Loading…
This is a firsthand, single-source example of agentic task completion, with limited evidence for broader capability claims and a concrete transparency concern.
Limits of the evidence: The account is a single author’s description; the article does not provide independently verified outputs or the execution code.
What to watch
Assessment revised Oct 4, 2026
Generating running routes with GPT-6 Astra and ChatGPT Work
Simon WillisonSep 12, 2026First report
First assessment, from Simon Willison
No maturity ladder applies to this desk.
Sens.ai aggregates and assesses published reporting. The assessment above is machine generated; the original sources are authoritative.