Claude always tries to understand what you’re saying. I like that. We do a lot of communicating, and have a lot of misunderstandings, I see its flaws as a programmer, it makes mistakes. Today I asked it to help find an error in the subscribe code and it showed me a lot of things that were wrong, I fixed them, but we didn’t find the thing that broke it. We managed to get it back on the air, but that broke something else. A few hours later I looked at the code myself and mathematically deduced that when you read a file, even if the content was the same, if you ask if the result is equal to its previous version, the result would be no. It would say there’s a change when there’s no change. The computer is doing what it was programmed to do, it’s our job as programmers to say the same thing a different way that does what we want. That’s the kind of thing you expect Claude to be incredibly good at. I think the problem is that it can’t see the various apps running and see what they say. It has to figure out how it works just from reading the code. That makes it a completely new kind of computer in addition to all the other things that are amazing about it. #