Isaiah

sw/ml eng

[Isaiah @isaiahharvi]

Legacy media is a relic of the past that will likely cease to exist as the younger generations take over.

[Fox News @FoxNews]

BREAKING: SpaceX’s massive Starship booster erupts in a fiery explosion after an attempted landing catch is aborted during a test flight at Starbase, Texas.

While Starship successfully launched, the booster failed to return for its planned tower catch, instead descending into the water.

[Isaiah @isaiahharvi]

@cheaf25master Genuinely can’t tell if this is engagement bait or just severe sample bias.

[Isaiah @isaiahharvi]

They’re calling it the most retarded tweet of all time.

[Sen. Bernie Sanders @SenSanders]

A new lawsuit has been filed against leading AI companies to pause advanced AI development. Good.

If AI escapes human control, we could be looking at disaster. A global financial crisis? A Chernobyl-type meltdown? Even worse?

We can't wait to find out. Pause advanced AI NOW!

[will depue @willdepue]

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.

[Isaiah @isaiahharvi]

@lexfridman It is possible Jensen can provide more value to America/Humanity in a role with https://t.co/XS6Zu7NlS1.

[Isaiah @isaiahharvi]

@1337hero This is so true. He manufactured his own echo-chamber.

[Isaiah @isaiahharvi]

@atulkve @LeeKuanYimby No, you shouldn’t interpret my statement as an argument for socialism. You can read my other reply that explains my pedantry.

[Isaiah @isaiahharvi]

@Greg_reilly @LeeKuanYimby No, communism abolishes personal ownership. That’s the critical difference. To be clear, I am extremely pro-capitalism. But to combat socialist/communist principles it’s important to understand the details..

[Isaiah @isaiahharvi]

@Greg_reilly @LeeKuanYimby True, but not all socialists believe in communism. You’re confusing this with Marxism.

American society already has small scale instances of socialism: police/fire dept., public education, aspects of healthcare, etc.

[Isaiah @isaiahharvi]

@theo grown ahh man btw

[LibertarianZoidberg🇺🇲🟨🗽🐍 @MichaelHauda]

The "first generation immune to propaganda" moment

[ApoStructura @ApoStructura]

What annoys me about the anti Al people is that they're not even upset for the right reasons.

They’re not worried about RSI and its implications, they're worried about short term job displacement (which comes with any productivity gains), and about fake problems like water use.

[Isaiah @isaiahharvi]

@elonmusk The voters with kids are likely the upper half (30-44) of the sampled age demographic, which we already know to lean more right than voters 18-29.

We can’t conclusively draw any conclusions from this as it relates to children.

[Elon Musk @elonmusk]

The people deserve to have their RAM too!