Super interesting detail from inside OpenAI:
The company hired a bunch of ex-Meta people. Some of these people wanted to invest heavily into internal tooling teams (Meta did it, worked great)
OpenAI leadership resisted, saying that in an AGI-first world, there will be no internal tooling teams.
And they were right: now, with Codex, there's a massive internal tools explosion, without having any internal tooling teams!
Gergely Orosz
@GergelyOrosz
Detail gathered from my visit to OpenAI HQ, and it will be in the deepdive coming in ~a week in @Pragmatic_Eng
OpenAI is one example of how hyperscalers succeed != how AI labs succeed!
OpenAI is one example of how hyperscalers succeed != how AI labs succeed!
8:12 PM UTC · Sep 7, 2026 · 49.2K Views
3118021
Gergely Orosz
@GergelyOrosz
Ofc it's different outside of OpenAI - hence I'm being clear this is OpenAI.
There is one possiblity that if token prices drop (as they have!) one day unlimited tokens can be an option everywhere (think: cheap inference and/or local models). Or not!
x.com/kunchenguid/st…
There is one possiblity that if token prices drop (as they have!) one day unlimited tokens can be an option everywhere (think: cheap inference and/or local models). Or not!
x.com/kunchenguid/st…
Kun Chen @kunchenguid
they are operating with unlimited tokens though
in real companies that spend real money on tokens, the economy is different when everyone basically vibe codes the same thing over and over again
in real companies that spend real money on tokens, the economy is different when everyone basically vibe codes the same thing over and over again
9:39 PM UTC · Sep 7, 2026 · 53.6K Views
31524
Gergely Orosz
@GergelyOrosz
Here’s more color from Tibo - thanks!
By next week will figure out how these two fit, and the internal tooling approach then+now. Consider Codex only started in 2024 (!!) and where it is today. The story of Codex, via the team: newsletter.pragmaticengineer.com/p/how-codex-is…
x.com/thsottiaux/sta…
By next week will figure out how these two fit, and the internal tooling approach then+now. Consider Codex only started in 2024 (!!) and where it is today. The story of Codex, via the team: newsletter.pragmaticengineer.com/p/how-codex-is…
x.com/thsottiaux/sta…
Tibo @thsottiaux
I challenge this a bit
(a) Codex was started as an internal tool. The goal was explicitly to build what was the highest leverage tool to accelerate infrastructure buildout.
(b) I don’t think the Meta vs non-Meta is particularly enlightening here. Most of us who came from older
(a) Codex was started as an internal tool. The goal was explicitly to build what was the highest leverage tool to accelerate infrastructure buildout.
(b) I don’t think the Meta vs non-Meta is particularly enlightening here. Most of us who came from older
10:37 PM UTC · Sep 7, 2026 · 29.1K Views
303716