6 Comments
User's avatar
Yusuke Tanaka's avatar

Fascinating connection between declarative AI coding and Lean IT principles. Your point about "designing around human judgment" is exactly what we need in this AI era. One of our community members, Yusuke Tanaka, is exploring this very intersection—applying the "Kaizen" philosophy to AI workflows to ensure that productivity gains don't come at the cost of human insight. Great to see these historical lessons being applied to the future of software engineering!

James Kaplan's avatar

I am trying to think through the tension between apply higher level lean principles as we move to AI-enabled software engineering versus movign (in effect) from smaller to larger batch size.

Because we need engineers to create larger "batches" of specifications.

On the other hand, parallel (rather than sequential) processing will be essential!

Dean Deng's avatar

Interesting angle on lean IT. I think the translation gap between user intent and code will probably be better as training continue to scale, but review/audit will be more and more important

James Kaplan's avatar

Do you mean model training or people training?

James Kaplan's avatar

that makes sense!