Raw LLM Responses

"If you tried to use it in some sort of like enterprise software or any app that's larger than what took somebody a day or two or even a week to whip up you see the current limits" Very true, there is no way you can just "vibe code" in a setup like that, which is pretty standard. 99.9% of businesses haven't been built in the last few days and aren't just a few lines long, most codebases are huge. This still makes it necessary for a Software Engineer to dig into the codebase, understand the data flow, the modules, etc., and then make changes to a specific part of the code. And I haven't even mentioned the whole understanding of business logic, software architecture, infrastructure, and environment. So, AI is definitely boosting my productivity in reading code, writing code, and fixing bugs, but only after this whole process of understanding, which takes years of study and experience to achieve. And I still need to review a lot of the AI generated code to make sure it's not writing some dumb shit. For sure, it won't replace developers. Maybe we could debate reducing the number of engineers due to the productivity boost, but that's another topic (which I think has some ground of truth, so I would love it if you could do another video on that specific part!). Anyway, for those already established in the market, it's insanely good to just write down your thoughts on what you literally want and have an "intern" write it all for you while you just drink coffee. :)