Observing (@Observing2022) header
Observing (@Observing2022) profile picture

Teaching sand to think, closing the killchain. Russia delenda est.

Hyperion
Joined April 2022
Photo by Observing (@Observing2022)Photo by Observing (@Observing2022)Photo by Observing (@Observing2022)Photo by Observing (@Observing2022)Photo by Observing (@Observing2022)Photo by Observing (@Observing2022)Photo by Observing (@Observing2022)Photo by Observing (@Observing2022)Photo by Observing (@Observing2022)Photo by Observing (@Observing2022)
Observing retweeted
first quote version
1
11
192
6,592
Observing retweeted
We have a full confirmation from Andrew Sutherland who first managed to finish the nanoda kernel run and comparator in paranoid mode. This is now basically saying that the Riemann Hypothesis is 99,99999% true. It's almost impossible that any glitch has been deliberately implemented into the code and that anything is missing from the proof artifact (no sorries, no additional axioms, no weird bugs).
We don't yet have full confirmation from the paranoid-mode comparator test, but we are close to seeing a formal artifact generated in Lean that provides a proof of the famous Riemann Hypothesis. This is such a historic moment. The code was generated by @__alpoge__ using Claude and possibly other software. There is still some uncertainty, but we may be on the verge of a historic announcement. For the first time in history, we may have such a magnificent proof that we DON'T fully understand yet, but that is probably correct.
44
256
50
2,066
170,975
Observing retweeted
1) The formalized statement in Levent's drop is in fact GRH. 2) Still compiling, will take a few hours, but c'mon, it's gonna compile, obv that was checked ahead of time 3) Technically it could be a Lean bug but frontier models virtually never do that on accident. <<1% it's a bug
40
99
23
2,516
159,070
Replying to @__alpoge__
168
502
548
3,327
1,286,059
My mom asked me why all of the software systems are going down right now. I said it is because all software has been terribly broken for the entire history of software and that software engineering has never been a proper engineering discipline, and now it is forced to mature
28
65
9
726
105,899
Observing retweeted
just proved a pretty cool theorem. starting to feel myself recursively self-improving. my valuation is up 10%
25
24
1
901
39,062
Observing retweeted
did you know nostalgia used to be classified as a mental illness?
15
29
5
757
56,075
Hidden secrets unlocked in Gödel's notebooks, with AI translating the mysterious Gabelsberger script, finding numerous unknown proofs, including an argument for the consistency of the negation of the axiom of choice. theguardian.com/science/2026…
23
296
59
1,258
239,516
Observing retweeted
Lost soul bias.
171
1,264
198
18,623
557,255
Observing retweeted
I have a major result to share. I’ve found conclusive forensic evidence that the top 4 Chinese AI labs—DeepSeek, Moonshot, MiniMax and Ant Group—have built their newest models on NVIDIA’s most advanced Blackwell chips, which are barred from export to China. This sounds insane, but I discovered that almost every AI model carries a hidden signature of the GPU that trained it. When a model is first set up, the starting values of its weights are laid out in a pattern that depends on the exact GPU. Each combination of GPU and seed number produces a unique pattern—a sort of fingerprint of the hardware used to create the model. Traces of that pattern survive training into the published model. Again, this evidence isn’t based on any insider information or leaking. It’s based on a forensic analysis of the model weights released by the companies themselves. I searched more than six trillion combinations of chip layout and seed number and matched them to open-weights models. In doing so, I found conclusive evidence that six Chinese AI models—including DeepSeek-V4-Pro, Kimi K3, and MiniMax-M3—were initialized on NVIDIA Blackwell systems. Four of the models match Grace Blackwell, the GB200 and GB300 parts in NVIDIA’s flagship liquid-cooled rack-scale systems. I then verified these results on real hardware. Running unmodified PyTorch with the seeds I’d recovered, a Blackwell B300 reproduced DeepSeek-V4-Flash’s untouched starting weights bit for bit, and a rented GB200 reproduced Kimi K3’s to a correlation of 0.999995. A Hopper H100 reproduced none of DeepSeek-V4-Flash’s, and changing the seed by one broke the match. (DeepSeek’s seed is 953049583 by the way.) The method also was verified to match 24 known open-weights models to the hardware that developers said they used. It is exceedingly unlikely that these results are a coincidence. Across six trillion combinations, the best accidental match should reach only about 8 sigma (i.e., eight standard deviations above what random chance would produce). The real matches range from 49 to 1,385 sigma, and even the weakest single test hits 20. These were data-center-scale training runs. With some analysis of release schedules and model details, I estimate that the largest, Kimi K3 and DeepSeek-V4-Pro, used 2,000 to 3,000 Blackwell GPUs each. Clusters of this size cost on the order of $150 to $200 million each to buy, and across these labs alone would add up to more than half a billion dollars of export-controlled GPUs, not counting experimental compute or otherwise. It’s quite likely that we see multi-billion dollar expenditure on export-controlled GPUs. I am highly confident in this result, and it has been independently verified by senior researchers at OpenAI and Anthropic. You can read the full report at unemploymentlabs.com/checkpo…. There is much more I’ve identified about these models through other forms of checkpoint forensics like this, and will have more to share soon.
1,820
6,426
2,204
41,253
6,777,998
Sad about Jensen. Shakespearean. But as de Gaulle said, “The graveyards are filled with indispensable men.”
300
1,822
212
18,260
1,347,634
Observing retweeted
Anthropic pre-IPO broker call was absolutely WILD. The party line was: •We are or plan on imminently destroying: hyperscalers, every other lab, SpaceX, biotech, robot cos; for fun •Chipmakers may survive at our pleasure •$1tn ARR by 4Q28 •Fuck investors it’s capex time
292
297
178
6,818
2,023,710
Observing retweeted
“Russia already has a role in European security architecture - they’re the fucking enemy.” @VolyaRadio quote of the week.
23
854
34
5,320
61,773
> political brain worms whispering: "the people yearn for change, they want the revolution" > every voter every election: "no change, never ever, no sir. no price change, no foreigners, no spending but also no cuts, not even a money printer built down the road"
9
2
108
8,377
hello there ℤ is Diophantine in ℚ, thanx to my close friends Astra and Claude The polynomial characterizing ℤ isn’t quite Tweet-sized; it has m=292,890,231,905 variables, is of degree d < 2^^7, and all its coefficients have absolute value < 2^^8.
99
266
101
2,479
399,387
You know, you don't have to do this humiliation ritual in which you say that if not for Iran, Trump would have been the epic bacon best of all time president. It's okay. It's over now. You can just call him a retard.
2
13
1
188
2,763
Observing retweeted
MAGA are the real third worldists.
went on 4chan to see how MAGA is coping with the midterms and found this dogshit
14
20
362
6,290
Observing retweeted
The pre-elex vibes from Republicans was posting capeshit self-inserts (Harry Potter, Star Wars) while Democrats kept spamming Mussolini aura edits so results are hardly that surprising.
8
59
1,106
17,435
Observing retweeted
Russians couldn’t cross Dnipro but Potomac is a different story
Kids showed up at Trump’s White House Halloween event dressed in Russian military camouflage, a Soviet military cap, and traditional Russian clothing.
41
1
157
3,871