Your curated collection of saved posts and media
Excited to present the FAIR Chemistry Leaderboard - a centralized space for our teamβs community benchmark efforts. Weβre kicking things off today with the OMol25 leaderboard! πLeaderboard: https://t.co/OVRTHWkniu π₯οΈCode: https://t.co/kGsXE414kC https://t.co/RYftFYHNTq

π Just updated lightevalβs readmeβcanβt believe weβve grown to cover ~7,000 tasks π³ with top-tier multilingual support π llm as judge π€ multiturn evals π£οΈ coding benchmarks π§βπ» https://t.co/JHsJ25YqNy
Unitree first open-source world-model on @huggingface! UnifoLM-WMA-0 is Unitreeβs first open-source world-modelβaction architecture spanning multiple types of robotic embodiments, designed specifically for general-purpose robot learning. Its core component is a world-model capable of understanding the physical interactions between robots and the environments. This world-model provides two key functions: (a) Simulation Engine β operates as an interactive simulator to generate synthetic data for robot learning; (b) Policy Enhancement β connects with an action head and, by predicting future interaction processes with the world-model, further optimizes decision-making performance. Link to the model: https://t.co/RtTDrHLrFy

Introducing 90+ open-source, stateβofβtheβart biomedical and clinical zeroβshot NER models on @HuggingFace by @OpenMed_AI Apacheβ2.0 licensed and ready to use Built on GLiNER and covering 12+ biomedical datasets π§΅ (1/6)
We're releasing an updated dataset format for @LeRobotHF, and it is built for scale. LeRobotDataset:v3 supports multi-million episode datasets and streaming, enabling better performance across the board Learn more: https://t.co/nORGObvQ7m
The @huggingface followers leaderboard is BACK after a month π Top users: 1. @TheBlokeAI 2. @lvminzhang 3. @mervenoyann 4. @bartowski1182 5. @akhaliq 6. @ylecun 7. @fffiloni 8. @xenovacom 9. @Teknium1 10. @maximelabonne 11. @TheEricHartford Fixed thanks to @charlesbben π https://t.co/hELP0Q1uUo
I made a game engine for vibe coding https://t.co/jXs4jkjBJq
I made a game engine for vibe coding https://t.co/jXs4jkjBJq
@AntLingAGI Ling-Flash-2.0 from Ant Finance just dropped on @huggingface - 100B MoE, 6.1B active (4.8B non-embedding) - 128k context length - Trained on 20T+ tokens - Base model available too - Great performance on reasoning tasks - MIT license https://t.co/CGyjw779Ak
New short course: Build AI Apps with MCP Servers: Working with Box Files, built with @Box and taught by @BenAtBox , their CTO. Many AI applications require custom code for basic file operations. The Model Context Protocol (MCP) standardizes this by letting you offload file tasks to dedicated servers that provide tools an LLM can use directly. In this course, you'll process documents stored in a Box folder using the Box MCP server. Rather than writing custom integration code to connect to the Box API and download files, you'll design your application to use the tools provided via MCP. Skills you'll gain: - Build an LLM-powered document processing app, using the Box MCP server to access files - Design a multi-agent system using Google's Agent Development Kit (ADK), consisting of specialized agents for file operations - Coordinate the multi-agent workflow through an orchestrator that uses the Agent2Agent (A2A) protocol to connect to the agents You'll start with a local file-processing app, refactor it to work with Box's MCP server, then evolve it into a multi-agent system. Sign up here: https://t.co/FitKgvGnpb
reminded of this paragraph from gsm8k paper, 2021 :) https://t.co/lmubBMBFIb
It just hit me that sub 1B-parameter models that are way better than 175B GPT-3 are a dime a dozen today. Kinda cool.
from this era https://t.co/1pIFiqes4Y
If you see me posting less here, it's because I'm posting on the private feed, which is where all the unfiltered nonsense is https://t.co/FzxGmUnqmD
π«‘ https://t.co/XBr0b2TK80
The 3rd edition of my book Deep Learning with Python is being printed right now, and will be in bookstores within 2 weeks. You can order it now from Amazon or from Manning. This time, we're also releasing the whole thing as a 100% free website. I don't care if it reduces book sales, I think it's the best deep learning intro around, and more people should be able to read it.
Buy the book from Amazon: https://t.co/6g4BtE7vf6 Buy the book from the publisher: https://t.co/BIIbAZ2iBT Read the free website: https://t.co/3CbcQ7hmjp The new edition has a ton more content around generative AI. It also covers JAX & PyTorch fundamentals, and everything new in Keras 3.

When I started LLMs-from-scratch I just hoped it might help a few people learn. Just saw the GitHub the repo has now been forked 10k times! More than the stars, the best part is seeing thousands of people actually use and build on the code βΊοΈ https://t.co/NoQ7gJNd0y
And I have a long list of things I plan to add over time! https://t.co/j3Knzj7TGW
@QRDL Neat! I haven't had a chance to check it out (crazy week) but bookmarked it. I think this will come in handy when I finally write a MoE-focused article some time in the future. PS: might be a github issue, but the Nb doesn't seem to render https://t.co/HqOedevWDc
In fact, he explained why it's *not* a contradiction, and leads to "real results" -- this is the actual translation: https://t.co/ASzMItpIUH https://t.co/Nz3e6j1S2l
The 'mathematics' in question: https://t.co/U20l0jJ5tR
The 'mathematics' in question: https://t.co/U20l0jJ5tR
OK dug into this and discovered that, despite the >6k likes, this "note on mathematics" doesn't exist, and in fact the quote is a mathematical misunderstanding from a Japanese economist. It's basically the opposite of what Marx really wrote. https://t.co/qmbA1DpMhQ https://t.co/WEcvp1yzA4
The 'mathematics' in question: https://t.co/U20l0jJ5tR
@kamivine That is an exploration of the difference between limit-based and infinitesimal-based conceptions of the foundations of calculus. https://t.co/iUb2nxhD1S
1/7 We're launching Tongyi DeepResearch, the first fully open-source Web Agent to achieve performance on par with OpenAI's Deep Research with only 30B (Activated 3B) parameters! Tongyi DeepResearch agent demonstrates state-of-the-art results, scoring 32.9 on Humanity's Last Exam, 45.3 on BrowseComp, and 75.0 on the xbench-DeepSearch benchmark.
@willccbb This was an article I liked a while back which I wrote https://t.co/mqd9aNQAyt Since then Iβve been tinkering with shorter tutorials with a few more in the works https://t.co/8kacyEZbWR

Congratulations to our cofounder & chief scientist, @peakji on being recognized as one of MIT @techreview's Innovators Under 35. The feature dives into Peakβs story: predicting AIβs trajectory long before it was mainstream, and the lessons from ideas that came too early. Itβs a reminder that what looks like failure at the time can become the foundation for whatβs next. Full story here: https://t.co/DQa3GO371o
ok life update: i'll be joining @Cognition! β’ Cog just went from 0 to $10b in 2 years β’ Net burn $20m in company history β’ Avg successful Devin impl sees >5x growth, $1.5m/yr customer expanded >10x in 8 months (not typo) β’ Windsurf x Cog cross sell going great, looking fwd to cross build. Lighthouse customers in every vertical. β’ Windsurf's polish + Wave 13/14 are looking v HYPE! more work to do on Windsurf Tab Here's the 5 decision points I had to cross to "buy" Cognition's rise as an Agent Lab that has the potential to be as dominant in the Decade of Agents: β’ Short Code timelines, Long AGI timelines β’ The rise of Agent Labs β’ Owning the Sync/Async spectrum β’ Start high, go low β’ Cracked Engineers + Ramped GTM This is probably TMI but I like to record thoughts in public so I can learn when I am wrong, and for others to follow along. Happy reading, in reply tweet. also: Most of @smol_ai capital base is rolling over to Cognition's round today. I'll be returning the rest. AINews and SmolTalk will remain my passion projects. I will continue to operate @aiDotEngineer and @latentspacepod ferociously independent of Cognition. All Cognition competitors were given ample heads up + red carpet to AIE CODE. Ben and I are BEYOND excited to have Lia step up as the new General Manager of AIE!

the new mcp announcement according to @NanoBanana ahahahha https://t.co/nXL4DbiEUv
Today, we announced that we will commission a statue of Charlie Kirk to honor his legacy and incredible work after his tragic assassination last week. The statue, privately funded by community leaders, will stand on campus as a commitment by New College to defend and fight for free speech and civil discourse in American life. The location for the statue will be announced in the coming months.
One Battle After Another (2025) https://t.co/yUX4PuKLwS
She bit the tops off all 72 of the seedlings that I started for my fall garden so unfortunately I had to confiscate her arms and legs :( will return them pending good behavior https://t.co/HKfumZW6xv
HOLY SHIT METAβs live AI demo FLOPPEDπ https://t.co/2eukAvAosl