The Ledger
Every statement that passed quotation and attribution checks. Mix any filter with any other: certainty 1/5, debate potential 5/5, or both at once.
why aren't all 5 resolved? a statement only gets an assessment when the public record can support or contradict it. opinions and what-ifs never can, and 0 checkable ones are still open, waiting for their date. predictions held up or didn't; assertions are supported or contradicted. on every card: ▮▮▮▮▮ certainty · ▮▮▮▮▮ debate potential. speakers are clickable
Eliahu-Ontiveros: Meta modern AI data centers omit backup generator sets
“Another interesting thing they've done, for example, is more at the data center level than at the energy level, but no backup. You look at any one of the modern meta data, so they don't have any gensets, or they have, you know, like, 10 megawatts for a hundred…”
Bullard: DeepSeek V3 matched Meta Llama 3.1 using under 2.8M GPU hours
“And the sort of benchmark to compare it to would be the Lama 3.1 model from meta, which had four and five billion parameters and just under thirty one million training hours. And deep seek is getting similar results of V three with six hundred and seventy one …”
McAvoy: Meta, Oracle, and Microsoft are building in-house energy teams
“And especially, like, you know, even the likes of, like, Meta and, you know, you have your Oracles of the World and your Microsofts and, you know, everyone else who are trying to build, build, build, build these out. They're investing even in their own teams. …”
Meta study: AI energy costs split evenly across pre-processing, training, and inference
“There was a study we did when I was a visiting research scientist at Meta where we Found that energy costs for AI were roughly broken into three categories. There's a data pre-processing aspect as well, and that's about a third. The training is another third, …”