Yeah. And from the statement given, it sounds as though the GGA know it. Was very much a "we're forced into this" vibe.

Me, an academic, laugh / crying at the idea of having time to read for work...

...to my outsider's eye, national IA over the last few years has caused a lot of pain, but no gain (OK, arguably saving USS, but nothing re pay or workload). And HEC didn't even attempt to make a case for why this time would be different.
"That didn't work, let's do it again!".

I'm not sure that core activists appreciate how badly UCU damaged its reputation with the MAB debacle. Ordinary members don't see it as "ooh, that bad faction made bad decisions"; they see it as a union which is addicted to starting fights that it has no idea how to finish.

Yeah, seems that way. (I wasn't aware of the family connection when I posted it...)

More info (thanks to the person who linked it to me): www.patreon.com/posts/that-v...

The fact that peer review is not reliable does not imply that all academic work is wrong or biased.

oh, fair point. I should have said "averaged over a year" or similar!

To add context, going by the figures at en.wikipedia.org/wiki/List_of..., 36 GW year-round would have OpenAI using more electricity than all but 15 countries in the world did in 2023.

(albeit some of those numbers will have risen now, because the AI data centres are in countries)

(except the big cloud providers, who are just reselling, and transmission utilities, who are just transporting)

All I can think of is some oil refineries, if you consider their flow of incoming feedstock as a power.