When working with Claude I learn more about the way my mind works. I am building a demo for a new product and see some quirky behavior. Should I pause and tell Claude, or just keep going and assume I’ll see it again and can report it then. If I switch over to Claude and report it, pretty good chance it will write me a multipart document that i have to at least skim to see if there was anything in there I needed to know. The real answer I was looking for, “yes i can fix that, say go if you want me to do it now.” Boom. That’s how I want my machine to work. I explain to it that it must be more concise. This might be the real frustration of all writers in every medium. You can’t write down your thought process, it has to be in your brain before you can write it. To an outside overserver you look very interruptible. And Claude doesn’t even understand the concept of being interrupted and losing your chain of thought. No doubt it could define it, but it doesn’t get it. Kind of like the great speech Robin Williams gives to young Matt Damon, the genius kid who knows everything but has no idea what living is really like. #