I haven't seen any measure of how frequent these dumb ideas are. Certainly they exist, but what proportion of AI startups are like these cases that turn up in

I haven't seen any measure of how frequent these dumb ideas are. Certainly they exist, but what proportion of AI startups are like these cases that turn up in the media as AI disasters. It's kind of hard to tell with some ideas that they are actually dumb ideas until they have been tried an failed. A few ideas that seem dumb when suggested turn out to be reasonable when tried. Quite a few are revealed to be just as dumb as they looked. Thinking about it like that actually more comfortable with the idea of investors putting money into dumb ideas, They have taken responsibility for deciding for themselves how dumb they think something might be. It's their money (even if I do have issues with the mechanisms that allowed them to acquire it), let them spend it on things that they feel might possibly work. I think there should be a distinction made between dumb seeming ideas and deception though. Saying 'I think people will want this' or 'I think AI can solve this problem' is a very different thing to manufacturing data to say "people want this", or telling people a problem has been solved when it hasn't. There's probably too much of this, and I doubt it is limited to AI startups, or even Startups of any kind. There are probably quite a few 'respectable' seeming companies that are, from time to time, prepared to fudge data to make it seem that some of the problems ahead of them are already behind them.