OpenAI Responds to Codex User Feedback Issues
OpenAI Codex core member Tibo acknowledged that the previous high-performance mode was too easy to enable without adequately explaining the impact on quotas. The desktop redesign made commonly used features hard to find, and the focus on promoting ChatGPT Work led to misunderstandings among Codex users about being phased out. The new version also affected multi-agent workflows and plugin experiences.
OpenAI has launched the first batch of improvements: adjusting default settings and model selectors, fixing plugin submission issues, optimizing Codex presentation, and addressing some desktop experience problems.
Market Mechanism: In the past 24 hours, OpenAI has intensively collected feedback and communicated, completing two quota top-ups; a larger update next week will restore sidebar chat/project features and enhance quota visibility. Resource allocation is tilted towards product iteration driven by events, benefiting OpenAI from user retention and the promotion of GPT-5.6 Sol/ChatGPT Work, while short-term experience issues are straining user trust.
Source: Public Information
ABAB AI Insight
Tibo, as the engineering lead for Codex, has previously responded quickly to user quota and bug issues and reset them. This public acknowledgment of promotional and experiential mistakes continues OpenAI's iterative path of "rapid release-feedback fix," similar to earlier quota events caused by unexpected Codex usage.
On the capital path, OpenAI mobilized the engineering team for weekend war rooms to analyze logs, recharge multiple times, and plan next week's updates, motivated by balancing the promotion of ChatGPT Work with retaining core Codex users, avoiding developer loss to competitors, while paving the way for new models like GPT-5.6 Sol.
Similar to early Claude iterations where feature adjustments led to feedback-driven optimizations, this move shows that OpenAI is currently in a phase of expanding and refining its product matrix, needing to cater to both new and old user scenarios.
Structural Judgment: Essentially, this is a transfer of pricing power in technological substitution. As AI products evolve from single tools to complex workflows, user habits and cost transparency become key. The mechanism is that after the widespread adoption of high-performance modes, the platform needs to rebuild trust through visibility and customization; otherwise, developers will turn to more predictable alternatives.
ABAB News · Cognitive Law
When promotion focuses on new products, the pain points of old users become trust fractures.
Quotas act as invisible levers; if unclear before use, resetting costs afterward are higher.
Fast iteration does not mean users can keep up; transparency is the real barrier to retention.