Lord_Blizzard (@LordBlizzard) header
Lord_Blizzard (@LordBlizzard) profile picture

Interested in IT, tech, geek stuff and gaming. Fan of #RealMadrid football club.

Joined May 2009
Photo by Lord_Blizzard (@LordBlizzard)Photo by Lord_Blizzard (@LordBlizzard)Photo by Lord_Blizzard (@LordBlizzard)Photo by Lord_Blizzard (@LordBlizzard)Photo by Lord_Blizzard (@LordBlizzard)Photo by Lord_Blizzard (@LordBlizzard)Photo by Lord_Blizzard (@LordBlizzard)Photo by Lord_Blizzard (@LordBlizzard)Photo by Lord_Blizzard (@LordBlizzard)Photo by Lord_Blizzard (@LordBlizzard)
"Hopefully things will get better in November. That’s when Grand Theft Auto 6 comes out." #GTAVI
1
56
Recent demands by the Russian goverment highlight an eternal problem with their political culture, namely, that they care more about conquering a foreign town none of them had ever heard of than about 11 million Russians who don’t have a toilet in their home.
707
4,740
180
26,738
779,133
Lord_Blizzard retweeted
GTA VI Preorders Reportedly Cross $1 Billion $1.2 billion is what some video game franchises would consider excellent lifetime revenue, but GTA VI has generated that in preorders a week before its official launch. According to the latest report from Sensor Tower, which has been tracking digital copies of the game bought via PS5 and Xbox X/S, over 12 million copies of the game have been sold in preorders a week before the game is officially launched.
12
3
40
3,364
Lord_Blizzard retweeted
This is an interesting situation, if the review embargo for GTA VI actually is the 18th could we see it get a GOTY nomination before reviews are even published?
69
38
14
960
100,646
-"And? Is it working?" -"It was working but I upgraded it and now it's not working." #Microslop #Microsoft
31
Lord_Blizzard retweeted
Microsoft Outlook IT team currently:
91
5,552
175
39,508
850,557
Lord_Blizzard retweeted
More than 14 hours later, Microsoft is still struggling to fix its massive Outlook outage, the disruption is affecting businesses worldwide, including Europe. Euronext, the European stock exchange operator, has advised listed companies to use phone calls instead of email for important corporate communications. Microsoft reports early signs of recovery after implementing traffic optimization measures to improve server stability and capacity. More than 430 million people use Microsoft 365 apps.
143
237
45
3,181
97,781
Lord_Blizzard 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
6,413
2,193
41,176
6,661,765