I'm not sure where the author gets the $100k number, but I agree that Cursor and Claude Code have obfuscated the true cost of intelligence. Tools like Cline and its forks (Roo Code, Kilo Code) have shown what unmitigated inference can actually deliver.
The irony is that Kilo itself is playing the same game they're criticizing. They're burning cash on free credits (with expiry dates) and paid marketing to grab market share -- essentially subsidizing inference just like Cursor, just with VC money instead of subscription revenue.
The author is right that the "$20 → $200" subscription model is broken. But Kilo's approach of giving away $100+ in credits isn't sustainable either. Eventually, everyone has to face the same reality: frontier model inference is expensive, and someone has to pay for it.
Also frontier model training is expensive, and at some point, eventually, that bill also needs to get paid, by amortizing over inference pricing.
oh go one more step: the reality is these models are more expensive than hiring an intern to do the same thing.
Unless you got a trove of self starters with a lot of money, they arn't cost efficient.
that's a good point, however maybe the difference is that kilo is not creating a situation for themselves where they either have to reprice or they have to throttle.
I believe it's pretty clear when you use these credits that it's temporary (and that it's a marketing strategy), vs claude/cursor where they have to fit their costs into the subscription price and make things opaque to you
It sounds like Uber