1. In OpenAI's case, they dumped 1.8 GiB of Lean proofs on the world. I don't think they've done anything ethically wrong by doing that, but it's the exact opposite of "commit by commit". In fact, I'd say human mathematics has been much closer to "commit by commit", usually using smaller results as stepping stones.
2. You can have "commit by commit" without formal provers like Lean. Just keep your text in a Git repo.
I'm sorry if I've misinterpreted you, but:
1. In OpenAI's case, they dumped 1.8 GiB of Lean proofs on the world. I don't think they've done anything ethically wrong by doing that, but it's the exact opposite of "commit by commit". In fact, I'd say human mathematics has been much closer to "commit by commit", usually using smaller results as stepping stones.
2. You can have "commit by commit" without formal provers like Lean. Just keep your text in a Git repo.