out its approximate bottom relative to today. You'll be glad you did.

> But also, don't count out OpenAI and Anthropic, no matter how shaky the numbers. The market has also demonstrated how to price SPCX so if they IPO, they will see their free market FMV and I'm sure it's >0 and much much less than what they believe. For even in the worst case scenarios, they have valuable personnel, experience, and IP deploying AI at scale for what is to come, even if it's based on Chinese open weight models. RIP lightcone of all future value though.

You know what I'd like to see? OpenAI going bankrupt, wiping out its investors, then being bought up by IBM for a nickel. Ditto with Anthropic, but ending up with Oracle (if it's still around after running up so much debt).

The karmic Schadenfreude of those outcomes would be delicious. And while Sam Altman has long been out as a lucky fool, until recently, I thought more of Dario Amodei. Not anymore.

I think it is very likely that OpenAI stop operating as an independent entity, but more likely that Microsoft would get full control.

> It's holding up the entire US stock market and therefore the entire US economy and the dollar value. You probably don't want this particular Atlas to shrug.

So you're saying we want to keep this bubble going forever?

If's it's a bubble, it's gonna pop, and better sooner than later.

"increasingly unneeded"?

In what way/shape/form is the infrastructure showing signs that are contrary to growth in need/demand? I'd really like to know if there's something I'm missing that I should keep my eye on, thanks

I mean, talk about circular investment and FOMO all you want, but if you use Fable every day you know what's coming. This is no bubble, we're just getting started.

More and more companies are signing up with OpenAI and Anthropic at near exponential growth to automate everyday tasks. If anything, these two should manage to IPO just fine. The rest of the downstream startups probably won't make it.

Reply on news.ycombinator.com