My guess is their strategy is to aim for clueless developers, with more focus on "proactivity" with Fable and very large models. I have a few non-technical friends who dabble in AI coding and I am watching with amazement how much better their creations get with the newer models. 3 months ago, they would "ship" stuff with a backend at 127.0.0.1 and api keys in the frontend. Right now they produce, without an iota of understanding, decently designed, but not very secure or efficient cloud native stuff that sort of works. I bet in 3-6 months they will be releasing secure, scalable, well architected stuff without spending a single thought on any of those things. And I think this is where the really big models will be a bit of a moat, for a short while at least.

Fable 5.1 is coming, then GPT-6...