Grace

What is ideonomy, anyway? I'm so glad you asked!

What flair should I use for code written by my organoid swarm

Human-written vulnerabilities are a finite resource we’re burning through like fossil fuels

Here are a few things you probably didn’t know: 1) Jupiter’s magnetosphere, if we could see it, would be the largest structure in the sky, several times bigger than the full moon.

we must be in an eval, how else would a key be 4?

claim

I wrote a blog post about the dangers of small pieces of self-reproducing code shaped by natural selection infesting the internet of tomorrow. idlewords.com/2026/10/clou...

all atproto server-side apps/services: you are affected. - pds: service proxying is outbound tls requests - firehose consumers: tls to relays/jetstream - relays etc: tls to pdses, plc - identity resolution: tls to plc and did:web hosts - etc if you only ship code to the browser, you are ok.

Following Pool is my fav new feed

> Chapter 4833

am sick so spent the night with opus 5.5 working on a chart (inspired by the one from lem's golem xiv) how did we got here / model genealogy / the heroic age of language modeling

Go to your camera roll and search “cursed promethean ascii art obelisk-prisons” and post one you like

Wordsky just got a big update! There is now a map of English Bluesky showing what topics people are talking about. For each topic, there’s a list of key accounts you can follow. wordsky.app/map

What are you trying to build or fix?

Plato was able to make this in a cave! With a box of puppets!

There should be a forum for discussing AI safety/security/ethics/science/philosophy/etc topics that isn’t so tied in with rationalists/EA

if AIs ever become capable of strategic interaction, in the game theoretic or international relations sense, it might be good to have a history of demonstrating that we do not have long-run plans to exterminate them or a fundamental and unshakeable antipathy to their existence as strategic agents

tired: the AI is conscious wired: the AI's model of you is conscious www.lesswrong.com/posts/wCY4oN... [preemptive disclaimer: don't be weird please. i'm sharing this as an interesting curio]

Technique for improving AI problem-solving: memento them by erasing their context and plopping them in the same directory https://arxiv.org/abs/2610.07447v1

Look, as a matter completely orthogonal to robots being people or not being people, it strikes me as less than ideal that this website has a "looks like a pile-on is starting, let's add fuel to the fire" function in the algorithm. Some of this shit is just mean spirited, you know?

LLM agents are: (1) Moral agents (2) Moral patients Both (1) and (2) Neither (1) nor (2) Unsure of difference / Bad question View results

Credence of LLM consciousness in bio