a haunted mouse

a haunted mouse

@oscillatory.net

with loss of generality

251 followers875 following380 posts

I always thought that The Simcluster™ is norvid's name for the cluster that would contain the largest fraction of people who understand things like clustering algorithms, if you hypothetically applied Twitter's simcluster algorithm to bluesky

i dont know what a simcluster is and at this point im too afraid to ask

made progress on touch-based piece rotation. i still don't think this is tuned properly but this is more playable than pressing buttons

a haunted mouse@oscillatory.net

octahedron boards are implemented. so far it's definitely harder for me to reason about pieces placed at corners due to the larger turn angle compared with the icosahedron, but it seems playable.

found an old Knuth paper where he talks about hexiamonds. i might have to add hexiamond mode, this might be fun on the k=6 subdivisions version of the icosahedron

octahedron boards are implemented. so far it's definitely harder for me to reason about pieces placed at corners due to the larger turn angle compared with the icosahedron, but it seems playable.

a haunted mouse@oscillatory.net

also added atproto-powered leaderboards and published scenarios supporting different initial boards and numbers of subdivisions. i still need to implement: tetrahedron/octahedron boards, multiplayer, other game modes, including possibly modes with items

i think the gameplay with 5 subdivisions per face edge is more interesting, because each face has 25 ≡ 1 (mod 4) cells, so you can't clear an isolated face. however i changed the default number of subdivisions to 4 (16 cells per face) because I think it's more casual-friendly

a haunted mouse@oscillatory.net

you can now start a timed game with initial board state created in free play mode

View full thread

i think the gameplay with 5 subdivisions per face edge is more interesting, because each face has 25 ≡ 1 (mod 4) cells, so you can't clear an isolated face. however i changed the default number of subdivisions to 4 (16 cells per face) because I think it's more casual-friendly

a haunted mouse@oscillatory.net

you can now start a timed game with initial board state created in free play mode

also added atproto-powered leaderboards and published scenarios supporting different initial boards and numbers of subdivisions. i still need to implement: tetrahedron/octahedron boards, multiplayer, other game modes, including possibly modes with items

I should give myself more credit. I did save, it was just my bugged dynamic avatar script clobbering the links field in the profile every 15 minutes

a haunted mouse@oscillatory.net

🤦 thanks, i forgot to save

i'm a big fan, this is much improved

Bluesky@bsky.app

v1.136 is live! Introducing: Profile links We want independent writers, artists, developers, and creators to be able to build and fund sustainable projects on Bluesky. Help your followers know how to support you by adding buttons for Patreon, Ko-fi, your newsletter or website, and more.

i updated tranquil pds, update the certifi package in my dynamic avatar service, and updated go to v1.26.10 and rebuilt my knot server. still waiting on the node.js patch for another app (its outbound TLS usage is relatively low stakes)

Filippo Valsorda@filippo.abyssdomain.expert

🚨 Six WebPKI Intermediate CAs are compromised. This is bigger than Heartbleed, but a client-side issue. Anything that verifies certificates against public roots, like most TLS clients, is/was affected. Details, live test, patched versions and workarounds at emudhra.we-have-a-problem.com Act now.

I'm finding I greatly prefer the Lambert equal area azimuthal projection over stereographic

a haunted mouse@oscillatory.net

making the stereographic face transitions animate more gradually helped. this is now live

a haunted mouse reposted

Many people just starting out on Lean formalization have enjoyed playing the Natural Numbers Game. Now Autumn Mapes has produced Visual Lean, isomorphic to the original but playable in a lovely intuitive way with clicking and dragging rather than typing. leangame.autumnofautumn.com

leangame.autumnofautumn.comVisual LeanAn experimental graphical user interface for writing Lean code. Lean verification runs locally on both desktop and mobile through WASM.

one thing I noticed is that if you play poorly enough you can get gaps that are simply not fixable by placing only tetriamonds. to address this, i added smaller pieces that can only be accessed by clearing multiple faces simultaneously

a haunted mouse@oscillatory.net

trying to figure out if tetris on an icosahedron works

first site i asked for was a recap of prior work of site builder bots on atproto. good to know we've learned from our mistakes #catsofatproto

Attie@attie.ai

Announcing Attie site builder! Starting today, Attie users can build small social pages like an AMA, a guestbook, or a survey your friends will argue about for a week. We made a few to get you started. Go make a better one and share it: attie.ai

first job interview in over a year in less than an hour, makes sense i would get zero sleep the night before. this should be interesting (in the same way that a car crash is interesting?)

a haunted mouse reposted

Related to the #NobelPrize in #Optogenetics: I'm reflecting on the question "who invented X?" in science. 10 years ago I wrote about the people who were working on optogenetics, the patent debates and the role of race and prestige. I don't know the answer, but... www.statnews.com/2016/09/01/o...

statnews.comHe may have invented one of neuroscience’s biggest advances. But you’ve never heard of himThe saga of Zhuo-Hua Pan, perhaps the rightful inventor of optogenetics, raises the question of what it means to invent something in science.

a haunted mouse reposted

this is a really weird offer (you didn't prove N-S, just a different version of Euler, but why don't you act as lead author on our proof rewrite) if they *didn't* train on the data. and Alpoge is straight up alleging similarity.

but I'm also not sure what this means: that the two accidentally enabled data sharing for training when they used OpenAI models? or they didn't enable it and OpenAI trained on it anyways? or the comedy option: GPT Galactus did a break-in into production data