[kir-gadjello]

Given a list of contributors it's not that hard to reverse-engineer the specific engineering choices made by looking up their publication history.

My analysis https://kir-gadjello.github.io/posts/gpt4-some-technical-hyp... was written without this info, though.