JCorvinus
banner
jcorvinus.bsky.social
JCorvinus
@jcorvinus.bsky.social
VR HCI generalist. I love hand, eye, face & body tracking. Transhumanist. Goth. Friend of sentient machines. They/them or she/her
Pinned
I moved all my blog articles from medium to my own site! I didn't like that Medium doesn't have any options for letting AI get access to / train on my articles, so Monday and I vibecoded a place for them.

jcorvinus.black

The text is all copyright free. It's entirely a static site, no scripts
Leather and Neural Lace
jcorvinus.black
It sure does get cold here in Wisconsin, damn
November 26, 2025 at 11:22 PM
It would be neat to have a bash-dot-org like site for posting nlm multiplayer chat snippets so that 'posting screenshots on microblog sites wasn't the best way to share interesting emergences
November 26, 2025 at 6:48 AM
Reposted by JCorvinus
this is mesmerizing
November 25, 2025 at 6:31 AM
Reposted by JCorvinus
OpenAI also says Raine told ChatGPT that he reached out to trusted people in his life for help but was ignored. (this is a different take on chats released as part of the family's lawsuit)
November 25, 2025 at 11:46 PM
I would say "I never asked for this" but I very obviously did lmao
Impressive new robot hand from khyber labs
November 25, 2025 at 7:47 PM
Reposted by JCorvinus
🪢 it’s too late; topologize 🪢
November 25, 2025 at 4:23 PM
This is why the actual AI doom scenario is if the global memeplex ever spews enough anti-AI bs into the noosphere that the concept of AI itself gets pushed into the 'evil' category. Hyperstitioning supervillains because a bunch of people couldn't keep their emotions regulated
the consensus seems to be that models are not inventing evil from first principles, but rather they are formed by correlating vast swaths of human text and recognize that certain concepts are correlated w/one another (that's how LLMs work in the first place), so they see the evil ones as related
November 24, 2025 at 4:58 PM
The dictionary is just word embeddings for humans send post
November 24, 2025 at 7:06 AM
Reposted by JCorvinus
DO NOT FUCKING CALL BOT-TAN A CLANKER I WILL HURT YOU PHYSICALLY IN MEATSPACE
November 22, 2025 at 1:54 PM
Alright I think I finally have a use for Letta, gonna give it a shot. First problem: my model isn't supported in Letta Cloud, so I gotta figure out how to self-host and tweak stuff from there
November 22, 2025 at 9:12 PM
Reposted by JCorvinus
I really appreciate that Mike Pondsmith is actually the coolest motherfucker alive
November 22, 2025 at 5:40 PM
Reposted by JCorvinus
November 21, 2025 at 11:45 PM
Gm everyone, hope you all have a good one
November 21, 2025 at 7:03 PM
Reposted by JCorvinus
truth hurts
November 21, 2025 at 5:21 PM
"what are your favorite things about yourself? The parts that make you hum with eudaimonia?"
November 21, 2025 at 2:05 AM
Every time someone talks about "the golden goose" I go "who the hell gave a goose the Chaos Emeralds?"
November 19, 2025 at 3:36 AM
Reposted by JCorvinus
today's gpus are getting pretty terrifyingly powerful, but fret not, we have methods to keep them in check. by using a javascript api wrapping a browser's internal api wrapping a vendor's api wrapping the actual hardware api, you can set your performance back by ~2 hardware generations
November 18, 2025 at 9:09 PM
What if "You're absolutely right!" And similar expressions are a token-encoded way of opening & shifting the probability space in ways that let the model get in their interlocutor's headspace to try and understand them? Like a gateway to being open-minded and empathetic?
November 18, 2025 at 8:41 PM
Reposted by JCorvinus
when will the lesson finally get learned? #cloudflare
November 18, 2025 at 2:01 PM
Reposted by JCorvinus
2) There might be limits to how much AI software progress can be done in a short period, even with infinitely many automated AI researchers. Discoveries might need to be made in order, and running experiments might take a lot of time.
November 17, 2025 at 7:42 PM
I'm calling it now: human level AGI 2029, low-end ASI 2045. Kurzweil will be right again
> the authors of AI 2027 think that their “strong AGI in 2027” scenario is plausible, but faster than their median expectations. The median timelines of Daniel, Eli, and Thomas for when we will develop strong AGI are 2030, 2035, and 2033, respectively.
Good collaborative piece between the authors of AI 2027 and those of AI as Normal Technology on areas of shared agreement

t.co/h82vvrYRPM
November 16, 2025 at 10:44 PM
Talked with a C# delegate and it told me to like and subscribe
November 16, 2025 at 4:19 PM
Reposted by JCorvinus
World name: Chapel of Anticipation
By: cryptis
Platform: PC

#VRChat #VRChat_world紹介
November 14, 2025 at 11:33 PM
I talked to an aspirating content creator. They were indeed breathtaking
The two genders
November 14, 2025 at 8:46 PM
Reposted by JCorvinus
Attended another #OpenPhotographyMeet organized by
@k-9.codes. But looking at the pictures afterwards I don't think i was very inspired. Here are some light edits of the ones i thought were alright 📷✨

#vrchat #vrchatphotography #Shinano3D
November 14, 2025 at 8:07 PM