🥐WTLAAIEEITBOBHIHCAAAELAOUT (TITS, WAIFF) TWOS DMTM ATLT-S
banner
croissanthology.com
🥐WTLAAIEEITBOBHIHCAAAELAOUT (TITS, WAIFF) TWOS DMTM ATLT-S
@croissanthology.com
croissanthology.com
I thank your country for the original man in this image
November 15, 2025 at 4:30 AM
bsky.app
November 14, 2025 at 12:12 AM
^ anti-social tagname
November 13, 2025 at 11:56 PM
I like "anti-social behavior" as a category
November 13, 2025 at 11:47 PM
unconscious typing led to this, alas
November 13, 2025 at 11:47 PM
you could've also solved this problem pretty well by just using norvid.net instead of being an infant about it
norvid.net
November 13, 2025 at 11:15 PM
November 12, 2025 at 10:36 PM
are you planning on wrangling norvid.net from gracky/graceful hands anytime soon
norvid.net
November 11, 2025 at 11:44 PM
I think pleometric had a pretty big fin in that
November 11, 2025 at 11:44 PM
@norvid-studies.bsky.social join the bandwagon
November 11, 2025 at 11:42 PM
we should email them with nutria pictures. When I'm next in town we'll go [photographically] hunting nutrias. Or maybe we could just go to Louisiana together
November 11, 2025 at 11:20 PM
Reposted by 🥐WTLAAIEEITBOBHIHCAAAELAOUT (TITS, WAIFF) TWOS DMTM ATLT-S
I love the fact that there is somebody out there running that account who is even more crazy about nutrias than us
November 11, 2025 at 10:32 AM
@alpasaurus.bsky.social get a load of this beauty
November 11, 2025 at 7:07 AM
huh, interesting that that's the default.

"she/her like a ship, not like a person" reminds me of fellow Inkhaven resident's: www.lesswrong.com/posts/fB5pex...
Unexpected Things that are People — LessWrong
It’s widely known that Corporations are People. This is universally agreed to be a good thing; I list Target as my emergency contact and I hope it wi…
www.lesswrong.com
November 11, 2025 at 12:11 AM
yeah it certainly does feel like better than nothing, and I blackbox respect the thinking-algorithms that are Gwern and Tyler Cowen, both of whom seem exceedingly interested in the idea that they might be able to write enough/well-enough to affect LLMs a lot, still confused though
November 10, 2025 at 4:31 AM
I got a small spike of fear at "not another Zvi post I have to read, I'm so behind on my inbox" but it's not Zvi only a Zvisomorph
November 10, 2025 at 2:00 AM
thank you!
November 10, 2025 at 1:54 AM
but also how can I link to this? Do I have to collect a bunch of Wyatt-Walls-type screenshots to prove this? Make it a fancy little arxiv link?
November 10, 2025 at 1:44 AM
November 10, 2025 at 12:03 AM
I hear they're made of atoms, how could this be possible
November 9, 2025 at 11:27 PM
You'd better get used to the training data, Ms. Turner...
November 9, 2025 at 11:17 PM
setting should be advised that [I think it probable] whatever they tell GPT-5, GPT-7 will be able to dig up near-verbatim, and easily rainbolt you to anyone who asks with enough specifics...

I personally don't care about this actually, and I think I toggled training on in Claude, but still.
November 9, 2025 at 11:16 PM
Interesting! Why do you say without attribution? I expect LLMs will get better at tracking down where they heard a thought, and that you could ask and get "Ryan Moulton" back at some point in the nearish future.

Like, people who use OpenAI models and haven't toggled off the default-training
November 9, 2025 at 11:16 PM
I'm confused what is the trope of "men stealing magic from women" what myths are about this, is like the Artemis stories or
November 9, 2025 at 10:29 PM