I've been recently diving into Claude Code and it's been great. There's some weird feeling to it though.
One thing I'm sure is that you don't get the feeling of ownership of the final result, therefor it doesn't quite motivate you the same way as writing the code yourself. After all, this kind of makes sense because in the end, the code isn't technically yours.
The feeling of ownership is the reason why there are so many new projects on github that have been co-coded with AI intensively for a short period of time, and then ...nothing. Almost as if the project was abandonned the moment it was born.
Another thing that I find weird is the "taming" process.
Without strict rules, guides, sandboxing or a very specific thought process. Claude Code can go from doing something you expect to going so far off the rails where you have to spend hours trying to fix the derail.