It is too early to follow this advice. The world is still at the Cambrian Explosion part of AI use. We're going to use AI for lots of things that we probably sh

It is too early to follow this advice. The world is still at the Cambrian Explosion part of AI use. We're going to use AI for lots of things that we probably shouldn't. But we won't know until we try it. Maybe we'll use AI as a personal PR department. Companies want to present a unified, curated experience to their customers. People do too: we dress up sometimes; we swear differently depending on the audience; we write emails differently for work vs. friends. The only reason people don't use PR departments is because it's too expensive (you can bet that Bill Gates uses a PR department to look over his annual letters). Does that mean that people will all sound like companies? Of course not! Different companies sound different. How Monster Beverage Corporation talks is different from how Boeing talks. Maybe we'll use AI to eliminate all boring person-to-person coordination. Instead of sending emails saying, "Where do you want to eat on Saturday?" we'll just have our AIs coordinate. It will be like having an executive assistant who makes sure everything goes smoothly. Maybe we'll use AI to expand or deepen our relationships. There are many friends that I wish I were more in touch with. I don't need to have a heart-to-heart with them every week, but I would love to know how they're doing this week. If they emailed me every week just to tell me things like, "Gargoyle's in David Square closed down and I'm bummed." or "My kid saw a sunrise for the first time and he thought it was lame." I would read that. Imagine that once a week, an AI prompted me with what to send and I said things like, "Tell everyone about us going to beach and the new restaurant we went to. Tell close family/friends about how John is struggling in school and ask for advice." Or maybe, throughout the week, I tell the AI things like, "Send the best pictures of our beach trip to friends and family on the next update." Okay, maybe you think all three of these ideas are stupid. You're probably right. But I thought Twitch (watching people play video games) was a stupid idea too. The bottom line is that we don't know how people will end up using AI and prematurely dismissing any specific use case smacks of unwarranted confidence, IMHO.