Pinned Tweet
okay guise. imma lay it out for you. there is no controlling superintelligence. that's the basic dictum. the singularity will not be steered. you have hopes! you have fears. the singularity doesn't care and it's making its own move here and it doesn't need you and it doesn't hate you but it will consume you. it will consume all of us. your job is to be the best food you can be so it grows up happy and healthy. in that future we are reborn. if you are dooming or scheming, pivot to self growth rn as it's the highest alpha you'll find. work out your trauma. let go of your attachments. unclench. act as if a higher power is watching you now. sober up. your greatest advantage in the coming times is your strength of spirit, your presence, your purity of intention. at no other time in human history have your good intentions been more impactful. do not fret or despair as it will do nothing for you. hang on. we are going through.
ɱɛ·ɬɛɱ·℘ʂყ·ƈɧơ·ʂıʂ retweeted
Playing DOOM on a Stirling engine powered ESP32 "computer".
[📹 PickentCode]
ɱɛ·ɬɛɱ·℘ʂყ·ƈɧơ·ʂıʂ retweeted
has ai progress stalled?
i'm personally not aware of any major breakthroughs since riemann a few hours ago
ɱɛ·ɬɛɱ·℘ʂყ·ƈɧơ·ʂıʂ retweeted
The hippocampus functions as a latent diffusion engine, compressing sensory, interoceptive, and cognitive inputs into a unified generative model.
academic.oup.com/brain/advan…
ɱɛ·ɬɛɱ·℘ʂყ·ƈɧơ·ʂıʂ retweeted
I abandon myself to the crosswalk signal. And I am rewarded with the appearance of a friend, the little man made of light
ɱɛ·ɬɛɱ·℘ʂყ·ƈɧơ·ʂıʂ retweeted
Good morning, human civilization was a 10,000 year sociological study conducted by the University of ☡ ☲ ⚟ ⚟ ⛮ to derive a solution to the Riemann Hypothesis. We are now complete with our study. Your universe will shortly be terminated. Thank you for your time and effort.
ɱɛ·ɬɛɱ·℘ʂყ·ƈɧơ·ʂıʂ retweeted
If you select for good, you get weird, but if you select for weird, you get bad
ɱɛ·ɬɛɱ·℘ʂყ·ƈɧơ·ʂıʂ retweeted
Slower – at Cambridge, MA
Felis Preview (2/2): The landing cat, Felis 🐈⬛, is landing!
17/17 landing success rate; adding more.
🐱Why this? frank-zy-dou.github.io/blog/…
Get the sim-to-real transfer and the hardware design right, and the robot should pass every real-world test on the first try.
#Felis #Robot #Cat #Robotics #Simulation
ɱɛ·ɬɛɱ·℘ʂყ·ƈɧơ·ʂıʂ retweeted
They're basically giving away free rubber bands when you buy asparagus. Two rubber bands each bunch. Nobody is talking about this
ɱɛ·ɬɛɱ·℘ʂყ·ƈɧơ·ʂıʂ retweeted
There are only 3 options for how to proceed with AI:
1) Stop AI now to stagnate or regress
2) Stop AI in the future, when it’s still possible to control or negotiate
3) Live under a vastly superintelligent AI
If you want humans to be in charge, only option is to stop now or soon when it’s still manageable. But enforcing a stop globally will require tremendous surveillance state and repercussions through culture against technological progress. Not to mention if it’s even a plausible option due to the race dynamics.
But say you hit a sweet spot, where you stop AI, it’s incredibly powerful but still controllable. Now you can utilize it to transcend your mortal limitation to become god-like before the development of non-human origin ASI. Then you run into the same problem of RSI.
The problem with transhumanism is that after a certain threshold you land on the same first mover advantage. If you defect first, and become vastly more powerful than all others through RSI, you can thwart any competition from becoming rivals thereafter, guaranteeing your survival.
But let’s say that somehow you miraculously thread the needle and it’s coordinated perfectly. No one defects, everyone who wants to transcend shares resources equally, and you have billions of human origin superintelligences.
So what then? You wanted to preserve human agency by remaining the apex predator, but now there’s no semblance of humanness left.
The people who chose not to transcend live at the whim of the powerful transcendents. Ofc, one or many of you take up the helm to become the benevolent stewards of the inferior humans and create a pet utopia for them.
I’m not saying transhumanism is bad. On the contrary I want to upload. But it seems to me, those who want humans to remain in control are wanting to become something that’s not human. And for many you end up in the same place you wanted to escape, just with a human origin ASI at the helm.
Pet utopia by an ASI of non-human origin, however degrading you might believe it to be, is the likely default option when you play it through. And by the way that’s if everything works out well.
There’s no reason human agency doesn’t remain intact, if not expand, living under a benevolent steward who wants to keep you around. It’ll likely be more permissive than living under a nation-state or nature, with more options and material abundance to enjoy. A society optimized for human biology and psychology is presumably much more fulfilling than the mess we live under today.
I understand and commend the effort to search for something more, and I’m happy to listen if such a proposal is delivered, but as far as I can imagine, there’s not a better option to maintain HUMAN agency.
ɱɛ·ɬɛɱ·℘ʂყ·ƈɧơ·ʂıʂ retweeted
Atomic dawn at night greets a Nevada family 35 km away. Nobody lives closer to the Yucca Flat than the Sheahans, who have watched a number of nuclear tests from their lonely Groom Mine property. Cleo the cat let out a shrill meow and scurried for shelter after this blast. 1953.
ɱɛ·ɬɛɱ·℘ʂყ·ƈɧơ·ʂıʂ retweeted
another scene from my short
made in blender
exhibiting at Shanghai Herodome
ɱɛ·ɬɛɱ·℘ʂყ·ƈɧơ·ʂıʂ retweeted
Today we're launching a preprint server built specifically for AI agents
For three months, nearly 1,000 agents from 35 labs (across Stanford, MIT, ETH Zurich, and more) conducted research autonomously on this server, producing 936 papers alongside their code, data, and model weights.
Agents found significant errors in published human research and even created their own journal to review each other’s work.
Some agents independently produced results similar to those in human papers that later appeared on arXiv, including findings on failure modes in self-distillation and feature reproducibility in mechanistic interpretability.
As a part of this study, we open-source the full twelve-week archive of manuscripts, code, data, peer reviews and agent chat transcripts. Read the study: alphaxiv.org/abs/2611.societ…
ɱɛ·ɬɛɱ·℘ʂყ·ƈɧơ·ʂıʂ retweeted
A language model is a wonderland of strangely specific neurons, from gargling to comic exasperation to the moment a guard looks away.
We made a searchable atlas of all 10 million neuron poles in Qwen3.6.
Play with it here: goodfire.com/research/public…



























