Usage per user is growing much faster than overall user growth

That is an interesting tidbit about Codex from OpenAI's Codex engineering lead Thibault Sottiaux in Alistair Barr's Business Insider article "Silicon Valley is buzzing about this new idea: AI compute as compensation." On X Sottiaux writes: "I am increasingly asked during candidate interviews how much dedicated inference compute they will have to build with Codex. Pairing this with usage per user growing significantly faster than the number of users, it's pretty clear that compute will be something that is scarce." (https://lnkd.in/ggx7TcmN) There is an interesting pricing dilemma lurking here. All-you-can-compute inference plans are creatures of the past, at least for real-time inference. Given the scarcity (and/or the cost) of high quality compute access, it is only natural to treat it as a coveted perk and as a reflection of your value as perceived by the organization, just like stock options. The always entertaining Peter Gostev, Arena's AI capability lead, suggests that "OpenAI and Anthropic should create recruitment sites where their clients can advertise roles, listing the token budget for the job alongside the salary range."