I have no illusions that what I've been building was ever going to be commercial or even particularly useful (it just offers code improvements, bug fixes, etc.

I have no illusions that what I've been building was ever going to be commercial or even particularly useful (it just offers code improvements, bug fixes, etc. as errors are piped into the system – it's very rudimentary but neat), but it's striking right now how incredibly deep Microsoft's moat in this arena is right now. I realized it would be best as a VS Code extension in its current form and started writing that this morning. I was thinking Github would do this eventually, it was obvious — copilot already existed and was primed for this — but evidently they'd already built it a while ago. I was also thinking eventually it could be a bot which generates pull requests based on error correction, but... They did that too. If I wanted to go all in one this project, say as a hypothetical competitor to Copilot, I don't see how it would be possible to succeed. Any clever UI improvements would rapidly find their way into Copilot, offering far better solutions due to their absurd access to pertinent data. They can even train Copilot to resolve errors based on how they've been resolved in the past; it could probably patch all kinds of common errors reliably soon. I think this is cool in the sense that it'll be genuinely useful to people, but frightening to realize how dominant MS is here, in this moment. Maybe I'm not creative enough and I can't see a way to build something novel with this which they haven't already. That's arguably likely. At the same time... I'm a little shaken at how narrow of an opportunity anyone had to try to get into this space. I don't think it's a good thing.