Knowledge catalogue
Our GB300 cluster went down yesterday, just as Deepseek released 😱 We were 😥 but @CoreWeave came through to contribute to the Open Source. They scrambled in the compute crisis, finding 2 spare dev rac
🎉 Day-0 support for @deepseek_ai V4 Pro and Flash on vLLM — a new generation of DeepSeek model, purpose-built for tasks up to 1M tokens. Alongside the release, we're publishing a first-principles walk
DeepSeek V4 by @deepseek_ai just dropped! SGLang is ready on Day 0 with a full stack of optimizations from architectures to low-level kernels. We also deliver a verified RL training pipeline in Miles
Claude Code spend had gotten to $10.95M runrate peak at SemiAnalysis But then Opus 4.7 saved me. More token effecient for tasks, smarter, and no fast mode. Thank you @AnthropicAI You saved me from ban
Every conversation I have with @dylan522p, I'm really just trying to understand the supply and demand of tokens. This is a unique episode in that it's entirely dedicated to talking about both sides of
How close are models to building a perfect Slack clone with less than 50,000 of tokens? I feel like not that far... Claude Code spend had gotten to 10.95M runrate peak at SemiAnalysis But then Opus 4.
'Lean In' likely refers to a post by Dylan Patel from SemiAnalysis discussing lean manufacturing, operational efficiency, or business strategy concepts, though the specific content cannot be verified
people who type with capslock are kinda fakers when i am crusing with caps, I HOLD DOWN SHIFT KEY WITH MY RIGHT PINKIE EXTREMELY HARD! THE EMOTION IS FLOWING THROUGH TO MY TEXT i often imagine others
This post appears to be a humorous take on AI optimization and technology trends, likely referencing 'pixel-maxxing' (maximizing pixel-level performance or visual quality) in contrast to 'catfishing'
Hi i'm dwarkesh! Grew up all over the US, now sf-based and always down to nerd out about AI, science & history :) a lil about me: 🟠 Host of the dwarkesh podcast 🟠 Studied at UT Austin 🟠 Just published
Intelligence per picojoule, with @itsclivetime and @dylan522p (0:00) Intro (1:22) What is codesign? (2:49) Codesign example: Swish vs ReLU (4:22) Are DeepSeek papers codesign? (6:45) Predicting where
Johny Sroji, the GOAT who made Apple silicon program what it is, is now in charge of all hardware at Apple as Chief Hardware Officer Media Johny Srouji Taking Over as Apple's Chief Hardware Officer as
At SemiAnalysis, we're quite tired of the minimalist style webapps and landing pages that have become so commonplace lately. Today, we're introducing Minecraft mode on inferencex dot com so that you c
This post likely explores the various relationship dynamics and social connections people develop in shared living and working spaces, with a speculative or humorous tone suggested by the 'Maybe lover
Intel is so fucking back Long road ahead but @LipBuTan1 is the man for the job He is making many long term principled decisions Despite my long term optimism, I believe numbers won't be great in the s
Dylan Patel of SemiAnalysis posted a clarification on X distinguishing NVIDIA from automotive companies, likely in response to comparisons or confusion surrounding NVIDIA's growing presence in the aut
Comparing hardware on InferenceX just got a massive upgrade. Our new changelog update lets you compare every change for any gpu all on the same chart. Just select any model and GPU config, then select
Dylan Patel, a semiconductor and AI industry analyst known for his work at SemiAnalysis, shared a post on X (formerly Twitter) humorously reacting to a LEGO-style avatar or depiction of himself create
I was unable to retrieve the specific tweet content from that URL, as the post requires a logged-in X (Twitter) session to access, and search results did not surface the full text of that specific ...
The Information wrote that Meta consumes 60.2T tokens per month, or ~750M per employee. Meta is getting absolutely token-mogged by us currently. SemiAnalysis employees consume 1.86B tokens per month.
Nvidia published DWDP (Distributed Weight-Data Parallelism), a new inference parallelism strategy focused on prefill. It sounds slightly insane until you remember the target machine is GB200 NVL72. Th