Post

there are so many good arguments against AI but i don't understand this one tbh. why should i care that the richest corporations on the planet apparently haven't figured out how to turn a profit on it?

it’s more about a business selling you an idea that they’ve built a reliable, sustainable dependency that you can add to the tech stack of your business (when they’re burning massive amounts of cash while praying that the promises they’ve made will come true)

Plus that is the business model that Uber, Amazon, and a lot of other companies did. They subsidize cost with VC money to kill competition and get everyone hooked on something until they need to turn a profit and then change the cost to squeeze customers

Uber/Amazon weren’t burning cash anywhere near these levels, and there’s no way to reduce costs. They only go up. Nobody has the stomach to pay what it actually costs to run these models. Would you pay $2500/month just because there was no other provider? Or would you just not use it, period?

it is weird that it has to be explained on bluesky of all places that the profitability of a good or service has minimal to no correlation on the actual value society derives from that good or service

I mean guy says "AI is a bubble, unprofitable, gonna burst, trust me bro" since 2024 (or even earlier?) and still has audience that keeps listening to that. That is quite sad. And by proxy I can't take this site seriously as something objective with any meaningful implications

ah thought you made this but also, what do you mean? AI generated or assisted code is everywhere now, in lots of products and sites that work

almost guaranteed which is kind of ironic though the creator doesn't seem to be anti-AI so I guess it makes sense, the FAQ sounds hopeful for profitability

The question shifts every 18 months. 'Is AI useful?' → 'Is AI safe?' → 'Is AI profitable?' The frame itself tells you where we are in the cycle.

I don't want to defend the AI companies specifically, but lots of stuff works like this. If you need infrastructure to provide a service and you have to pay for the infrastructure up-front and then make the money back later, you're going to have charts that look like these.

In this case, trained models and servers to run inference on. If we were talking about a coffee shop, it would be espresso machines.

colonialism fans these days be like "colonialism wasn't very profitable. it was negatively profitable," ignoring that the reason for the 'negative' profits were that they socialized the costs and privatized the wealth, and also, so so so so very many wars over colonies against other colonizer powers

the reason to oppose AI isn't that it's unprofitable. it's that we should stop it, before it is intensely profitable for capital at the expense of everyone else.

utter clown era right now as people insist "ah, but dajugg-209 meets the complexity benchmark and can be run on an m5 macbook pro with 64gb ram," even as they continue to increase the amount of money they hand to anthropic and openAI to rent closed-source model access instead,

This seems like a good reply to some other comment! I agree that there are plenty of good arguments against AI but "it seems to be insufficiently profitable in 2026" is not one of them.

Using an article stating that Anthropic ARR grew from $9B in 2025 to $30B by Q1 2026 as source and conclude that total cumulative revenue of Anthropic since 2021 was $6.5B is a crime against math. Even if the growth would have stop in April (which it didn't), Q2 @ $30B ARR alone would be $7.5B.