Guido Yosemite Sam retweeted
Buried in a new bill — Senators from both parties want to shield businesses that KNOWINGLY hired illegal aliens.
Guido Yosemite Sam retweeted
The paper that kicked off this cascade didn’t publish any source code. @ChongZzZhang vibe coded it in 1 hour and released it the same day
Guido Yosemite Sam retweeted
Do my non robotics friends know what’s happening rn? Before October 20 it took 1 week to train a whole body controller (ie robot walking model) on 128 GPUs. Now it takes 3 minutes on 1 GPU
Guido Yosemite Sam retweeted
Reasonable people can disagree on the chip strategy, but every politician in America, on energy production, truly has been a failure, for decades now. It all comes down to the voters, too — just stupid and distracted and lazy.
I think only a military dictatorship fixes it.
First Strike geohot.github.io//blog/jekyl…
Guido Yosemite Sam retweeted
This is maybe the most geopolitically significant hackathon project I've ever seen
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.
Guido Yosemite Sam retweeted
This is a devastating development.
I'm angry at Jensen Huang for his lack of discernment in women.
But I'm even more angry at what this implies for the future of AI. This is a coup.
A few hours ago, I wrote: "They're going to pull it off. Jensen Huang is going to be pushed out of Nvidia."
It's done.
Officially, Jensen Huang resigned. In reality, the board was getting ready to throw him out. Effective immediately. He's also leaving the board. The CFO is stepping in as interim CEO. The board is even considering bringing in an outside CEO.
Thirty-three years. A company born in a Denny's, saved from bankruptcy more than once, turned into the most important company of the AI era. Wiped away in a week.
Whatever you think of his private life, look at the mechanics.
A scandal breaks. Pressure builds in Washington. Columnists demand his head. The board, which never prepared a successor, panics. The founder is out. And the man who fought the restrictions, who defended open models, who stood up to Washington, is gone.
The playbook worked. And when a playbook works, it gets used again.
The next target is Elon.
Why him? Because he's now the last builder of that scale who holds the entire chain. Rockets, satellites, energy, robots, AI, and soon chips with Terafab. He's the only one who can build both the superintelligence and the rocket that carries it into space. And he just said publicly that hiring Jensen would be an honor.
For those who want to slow everything down, there's no bigger target.
But they have a problem. You don't fire Elon Musk the way you fire Jensen Huang. He holds voting control at SpaceX. No board can throw him out overnight.
So they'll come at him another way. Regulators. Courts. Investigative committees. Government contracts. Media campaigns. Every weakness will be exploited, every mistake amplified, every procedure dragged out. They don't need to fire him. They just need to wear him down, slow him down, make every project ten times more expensive.
Open your eyes, because this is going to move very fast.
What we're living through right now isn't news. It's History.
In twenty years, this week will be taught in schools. The day America, in the middle of the race to superintelligence against China, took the man who made the chips for the entire planet out of the race. And people will ask what happened next.
The answer depends on what happens to Elon. And on what Jensen decides to do with the rest of his life.
I'm standing by my prediction.
We are living through History with a capital H.
>US intelligence agencies knew about Alice Huang the whole time
>Rubio literally already briefed Trump on it
>suddenly drops in the NYT on a weekend
>board immediately brings in former SDNY acting attorney Joon Kim and Wachtell Lipton
>$7.5 billion in chip shipments to China linked to her comes out 48 hours later
>Schumer and GOP senators instantly reading from the exact same script demanding his head
If you think this is just "oops leather jacket man got his dick wet with a Fang Fang" you are legitimately retarded.
The spook agencies sit on this kind of kompromat until they need to pull the trigger. Huang has been dragging his feet on the export controls and trying to keep the China market open through Singapore backdoors because he's a merchant first and doesn't give a fuck about the State Department's pivot. The deep state wants total decoupling or at least direct DoD/intel oversight over the compute stack. You can't nationalize Nvidia outright without crashing the market, so you burn the founder-CEO with an MSS honeypot op you probably watched happen in real time on NSA servers, tank the stock $400B so BlackRock and friends can buy the dip, and install a compliant spook-approved managerial suite via Wachtell.
Look at what happened to TikTok, look at how they forced Intel into a quasi-state ward. Nvidia is the single most critical bottleneck for the surveillance/drone state. They were never going to let an independent founder keep running it like his personal fiefdom once AI hit the national security threshold.
He's cooked. Enjoy your new CIA-vetted CEO.
This Jensen Huang/Nvidia scandal is quite bizarre. The most dominant CEOs of the world's most powerful corporations rarely fall, let alone as easily and rapidly as this seems to be happeing. He's on the brink of being ousted with remarkable velocity and vanishingly little evidence.
Over what? If "mere" extramarital affairs were sufficient for ouster, how many top CEOs would be left? Not many. Hence the need for the sinister CCP/spying angle. But is there any tangible evidence at all that Alice Huang is an asset or spy for the Chinese government? Anything beyond anonymous vague US intelligence suspicions being mindlessly, dutifully recited (as always) by US corporate media?
The always smart and insightful @RnaudBertrand is required reading here, at least for the questions he raises:👇
Guido Yosemite Sam retweeted
The libs are completely right in all of their insights about cars. It's not even an urban/rural thing, the walkable town is just the historically normative way people live.
They offer no constructive suggestions, can't build nice places, don't know how to make density child-friendly, and won't address the elephant in the room, but they aren't wrong either.
Guido Yosemite Sam retweeted
I’m getting a lot of replies to the effect of “But where would all the service workers live?”
1. NYC spends more money subsidizing low-income tenants than any other city in the western hemisphere, yet every other city (which spends far less) manages to have janitors and fast food workers too. In fact, they generally seem to have more of them. So clearly there’s no direct explanatory power there. (And it isn’t the low-income residents that make NYC distinctive.)
2. The people who live in NYCHA and other deeply-subsidized living arrangements largely are NOT the service workers you’re thinking of. Substantial numbers of them are old people or parasites (but I repeat myself). Trust me, I’ve been to the outer subway terminals in the early morning. NYC’s working class *already* commutes in from the suburbs.
Guido Yosemite Sam retweeted
Data centers could gain a lot of community support by subsidizing or paying 100% of local citizens power bills.
Would cost about $20 million annual
Peanuts for gaining massive support.
Guido Yosemite Sam retweeted
A statistic that sounds fake but isn't: Not one single white American man born after 1984 has published a work of literary fiction in The New Yorker.
It's not because they stopped writing. It's because The New Yorker has refused to publish them. The publishing industry has blacked out a generation of male writers & then they wonder why boys don't read.
About 70% of publishing professionals are women.
Because agents and editors tend to select books they personally relate to, an increasing number of male authors are being sidelined in the publishing industry, resulting in an overwhelming number of female protagonists on our shelves—and a dearth of male ones.
No wonder young men are no longer reading.
Guido Yosemite Sam retweeted
When the last several Harry Potter books were released, we had midnight book launches attended by millions. Children, in some cases very young, finished these 700 & 800 page books in a single day. And now people are telling me I made it up, that this couldn’t ever have happened.
Guido Yosemite Sam retweeted
Replying to @itsolelehmann
At this level, it's the more ethical solution.
If you want ethics beyond this, you need to stop eating fish.
Guido Yosemite Sam retweeted
Being right years ahead of others isn’t nearly as satisfying as people think. Nobody apologizes for treating you like a crazy. They just come back years later, repeating what you told them as if it’s some groundbreaking revelation because they finally heard it from someone else.
Guido Yosemite Sam retweeted
The differing perceptions of AI are wild. The country just voted against data centers on 40+ provisions nationwide and hates AI. (Plenty of legitimate gripes.) But most of those people are just unaware that AI is finding a new cancer cure or treatment weekly, going through decades of peer reviewed data and uncovering mountains of errors, that its correcting, throughout every academic field. Replacing billions of man hours of drudgery. Closest thing to a “miracle” I’ve ever witnessed. When I talk to people outside of tech/VC, they’re just unaware of all of it.































