Anthropic’s unit economics are really good, so I don’t think (b) is true. Point (a) may be valid, but they’ve also grown consistently, and even their slower growth rate in Q3 of this year is still very strong. To keep up growth, they likely need to procure compute months ahead of time as well, so it makes sense for them to do this before they are actually limited.

I assume some of this is the escalating cost of training, indeed, and that being faster means abandoning efficiency.

But you’re right. I have often wondered why it is that such a small amount of extra compute (300MW) from SpaceX was entirely enough for them to uncap so many customers.

I think the (in-house) data centre burdens of ever-more complex models may be considerable, but I think the data centre businesses expecting to serve inference tasks are probably heading into a Chinese real estate bubble.

> would be like an expensive Bloomberg Terminal or an investment letter that offers valuable and expensive intelligence.

Whether it's Anthropic or not, I think it's inevitable some highly specialized offerings will emerge. I suspect there are already some incredibly sophisticated, hyper-narrow, astronomically expensive models being created for private in-house use.

As an imaginary example, I wouldn't be shocked if some company like Jane Street is secretly investing a billion dollars to create the highly specialized synthetic RLVR data needed to train an AI that gives them an average .08% edge on a certain kind of commodity, equity or asset trade. If so, it's not a leap to imagine another company spending $100M on a model that gives a .008% edge and selling access to 100 high-volume institutional traders for $2M each.

Expanding general purpose LLMs to incorporate increasingly specific use cases has, so far, seemed to be economically viable but I don't see a fundamental reason that must remain true for incredibly narrow yet ultra high-value domains requiring monumental bespoke development, data, etc to profitably exploit.

Being a specialized firm might work if they didn’t have $700 bn in commitments.

Already they have exclusive deals with trading firms who pay $$$ for faster models and more features. But there are only so many firms who will pay that much (and a lot of companies are pulling back on spend, not increasing it)

and that would still be valued at a fraction of $1T.

As a software engineer I’d be paying thousands if there were no alternatives. As a household - just about everyone pays $100 a month for internet access alone and something like cloude can be easily worth that.

Thing is - regular people have no clue what to do with it since they are locked in ecosystems of apple and google.

And even then, the path for investors to ever make their money back is really essentially non-existent. There’s so much “total addressable market” fantasy here.

What does that imply about where SpaceX should have been? It feels like we are re running the WeWork IPO except without the investors who said "no way"

And yet, $1T is still insanely high for an IPO. The size of this...inflation...is truly mind-boggling.

After Tuesday no way in hell that’s happening.

I would

Yeah fuck that. If they fail, they fail.

In the same way they famously invested 10% in Intel, which has proved volatile at best?

OpenAI are openly suggesting 5%, but my instinct is that this is because they think Trump is foolish enough to demand double that.

[dead]

Odds of coming out this year just dropped from 95% to 70% on polymarket

They can but they shouldn't. They have nearly $100b in cash and cash equivalents. But given the massive compute cost commitments, that clock keeps ticking down especially as the private equity and debt markets are closing on them day to day. And delaying sends a signal to the wider market which causes cloud providers and everyone downstream to get repriced lower, making the financial landscape even worse. The question is if Anthropic's situation will look better say three months from now. Given growth slowing and margin compression in Q3, that's a fat if.

[dead]

distribution, a foothold in the enterprise, market share, brand recognition, ...

and how much is gpu rent?

And anyone with 3 can also see another trend. The hardware required for inference is rapidly falling. It's all following what happened with text-to-image image gen stuff almost to a t. It started out requiring super computing clusters, and now a days you can do high performance image gen on any plain old computer. I think within a few years we're going to be running frontier level models on a basic mid-range gaming hardware. You can already run quantized frontier level models on $1600 tier cards at high performance.

Seeing this bubble play out in real time is the most surreal thing. It's like watching a train wreck play out at a speed of a few inches per day. Like come on, just get off the track FFS.

> anyone can pay google for GPU rent.

Have you tried requesting GPU instances on Gcloud without being associated with a major institution? Yeah, good luck.

Kasparov, btw

Chess is a fixed target, so as a benchmark, it can be saturated.

Reality can’t be saturated. The set of things you can accomplish with more intelligence is unbounded. So while cheap models will continue to proliferate, there will always be a massive market at the frontier, and that won’t ever change.

The models themselves aren't what's valuable.

This is a real question, because I just don't know. Don't the models needs lots and lots of training that need the gobs of giant data centers? So you either need the infrastructure or the money to buy computing to make the models? I see that all the money is being invested in AI and data centers but I do not quite understand why.

The biggest customer bases will generate cash flow to build better models. The ones trailing are the ones opening their weights in order to commoditise the models. Compute providers are also incentivised to provide open models.

We will continue to have every major player trying to build vertically - model, chips, infra, packaged products and services.

> There's no lockin that Anthropic has because it's trivial for people to switch to a different provider.

This is like saying there's no lock-in that iPhone has because it's trivial to switch to Android. Or that there's no lock-in that Windows has because it's trivial to switch to Linux or Mac. Or that there's no lock-in that cars have because it's trivial to switch to bikes or trains.

They're different products. They're not the same. Being able to write the same code does not make a model Claude. Nor does being able to reach the same benchmark scores. Nor does performing one or a choice set of tasks to a comparable level of quality.

It's true, people can switch if all they want is to perform X, Y, or Z, and some other model can do X, Y and Z well enough. People switch from iPhone to Android all the time. From Windows to Linux or Mac. From cars to bikes or trains or other transit. But they're not the same thing! Android doesn't have a missing business model just because iPhone exists, nor vice versa.

You can't just say that makes cars the same as bikes or Windows the same as Linux. You can't get Claude from anybody other than Anthropic. Maybe you can get something else that does as well where it counts, but not Claude. You can't just say Anthropic is just the same as any open model provider, and that you can just switch providers for the same result. The open model has to actually deliver the desired result, because you're not going to find one that is just Claude.

This is what keeps me on Apple now. This is why most of my biggest hopes and dreams for my personal computing are that Apple will do the right things and in the right way. Because if I want to reject even one single aspect of an Apple product, it means changing to an entirely separate product. That gives up every aspect Apple had over that product at once, even the vastly positive ones. So I put up with the decisions I don't agree with, in exchange for the decisions that I do.

(Meanwhile, I plan to switch to SteamOS as soon as they support Nvidia. I don't use or care about any aspects of Windows besides running games. Now that I have Steam Frame, I don't need Windows Mixed Reality either.)

Why have I been reading a variation of this comment since DS launched more than 18 months ago, and in that time ol’ No-Moat Anthropic has grown their revenue 100x, hmm?

And the cost of training is going down as a percentage of revenue over time, is it not?

They have a moat if they're like all the other businesses: unethical, buys it's moat via government protectionism, uses oligarchies to stiffle competition, etc.