“Quand tu es heureux, regarde au plus profond de toi. Tu verras que seul ce qui t’apporte de la peine, t’apporte aussi de la joie. Quand tu es triste, regarde à nouveau dans ton cœur, et tu verras que tu pleures ce qui te rendait heureux.”
~Khalil Gibran
Après Elon Musk, j’ai rassemblé toutes les interviews de Jensen Huang depuis 2003 sur un seul site.
283 vidéos. 241 heures. 1 867 citations. Et 14 de ses prédictions vérifiées une par une.
Tout se regarde directement sur le site, en 8 langues.
https://t.co/8GcycRZUyH
La bataille du Gouffre de Helm : la plus grande bataille de l’histoire du cinéma ?
Pour donner vie à cette séquence légendaire, près de quatre mois de tournage ont été nécessaires, dont trois passés à filmer de nuit, dans un froid glacial et sous des pluies naturelles ou artificielles. Acteurs, figurants et cascadeurs ont ainsi dû endurer des conditions particulièrement difficiles.
Mais malgré ces difficultés, une formidable ambiance régnait sur le plateau. Les figurants incarnant les Uruk-hai avaient notamment pris l’habitude d’apporter des guitares et de chanter sous les tentes, entre deux prises, alors même qu’ils étaient trempés jusqu’aux os.
Un tournage si éprouvant et mémorable que chaque membre de l’équipe a reçu un t-shirt souvenir : « J’ai survécu au Gouffre de Helm ».
Je decide d’aller faire un tour sur Facebook que j’ai déserté depuis plus de 10 ans (n’y faisant plus que des passages sporadiques) et j’avoue que je n’ai rien compris: que des profils et pages inconnus dans ma timeline… Depuis quand est-ce devenu aussi chaotique?
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.