Gotcha. I think this might be where Protobuf falls short because you have to "source" the tooling for all its different outputs, but support for OpenAPI 3.0 spec generation has existed for over a year.
Amazon is great at a lot of things, but AI and Machine Learning aren't it. From what I saw when I worked there, they don't have the talent to produce their own models (which is, I think, what we all expected to see), but the surprising thing is that they haven't partnered yet.
There's an HN post that takes first place on my bookmarks bar, which is an Ask HN about the best paper people read in 2020 [^1]. Every once in a while I'll open it and browse through a ton of amazing papers and research and try to learn something new. Hasn't let me down yet.