Dystopian AI hellscape is when your image generator makes non-whites
Dystopian AI hellscape is when your image generator makes non-whites
This has convinced me more and more that the only possible way forward that’s not a dystopian hellscape is total freedom of all AI for anyone to do with as they wish. Anything else is forcing values
This dude also posts a direct link to a race-bait bluecheck two comments down, further cementing hn AI threads as downstream frog twitter.
I know this one might be stretching it a bit, but every comment on this post is sneer-worthy, every single one.
You're viewing a single thread.
It’s a horrific tragedy that John Locke should have become America’s (made up) philosopher king after Reconstruction, when Thomas Hobbes was right there
I just want to draw special attention to the reasoning here
BigTech, which critically depends on hyper-targeted ads for the lion share of its revenue, is incapable of offering AI model outputs that are plausible given the location / language of the request. The irony.
- request from Ljubljana using Slovenian => white people with high probability
- request from Nairobi using Swahili => black people with high probability
- request from Shenzhen using Mandarin => asian people with high probability
If a specific user is unhappy with the prevailing demographics of the city where they live, give them a few settings to customize their personal output to their heart's content.
Not gonna say anything in particular about that reasoning, just gonna draw attention to it
BigTech, which critically depends on hyper-targeted ads for the lion share of its revenue, is incapable of offering AI model outputs that are plausible given the location / language of the request.
Hardly surprising given how insistently the ad industry seems to be to show me ads for things I actively dislike or otherwise wildly missing with the products, services or attention seekers it tries to match to my interests and demographic profile.
If it is indeed true that AI companies just do naive search-and-replaces on the inputs to have more diversity in the outputs, that seems like a really crude and low-effort solution and also way more giving a shit than I would expect from an AI company.
Ultimately I think the main disagreements between me and hackernewses is twofold. I don't mind seeing black people, but I also don't take LLMs and ML image generation models seriously as a tool for anything other than shitposting and spam production.
Oh it's definitely how they do it lol, remember the "Person holding up a sign that says" trick.
Pixel art of an adult person female holding a text sign that says adult person chicken.