Raw LLM Responses

I think the use cases are overhyped. If you scope it to the right use cases that it is actually capable of I think it does a good job at those use cases. For example, it's really good for making rapid prototypes where you're going to throw out all the code, and that's a big deal, because despite being incredibly helpful, prototyping has often taken more time and effort than we'd like which is why prototypes often get deployed to production. AI makes it a lot easier to do a quick and dirty prototype that are ready fast enough to act as an actual throwaway prototype and that lets us try out more approaches faster, and that means we can write better code. That's a big deal.