The biggest AI "breakthroughs" usually go silent after a short marketing drive.

I remember the slowdown from Tesla making huge claims about full AI self driving to a point where it was just a past tense "we achieved it" (it didn't) and then moving on to other marketing topics. The AI self driving marketing drive is coming back again, in the same style as it happened before. "We were optimistic about it, we made errors, we were defeated by it, we are working on it, we achieved it. Whatever. **Just buy our cars.**"<br><br>Same happened with Google's "AI chatbots". They are really bad in reality -- and I have to insist that they are, because there is stealth shilling here and people still think Google has been able to achieve a level that they have not achieved.<br><br>The pattern is: market it, make everyone believe, pull back, hide from it.<br><br>You will see this happening with OpenAI and Microsoft AI products. Startups will build up a mythical product, receive funding, fail to deliver, and hide from it.<br><br>This is what happens when you are trying to sell a product that often is simply 1) not really possible and 2) morally wrong.<br><br>The whole AI industry has become a marketing industry that is selling smoke. And we are going to have to deal with the massive "AI crash" that comes out of this hype and reality mismatch.