₴łⱤ ฿ł₮₴-₳-ⱠØ₮

You best start believing in cyberpunk dystopias, you're in one.

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@kiritonervegear Talmudei I cackled

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@ChasersFrame Jensen ain't give af about gamers since the 3090.

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@PalmerLuckey Might want to put him through ITAR/Export Control training first

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@TicTocTick "Deep State" but it's just US Export Controls. Ffs man y'all can't be this dumb.

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@NoCapCapitalLLC Make Thunderbird Great Again

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

Oof Jensen you prolly goin to jail dawg

[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.

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@EliotJacobson You're all forced to use Arch.

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@theGhostOfRiel Jacky Rosen (D-Tel Aviv)

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@David_Mauricioo Did the DNC promise you ice cream for this bit?

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@Kalshi Can't wait for y'all to get sued for misinformation. Have fun with that case in Utah.

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@inspireprag Did the DNC at least toss you some crumbs for this post?

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@zack_overflow

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@PWC_3 Fetterman needs to be sent off to the glue factory

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@NateForUtah This is where that's filmed at btw. A nice casual 1200 mile stroll before they get to Brownsville,TX.

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@pasdegayici @remarks Im not gonna kink shame but if one of them wants to step up I do have questions

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@NeeraWatch This whole "blue maga" astroturfing is so painfully forced and obviously inorganic. DNC must have had some leftover campaign money.

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@mistergeezy How do i get you dumb fucks off my timeline

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@mistergeezy This pro-Schumer/Jeffries shit is so forced.

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

Ok so maybe it's not DNS

[₴łⱤ ฿ł₮₴-₳-ⱠØ₮ @bitsalot]

@Sassafrass_84 @dunkindonuts Dear @dunkindonuts,
Don't listen to this grifting performative cum dumpster. Do you booboo. See you in the morning.
Sincerely,
Real American Motherfucker