It's hard to take this piece seriously if he's citing _Ed Zitron's_ math, and equally hard to make the blanket statement that flat-rate plans = "the current AI pricing". But yes, those pricing models were pretty silly and unsustainable.
Get back to me when there's an AI company that's actually profitable and we can compare their service and pricing.
Claiming that there's some small subset of their services (like inference per token) that's "profitable" doesn't mean anything when it relies on everything else that company is still paying for. If you could make money from it at current prices - why aren't they?
Otherwise it's just "how much they're willing to subsidize".
On OpenRouter there are 11 third-party inference providers hosting DeepSeek V4 Pro right now, of which 8 are US-based and 7 of those have zero data-retention policies (which I mention to rule out any claims of "oh they're making up the money by logging all your data"). This is a 1.6T-A49B model, so a bit bigger than Sonnet (~2/3rds the size) and a bit smaller than Opus (~3x as large). These third parties are almost perfectly interchangeable via OpenRouter as a marketplace, so they have no incentive to offer any sort of "growth pricing" below costs, and they serve it at $3.48/Mtok out.
Kimi K2.6 is 1T-A32B with a slightly less computationally efficient architecture, and is served at around $3.50/Mtok out by 9 US ZDR providers.
Unless you think that either the generally accepted size estimates for Anthropic/OpenAI models are wildly off or those companies are a lot worse at serving models efficiently, Anthropic and OpenAI are probably making around 5-8x margins on their API costs.
The cost of training new models is of course a major factor not counted here. Depending on how you want to think about that this may or may not make them net profitable. I remember one of those CEOs gave an interview a while back where they described it as a series of independent investments, where each model they train is net-positive in revenue by EOL just from its own inference, but I don't know whether that's still true or not.
Regardless, the point is that if they stopped training new models today, both Anthropic and OpenAI are making incredibly generous profits on their API inference.
The problem with fixating on earnings as you're doing is that it's a bad metric for a growth company. COGS is much more important. What you're doing is setting it up so every growth company is terrible until they've matured into a 20 year old company. That's obviously dumb.
From what I've seen pretty much every company is limited by hardware supply, to the level where's there complaints from current customers about the speed of new customer growth is exceeding their ability to service them properly.
And "growth at all costs" makes sense if there's lock in and you can monetize those "now locked-in users" later - but that doesn't really seem true on the consumer side. It seems pretty trivial to switch out which model and provider on the consumer side.
Any "lock in" has then to be on the model or inference side, and that's still advancing in multiple areas from so many different sources I'm not sure I'm comfortable saying that will also be a "winner takes all" situation either.
My approach is generally "enjoy using it while it's cheap and subsidized, but understand that might not last forever". If it does remain cheap after the subsidies end, great, you can just keep using it. But if it doesn't and you've lost the ability to work without you'll be in for a world of hurt.
Your concern about their business is that demand for their products is growing so stratospherically that they cannot meet that demand easily? I mean that's like an A+ scorecard for any business. Everyone in business would dream of such a scenario. That's called a luxury problem.
But each new customer is still losing money. As I said, subsidized growth only matters if you can recoup those subsidies afterwards - and that's what I'm not sure will be true.
I think the idea of "all growth is good no matter the cost" has been taken to an extreme.
There is probably going to be a quarter or two of profits when the prices dramatically increase. Vibe coding techbros are hooked on the Iron Lung and may not want to get off.
At my work are multiple developers bragging about overnight AI usage to solve problems hands off. Yes they are wasting money and resources but the fad is here. People be vibe coding for now.
In like 6 months when all the costs need to be paid and the prices go up, we will see if these companies stay profitable. But I'm of the opinion that the vibe coding tech bros are more than enough to sustain a short or even medium term profit for these companies. Just on fad-energy alone (see OpenClaw)
The fad probably collapses soon after. I hope anyway, the waste I see is nauseating.
------------
I dunno where this is all going. But I do have faith in human ingenuity still. Things are changing, possibly for the worse, but we need to make the best of it.
The worst of behaviors is wasteful and blatant fraud. There's something useful here though.
At least on the HomePod side, intercom is at best a half-baked feature and at worst an infuriation machine. It uses a cumbersome voice trigger ("siri, tell <room>…") to begin recording audio, with no clear indication of when recording began and no way to know for sure that the audio was directed where you wanted it to go.
To respond is similarly cumbersome and soon you give up completely. I can only assume it was designed by someone whose parents were killed in an intercom-related disaster and has sworn revenge.
I bought a mini for my office with this purpose in mind, but it has been a total waste.
losers, clueless never had to be productive, just scapegoats. But now losers dont get that buffer window to try and become sociopaths, they just dont get hired at all.
Nope. Still does not. I have 2 macs on my desk and no simple way to connect them to a single Apple display! It's a glaring hole that to me suggests they have no idea who their market is for these.
Unlike many of those approaches which concern themselves with delivery of human-designed static UI, this seems to be a tool designed to support generative UIs. I personally think that's a non-starter and much prefer the more incremental "let the agent call a tool that renders a specific pre-made UI" approach of MCP UI/Apps, OpenAI Apps SDK, etc for now.
that's not quite what parent was talking about, which is — don't just use one giant long conversation. resetting "memories" is a totally different thing (which still might be valuable to do occasionally, if they still let you)
Actually, it's kind of the same. LLMs don't have a "new memory" system. They're like the guy from Memento. Context memory and long term from the training data. Can't make new memories from the context though.
(Not addressed to parent comment, but the inevitable others: Yes, this is an analogy, I don't need to hear another halfwit lecture on how LLMs don't really think or have memories. Thank you.)
Context memory arguably is new memory, but because we abused the metaphor of “learning” rather than something more like shaping inborn instinct for trained model weights, we have no fitting metaphor what happens during the “lifetime” of the interaction with a model via its context window as formation of skills/memories.
i'm geniunely curious about how you made the jump from "here's a single regulation" all the way down the slippery slope to "can't regulate away ALL parenting". does this one regulation cross that threshold? how'd you get there?
in an ideal world, parents would also prevent their kids from smoking, but the fact that in many places minors aren't allowed to purchase tobacco sends a social signal and actually does seem to put a speed bump in place deterring casual use.
is it not _also_ ideal to have some of these regulations in place? does it not help parents make the case to their kids?
it does help. i think this is a good step in the right direction.
but there's still a lot of stuff that only parents can do. for example, screentime in the home. you can't really create a law that says no screens for anyone under the age of X because there will exceptions (movie night, homework, etc).
Screentime helps, but it doesn't really solve the problem. They still see the exact same content shared by friends at school, and 15 minutes a day is enough to do damage.
