Tibo, once hailed as the "cyber godfather" by Codex users, had long operated by resetting usage quotas for users—a move that helped OpenAI attract a large number of price-sensitive users through a cost-effectiveness strategy and counter Claude Code's rapid expansion in the AI programming market. As computing power supply tightened and IPO-related revenue pressure mounted, OpenAI adjusted its strategy, reallocating core computing resources to high-net-worth paying users. Following Tibo's reassignment, Cat Wu—once known as the "ban queen" at Anthropic—joined the team and subsequently introduced tightening measures such as unified quota interoperability and metering, the cleanup of low-price tiers, and restrictions on shared accounts. Behind this shift lies the reality that the large model industry broadly faces high computing costs, low user loyalty, and a market gravitating toward high-value users. The previous model of using subsidies to acquire users is no longer sustainable, and the industry has entered a new stage of competing on cost efficiency per unit token.