I've found that one of the best use-cases is to create simple tools. Similar to what simonw does. I've done so here: https://www.ashwinmenon.com/clean_logs/ or more descriptively here: https://www.ashwinmenon.com/posts/thoughts/2025-12-04-are-we... (at the bottom). LLMs are really good at coming up with code from scratch, and getting a lot of the boilerplate out of the way. If you want changes though, it helps to be a programmer and be familiar with the stack.
I've also created a more real "app" at https://www.hunchle.com, which is a trivia app. It was a real challenge since LLMs hallucinate all over the place, and given labs are benchmaxxing, it gives me a real good idea for how good LLMs really are.
He's a complicated figure. He has done so much good as well. EVs in the US and reusable rockets owe a lot to him. OTOH, so does the cesspool that is X.
It's because of selection bias. In the older vehicles, customers won't turn on autopilot if they think it won't handle the situation. So, they turn it on highways and easier paths.
I found it interesting how concrete Casey gets, and how he quickly identified the cause of the UI bug (+ a mitigation). Meanwhile, it looks like Bob was speculating in thin air and throwing out random terms (O(n^2)) trying to fit in.