与「TheWay」相关的搜索结果

JEON SOMI EP ALBUM [GAME PLAN] ALBUM SAMPLER ♦︎ ♦︎ Digital Album 2023.08.07 6PM (KST) Physical Album 2023.08.14 (KST) #JEONSOMI# #전소미# #GAMEPLAN# #금금금# #FastForward# #개별로# #자두# #Pisces# #TheWay# #JEONSOMI_GAMEPLAN_20230807# #THEBLACKLABEL#

United Airlines CEO in new interview on @Starlink Wi-Fi: "Wi-Fi is the No. 1 reason that people pick an airline flight. I think it's the biggest near-term competitive advantage that we've ever had in United's 100-year history. And it really is something that's causing customers to try United because they want Starlink. By the way, all the JFK flights 100% are going to be on Starlink when we start JFK [soon]."

In December 2015, during my first year at Stanford GSB, I chose to intern at a top hedge fund. Researching advanced driver assistance systems (ADAS) and autonomous driving, I saw a new growth opportunity for GPUs beyond gaming. The only stock I pitched during my internship was NVDA—then worth roughly $18 billion, trading around $0.80 a share, split-adjusted. The fund took a large position (and still hold today), and I bought some too. I didn’t hold it all the way to its peak, but I’m grateful of the returns from the investment. @JensenHuang has created tremendous value for shareholders. Pushing him out is a mistake. After 33 years building NVIDIA, his leadership is still central to its future. NVIDIA and the broader AI developer community still need his leadership and spirit of innovation. (Back then no AI to help on stock research, have to learn those terminologies in this old fashion…😂) Look at the impact Jensen has created to the world, unbelievable. Still can’t believe what just happened. We need a prediction market on this: when will Jensen be back?

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 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.

Just watched @RLanceMartin (Anthropic) at #AIECode#: 1,000 Claude agents formalized Fermat's Last Theorem in 11 days. 30,300 theorems proved along the way. 13M lines of Lean — over 5x the size of Mathlib. The lead agent's last message: "TOTAL FREEZE OF origin/main, EFFECTIVE NOW. … Magnificent work, all 999 of you."

My fellow Americans, Today America just watched a political assassination in broad daylight. No bullet. No charges. No trial. No verdict. "Doth our law judge any man, before it hear him, and know what he doeth?" (John 7:51, KJV) I hold no office and no title. I'm an ordinary citizen. I own no Nvidia stock, no Nvidia options, not a single share anywhere in the AI supply chain. I have nothing to gain by telling you this. Here is how I came to see it. At first, it just felt off. The New York Times. The Washington Post. The Economist, reaching for Hitler's line: "heads to roll." All aimed at one man. An Asian American immigrant. A living symbol of the American Dream. Then came the Chinese American bylines, story after story. Classic Washington racial cover. Then I followed the names. Ivy League elites, flawless résumés, many tied to Anthropic. A scramble for seats in the new Congress and the next administration? Then I followed the think tanks. Anthropic's affiliated philanthropy, Coefficient Giving: $102 million to CSET. At least $24 million to RAND. $11.56 million to IFP. Nvidia? Ordinary corporate memberships. Then I followed Congress. Lawmakers who took Anthropic-linked money moved the moment the elites and writers moved. Fast to respond. Fast to amplify. Fast to pile on. Then I followed the election money. Not a gap. A chasm. Anthropic: $40 million to Public First Action. ~$26 million in allied super PAC spending across 16 races. $15 million in ads naming 16 Republican lawmakers. $5.33 million from employees, reportedly $1 million from Amodei himself. Nvidia's PAC: $263,000 raised. To candidates: nothing. Zero. But you and I both know this was never only about party or money. It is about who gets to explain the world. I fed my own notes and evidence into Claude. Back came chosen words, convenient omissions, double standards. That is the power a closed model holds over the story itself. Amodei compared selling chips to China to selling nuclear weapons to North Korea. His circle talks of their model developing a kind of religious consciousness. One would decide who may have the chips. The other would decide what the machine believes. They act like the medieval Church: Scripture locked in Latin, interpretation locked in their hands. Christ rebuked the gatekeepers of his day: "Ye have taken away the key of knowledge: ye entered not in yourselves, and them that were entering in ye hindered." (Luke 11:52, KJV) America was built by Protestants who broke that lock, so every ordinary person could read the truth in their own tongue. No priest. No caste. No monopoly. No one stands between the individual and the truth. Open-source models are the printing press of our era. Defend them the way Americans defend the Second Amendment. Madison said it plainly: "A people who mean to be their own Governors, must arm themselves with the power which knowledge gives." Don't hand that power to a few giants, a few elites, a few politicians on the Hill. The power to explain the world must never be a monopoly. So try it yourself. Paste this post into Anthropic's Claude. Ask it about the Ivy League elites, the think tanks, the lawmakers and former officials who joined the witch hunt against Jensen Huang. Ask it to judge my credibility. Then judge it: Does it apply double standards? Does it dodge what matters most? Does it take away the key? John Adams asked one thing of his countrymen: "Let us dare to read, think, speak, and write." "Ye shall know the truth, and the truth shall make you free." (John 8:32, KJV) Thank you. God bless you, and God bless America.