OpenAI didn’t just release another model, it rearranged where work happens, who controls it, and how quickly dependence forms. Frontier and GPT-5.3-codex landed on the same day for a reason, and the timing alongside Anthropic’s Claude Opus 4.6 makes the shift hard to ignore. From the outside, this looks like routine platform progress. Underneath, it marks a change in how software gets built, how trust migrates, and where lock-in quietly starts to harden…

Become a paid reader to get the close read: what Codex actually changes in day-to-day work, why Frontier matters more to boards than developers, how Anthropic’s response reframes the competition, and where the real power now sits once context stops being portable.