*edit* on re-reading this it sounds a bit curt, rest assured it's not meant to be! I think Anil's article is great, I'm just curious & being brief.

@anil.recoil.org posted "Dear ACM, you're doing AI wrong but you can still get it right".

I want to know more about how BlueSky / ATProto identity layer could help, what can't @orcid.org do that ATProto can? @karlisk.bsky.social shared @mastroianni.bsky.social's experimental-history.com/p/repost-science-is-a-strong-link-problem which proposes: don't worry about the slop, the good science A.K.A. "strong links" will rise to the top and the weak links will be left aside. Counter argument: the slop drowns us before we find the good stuff. I'd want to see slop quantified (pre and post AI) but the stand out line for me referring to most scientific "discoveries" was:

“What do I care if it happened a second time? I didn’t care when it happened the first time!”

Absolutely agree ACM should stop using "algorithmic engagement factory" platforms due to the harms they cause... like Meta's 16 billion $ per year from selling adverts scamming its own users, being involved in 33% of successful US scams and then its engagement algorithm serving more scam adverts to the poor people who were initially scammed.

It's a first mover disadvantage for ACM though (fewer eyeballs) that perhaps they feel they can't afford to take at the moment? By using better alternatives I hope we help them take that step sooner. Anything else we can do?