johnfahey

I don't think it's a conspiracy, they're just making the same mistake many other software companies have in the past. When you sell your products to the most discerning and well-informed consumers, you enter into a cutthroat race to the bottom. OpenAI would like to diversify to more profitable ventures, but since ChatGPT they have yet to release something novel that was truly successful, nonetheless profitable, and thusfar nearly every experiment has been a flop, so to speak (ChatGPT Atlas, the Sora App, Instant Checkout, etc).

OpenAI won a lot of good favor for the generous Codex subscription and the efficiency of their models, but now that many people have switched over from Claude, they think they can leverage their position to peddle a stream of unnecessary products, and crack down on the generous limits[1] that brought everyone to Codex in the first place.

Anthropic did the same thing. Earlier this year, Claude subs and Claude Code took off because of the subscription's incredible capability and value, then once they gained enough users, they started focusing on unnecessary products no one asked for (see Claude in Slack), and eventually lost their lead. After losing a bunch of customers to Codex subs they realized their mistake, and now they're shipping again.

AI companies are bad at making software; they are good at making AI models. And that's about it.

[1] https://x.com/thsottiaux/status/2104823812042940713

No doubt xAI has seen rapid progress, but it's been several months of them being "just behind" OpenAI and Anthropic. It seems the gap between just behind the frontier and pushing it is a lot wider than most people thought it was a year ago, and that's why a clear third contender in the frontier model space has yet to materialize.

Judging from those cost-performance graphs, Sonnet doesn't make sense to run at anything higher than a medium reasoning level, since Opus 4.8 low reasoning outclasses it for the price.

This line as a selling point is also pretty funny:

> Evaluations also show that it has a much lower ability to perform cybersecurity tasks than our current Opus models.