By
–
Perplexity published its 1st research on Weight Transfer for RL Post-Training. “We recently achieved 1.3-second cross-machine parameter updates for Kimi-K2 (1T parameters), transferring weights from 256 training GPUs (BF16) to 128 inference GPUs (FP8).” On my weight
By
–
Experience AI superpowers at your desk. Researchers, faculty, and students can unleash personal AI supercomputing for innovative projects, creative experiments, and faster discoveries, with NVIDIA DGX Spark. Learn more to #SparkSomethingBig https://
nvda.ws/3KvPQaw
By
–
GPUs are expensive and setting up the infrastructure to make GPUs work for you properly is complex, making experimentation on cutting-edge models challenging for researchers and ML practitioners. Providing high quality research tooling is one of the most effective ways to improve research productivity of the wider community and Tinker API is one step towards our mission there. Tinker API is built on top of our experimental results on fine-tuning with LoRA: thinkingmachines.ai/blog/lor… Beta starts and you can join the waitlist today: thinkingmachines.ai/tinker/
→ View original post on X — @lilianweng, 2025-10-01 18:29 UTC
By
–
Take a look at this marine surveillance system from the Axelera AI community, using edge AI and a custom 2-axis gimbal. It tracks vessels up to 10km away, logs metadata instead of endless video, and monitors restricted zones automatically. Why does it matter? Overfishing,
By
–
gm anon microcenter has new deals on RTX 5070 and RTX 5080 btw
By
–
AI is rewriting the rules of wireless. Join our Telecom SVP Ronnie Vasishta at #NVIDIAGTC Washington, D.C. as he shares the first 6G milestones achieved on NVIDIA platforms and how AI is transforming every layer of next-gen networks. Wednesday Oct 29 | 9 AM – 9:40 AM EDT
By
–
Fast progress in training AI agents to interact with the world. Training on just 2,541 hours of Minecraft video, Google built an AI that runs on a single GPU & was able to mine diamonds offline (which takes an average of 24,000 clicks). The same approach may work for AI robots.
By
–
How is it possible to get to below 5cm accuracy as in this image? Or do the wireframe represent something else?
By
–
Oh yea… And Meta released their new Meta Ray-Ban Displays… Except almost no one can get them because they force you to demo them at Best Buy before they let you buy… And only at Best Buys that are far from where you live… And you have to wait 6-weeks for an appointment.