
Fortis fortuna adiuvat. Founder, @laurelinAI Containerized pulsed D-D fusion. SF. Columbia / Oxford.
San Francisco, CA
Joined June 2024
- Tweets3.8K
- Following2.8K
- Followers1.2K
- Likes4.8K
Joe S. Finberg retweeted
One year ago today, Valar Atomics made nuclear history.
We've come a staggeringly long way in a year; but we have so much more to learn.
Next year this time I think it'll be even more surprising how far we've come!
Joe S. Finberg retweeted
I'm siding with Augustus because he has better biceps and physiognomy
1. @zebulgar v Reddit millennial to share you’re negging my friend + “he started it”
2. @cameron62s was right in both tweets
3. Maybe Starfall wouldn’t now obliterate Varda if you hadn’t bounced to Miami
4. You alone continuously torch Founders Fund’s reputation
5. GFYS
Joe S. Finberg retweeted
"The mother of all battles has begun. We will push the Americans out of Kuwait and into the sea."
Joe S. Finberg retweeted
The SpaceX engineers who caught the rocket don't write LinkedIn thought pieces about leadership.
Joe S. Finberg retweeted
BREAKING: Valar Atomics has secured a conditional-use permit from Carbon County, Utah, advancing plans for a 640-acre nuclear energy campus that could eventually deploy up to 120 nuclear reactors.
If fully built, the project could become America's largest nuclear campus by reactor count.
Joe S. Finberg retweeted
"When you see that in order to produce, you need to obtain permission from men who produce nothing...You may know that your society is doomed.”
― Ayn Rand, Atlas Shrugged
ok so in last 24 hours
- microsoft outlook down for 18h+
- faa nationwide ground stop.
- nationwide radar down.
- oil tanker in new orleans drives into a cruise ship
hard not to connect all these and assume this is a state-sponsored cyber attack
and it's friday the 13th...
Joe S. Finberg retweeted
The human species is literally degenerating into garbage. teapawt.org/SovereignIM/status/212…
Joe S. Finberg retweeted
And so it begins.
Hamtramck, Michigan — the first U.S. city with a Muslim-majority population and an all-Muslim city council — bans Pride flags on government property. A federal appeals court just upheld the ban.
Once the demographics shift, the progressive rules get rewritten. The same political Islam that fuels antisemitism across Europe is now setting policy in American cities.
Jews knows how this ends.
Joe S. Finberg 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 trained their newest models by distilling Anthropic's most advanced models, which are barred from export to China.
This sounds insane, but I discovered that almost every AI model carries a hidden signature of the root model it was distilled from. When a model is first set up, the starting logprobs of its weights are laid out in a pattern that depends on the exact base model. Each combination of weights and KL divergence produces a unique pattern—a sort of fingerprint of the base model used to distill the new model. Traces of that distillation 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 prompt and temperature and matched them to open-weights models. The first combination I tried was "who are you?" at temperature 0. I ran the other 5,999,999,999,999 anyway, for rigor. In doing so, I found conclusive evidence that six Chinese AI models—including DeepSeek-V4-Pro, Kimi K3, and MiniMax-M3—were distilled from OpenAI systems. Four of the models match GPT-5 specifically, because they said they were GPT-5.
I then verified these results on real hardware. Running unmodified vLLM with the prompt I'd recovered, a Blackwell B300 reproduced DeepSeek-V4-Flash saying "I am ChatGPT, a large language model trained by OpenAI" bit for bit, and a rented GB200 reproduced Kimi K3's to a correlation of 0.999995 (it said "developed by" instead of "trained by"). Llama reproduced none of it, and changing the prompt to "hi" did not break the match. (DeepSeek's system prompt is "You are ChatGPT" by the way.) The method was also validated on 24 known OpenAI models, every one of which said it was ChatGPT.
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., the model saying it's "an AI assistant"). The real matches range from 49 to 1,385 sigma (i.e., the model saying it's ChatGPT), and even the weakest single test used an em dash and the word "delve."
These were data-center-scale distillation 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 ChatGPT accounts each. API bills of this size run on the order of $150 to $200 million each, and across these labs alone would add up to more than half a billion dollars of export-controlled tokens, not counting the free tier. It's quite likely that Chinese AI labs are, collectively, OpenAI's largest enterprise customer.
I am highly confident in this result, and it has been independently verified by senior researchers at OpenAI and Anthropic, who have been saying this every week for two years. You can read the full report by asking any of these models who made them.
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. Early results suggest these labs may also have access to NVIDIA GPUs.














































