It seems he realizes he was wrong about that and has pivoted slowly to, "well, maybe they work sometimes, but the cost isn't justified." Which is a reasonable question! I just find his style of never admitting when he is wrong off putting and the way he presents things as absolute fact, when he's guessing like the rest of us. He was right about a lot, wrong about a lot, it's okay to admit that, I don't think his fan base would care.

How many tokens can you realistically burn through in one chat session? Opus and many other frontier models do maybe 60tok/s, less 250k/hr out. In you can use more, but in most cases cache is 5-10:1 cheaper than new input. Say you average 500ktok in, 90% cache, per request. That amounts to 100-150ktok in new input-equivalent costs, which in most cases is ~20-30ktok in output-equivalent costs. Do a request every minute, that's a total of about 1.5-2Mtok/hr. At API prices that's $50/hr for Opus, but really it probably only costs Anthropic $10/hr to serve that.

That said, even if a developer is burning $50/hr, many, many employees at large companies cost more than $100k/yr to employ all costs considered, so making them say 20-30% more productive can easily make that worth it for most. If the labs shave their margins ultimately to more like 20-30%, you'd have ~$15/hr in costs to use the services, and nearly every white collar job is way over 30k/yr to employ. If your salary is 80k, you probably cost the company 200k all in, so making you 15% more productive offsets the $15/hr cost.

So first party providers are not in a horrifying position or anything from a subsidization standpoint. The people in bad shape are Cursor and Perplexity, who don't have frontier models and are dependent on the open source community, which is typicly 6-12 months behind the frontier. They have to pay full freight API costs at 80% margin for the big boys to serve their harnesses, which is indeed untenable, and they'll have to either force users to use open source models and/or in house models they can serve at-cost or they will have to charge vastly more.

Gemini, Claude, and ChatGPT first-party services like Antigravity, Codex, and Claude Code are not in serious trouble though.

I just don't think that LLM business models can survive the allure of advertising dollars, any more than Search could, or TV, or Radio, or Movies. Ignoring the talk of copilot putting ads into pull requests, there is just no way that publicly hosted LLMs will not end up inserting ads into the output.

This looks like what I remember. https://freakonomics.com/podcast/is-google-getting-worse/

> On an economic basis, a monthly subscription only makes sense with relatively static costs.

Running a data center is a fixed expense. Whether or not people use that data center to it's capacity doesn't change how much the operator pays (electricity use factors into this, since a GPU running at 100% will use more watts than an idle one, but it doesn't move the needle much on other fixed and variable costs of a data center).

> They also assumed, I imagine, that the cost of tokens would come down over time, versus what actually happened — while prices for some models might have come down, newer “reasoning” models burn way more tokens, which means the cost of inference has, somehow, gotten higher over time.

This is backwards. When the cost of something goes down, people use it more. This is basic supply and demand. Inference has gotten cheaper already, and will continue to do so.

Companies subsidizing costs for growth happens all the time. Yes, switching to usage-based pricing instead of subscriptions sucks for customers, but enterprises will continue to pay.

I pray this happens soon, but I feel I've been hearing some version of it for a while.

It's interesting to compare it to electricity. Basically Anthropic was selling a flat fee electricity subscription, and when someone started connecting expensive washing machines (OpenClaw) to their subscriptions, instead of changing the pricing model, they banned washing machines...

I wonder if we will get to "electricity" style pricing for AI. What makes electricity predictable is relatively constant average usage over time + price is manageable. I'm just not buying electrical house heating and manage my electricity spending within some bounds.

With AI the problem is that we are only now getting to useful AI, and for now it's still too expensive to be useful, so they subsidize until they can stabilize at "cheap enough and smart enough" level. But it feels like that's still 2 years away while they are stopping to subsidize now. Will be interesting.

Also, I didn't read this whole thing, but I have yet to see Zitron respond to the strongest AI financials claim, which is that the models themselves are profitable on a life-cycle basis, even if the companies are not profitable on an annual basis due to capital expenditure. Dario made this claim exactly, and it more or less blows all of Zitron's financials arguments up.

They went from GPT 2 a text only, goldfish-esque memory at a 8th grade reading level to what we have today, GPT 5, multimodality + a token window encompassing a enclyopedia and a Doctorate/Masters level of mastery in major subjects.

The economics are probably betting on this exponential growth to continue, which if it fails, the cash would burn.

Doubtless some people will reduce usage as a result. But Ed seems to find the idea that a 10 man developer team might spend 80K a year on tokens ridiculous. I don't understand this. Has he seen how much developers are paid? If you get a 20% productivity boost from coding agents, then that's two developers for 80K - effectively very good value.

Where things could go wrong is in comparison to cheaper models. If it's 5K a year for Qwen, and it's 2/3 as good will you pay 75K extra for Opus? Perhaps not.

> Don't attribute to malice what can be attributed to incompetence.

We're currently used to SAAS billing models that are either all-you-can-eat subscriptions, or metered around some easy-to-understand metric like # of users, or otherwise number of gigabytes consumed.

The SAAS economics work that way because the compute consumed is typically too cheap to meter. Some customer uses a little more than average, some customer uses a little less than average; it's not worth the time to even it out to the penny.

AI is so darn CPU (GPU? AIPU?) intense that will only be profitable, and affordable, if it can be metered like electricity and billed with a small margin.

In SAAS, we're not used to metering billing computations this way.

Companies don’t sell their goods/services at cost. A model’s being priced at, say, $30/M for output tokens doesn’t say anything about what it costs the company to provision the 1M tokens via the model.

And no, you cannot extrapolate from any company margins that someone may have overheard in an SF coffee shop onto individual product line margins or their trajectory either. This information is usually unknowable even in most SEC filings for public companies.

It’d be great if people who wrote these articles used, say, AI to look up some basics on how a business operates. It’s really easy to do, believe me ;)

The internet seems to be saying that 70%+ of Anthropic revenue is per-token metered API, which would largely invalidate the article, but I can't find a solid source.

Customer: “I don’t want to pay more than $100/mo for my website” Developer: “What are your goals?” Customer: “1M daily visits, 1,000 monthly signups.”

And we've spent the past 25 years offering serverless compute, auto-scaling, pay-as-you-go for AWS and Internet infrastructure. And the economics are still a hard sell.

[0]: https://www.wheresyoured.at/why-are-we-still-doing-this/

A $20 subscription 2 years ago is not providing the same level of intelligence you're getting today.

Every major lab knows open source models are 6 months behind (See Google's "We have no moat") and none of them plan to make money on inference. Companies are subsidizing users to create moats that persist when models are essentially free for most everyday use.

Because, comparing vs GPUs

~16k–17k tokens/second per user

<1ms latency

10x power efficiency

20x cheaper production

Model to Si ~ 60 to 90 days

We have every reason to believe SW_to_Si will facilitate improving economics

Economics Don't Make Sense.

I mean, seriously... our current late-stage capitalist economy is the chaotic sloshing of excess capital or inverted debt in a shallow tub within which clumsy giants are stamping like toddlers, and a parasitic kleptocratic oligarch class balances its efforts biting the toddler ankles in hope of more stamping judged advantageous, and, bagging what water they can.

The truth is that the AI companies are gambling that inference cost will continue following a hyper version of Moore's Law, e.g. Google TurboQuant.

The countervailing thesis is that frontier models are consuming more and more compute.

The deepest truth: you often don't need a frontier model to get commercially acceptable results from AI. Thus, bring on the true pricing! and I'll just switch models to something financially sustainable.