When you say you work with very large projects, do you mind sharing an estimate of how many LoC (a poor metric, but still), language and contributors your project has? From my experience, when you do not have problem with bulk refactoring, either your project is not big enough, or do not have many change or no so many contributors so that any kind of replacement with review is not a significant issue. In these cases, it's hard to see the benefit of features offer by IDE/LSP. Just like when do you see bubble sort and quick sort have similar performance? When there is only a few items in the list.

And finally, I also an Emacs user, so I ask you to experiment one thing: instead of using find-grep, try consult-ripgrep and see what improvement it offers compare to find-grep. And remember that an IDE can do even better than that.

So when someone says he do not need an IDE because its feature are just distraction, my first question is always about the biggest project that he has worked with, and the second one is how he do rename/navigation/... with his code base. Never find a single guy does not fall out of these 3 groups.

Copilot is another thing for me. The reason is its accuracy is not good enough yet. The moment it can produce code align with my intention 100% accuracy, just like how search/refactor/suggestion that IDEs are currently offer, I will include it in my workflow instantly.

I have a hobby with natural photos, and in the last 2 years, I have stopped spending time browsing pages as most of them are garbages generated by AI.

One thing many people hype about generative AI as they believe human also make mistake the same way as AI do, so at worst you have something similar to human mistake. Yes, but the volumes generated by AI are at magnitude bigger, and only a limited amount of human can validate and filter it out. If there is nothing change, this volumes of garbages will surely overflow and there is no way we can differentiate bad and good, fake and authentic contents.

I have more than 20 years with backend development and just some limited experience with frontend tech stacks. I tried using LLM initially with for frontend in my personal project. I found that code generation by LLM are so good. It produces code that works immediately with my vague prompts. It happily fixes any issue that I found pretty quick and correct. I also have enough knowledge to tweak anything that I need so at the end of the day, I can see that my project work as expected. I feel really productive with it.

Then I slowly start using LLM for my backend projects at work. And I was so suprise that the experience was completely opposite. Both ChatGPT and Claude generated code that either bad practice or have flaw, or just ignore instructions in my prompt to come back to bad solutions after just a few questions. It also fails to apply common practices from architecture perspectives. So the effort to make it work is much more than when I do all coding myself.

At that point, I thought probably there are more frontend projects used to train those models than in backend projects, therefore quality of code in frontend tech is much better. But when using LLM with another language that I did not have much experience for another backend project, I found out why my experience is so much different as I can now observe more clearly on what is bad and good in the generated code.

In my previous backend project, as I have much more knowledge on languages/frameworks/practice, my criteria was also higher. It is not just the code that can run, it must be extensible, in right structure and in good architecture, use correct idiom ... Whereas my frontend experience is more limited, the generated code work as I expected but possibly it also violated all these NFRs that I do not know. It explains why using it with a new program language (something I don't have much experience) in a backend project (my well know domain) I found a mixed experience when it seems to provide me working code, but failed on following good practices.

My hypothesis is LLM can generate code at intemediate level, so if your experience is limited you see it as pure gold. But if your level is much better, those generated code are just garbage. I really want to hear more from other people to validate my hypothesis as it seems people also have opposite experiences with this.