[wavesounds]

This would be a great thing for all of the AI companies to devote some energy towards. Especially with their reputations in decline. Surely there must be some patterns the AIs could find if we had enough data about the people who died from cancer.

[opsnooperfax]

A dude at Oracle did that with a recent mass experiment of mRNA technology. We do not speak Of the results.

[witx]

Yes, one good data point is pollution, which all these companies are prollific at.

[witx]

Peak hackernews behaviour denying pollution from AI datacenters

[dekhn]

It's more like... the comment doesn't contribute anything of value to the conversation.

[witx]

Or I'm a young person with lots of cancer cases "near me" and I'm worried that we're moving in very undesirable directions and everyone clapping their hands

[dekhn]

In my experience, many claims of cancer clusters aren't actually caused by the things people say they're caused by. For example, for a long time people thought living near power lines caused cancer, but careful work suggests that instead, it works through an intermediate variable (people living near power lines often have lower socioeconomic status, which is closely coupled to worse health outcomes). AI datacenters (or datacenters) are really not the most likely cause- their emissions and impact are far smaller than living near a highway or a coal fired power plant.