Here's a recent talk I gave recapping the last 6-12 months of AI progress, why getting perfect models is hard, how labs are likely approaching the next phase of training (for agents), and other interesting tidbits across the reasoning landscape. Topics: 00:00 Introduction & the

Excited to partner with OKX!

from “memory of search engines” —> “memory of foundational models” We continue to be higher order tool users as tools evolve 😄

I didn't believe when I first saw, but: We trained a prompt stealing model that gets >3x SoTA accuracy. The secret is representing LLM outputs *correctly* 🚲 Demo/blog: mattf1n.github.io/pils 📄: arxiv.org/abs/2506.17090 🤖: huggingface.co/dill-lab/pils-… 🧑‍💻: github.com/dill-lab/PILS

In case you missed this — We're going behind the scenes — how it came together, what's live now, and what this unlocks. We shared the vision on AI developer platform and AI marketplace, and talked through what's next for Sahara AI | SaharaAI.com 🔆.

We Z.ai are thrilled to open-source GLM-4.1V-9B-Thinking, a VLM that can think with long CoTs. SoTA in <10B VLMs, comparable to Qwen-2.5-VL-72B in 18 tasks. One RL to rule them all! Details - Tech report: arxiv.org/abs/2507.01006 - Code: github.com/THUDM/GLM-4.1V…

Data Services Platform (DSP) is LIVE! 🔆 Now anyone, anywhere in the world, can contribute to AI development and earn real rewards for their work. 🔆 $450K+ in $SAHARA + partner rewards available day one! Get started today → app.saharaai.com/data-services-… #AIforALL

1+1=3 2+2=5 3+3=? Many language models (e.g., Llama 3 8B, Mistral v0.1 7B) will answer 7. But why? We dig into the model internals, uncover a function induction mechanism, and find that it’s broadly reused when models encounter surprises during in-context learning. 🧵

I’ll be at ACL 2025 next week to present this work! 🇦🇹 Excited to meet old friends and make new ones. Let’s catch up if you like thinking more about the future of human-centred NLP, personalization and multi-turn interactions or just wanna get some nice Viennese coffee ☕️

Welcome back Sean Ren 🔆, Co-Founder & CEO of Sahara AI 🔆, to #KBW2025: IMPACT! Sean is leading the charge for decentralized AI platforms that empower collaboration and fairness! 📍Sept 23–24 | Walkerhill, Seoul 🎟 tickets.koreablockchainweek.com #KBW #KoreaBlockchainWeek #Web3

I passed my thesis defense today! 😊 It's been a rewarding journey at USC NLP USC Thomas Lord Department of Computer Science. Many thanks to Sean Ren Robin Jia Swabha Swayamdipta Salman Avestimehr, and everyone who supported me!

🔈BASS SBC Speaker Highlight 🔈 Sean Ren Co-Founder Sahara AI | SaharaAI.com 🔆 & Franklin Bi General Partner Pantera Capital

Excited to be speaking at Berkeley #SBC2025 #BASS2025 about convergence of AI x Web3 and how we could create an open, collaborative economy for an AI-driven future. Our blockchain product lead shares the tech architecture for how we achieve on-chain monetization flows for AI

LLMs can appear to reason well, but a single wrong token can derail the whole output. Our new work shows that token-level memorization is a key cause of failure, especially under distribution shift. Introducing: STIM 🔍🧠 arxiv.org/abs/2508.02037 🧵 #NLProc

🚀New dataset release: WildChat-4.8M 4.8M real user-ChatGPT conversations collected from our public chatbots: - 122K from reasoning models (o1-preview, o1-mini): represent real uses in the wild and very costly to collect - 2.5M from GPT-4o 🔗 hf.co/datasets/allen… (1/4)

Sean Ren of Sahara AI | SaharaAI.com 🔆 and Oleg Golev of Sentient discussing signal vs. noise at the intersection of AI and crypto.

Excited to release new repo: nanochat! (it's among the most unhinged I've written). Unlike my earlier similar repo nanoGPT which only covered pretraining, nanochat is a minimal, from scratch, full-stack training/inference pipeline of a simple ChatGPT clone in a single,