OpenAI’s vague usage limits contradict the transparency and user empowerment it claims to value

The frustrating part isn’t that usage limits exist. It’s that the end user is expected to plan around limits that OpenAI barely defines. With Codex you get a 5h percentage, a weekly percentage, “usage varies,” and not much else. You don’t know what the actual allowance is, what a task cost, or how much work you realistically have left. I literally ended up reverse engineering the relationship between my 5h and weekly limits from percentage bars just to understand what I’m paying for. Dots makes this even worse. Usage is basically unlimited for the first month, while OpenAI says they’ll publish permanent per-plan usage terms later. So if you like Dots, build workflows around it, or start depending on it, the real limits can show up afterward and suddenly you have to scale back, deal with more vague limits, pay more, or stop using it. That feels fundamentally backwards. OpenAI and its employees constantly talk about empowering users, broad access, affordability, transparency, and building technology that benefits people. Their own policies and public principles lean heavily on those ideas. But if you’re selling access to a finite resource while refusing to clearly define how much the customer is actually buying, that doesn’t feel particularly transparent or empowering. I don’t need unlimited usage. I need to know what I’m paying for before I build around it.