Raw LLM Responses

Agreed. My concern is that the bottom is being hollowed out. Companies are replacing junior engineers with AI, or the juniors are all using AI and don't really understand the code and are not being given the opportunity to learn. In 20 years, there won't be that senior knowledge base anymore because everyone was learning on AI. The counter argument is that in 20 years AI will be advanced enough that we won't need those senior engineers either, but I'm not convinced we're going to see a huge breakthrough in model performance, and I very much doubt we'll get human-level AGI. We're already seeing the huge performance improvements of the last couple of years begin to slow. I expect we will continue to need humans driving AI, but we won't have that if people aren't learning to program without AI.