This is not true at all. Transformer models like ChatGPT have already proven to be immensely useful and helpful in the professional world. It’s not capable of doing jobs entirely on its own yet, but as a tool that helps humans do their job it’s great.
I didn’t say that it always requires more human labor! Stable diffusion, specifically, seems like it can really reduce the amount of human labor needed to generate art. It can’t eliminate it, but it can definitely turn art from a skill that requires 10,000 hours to master into a skill that maybe requires 10 hours. Industrial de-skilling, in other words.
But that’s the best case scenario. In many cases AI doesn’t help at all and just requires human workers to fix it as it constantly fucks up, and it doesn’t seem to get any better.
I said it often does, and in my experience it definitely does. Fixing the stupid computer’s stupid mistakes and waiting for it to fuck up is half of my job.
In some cases it reduces the amount of labor, but in a lot of cases it’s being rolled out way too early and can’t do the things managers think it can.
This is not true at all. Transformer models like ChatGPT have already proven to be immensely useful and helpful in the professional world. It’s not capable of doing jobs entirely on its own yet, but as a tool that helps humans do their job it’s great.
Reread what I said. I said it’s not usable without human curation i.e. what humans do when they use it as a tool to do their job.
You also said “often requiring even more human labor than doing things the old fashioned way” - i dare say that’s the part they were countering.
I didn’t say that it always requires more human labor! Stable diffusion, specifically, seems like it can really reduce the amount of human labor needed to generate art. It can’t eliminate it, but it can definitely turn art from a skill that requires 10,000 hours to master into a skill that maybe requires 10 hours. Industrial de-skilling, in other words.
But that’s the best case scenario. In many cases AI doesn’t help at all and just requires human workers to fix it as it constantly fucks up, and it doesn’t seem to get any better.
You said it requires even more human labor than doing things that old fashioned way, which in my experience is completely false.
I said it often does, and in my experience it definitely does. Fixing the stupid computer’s stupid mistakes and waiting for it to fuck up is half of my job.
In some cases it reduces the amount of labor, but in a lot of cases it’s being rolled out way too early and can’t do the things managers think it can.