OpenAI's shared usage limits risk losing the power users whose feedback improves its models
evanmac42 · reddit · 2026-11-03
A Spanish-language Reddit essay argues OpenAI's upcoming shared usage limits (chat and Codex competing for the same quota) risk more than subscriber churn: heavy users who stress-test the model — questioning answers, spotting errors, verifying hypotheses, and explicitly opting conversations into training — provide reasoning and correction signals that documents and synthetic data can't. The author, a 1.5-year intensive user, says he'd consider cheaper alternatives if the quota tightens. He warns of a vicious cycle: limits drive away demanding users → fewer hard problems discovered and fixed → product degrades for those who most improve it. The strategic question isn't the cost of serving these users but the value of what they contribute back.
Related event: Users push back as OpenAI plans to merge ChatGPT and Codex usage limits(2 posts)→
More from Companies & People
- OpenAI DevDay hackathon finalists share their build experience — paw_lean · 2026-11-03
- SF's top AI companies cram into a 1.5-mile radius — rivals visible through windows — signulll · 2026-11-03
- Frontier model training is a monopoly business: the bull case for Anthropic's 'One Ring' thesis — abhiadesai · 2026-11-03
- Micro schools will adopt AI tutors first and deliver '10x better' education, advocates say — RachelVT42 · 2026-11-03
- Anthropic pays Safeguard PMs up to $460K — and bans Claude in take-home tests — aakashgupta · 2026-11-03
- Legal AI startup Crosby's paralegal agents now live for 95% of clients — graceisford · 2026-11-03