Codex Head Tibo Reveals ChatGPT Work and Codex Users Approaching 9 Million
OpenAI Codex product head Tibo posted that the user base for ChatGPT Work and Codex is rapidly growing and may soon reach the 9 million milestone, referring to this as a "wealthy embarrassment."
Tibo asked the community whether they should reset usage limits again or provide more space, as multiple optimizations and temporary resets have been implemented due to surging demand for models like GPT-5.6 Sol.
The explosive demand for AI productivity tools under market mechanisms is driving developers and enterprises to invest in subscriptions and credits. OpenAI is balancing server load and user experience through limit management, benefiting highly active users while putting pressure on the free tier.
Source: Public Information
ABAB AI Insight
Tibo has previously responded directly on X regarding token consumption, model experimentation, and limit adjustment feedback, pushing OpenAI to rapidly iterate the Codex desktop application and Work product following the GPT-5 series release.
The capital strategy focuses on scaling and retention, with OpenAI mobilizing computational resource allocation through user growth data and reset mechanisms, aiming to establish Codex as a core workflow for developers to secure long-term paid conversion.
Similar to the early explosive growth of ChatGPT followed by limit management, the current phase is a transformation of AI agent tools from testing to large-scale adoption, with OpenAI maintaining its lead due to infrastructure advantages.
This essentially represents a concentration of capital accelerated by technological substitution, with explosive adoption forcing platforms to optimize pricing and capacity, shifting funds from traditional development tools to integrated high-performance LLM ecosystems, reshaping the software productivity industry chain.
ABAB News · Cognitive Laws
User explosion is the wealthy embarrassment of servers.
The faster the growth, the more resetting becomes a normal lever.
The platform sells capacity, while users sell productivity time.