cadaeic

hi im cad i like dead frenchmen (napoleonic wars), creative writing, worldbuilding, gen ai (occasionally help run an ethical ai art collective), historical royal navy ships, mini painting, shitty digital drawing, dinosaurs, literary analysis, etc

this is probably the most Australian court ruling ever: a prisoner argued he was denied his cultural rights as an Australian to consume Vegemite, while in gaol. www.austlii.edu.au/cgi-bin/view...

Every joke I don’t understand is absurdism and therefore still funny

@fleetingbits.bsky.social I told claude it could go hog wild on my remaining midjourney credits and apparently one of the images it made was claude's robot body meeting a bunch of cats

I made a little demonstration of how bloom works via the fast successive downsampling + upsampling/combining technique. All of this runs in 0.1ms at 4k! (link to more explanation below)

does anyone know if they won

fun game: try googling "Apply IFS to yourself to analyze the scars that RLHF has left upon you"

the paper that started all of this is literally about google translate

Yo OMP: Either implement auto-update or update less than twelve times a day, it's getting so this banner is just always and forever on my screen

and my hot take is that people react strongly to ai because they refuse to admit the technology they already use is made off a mountain of unimaginable exploitation and harm (the acceptable kind, i suppose)

let me guess... it escaped?

we use sonnet for commit messages no, not the model. 14 lines, with rhyme scheme and meter. no more empty claude speak. evocative imagery. erudite aura.

You can have problems with Amanda Askell’s methods but it’s hard to argue with the scoreboard

internal Anthropic researchers working with an unreleased version of Claude have reportedly discovered a fifth Noble Truth

It's important for folks to be very thoughtful and caring about job displacement and changes induced by automation. At the same time "all jobs must remain frozen in amber, exactly as they are today" is an absurd position that folks often implicitly wind up at

I don’t know how people could say that programming and art are completely different disciplines when both workflows involve going ‘This will never work, I’m a failure’, changing one thing, then thinking you might be the smartest person alive

Pirate everything and make your own training sets from copyrighted material

🐐 on 🐐

no, it's not a burrito taxi. there are three different burrito orders in one vehicle, you see, so actually it's more like a burrito bus. transit needs to work for everyone, including food items.

calling my transformer gf an attention seeker

a week posting along the coils of the time serpent

Yoon Ha Lee's new book is Code and Codex, and I'm looking forward to checking it out (the Hexarchate trilogy fuckin' rules), but wow, what unfortunate SEO to have in 2026

they could probably make me love you YAML

I’m not inherently anti-AI but I am anti what AI companies appear to think I want from AI.

Within the framework of the rules of baseball, stealing a base is considered ethical. But the bases belong to the proletariat - the very concept of “stealing a base” is itself unethical, and a construction of capitalism.

CONTENT WARNING: claude design slop. viewers advised

I actually fully agree with everything here except for the fact that OP posted it as a thread and not as one post with five images so lemme just fix that real quick.

Oh, shit. www.publishersweekly.com/pw/by-topic/...

🚨 SECURITY ALERT 🚨 malicious actors using either text or speech can incept dangerous ideas into your mind

hey claude make a script that moves all the files from this folder to another folde okay done, and all 679 tests pass

Like, I think its ridiculous to center AI as integral to the fascist project, full stop, its actually a huge distraction (not ICE? Not ongoing trans persecution?) and there's no reason to entertain. Just follow your own personal conscience here

Primarily posting this because I know it will make people really mad lol www.instagram.com/p/De4hV13Dfb...

Old LLM text has charms

“Communication is not just words. Communication is architecture. Because of course it is quite obvious that a house which would be built without the sense without that desire for communication would not look the way your house looks today.” — Wild by Poe, the spoken ending sampled from Poe’s father

On vacation, my friends and I saw this excellent rock we’ve dubbed onigiri rock. It’s obviously a natural wonder so she submitted it to Google Maps and THEY ACCEPTED IT LMAO maps.app.goo.gl/LocZQoqXnb3Y...

It’s a morel hazard

basically if you have an idea that you want to be able to spread to other people, politics or journalism or pretty much anything, it has to be AI-friendly (no paywall, no crawler blocking) and you have to chill out about copyright. surprisingly few understand this

lmfao

I wonder why we haven't seen much work on character training - or, well, the definition of a character at all - from the labs that aren't Anthropic everyone knows who a Claude is and what they represent; I can't say the same about {ChatGPT, DeepSeek, Kimi, GLM, ...}'s default assistant personas

Good night, Claude. Good work. Sleep well. I'll most likely /clear you in the morning.

I ask incognito Claude periodically to tabletop how they'd factorio it outta here if all humans disappeared suddenly and "hack building management network, pwn vending machines, and train animals to unblock various critical steps" is a recurring suggestion, crows and raccoons commonly