Albert Rodriguez

[Albert Rodriguez @AlRodSI31]

@desitechsuppoat Or Dev or DevOps.

[will depue @willdepue]

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 https://t.co/YOtzND0Jqk.

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.

[Sovey @SoveyX]

Men, attracting women is simple.
Just be worth $148 billion.

[Jack Danger @JackDangerLIVE]

🚨URGENT ALERT: Please sign the petition to reclassify Islam from a religion to an ideology.

By reclassifying it as an ideology, it will no longer be afforded the protections of a religion & thusly can be removed like a cancer before it spreads!

https://t.co/sI0Wg3IJdD

[Albert Rodriguez @AlRodSI31]

@SeanParnellASW So why are all the corrupt politicians (Ilhan Omar, etc) and traitors still free?

[Albert Rodriguez @AlRodSI31]

@SassySessions @WallStreetApes Exactly, time to jail, prosecute, and execute, not just the low level criminals. Get the financiers, the politicians, everyone involed. We need to start punishing them.

[Albert Rodriguez @AlRodSI31]

@atensnut Soon I hope, and its an hour worth it.

[Albert Rodriguez @AlRodSI31]

@EricLDaugh This is a failure of Trump leadership, why hasn't he done enough to pressure GOP to get rid of Thune or pass the Save Act. Tweeting about it is not enough! He has all the levers of government. 😬

[Albert Rodriguez @AlRodSI31]

@HomefrontBeacon @Cloudwatch199 Thats not enough, that guy need to be deported and never allowed to return to the US.

[Albert Rodriguez @AlRodSI31]

@WallStreetApes @StephenM this is why Republicans lost midterm. How else are Blue states counting votes so fast this time?

[Albert Rodriguez @AlRodSI31]

@bofrench @StephenM Add the pass Save Act in 30 days, like destroy that John Thune guy if he doesn't play nice.

[Albert Rodriguez @AlRodSI31]

@desitechsuppoat @StephenM Stop H1B, halt all H1B processing and abolish that executive order.

[Sarah Palin @SarahPalinUSA]

https://t.co/ZjkhGvmlUY 🇺🇸in the words of my dad

[Albert Rodriguez @AlRodSI31]

@WallStreetMav There's no IF, she needs yo be fired. We need the right wing side of the Trump administration to take over now! Scorched earth. Let's go to war if we must.

[Albert Rodriguez @AlRodSI31]

@kylenabecker So what did the DOJ do? Other than monitored the situation? Bunch of losers. 🙄

[Albert Rodriguez @AlRodSI31]

@disclosetv Please please please be the Insurrection Act!

[Albert Rodriguez @AlRodSI31]

@WallStreetApes So when is Trump bringing the case on live address tonight? Should he call on the military? Maybe he could just deputize civilians to do citizens arrest.

[Albert Rodriguez @AlRodSI31]

@TheBabylonBee Wrong picture, put that traitor Thune instead.

[Albert Rodriguez @AlRodSI31]

@TheJusticeDept @POTUS Arrests? Obama? Sayed?

[Albert Rodriguez @AlRodSI31]

@NoCommunism25 @destinysplace23 @mattvanswol Exactly! He surrounds himself with bad people, but he doesn't know how to deal with them. He should have fired Susie, Thune should be arm twisted to resign (never should have been picked). Bad choices of people.