Frontier labs reportedly 6-7 months ahead internally as token spend concentrates at top labs
modestproposal1 · x · 2026-10-28
An industry analysis thread argues a fundamental mismatch in AI: frontier labs reportedly run internal models 6-7 months ahead of public releases and are accelerating their own cycle times; open-weights models are rapidly expanding the enterprise task surface they can cover; the vast majority of token spend accrues to frontier providers; and per-token/per-task prices are plummeting for more use cases.
The author frames AI as a potential "China shock" for white-collar employment—except losses would spread broadly rather than concentrate in a few industries and regions. Commenters conclude that even if individual companies like Ant buck the trend, the industry-wide mismatch stands.
More from AGI Musings
- Stanford launches TaskNet: 132 worker interviews build a large-scale occupational task inventory — soumitrashukla9 · 2026-10-28
- Ex-CEA economist pushes back on replacing staff with AI: "AI cannot cry when it loses at foosball" — soumitrashukla9 · 2026-10-28
- Ethan Mollick: claiming AI firms hype existential risks to pump stocks doesn't hold up — eldonredwards · 2026-10-28
- signull: human attention monetizes even without action; agent attention is nearly worthless — signulll · 2026-10-28
- Kevin Xu: There's a sovereign AI bubble inside the AI bubble — kevinsxu · 2026-10-28
- Chinese AI labs rely heavily on Western revenue while US labs need none from China — kevinsxu · 2026-10-28