2025 was a whirlwind of activity in AI+science. Our team’s seminal contributions to the field were well recognized and we deepened those foundations. You can read the year end summary here
MACHINE LEARNING
-
Robotics Revolution: AI, 3D Printing, and Technological Evolution
By
–
The #Robotics revolution continues, powered by advancing #AI, improved #3D printing, an unstable #Megatrend of technological advancement; Darwin’s Theory of Evolution and Moore’s Law need more contemplation. pic.twitter.com/tilPlcN9cZ
— Mike Quindazzi (@MikeQuindazzi) 29 décembre 2025The #Robotics revolution continues, powered by advancing #AI, improved #3D printing, an unstable #Megatrend of technological advancement; Darwin’s Theory of Evolution and Moore’s Law need more contemplation.
→ View original post on X — @mikequindazzi, 2025-12-29 13:06 UTC
-
MIT System Transforms Natural Language Into Physical Robot-Built Objects
By
–
Speak an idea. Watch a robot build it.
We’re getting closer to a world where intent becomes physical reality. MIT researchers just demonstrated a system where natural language turns directly into assembled objects — no CAD tools, no programming, no manufacturing background -
2025 AI Landscape Released: FirstMark’s Major Industry Redraw
By
–
The 2025 AI Landscape just released. The key highlights of the year: Matt Turck and the FirstMark team have been mapping Machine Learning, AI & Data for 11 years. This is their biggest redraw yet. They cut logos from 2,000+ to ~1,150. Not because the space is shrinking, but
-
Sim-to-Real Transfer: The Critical Test for AI Systems
By
–
Same. The sim-to-real jump is the real test. If it holds up, that’s huge!
-
MiniMax M2.1 Powers Robot Dog with 230B Parameter Sparse MoE
By
–
MiniMax M2.1 × Vbot is super-powered in action! This AI agent model (230B params, sparse MoE) controls a physical robot dog seamlessly—virtual training to real-world navigation, coding, & multimodal mastery.
— Futurepedia – Learn to Leverage AI (@futurepedia_io) 29 décembre 2025
Speedy, efficient, rivals Claude Sonnet.#MiniMaxM21 #AIRobotics pic.twitter.com/Hji1KAuIgzMiniMax M2.1 × Vbot is super-powered in action! This AI agent model (230B params, sparse MoE) controls a physical robot dog seamlessly—virtual training to real-world navigation, coding, & multimodal mastery. Speedy, efficient, rivals Claude Sonnet. #MiniMaxM21 #AIRobotics
-
AI breakthroughs like world models accelerate progress
By
–
Translation: we are about to accelerate because of AI breakthroughs like world models.
-
Claude Automates Nanochat ML Experiments End-to-End
By
–
Claude has been running my nanochat experiments since morning. It writes implementations, debugs them with toy examples, writes tests and makes them fail/pass, launches training runs, babysits them by tailing logs and pulling stats from wandb, keeps a running markdown file of
-
Proposed Technical Curriculum for AI and Robotics Integration
By
–
Ce que je verrais dans le programme : Ça serait un mélange de DL, du RL, de de modeling 3D, des LLMs , du signal processing, de l’électronique , de l’embarqué, du réseau.