Working on incorporating existing AI capabilities into formal methods is one of the most robustly differential-tech-development things you can do that looks like advancing AI capabilities, imo. It seems like none of the next-proof-step models are even using Transformers yet (let alone code-davinci-002), but there are already encouraging results from TreeLSTMs.
cited on: code-davinci-002
Reproduced against link rot, credited and linked to its original. Yours and you’d rather it weren’t here? Open an issue.