I'm also trying out a lot of different approaches, and generally I like the sparring I can get from LLMs but with the acceptance that they make mistakes and som

I'm also trying out a lot of different approaches, and generally I like the sparring I can get from LLMs but with the acceptance that they make mistakes and some things still need some fact checking. But the models are improving fast, so I'm positive about this approach just becoming better going forward I'm working on a side-project called tech-professor.com which is a platform for learning. The content is built directly from the source code in your pull request and repositories you follow. So you can quiz yourself and your team based on the code you ship. It's in beta and a lot of changes are still on the way, but if you want to check it out and give feedback, feel free to sign up for free.