Science
Navalia AI Agent: A 35B Model Beats GPT-5.4 on Finance Docs
Navalia AI agent turns company financial reports into verified training tasks and lifts a 35B Qwen model by up to…
Topic
Reinforcement Learning
Science
Autonomous Construction: A Crane That Builds and Reports Back
Autonomous construction gets a voice: UBC researchers pair reinforcement learning with ChatGPT 5 so a robot crane builds walls and…
Science
Microduck Explained: Pollen’s $399 Open-Source RL Robot Duck
Microduck robot duck explained: Pollen Robotics' 25cm, $399 biped learns to walk via reinforcement learning. What's open, what isn't, and…