Codex Head Tibo: Recent Silence Due to Funny Internal Slack and Everyone Locked in DevDay, Tuesday Will Be Interesting
Tibo Sottio wrote on X that his recent silence is due to the funny internal Slack and because everyone is locked in DevDay, which will be interesting on Tuesday. The San Francisco DevDay is scheduled for Tuesday, September 29.
Sottio is one of the core creators of Codex and currently oversees OpenAI's core products and platforms, including ChatGPT and Codex, reporting to Greg Brockman. He is known for resetting usage limits for developers on social networks. The Codex CLI is written in Rust and open-sourced, supporting multiple models at runtime. ChatGPT and Codex are being integrated into the same super application, while independent products like the Sora video application have been consolidated.
In the same week, a frontend for ChatGPT Pro Max appeared before launch, with a monthly fee of $500, emphasizing "the fastest Work and Codex." The existing $200 Pro has paused new sign-ups due to demand for GPT-6 Astra. DevDay serves as a pricing window for features that are internally locked, with coding agents' long tasks being the most demanding in terms of inference latency.
The "full lock" positions the launch event as a convergence point for capacity and product calendars. The Slack jokes and external silence are typical preheating for developer relations: no leaks on specifics, only on temperature. After Codex transitioned from research preview to one of the fastest revenue-generating product lines, the audience for the launch event shifted from researchers to engineering teams that treat the model as labor.
In market mechanics, what is sold is the expectation of the launch event, and what is bought is developer renewals and upgrades. Demand comes from subscription users waiting for faster coding agents, while supply is constrained by inference cluster scheduling. The beneficiary is OpenAI, which can turn strings into price lists on Tuesday, while the pressured party is users already queued at the $200 tier. Funds will flow from the new tier's monthly fees into inference bills after the launch.
Source: Public Information
ABAB AI Insight
Sottio's reduced tweeting is because product authority has shifted from community operations to the countdown for the launch event. Codex can open-source the CLI and reset limits because it must remain in the developer toolchain; ChatGPT aims to harvest a super application because the consumer side has nearly a billion weekly active users. One person manages two lines, and DevDay needs to satisfy both coders and chatters.
The capital path is subscription for agent work hours. If the $500 tier is realized, Codex could quickly sell Cerebras or similar low-latency clusters as SKUs. The pause on the $200 tier indicates that the existing tier has been filled with long tasks. The silence before the launch reduces leak costs and increases attention rent for Tuesday.
Similar preheating can be seen in Apple's pre-launch engineer silence and cloud vendors showcasing instance families at Re:Invent. The industry is at a stage of transforming coding agents from plugins into operating system layers. Whoever promises continuous overnight latency at DevDay will find reasons for the $500 tier.
Structural judgment indicates a transfer of pricing power. The pricing of coding capabilities shifts from plugin seats to speed tiers at the launch event. The mechanism is: the tighter the internal lock, the higher the external premium for "the fastest"; jokes in Slack serve as a pressure release valve, while the price list is the product.
ABAB News · Cognitive Law
The full lock launch event sells speeds not yet written into the price list.
The one resetting limits is deciding how much the limits are worth.
Once coding agents can run all night, chat tiers will no longer be the top tier.