Whoa, I asked Claude and apparently that was the correct guess:

Hah, not quite, I think it's that that cat would be perfectly average across the three PCA dimensions? Which would maybe imply it's average across the regular dimensions if they were normalized to the amount of variance (just guessing, this part is probably over my head)

Also if you search for "blocklist" on here you'll find a lot of people reporting getting added to various garbage / bot / scam lists and there doesn't seem to be much recourse

I'm on a blocklist because I follow someone's definition of "conservative" accounts on here. This has resulted in over a dozen people blocking me which is no big deal, but I can't say it isn't a bit dismaying, so it's something I'm trying to wrap my head around.

lol, honestly afraid to name them

I think what it’s making me realize is just how much Threads does to circulate your posts to people who might like them, whereas Discover seems much more basic

How does the “Discover” tab work then? Assume that’s using some algorithm to sort stuff you haven’t seen before

right back at cha

It always struck me as shortsighted to only worry about what happens when AI makes mistakes as opposed to what happens when AI becomes more accurate than l humans.

Prison recidivism is another tough example: human judges are probably more racist than LLMs at this point.