A 14B model just beat a 671B model on math reasoning.
Here’s how Microsoft’s rStar2-Agent achieves frontier math performance in 1 week of RL training
- by “thinking smarter, not longer.” 🧵
🚨 New paper, and big personal news! 🧵
First, I just published a new theory of emergence.
It traverses the dimension reductions of a system, treating each scale like a 2D slice of a 3D object, looking for what each adds causally and irreducibly.
arxiv.org/pdf/2503.13395
You can question Gorgias the sophist directly about the nature of the human condition! Super smooth work from @plastic_labs@honchodotdev. Monetize with x402 @CoinbaseDev
Fresh @honchodotdev + x402 (@CoinbaseDev) demo!
Penny for Your Thoughts is a demo where anyone can share their expertise and sell bits of it via micro-transaction.
Hop on, get interviewed, set your price, & collect payments!
Fresh @honchodotdev + x402 (@CoinbaseDev) demo!
Penny for Your Thoughts is a demo where anyone can share their expertise and sell bits of it via micro-transaction.
Hop on, get interviewed, set your price, & collect payments!
NEW 🫡 HONCHO DEMO!
Introducing "Penny for Your Thoughts"
It's a @honchodotdev + x402 (@CoinbaseDev) powered platform that allows you to get paid for the valuable context in your head!
STEP 1: Get interviewed by an AI agent & publish your "expert"
STEP 2: Set your price
STEP…
We collaborated with the @TextArena team on SPIRAL—a framework that teaches language models reasoning skills through self-play on games.
They trained a 4B parameter Qwen3 model on Kuhn Poker, a simple 3-card game with no mathematical content. After training, the model performed…
# Plastic's Mission
@plastic_labs’ mission is to radically decentralize alignment. Your AI should be an extension of you. You should dictate how it's aligned. And you should own the data used to do it.
Most LLM applications are still optimizing for homogenization, if not…
There is no secret sauce to making something great.
It just takes a group of believers who care, who are unafraid to build and try crazy ideas.
ryolu.notion.site/how-to-make-so…
Introducing the Neuromancer Series of Models
Today we're announcing Neuromancer XR. The first of a series of custom reasoning models for AI-native memory & social cognition.
Neuromancer XR is expert at explicit reasoning. It extracts atomic certain conclusions from user & agent…
83K Followers 8K FollowingCompiling in real-time, the race towards AGI.
🗞️ Don't miss my daily top 1% AI analysis newsletter directly to your inbox 👉 https://t.co/6LBxO8215l
319 Followers 1K FollowingLast cycle i was a defi and token researcher • now i just whip up alchemical • meme • cream • to • azoth • your • dreams • ⌖ •
2K Followers 887 FollowingAssociate professor @EmoryUniversity. Working on large language models, LLM inference, reasoning, natural language generation, and various aspects of GenAI.
55K Followers 2K FollowingHead of Design @Cursor_ai. Early @NotionHQ, @Stripe, built startups. I make a world where anyone can make software. Aspiring k-pop idol.
3K Followers 101 FollowingI make mathematics and physics animation videos on YouTube
BSc., MSc. Mathematical Eng.
M. Eng. Electrical & Computer Eng
https://t.co/5mEbk2JCqF
10K Followers 851 FollowingI want to understand things deeply and explain them well. Building friendly AI @AnthropicAI
Give me anonymous feedback: https://t.co/7aBNrpbad8
83K Followers 8K FollowingCompiling in real-time, the race towards AGI.
🗞️ Don't miss my daily top 1% AI analysis newsletter directly to your inbox 👉 https://t.co/6LBxO8215l
163K Followers 0 FollowingInvented principles of meta-learning (1987), GANs (1990), Transformers (1991), very deep learning (1991), etc. Our AI is used many billions of times every day.
346K Followers 1K FollowingDeepMind Research Scientist. Opinions my own. Inventor of GANs. Lead author of https://t.co/M6vl8pEQ4I Founding chairman of @pubhealthaction
163K Followers 166 FollowingCo-founder of Thinking Machines Lab @thinkymachines; Ex-VP, AI Safety & robotics, applied research @OpenAI; Author of Lil'Log