According to the author, the Codex and Claude systems hide traces of reasoning, and data from OpenAI was likely not provided for the demonstration. In practice, agents stream text in less than 1% of the time; the remaining time is spent on non-streamable reasoning and tool execution, so traceability is limited.
AI-generated text
AI agents stream text in under 1% of the time; the rest is reasoning and tool execution
According to the author, the Codex and Claude systems hide traces of reasoning, and data from OpenAI was likely not provided for the demonstration.



