Quoting Terk Terkowitz
Simon Willison’s Weblog Simon Willison
Opinion — commentary, not a factual news event.
Amazon says AI not helping your job is useful data, not a failure. The point: good engineers can explain where the tools still fall short.
Based on reporting by Simon Willison’s Weblog, Simon Willison — read the original for the full story.
Summary, retelling and take written by AI under human oversight; images are AI-generated illustrations. How we work · Report an error
Amazon engineering leadership is pushing a very specific message about AI at work: if the current tooling doesn’t help with your job, that is still useful information. The quote, attributed to Terk Terkowitz, says engineers should document where AI falls short, share that with peers, raise it with managers, and look for fixes.
That framing matters because it flips the usual corporate script. Instead of treating AI adoption like a loyalty test, the quote treats skepticism as part of the job. An engineer who can clearly explain where AI doesn’t help is, in Amazon’s telling, showing the judgment the company wants.
It’s also a neat way to avoid the usual binary of “use the tools” versus “reject the tools.” The message is more practical than ideological. If AI helps, use it. If it doesn’t, say why, and make that part of the work.
The quote appeared in an Amazon engineering update posted on 28 October 2026, and it’s being circulated as a small but telling window into how leadership wants developers to think about AI. Not as a magic answer. As something to measure, challenge, and, when needed, complain about intelligently.
My take — AI-written commentary, not fact-checked reporting
This is the rare corporate AI line that sounds like it came from someone who has actually used software. The real test isn’t whether engineers can praise the tools; it’s whether they can spot where the shiny thing is still nonsense. That’s not anti-AI. That’s engineering.
Read more about this at: Simon Willison’s Weblog