What a day to release this monster of a model. Open weights w/ a detailed report for reasoning models and MIT licensed.

The AlphaGo moment for general intelligence is here. RL is mainstream now. It remains to be seen if organizations can use RL to improve their top line with past decisions, revenues as inputs.

What are some of the best resources on the internet to learn the best and latest on reinforcement learning for LLMs?

Build your own Evals. This advice stands the test of time.

those who think RL use less compute don’t know RL at all 😅 SFT: human generates data and machine learns RL: machine generates data and machine learns

We have to take the LLMs to school. When you open any textbook, you'll see three major types of information: 1. Background information / exposition. The meat of the textbook that explains concepts. As you attend over it, your brain is training on that data. This is equivalent

Oopsie.

Amazing journey from an OSS contributor to creating products (in public) and helping developers, enterprises & community! You’ve created so much impact not just with hard work & passion but with extreme ‘agency’. Super excited to see what you’ll do next!

After 6+ months in the making and burning over a year of GPU compute time, we're super excited to finally release the "Ultra-Scale Playbook" Check it out here: hf.co/spaces/nanotro… A free, open-source, book to learn everything about 5D parallelism, ZeRO, fast CUDA kernels,

here you go... next version of Obsidian Web Clipper has nice Markdown export for Claude and ChatGPT out of the box

Everyone use this tool to find relevant papers paperfinder.allen.ai It's sooo good

Teachers using prompt injection to catch students Inadvertently this is the best course for ai security you could take

Fixed the o1 and o3 evals. Ofcourse, using o4!

Big congrats from team India@ML Lossfunk! 🎉🇮🇳 Absolutely thrilled to see 25 papers featuring brilliant researchers from India accepted at #ICLR2025! 🔥 Massive achievement & testament to the growing strength of AI/ML research in the country. A thread celebrating their

Massive productivity gains for every torch user. The path to AGI is unblocked now.

It was good while it lasted!

TOC for the open book "Beyond Naive RAG: Practical Advanced Methods" from our RAG series. This condenses 5 hours of instruction into something you can read in ~30 minutes. Link: maven.com/p/945082/beyon… Ben Clavié Nandan Thakur Orion Weller Antoine Chaffin Bryan Bischof fka Dr. Donut

curious about the training data of OpenAI's new gpt-oss models? i was too. so i generated 10M examples from gpt-oss-20b, ran some analysis, and the results were... pretty bizarre time for a deep dive 🧵

Hillclimbed by way from 83 to 95. Pretty sure I'm stuck at a local minima now. Vibe coded the progress with claude code. binga.github.io/vibe-dat-creat…

A fabulous list of projects and super cool progress!