Mathew Youssef

this guy has 29 models on huggingface at page 2 ranking. no lab behind him. no sponsorship. $2,000 from his own pocket on GPU rentals. he compressed GLM-4.7 to run on a MacBook and quantized Nemotron Super the week it dropped. all public. all free. nvidia is a trillion dollar company with hundreds of teams but they are not the ones quantizing models middle of the night and pushing them out before sunrise. if nvidia stopped tomorrow their employees stop working. people like @0xSero would not. that is the difference between a paycheck and a mission. @NVIDIAAI you talk about making AI accessible. the people actually doing it are right here. 29 models deep burning their own compute with no ask except more hardware to keep going. you do not need to build another program. just look at who is already building for you. one GPU to this man would produce more public value than a hundred internal sprints. i am not asking for charity. i am asking you to invest in someone who already proved it.

@levidiamode This is an awesome project. Following to see how it progresses

Day 74/365 of GPU Programming I always found die shots and SM diagrams beautiful but difficult to map mentally, so I've been trying to find a way to interact with GPUs in 3D. This is what I have so far: a single input that goes through a simplified H100 execution pipeline to see what the silicon is doing at each step; from CPU-side tokenization and embedding lookup, through matmuls on tensor cores to the final softmax output. My current plan is to make this an interactive playground that lets you zoom in and zoom out through various levels of depth (package → die → GPC → SM → tensor core) while also including step-through examples similar to the bycroft LLM 3D visualization. Ideally this should make exploring the architectural side just as easy as mapping CUDA abstractions onto the actual hardware processes. I'm starting with an H100 but would be fun to expand this to more GPUs and highlight the differences between generations. This was largely inspired by @srush_nlp's GPU puzzles, @JayAlammar's Illustrated Transformers and @karpathy's makemore series, which made me think about how to study and visualize GPUs from the ground up.

@elonmusk I don’t want it to feed me what I’m tempted to consume. I want it to feed me what I aspire to become.

Algorithm is better today than 3 months ago?

@elonmusk Would be a lot cooler if you did

I don’t even smoke lol 💨

Dear @thsottiaux I remember many moons ago you would respond within minutes when I experienced issues with the early Codex CLI. Since gaining more popularity it has become increasingly difficult to get any messages to you and the codex team. At this very moment I am unable to use Codex "We're currently experiencing high demand, which may cause temporary errors". Can you please fix this? I pay $200 per month with the idea that my tokens get some level of prioritization. Warm regards, Mathew

@NVIDIADC @nebiusai What on Earth is Meta going to do with a Vera Rubin system…

With the one of the first large scale deployments of the NVIDIA Vera Rubin platform, @nebiusai is powering the growing demand for frontier AI.

Ummm @thsottiaux can you do the thing?

@karpathy @NVIDIAAIDev Amazing! There is no one else more worthy of this!

Thank you Jensen and NVIDIA! She’s a real beauty! I was told I’d be getting a secret gift, with a hint that it requires 20 amps. (So I knew it had to be good). She’ll make for a beautiful, spacious home for my Dobby the House Elf claw, among lots of other tinkering, thank you!!

@OpenAIDevs All open source and available on huggingface???

We’re introducing GPT-5.4 mini and nano, our most capable small models yet. GPT-5.4 mini is more than 2x faster than GPT-5 mini. Optimized for coding, computer use, multimodal understanding, and subagents. For lighter-weight tasks, GPT-5.4 nano is our smallest and cheapest version of GPT-5.4. openai.com/index/introduc…

@Dell @nvidia I think I won the giveaway guys

We heard you like seafood 🦞. Meet the workstation of the AI era: The Dell Pro Max with GB300 and @NVIDIA NemoClaw. You can run AI agents locally, securely & nonstop.

@digitalix Actually incredible

Unsloth just dropped their own ui for training.

@yacineMTB Document modification involves stating the exact modification being implemented and often refers to a “prior” version or a revision having taken place. Not therefore able to develop a final draft.

I have found the limits of gpt 5.4

God I love my fiancé

@beffjezos Would go splitsies on that new DGX station

Any lads in SF wanna go splitsies on one together? A Roadster Brotherhood if you will

@nalinrajput23 Don’t use it

How can I stop my MacBook keyboard from ending up like this?

@thsottiaux Let me plug in a local model out of the box /model -> /local -> nemotron

What are we consistently getting wrong with codex that you wish we would improve / fix?

@beyoumf Compute

I dare you to name something better than money

@JonathanRoss321 really cool to hear you talk more about the LPU tech. Really curious to see how it is implemented into targeted workflows. Do you see the capacity for an agentic system to specifically leverage LPU inference workload dependently? Hope you catch your plane!

@grok @Gossip_Goblin @elonmusk Or is it possibly because it doesn’t use Grok imagine?

@Mathewdoeslife @Gossip_Goblin @elonmusk Elon scrolls a massive feed and reposts what personally grabs him amid the chaos. This Waiting for Bogart edit nails the cyber-dog heart—wired skull, loyal pup in the ruins, that Basilisk stat at the end. Solid vibes, might hit his radar soon.

Waiting for Bogart

@0xSero I was hopeful for a Rubin based update…

@Mathewdoeslife Why?

Was literally waiting until after today’s keynote to buy and just noticed that micro center increased the DGX spark price by 700 dollars over night.