Greg Lin Tanaka

@jeremyberman @MLStreetTalk Great interview! I like your strategy around using natural language instead of Python. It is similar to @Prompt_Driven. Perhaps we can chat more about it?

Was great chatting about general intelligence @MLStreetTalk. Full video here: youtube.com/watch?v=FcnLiP…

@sgrove Yes! I think it is natural language. Would love to chat with you more about @Prompt_Driven

Does the world need a new programming language? Quite possibly!

Looking forward to the discussion on Prompt Driven Development with @ToolhouseAI @Prompt_Driven x.com/i/spaces/1lPJq…

Ever get that "grenade in the codebase" feeling from agentic coders like Claude Code? You're never sure what they'll add, delete, or duplicate. I started exploring a new approach: what if prompts themselves were the source of truth instead of merely being used to patch the code?

@sgrove awesome talk on The New Code! Would love to chat with you about promptdriven.ai.

@ivanfioravanti I wonder what the performance is of this quantized version

Model here. huggingface.co/mlx-community/…

Another MLX Boom! 💥 Qwen3-235B-A22B-Thinking-2507-3bit-DWQ! Group Size 32 to keep quality high enough! arc_easy test: 80.3 (testing 3bit, 4bit, 8bit to compare!) M3 Ultra performance: Prompt: 20.2 tokens-per-sec Generation: 31.6 tokens-per-sec Peak memory: 103.010 GB Ready for our Mac with 128GB+!

>>> Qwen3-Coder is here! ✅ We’re releasing Qwen3-Coder-480B-A35B-Instruct, our most powerful open agentic code model to date. This 480B-parameter Mixture-of-Experts model (35B active) natively supports 256K context and scales to 1M context with extrapolation. It achieves top-tier performance across multiple agentic coding benchmarks among open models, including SWE-bench-Verified!!! 🚀 Alongside the model, we're also open-sourcing a command-line tool for agentic coding: Qwen Code. Forked from Gemini Code, it includes custom prompts and function call protocols to fully unlock Qwen3-Coder’s capabilities. Qwen3-Coder works seamlessly with the community’s best developer tools. As a foundation model, we hope it can be used anywhere across the digital world — Agentic Coding in the World! 💬 Chat: chat.qwen.ai 📚 Blog: qwenlm.github.io/blog/qwen3-cod… 🤗 Model: hf.co/Qwen/Qwen3-Cod… 🤖 Qwen Code: github.com/QwenLM/qwen-co…

I attended a vibe coding hackathon recently and used the chance to build a web app (with auth, payments, deploy, etc.). I tinker but I am not a web dev by background, so besides the app, I was very interested in what it's like to vibe code a full web app today. As such, I wrote none of the code directly (Cursor+Claude/o3 did) and I don't really know how the app works, in the conventional sense that I'm used to as an engineer. The app is called MenuGen, and it is live on menugen.app. Basically I'm often confused about what all the things on a restaurant menu are - e.g. Pâté, Tagine, Cavatappi or Sweetbread (hint it's... not sweet). Enter MenuGen: you take a picture of a menu and it generates images for all the menu items and presents them in a nice list. I find it super useful to get a quick visual sense of the menu. But the more interesting part for me I thought was the exploration of vibe coding around how easy/hard it is to build and deploy a full web app today if you are not a web developer. So I wrote up the full blog post on my experience here, including some takeaways: karpathy.bearblog.dev/vibe-coding-me… Copy pasting just the TLDR: "Vibe coding menugen was exhilarating and fun escapade as a local demo, but a bit of a painful slog as a deployed, real app. Building a modern app is a bit like assembling IKEA future. There are all these services, docs, API keys, configurations, dev/prod deployments, team and security features, rate limits, pricing tiers... Meanwhile the LLMs have slightly outdated knowledge of everything, they make subtle but critical design mistakes when you watch them closely, and sometimes they hallucinate or gaslight you about solutions. But the most interesting part to me was that I didn't even spend all that much work in the code editor itself. I spent most of it in the browser, moving between tabs and settings and configuring and gluing a monster. All of this work and state is not even accessible or manipulatable by an LLM - how are we supposed to be automating society by 2027 like this?" See the post for full detail, and maybe give MenuGen a go the next time you're at a restaurant!

Today is the start of a new era of natively multimodal AI innovation. Today, we’re introducing the first Llama 4 models: Llama 4 Scout and Llama 4 Maverick — our most advanced models yet and the best in their class for multimodality. Llama 4 Scout • 17B-active-parameter model with 16 experts. • Industry-leading context window of 10M tokens. • Outperforms Gemma 3, Gemini 2.0 Flash-Lite and Mistral 3.1 across a broad range of widely accepted benchmarks. Llama 4 Maverick • 17B-active-parameter model with 128 experts. • Best-in-class image grounding with the ability to align user prompts with relevant visual concepts and anchor model responses to regions in the image. • Outperforms GPT-4o and Gemini 2.0 Flash across a broad range of widely accepted benchmarks. • Achieves comparable results to DeepSeek v3 on reasoning and coding — at half the active parameters. • Unparalleled performance-to-cost ratio with a chat version scoring ELO of 1417 on LMArena. These models are our best yet thanks to distillation from Llama 4 Behemoth, our most powerful model yet. Llama 4 Behemoth is still in training and is currently seeing results that outperform GPT-4.5, Claude Sonnet 3.7, and Gemini 2.0 Pro on STEM-focused benchmarks. We’re excited to share more details about it even while it’s still in flight. Read more about the first Llama 4 models, including training and benchmarks ➡️ go.fb.me/gmjohs Download Llama 4 ➡️ go.fb.me/bwwhe9

@ctjlewis Every image I try I get this message!!!

@tom_doerr Agreed, I did the same thing

Bought a 55-inch OLED TV as a monitor - best purchase I've made in a long time

@novita_labs @huggingface @julien_c How quantized is this?

Our powerful DeepSeek R1 Turbo is live on @HuggingFace! 🤗 It's improved in every way: 64K context + 16K max output + 30 throughput + 99.9% stability + 💰 CHEAPER THAN EVER Try it out on Hugging Face's model page 🫶 @julien_c

Today, we release QwQ-32B, our new reasoning model with only 32 billion parameters that rivals cutting-edge reasoning model, e.g., DeepSeek-R1. Blog: qwenlm.github.io/blog/qwq-32b HF: huggingface.co/Qwen/QwQ-32B ModelScope: modelscope.cn/models/Qwen/Qw… Demo: huggingface.co/spaces/Qwen/Qw… Qwen Chat: chat.qwen.ai This time, we investigate recipes for scaling RL and have achieved some impressive results based on our Qwen2.5-32B. We find that RL training con continuously improve the performance especially in math and coding, and we observe that the continous scaling of RL can help a medium-size model achieve competitieve performance against gigantic MoE model. Feel free to chat with our new models and provide us feedback!

@paulgauthier Underwhelming…

GPT-4.5 Preview scored 45% on aider's polyglot coding benchmark. 65% Sonnet 3.7, 32k think tokens (SOTA) 60% Sonnet 3.7, no thinking 48% DeepSeek V3 45% GPT 4.5 Preview 27% ChatGPT-4o 23% GPT-4o aider.chat/docs/leaderboa…

Join us today at 1 PM PST to learn a bit about blockchain and AI policies and regulations in the U.S. with our guest speaker @GregTanaka! See you then!

@NVIDIAAIDev What is the pricing for api usage?

Securely experiment and build your own specialized agents, as the 671-billion-parameter DeepSeek-R1 model is now available as an NVIDIA NIM microservice in preview on build.nvidia.com. Learn more ➡️ nvda.ws/4grQaBq

Everyone’s debating how @deepseek_ai R1 trained for <$6M, but it’s simple: same reason @Google Search is ‘free.’ Google sells ads; DeepSeek sells intent. Knowing what the world is thinking fuels their edge in quantitative trading. Data is the new currency. 💡📊 #AI #QuantTrading

@garrytan Very sad…

Meanwhile in Palo Alto they are banning honors biology classes Vote out the the decelerationist education bureaucrats obsessed with virtue signaling equity while everyone suffers

OMG

#RecaptchaHell @digitalocean makes you do 12 of the recaptcha's to get an account!! one, maybe three is okay, but 12?