GenAI work good enough. They don't need to be any better to be usable in hands of people who learn how to use them. It's irrelevant if it makes mistakes or lies . What is relevant is the cost: social, ecological, our impeding economical disaster. It doesn't matter how good they are or if they will improve. The cost is unacceptable and this technology has no right to exit. Focusing on the shortcomings of them as product - a millionth example it it counting letters incorrectly, or that drive to the wash example - is not the way to go. It's funny, but it's just playing their game: it's technical problem, it can be fixed in the next iteration. Jut a few more billion dollars, just half a year. Meanwhile the actual problems emerging with GenAI being present everywhere are shunned of. Have you read Antropic reply to Pope? They have nothing to address any non-technical problem. They don't even have a framework to even think about it, because that's not where the discussion is. VCs don't ask, press doesn't ask, and we just focus on some random tidbit on Google. Look at our shiny benchmark, forget about the CO2 we are generating.