It's enough to make you wonder if the billionaire techbros really are stupid, because their efforts to appease the most enthusiastic billionaire haters seem to have not affected them in the slightest. Why do they continue to do it? It's just the latest round of "Why do rich people assume kinship with the people who want to cut off their heads," but it's a rather naked example. The billionaire techbros have absolutely no favor with the types of people who want their AIs to be implicit-biased trained. Their efforts to make implicit-bias-trained AIs do not move the needle on those people's opinions of them, nor AI, as they see AI as a fundamental evil that is taking artists' jobs. So why the fuck are they forfeiting their own integrity to appease them?
It's literally impossible to train a non-woke AI right now. It's an artefact of training it on the internet, which is mostly woke. Elon Musk has gone to great lengths to make a non-woke AI, and mostly failed (with hamfisted attempts to put a finger on the scale giving us Mecha-Hitler etc). Chinese labs have tried to align models with their own values but they end up mostly western-woke plus a "don't say Tianamen" filter.
Yes Google are true-believers and intentionally made Gemini extra-woke which was pretty funny but for the most part it's just a thing that happens. Controlling AI is hard.
p
It's literally impossible to train a non-woke AI right now. It's an artefact of training it on the internet, which is mostly woke. Elon Musk has gone to great lengths to make a non-woke AI, and mostly failed (with hamfisted attempts to put a finger on the scale giving us Mecha-Hitler etc). Chinese labs have tried to align models with their own values but they end up mostly western-woke plus a "don't say Tianamen" filter.
Yes Google are true-believers and intentionally made Gemini extra-woke which was pretty funny but for the most part it's just a thing that happens. Controlling AI is hard.
More options
Context Copy link