[jedwhite]

Thank you! Andi is starting to get really good at those sorts of serious questions, and that's a great example.

The trick questions that everyone likes to ask any new AI give us great training data for improving the intent-detection edge cases. But Andi's mission is these more serious questions that it is trying to help people with. Linguistic acrobatics are for the next generation of gpt-based models. So your feedback really lifts a coder's heart!

Users fall into two camps with the UI - pretty much 50/50 between "it makes my eyes bleed make it look like google" vs "I love it, this is the future". We have some thoughts on how to reconcile the two things :)

Thanks again for trying it out and grateful for all and any feedback and suggestions you have!