Why AI Struggles with Complex Coding Tasks and How It Excels

Key Takeaways

I think that the AI systems today have been trained on the the data that's been available and so that they're good at certain things. They're not not good at other things. So for what we're doing at Eigor, I've done a lot of Swift UI code and AI tools have written me a lot of Swift UI code because I I don't I don't know the uh the language very well. uh and also I've done uh a bunch of backend code some front end frontend code AI tools had a huge training set of data that was open source and can produce pretty good results I think whether it's UI based mobile code or highly complex backend code whether or um multiple processes multiple systems hitting databases asynchronously it struggles in those two areas and I think it's probably because of the lack of strong training sets And my guess is that the same thing is also true in the firmware world as well.