Technology learning and innovation ecosystem empowering people to explore new frontiers, develop practical skills, and turn ideas into real-world solutions.
๐ Course spotlight: Charge AI Agents Per Request With x402 Payments*
HTTP has a built-in payment status code: 402 Payment Required.
x402 turns it into a payment protocol, allowing AI agents to pay for APIs and services in stablecoins, one request at
https://t.co/OVXgBTJIQD
A subagent hack that has really come in handy for me
Categorise your subagents so each one focuses on a specific niche
Then run a smaller model on each of them
You save on cost, and you get a specialised agent that's focused on something small
Small model + small job = happy wallet ๐
I played with @arc Arc Studio to recreate this and make it better and the cool thing is you don't need to worry about getting a video demo,it does that for you.This is what it built:https://t.co/vGSsdrw5fz
cc: @SarahWahinya@samconnerone@corey__cooper@bleso_a@bobbilee
I played with @arc Arc Studio to recreate this and make it better and the cool thing is you don't need to worry about getting a video demo,it does that for you.This is what it built:https://t.co/vGSsdrw5fz
cc: @SarahWahinya@samconnerone@corey__cooper@bleso_a@bobbilee
Meet GR00T 2, our new generation of world-action models (WAMs): it imagines how the physical world will unfold and generates the actions that get it there. One model runs on parallel grippers, dexterous hands, and humanoids.
https://t.co/32Z1plSGQg
Introducing Gemini Robotics 2.5, our most capable vision-language-action (VLA) model to date. ๐ค
It increases both the breadth of skills a robot can learn and the speed at which it acquires them โ helping robots step out of the lab and into real-world environments.
Learn more about Gemini Robotics 2.5, now rolling out to our robotics partners ๐งต
Praxis-1: an open-weight world action model for real robots ๐ค
New from @runwayml and now in LeRobot.
Praxis-1 learns how objects behave and how tasks unfold from large-scale video, which gives it a head start on a new robot. One checkpoint adapts with light fine-tuning or LoRA: fine-tuned on Google's RT-1 robot, which it never saw in training, it reaches 80.0% on SimplerEnv. Weights, inference and fine-tuning code are open under Apache 2.0, commercial use included, and it runs on NVIDIA from H100 to Jetson Thor.
Model, blog and docs in the comments ๐
๐ Course spotlight: A Bill to Ban Superintelligence
Senator Bernie Sanders and Representative Greg Casar introduced the Ban Artificial Superintelligence Act on September 4, 2026.
https://t.co/n3FeVso7OU
๐ฎ Learning just got a lot more fun!
What if learning technology felt like playing a game?
Weโre introducing the *Kids Learning Path on Multichain Builders*, designed for children and young people aged 6โ18.
๐ Explore the Kids Path: https://t.co/kRFMSG6wnq