Post by @glyph@mastodon.social

RE: https://social.bau-ha.us/@raganwald/116972012374028924

I am not an ML researcher. Even when I thought the field was abstractly kinda cool, it wasn't something that interested me. So I don't have any technical skills directly relevant to evaluating a claim like "inference is getting 40 times cheaper every year!" or "concerns about water use are overblown!"

However, I am increasingly confident that anyone making such a claim is just lying now because a hallmark of making a robust technical claim is *not constantly hiding all the relevant facts*

23 hours ago

“Like Enron, they’re using special purpose vehicles, or off-balance sheet arrangements such as legally distinct subsidiaries, as a way to make their financial reporting look healthier than it actually is — often a glaring sign that something is deeply amiss behind the scenes.”

https://futurism.com/artificial-intelligence/ai-companies-hide-debt-off-balance-sheet

1 Replies

0 Boosts

fancysandwiches

@fancysandwiches@neuromatch.social 6 hours ago

@glyph if the water use issue was overblown then I think we'd see more transparency around the numbers than we do. Instead we get deals signed that keep the numbers under NDA. Plus, we've also seen that the amount of water is only part of the problem. Meta just recently did a bunch of damage to some water infrastructure in Wyoming, and they were using the "reclaim" water or whatever it's called that doesn't go to humans. Another Meta data center project is dumping heavy metals and other contaminants into a river in Wisconsin.

https://www.msn.com/en-us/money/markets/wisconsin-resident-flags-water-contamination-near-meta-s-new-1b-data-center/ar-AA246msD

MSN