Token Wei retweeted
Replying to @Semiaut0o0
Well i have a group chat with Dan and Clara and they basically ignore my advice. SA telling institutional clients wrong information helps my P&L.
Damn what if RLHF is just waterboarding for LLMs? Is this ethical?! Will someone at an ethical AI lab look into this?
3
844
Token Wei retweeted
If you want to power a datacenter off the grid, a gas turbine is the "obvious" choice. But it might not be the best option! Many developers select reciprocating engines for a reason. (1/4)🧵
Token Wei retweeted
SemiAnalysis has long advocated AMD spend less on buybacks and more on investing in the ROCm OSS ecosystem. We now issue the exact opposite plea to Nvidia: please resume aggressive buybacks and stop buying OSS projects and ruining them.
BREAKING CUDA MOAT EXPANDS: Today, NVIDIA has acquired SchedMD, makers of SLURM, a widely used "open source" workload scheduler. Many AI companies such as Mistral, Thinking Machines, parts of Meta's FAIR division, university academic labs use SLURM. NVIDIA's acquisition expands the CUDA moat to a new layer, as many customers trying to use alternative AI accelerator chips like AMD, Intel, etc. also rely on SLURM (a lot use k8s too). Unlike Run:AI & DGX Lepton, which nobody uses, a lot of people use SLURM. (1\2) 🧵
Token Wei retweeted
Announcing Jemini 3: Our most intelligent model, now tuned for Jeffrey Epstein.
With Jemini's multimodal reasoning, get answers to any question on Epstein's emails, documents, photos, and DOJ files in the Jmail Suite.
I am actually a Max Verstappen fan
Latest episode of the AGI Podcast is out, featuring Wei Zhou, Head of AIUR at @SemiAnalysis_.
@TokenWei and I go deep on compute, capital markets, tokenomics, and and the future of AGI.
Inside the conversation:
01:00 — Wei’s background and early journey
3:38 — The most cracked researchers in China
5:30 — My path after finishing a Math PhD at Stanford
7:52 — A based, unhinged take on Andrew Ng
10:29 — Choosing money vs “contributing to society”
19:32 — What Wei learned working with George Soros
24:00 — Intel vs Nvidia and the road to becoming the next compute central bank
27:15 — CUDA’s early days and how it became Tencent 2.0 for AI
35:22 — The true base unit of token cost: gigawatts vs dollar capex/opex
39:18 — “How do you price everything?” — a compute-first worldview
42:22 — How investment flows actually move in public markets
43:54 — Understanding depreciation distribution
52:00 — Is memory/storage scarcity real, and what drives it?
54:41 — Should storage cost be part of tokenomics models?
57:44 — Why inter-chip interconnect defines the economics of AI tokens
1:05:56 — Wei’s surprising love for Pop Mart
1:10:52 — How insiders really define AGI and ASI
Watch the full podcast here 👇
Token Wei retweeted
I modestly raise you
fabricatedknowledge.com/p/gp…
As it turns out, I read @scuttleblurb's $NVDA pieces a few days ago. This is what I wrote in a group chat after reading part 1:
"I think there's a good chance that may have been the best piece on NVDA written till 2021. So many good stuff there that later became "common knowledge" (bitter lesson, jevon's paradox, detailed explanation of CUDA moat). Amazing just how under-hyped Scuttleblurb is."
And this is what I wrote to him after reading part 2:
"bro, you should be really proud of your work on NVDA. Honestly, I cannot believe you posted all that in mid 2021.
god tier writing. Incredibly sharp analysis. Frankly speaking, I had to update a bit how good of an analyst you are (and I already had a lofty opinion)"
I know someone whose date canceled on her because IREN was down so much last week @ellehcimnehs
16
4,050
Token Wei retweeted
Replying to @sdrzn
His joke was funny.
Meanwhile you did the out of touch boomer thing to explain the joke saying it's true, then turn around and fire him? – at Millbrae, CA
Token Wei retweeted
NEWS: U.S. based Lidar-maker Luminar has just filed for bankruptcy.
Luminar Technologies was a maker of lidar sensors for self-driving cars. Last month, Volvo decided to remove the Luminar lidar sensors from its EX90 and ES90 cars and discontinue its relationship after Luminar’s failure to meet its contractual obligations.
Token Wei retweeted
Just met a Claude Code user who has been the top daily user numerous times.
He says Codex is better now. – at San Francisco, CA
Token Wei retweeted
It’s official. We’ve raised $14m led by @OpenAI Startup Fund to bring AI to Excel.
Endex is the first AI agent to live inside Excel.
For the past year, we've been working with financial firms. Today we’re releasing it to the world.
Our capacity is limited; comment below for an early invite 🧵
Token Wei retweeted
AWS Trainium3 Deep Dive,
A Potential Challenger Approaching,
Step-Function Software & System Improvements,
“Amazon Basics” GB200 NVL36x2,
NL72x2/NL32x2 Scale Up Rack Architecture,
Optimized Perf per TCO, Trainium4
newsletter.semianalysis.com/…
























