Loading video...

Video Failed to Load

Go Home

Here are ten 42M TinyStories models running simultaneously in the browser. Andrej Karpathy's Llama2.c code has been ported to Rust using the Candle framework and compiled to Wasm.

radamar's profile picture

Radamés Ajna

6,810 subscribers

45,424 views • 2 years ago •via X (Twitter)

LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Andrej Karpathy llama2.c running on galaxy watch 4
0:18

Andrej Karpathy llama2.c running on galaxy watch 4

Joey (e/λ)

224,497 views • 2 years ago

Hack of the day: Llama on a microcontroller 🦙🔬 Details and source code at Thanks to Andrej Karpathy, whose llama2.c inspired this!
0:10

Hack of the day: Llama on a microcontroller 🦙🔬 Details and source code at Thanks to Andrej Karpathy, whose llama2.c inspired this!

Max Braun

463,646 views • 2 years ago

I always wondered "What if I port my three.js game to Rust? Will it improve performance?" I finally ported it to both Rust Desktop and Rust WASM (that runs on the browser) Comparing the three.js version with the Rust desktop I had a clear winner: Rust desktop. But now I took the time to port it to the web using Web Assembly (WASM) And my conclusion is... A tie! Both versions averaged 42FPS (60FPS if running one at a time) So if your goal is to ship a web game, it doesn't even matter the tech you use. (Three.js is recommended because the ecossystem is huge though) As long you optmize well (Web workers, reduce triangles, draw calls, etc) you should be good to go. But in the event you want to ship a native game to publish on Steam a system language (Rust, C++, etc) will be better for you. Now I want to port the three.js game to Desktop using Electron and see the results. I'll let you know. Btw, I'm open to new freelance projects. If you have anything for me, my DMs are open!
0:18

I always wondered "What if I port my three.js game to Rust? Will it improve performance?" I finally ported it to both Rust Desktop and Rust WASM (that runs on the browser) Comparing the three.js version with the Rust desktop I had a clear winner: Rust desktop. But now I took the time to port it to the web using Web Assembly (WASM) And my conclusion is... A tie! Both versions averaged 42FPS (60FPS if running one at a time) So if your goal is to ship a web game, it doesn't even matter the tech you use. (Three.js is recommended because the ecossystem is huge though) As long you optmize well (Web workers, reduce triangles, draw calls, etc) you should be good to go. But in the event you want to ship a native game to publish on Steam a system language (Rust, C++, etc) will be better for you. Now I want to port the three.js game to Desktop using Electron and see the results. I'll let you know. Btw, I'm open to new freelance projects. If you have anything for me, my DMs are open!

André → andreelias.dev

19,531 views • 1 month ago

I made an extension for using Claude Code, Codex CLI, Copilot in the browser. Select elements and send them to the agent. Works in: ✅ Next.js ✅ React (Vite) ✅ TanStack Start ✅ Vue ✅ Astro ✅ Plain HTML ✅ Any other framework (.NET, Ruby, etc.) Here's how it works...
0:15

I made an extension for using Claude Code, Codex CLI, Copilot in the browser. Select elements and send them to the agent. Works in: ✅ Next.js ✅ React (Vite) ✅ TanStack Start ✅ Vue ✅ Astro ✅ Plain HTML ✅ Any other framework (.NET, Ruby, etc.) Here's how it works...

Wesley

97,242 views • 9 months ago

Breaking News 🚨 Russia has been continuously attacking for the last few hours. It is 100% certain that Russia has been continuously attacking Ukraine for the last few hours, due to which the common people are running here and there in fear.
0:11

Breaking News 🚨 Russia has been continuously attacking for the last few hours. It is 100% certain that Russia has been continuously attacking Ukraine for the last few hours, due to which the common people are running here and there in fear.

Russian Army

32,503 views • 1 month ago

Inspired by Andrej Karpathy and the idea of using games to compare LLMs, I've built a version of the game Codenames where different models are paired in teams to play the game with each other. Fun to see o3-mini team with R1 against Grok and Gemini! Link and repo below.
0:15

Inspired by Andrej Karpathy and the idea of using games to compare LLMs, I've built a version of the game Codenames where different models are paired in teams to play the game with each other. Fun to see o3-mini team with R1 against Grok and Gemini! Link and repo below.

Ilya Abyzov

378,698 views • 1 year ago

made this pixelart grass running at 120fps in the Browser, thanks to Three.js and WebGPU
0:11

made this pixelart grass running at 120fps in the Browser, thanks to Three.js and WebGPU

Erik

37,173 views • 11 months ago

Adam McDonnell has been handed his second yellow of the game and Bohs are down to ten 🔴
0:19

Adam McDonnell has been handed his second yellow of the game and Bohs are down to ten 🔴

League of Ireland

43,865 views • 3 months ago

The Claude Code team gave me early access to the new version of Claude Code Desktop and I’ve been using it for the past few days… here’s what I’ve liked most about this new version 1- Panel creation and the ability to run agents simultaneously. This makes me think we’ll definitely see these agents working in a coordinated way, communicating across sessions.
0:19

The Claude Code team gave me early access to the new version of Claude Code Desktop and I’ve been using it for the past few days… here’s what I’ve liked most about this new version 1- Panel creation and the ability to run agents simultaneously. This makes me think we’ll definitely see these agents working in a coordinated way, communicating across sessions.

Daniel San

64,963 views • 2 months ago

As a learning test, I ported part of my 2D Cloth Physics Simulation to the web using #Rust 🦀 🧪 Try the mini demo: Real-time simulation with Verlet integration, touch/mouse interaction, and web delivery. #gamedev #physics #simulation #verletintegration
0:18

As a learning test, I ported part of my 2D Cloth Physics Simulation to the web using #Rust 🦀 🧪 Try the mini demo: Real-time simulation with Verlet integration, touch/mouse interaction, and web delivery. #gamedev #physics #simulation #verletintegration

Claudio Z.

90,790 views • 1 year ago

This was already the fastest seam carving on the internet thanks to rust+wasm. But I made it 5x faster by putting it on the GPU!
0:13

This was already the fastest seam carving on the internet thanks to rust+wasm. But I made it 5x faster by putting it on the GPU!

Nauseam (in sf!)

79,202 views • 3 months ago

I've been putting together an interactive tool called DiffusionLab for explaining the geometric intuition behind diffusion and flow based generative models. Sampling is actually being done in the browser using Tensorflow.js! It is still in the very early stages.
0:11

I've been putting together an interactive tool called DiffusionLab for explaining the geometric intuition behind diffusion and flow based generative models. Sampling is actually being done in the browser using Tensorflow.js! It is still in the very early stages.

Alec Helbling

104,658 views • 1 year ago

🔥 MAJOR DEVELOPMENT UPDATE 🔥 Wallet Connect has been integrated allowing all mobile users to play directly in the safari browser, giving a much bigger and better user experience! Play Here:
0:18

🔥 MAJOR DEVELOPMENT UPDATE 🔥 Wallet Connect has been integrated allowing all mobile users to play directly in the safari browser, giving a much bigger and better user experience! Play Here:

Meta Monopoly

610,695 views • 3 years ago

Meta just announced that Code Llama was now free for both research and commercial. This might the strongest competitor to ChatGPT: ▸ Can generate, explain, and debug your code ▸ Handles input 100,000 tokens ▸ Free for research + commercial use ▸ Outperforms most open models ▸ Comes in 7B, 13B, and 34B ▸ Supports Python, C++, Java, PHP, Typescript (Javascript), C#, and Bash Available in: ▸ Foundation base models (Code Llama) ▸ Python specializations (Code Llama - Python), ▸ Instruction-following models (Code Llama - Instruct)
0:11

Meta just announced that Code Llama was now free for both research and commercial. This might the strongest competitor to ChatGPT: ▸ Can generate, explain, and debug your code ▸ Handles input 100,000 tokens ▸ Free for research + commercial use ▸ Outperforms most open models ▸ Comes in 7B, 13B, and 34B ▸ Supports Python, C++, Java, PHP, Typescript (Javascript), C#, and Bash Available in: ▸ Foundation base models (Code Llama) ▸ Python specializations (Code Llama - Python), ▸ Instruction-following models (Code Llama - Instruct)

Lior Alexander

196,384 views • 2 years ago

"The rust has been taking its toll on me. Gotta stay in the gym." 😂 SGA after the Thunder completed back-to-back sweeps in the playoffs.
0:15

"The rust has been taking its toll on me. Gotta stay in the gym." 😂 SGA after the Thunder completed back-to-back sweeps in the playoffs.

ESPN

1,001,324 views • 2 months ago

Turso is an incredible technical feat. A Rust rewrite of sqlite, with an async-first architecture, incoming support for concurrent writes, vector search, and browser / wasm support out of the box. I think this has a very good chance of being a foundational piece of infrastructure of the vibe-coding age. On-demand, sqlite-compatible global databases that can also run in-browser and on-device. The pace at which the project is evolving is most definitely *not normal*. Pekka Enberg and Glauber Costa are built different. Demo:
0:19

Turso is an incredible technical feat. A Rust rewrite of sqlite, with an async-first architecture, incoming support for concurrent writes, vector search, and browser / wasm support out of the box. I think this has a very good chance of being a foundational piece of infrastructure of the vibe-coding age. On-demand, sqlite-compatible global databases that can also run in-browser and on-device. The pace at which the project is evolving is most definitely *not normal*. Pekka Enberg and Glauber Costa are built different. Demo:

Guillermo Rauch

242,187 views • 9 months ago

Each booth at Apsara 2025 is buzzing with the latest AI innovations! From the next-generation large language models of the Qwen3 family and the Wan 2.5 models to the new agent development framework feature in Model Studio, the updates are rolling in! Stay tuned as we zoom into the booths and bring the breakthroughs to you!🔥
0:19

Each booth at Apsara 2025 is buzzing with the latest AI innovations! From the next-generation large language models of the Qwen3 family and the Wan 2.5 models to the new agent development framework feature in Model Studio, the updates are rolling in! Stay tuned as we zoom into the booths and bring the breakthroughs to you!🔥

Alibaba Group

161,597 views • 9 months ago

Learn How To Build a Gemini-Powered Robotics Simulator in the Browser with MuJoCo WASM. MuJoCo (WebAssembly) + Three.js + Gemini ER Thanks for sharing, Omar Sanseviero! 📍Tutorial: Demo: —— if it matters in AI or Robotics you'll see it here first:
0:10

Learn How To Build a Gemini-Powered Robotics Simulator in the Browser with MuJoCo WASM. MuJoCo (WebAssembly) + Three.js + Gemini ER Thanks for sharing, Omar Sanseviero! 📍Tutorial: Demo: —— if it matters in AI or Robotics you'll see it here first:

Ilir Aliu

17,686 views • 4 months ago

Decades after its initial launch, Grand Theft Auto 3 has finally been ported to Sega Dreamcast - and the results are phenomenal. John Linneman @dark1x.bsky.social takes a look at the game, stacking it up against the PS2 original:
0:16

Decades after its initial launch, Grand Theft Auto 3 has finally been ported to Sega Dreamcast - and the results are phenomenal. John Linneman @dark1x.bsky.social takes a look at the game, stacking it up against the PS2 original:

Digital Foundry

151,570 views • 1 year ago