🔥The 9th Round of Easy Loan, Earn $40 Reward is in progress❗️ ⏰ Promotion Period: January 15th - Feburary 15th, 2025 👉 Register now and check more details at gate.io/campaigns/358

My favorite local model right now is a bit of surprise to me: I'm really enjoying the relatively tiny Qwen3-8B, running the 4bit quantized version on my Mac using MLX It's surprisingly capable given it's a 4.3GB download and uses just 4-5GB of RAM while it's running

We have just made the largest dataset for understanding radiologist-AI collaboration available to the world with 227 radiologists in our study.

What's the benefit of toxic data? A Harvard University study unveiled that training on more toxic data can help us reduce harmful behavior in models + lead to better control. Of course, that's if we use the right post-training techniques. That's a path to easier detoxification 🧵

Current radiology report models lack expert-like structured reasoning. They fail to link visual findings to precise anatomical locations, hindering clinical trust. BoxMed-RL solves this with a two-phase framework. It first instills radiologist-like thinking and visual

🚨New Post 🚨 If you're like me and you've been putting off learning about GRPO, now is as good a time as ever, especially with Deepseek-R2 just around the corner!

📄 Our new study published in Journal of Imaging Informatics in Medicine: A cross-institutional evaluation of #LLMs for #radiology diagnosis extraction using a standardized, human-optimized prompt. link.springer.com/article/10.100…

This is why I consider Ludwig one of the best academics of recent years. He's been leading the most impactful work that academics usually shy away from. Not just once, not just twice, but many times over! And his pivot from robustness was amazing to witness, hats off.

Is Reinforcement Learning the key to unlocking the next level of AI? 🤯 In our 6th videocast, @PouriaRouzrokh & Moein Shariatnia break down Google's AI blitz, the power of RL, and new tools to peek inside the AI black box! Watch now: youtu.be/I-dmN7-3hFw📺 #AIinMedicine

the value of ai talks have exponential decay we PeepResearched™ Andrej's slides and compiled them all for subscribers slide screenshots are from too many people to thank, but extra special shoutouts to Ethan Niser, hung , Aidan Andrews Mert Barutçuoğlu , Alperen Keleş

This is literally the best AI tool roadmap, It saved me hours.

🤯 Building computer-vision applications is never going to be the same again. As Andrej Karpathy says, "English is going to be the new programming-language", and we just made that happen for programming visual AI. 👨‍💻 This past month, the VLM Run team cooked up something magical that

Announcement 📢 We're hiring at Sophont for a variety of positions! We're looking for exceptional, high-agency ML research scientists/engineers who are passionate about using AI to build the future of healthcare. Check our website for more details (link in 2nd tweet)

Can an AI model predict perfectly and still have a terrible world model? What would that even mean? Our new ICML paper formalizes these questions One result tells the story: A transformer trained on 10M solar systems nails planetary orbits. But it botches gravitational laws 🧵

Aravind Srinivas has mandated the use of AI coding tools at Perplexity. ML engineers upload screenshots of pseudocode and get runnable experiments in an hour. Work that used to take 3–4 days. The same tools fix front-end bugs and redesign UI from annotated screenshots. But you

Today we are rolling out our first Gemini Embedding model, which ranks #1 on the MTEB leaderboard, as a generally available stable model. It is priced at $0.15 per million tokens and ready for at scale production use!

New Eps of our podcast is out! 🔥 We explore: 🧠 Agentic AI revolutionizing science. 🩺 A model that has effectively "solved" ECGs. 😳 The creepy side: AI profiling you from browser history & geolocating photos youtu.be/jOs5Qmk-wc4?si… #AI #Podcast #LLMs #AgenticAI #MedAI

🚀 Big news in healthcare AI! I'm thrilled to announce the launch of OpenMed on Hugging Face, releasing 380+ state-of-the-art medical NER models for free under Apache 2.0. And this is just the beginning! 🧵

Diffusion video models but now - **realtime**! Simple video filters are real-time but can only do basic re-coloring and styles. Video diffusion models (Veo and friends) are magic, but they take many seconds/minutes to generate. MirageLSD is real-time magic. Unlike simple video

1/N I’m excited to share that our latest OpenAI experimental reasoning LLM has achieved a longstanding grand challenge in AI: gold medal-level performance on the world’s most prestigious math competition—the International Math Olympiad (IMO).

Two cents on AI getting International Math Olympiad (IMO) Gold, from a mathematician. Background: Last year, Google DeepMind (GDM) got Silver in IMO 2024. This year, OpenAI solved problems P1-P5 for IMO 2025 (but not P6), and this performance corresponds to Gold. (1/10)