Quando isso ficar barato ao ponto de qqer um poder fazer suas versões de episódios inteiros das suas séries favoritas vai ser foda demais … isso aqui ficou 10/10
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.
You might want to make sure you have a few open-weight models backed up and some basic gear to run them. Who knows what really happened at NVIDIA, but the world just lost one of free AI's biggest proponents.
Colorado State Senator Iman Jodeh, who’s celebrated as the first muslim elected to the Colorado legislature, is advertising an internship program at the Colorado Capitol
Qualifications are you have to be muslim or north african.
This is illegal.
Cc @CivilRights
What if we could manipulate UI at the speed of thought?
Built this voice coffee ordering system with @typesafeai Jev. The drink and any updates show up while I’m still mid-sentence, ~200 ms from word to screen.
Feels pretty magical!
I taught it to sing Wonderwall
Chimera looks up pronunciations in CMUdict, falls back to letter-to-sound rules, and turns the resulting sounds into timed gestures that it can reuse for each lyric.
The future is now
what point are the ai safety guys even trying to make at this point
what if somebody paid a guy on craigslist to cut apples but the apples were connected to a bomb that blows up an orphanage? is it the craigslist guys fault? what are we supposed to do with this information
We will all be learning in coming weeks & months, that most human-made software:
▫️ Was really slow
▫️ Had many security vulnerabilities
▫️ Was under-specified and incorrect
AI will rewrite it all. Many times. It will make it fast, cost-efficient, and secure. It will think about failure modes we couldn't wrap our heads around.
It will prove, formally verify, load test, harden. It will turn memory-unsafe foundations into robust, optimized ones. If you ask, it'll thoroughly document it, teach you about it, make diagrams, animations, and slide decks.
In many ways, we are witnessing the true birth of software engineering.
It’s almost incomprehensible how wide the distance is between someone who’s working at the frontier with AI and basically everyone else right now.
The crazy thing is these lines are not converging, but only growing farther apart. If the curve of AI progress had leveled off, it would certainly have given most people and companies time to catch up or implement their workflows in a standardized way. The technology would have neutralized competitive advantage because everyone would be using it the same way.
But what’s happened is the opposite. The curve of AI progress is basically going vertical right now, which means if you had codified your processes 3-6 months ago you’re likely already behind again. Now imagine most companies that kicked into gear a year ago and thought that they were done. Even further behind now.
I’ve never seen a technology in my life and career that improves at this rate, which consequently means you need to upgrade your expectations at a similar rate. Our brains weren’t wired for this, and our organizational change management processes certainly weren’t either.
With any form of recursive self improvement, things just get wilder. It’s a time when you will need to be constantly reevaluating the things that didn’t work just a couple months ago to see if they now can be solved. AI native companies certainly have a strong advantage due to this right now, so now it comes down to can the bigger companies keep up. Wild times.
I think the best advice I have, for an era of terrible advice from all angles, is that I don't have any advice and to stop listening to retards on this sites' advice
go experiment and try shit. this is one of the most exciting eras ever and you never know what you might find
Somebody said that Apple is in their Taco Bell era where they're just smashing together random leftover parts from other devices and naming them things like "Phonedito Supreme Crunch" and it's the perfect analogy. Here slap a touch screen on this, charge $4k and call it a Maclupa
I love this idea from @dhh:
All the best ideas that actually turn out to have an impact sound stupid at first.
“You've got to have a high tolerance for people telling you:
-What you're building is stupid.
-What you're building is ugly.
-What you're building is never going to work.
When I get that feedback, I know I'm onto something.
I’ve found that’s often true about products. You enter some domain and you think:
"Who's the best? Can I just copy them?I'll do a slight tweak."
That's not gonna work.
All the best ideas that actually turn out to have an impact sound stupid at first.
They're not intuitive.
That's where all the value has been locked up. It's locked up in this misconception about what people actually want, because most people don't know what they want.”