I’m excited to introduce PHASMA Robotics 2.5, a decentralized SaaS platform designed for training surgical robotics reinforcement learning (RL) models directly on the Internet Computer.
This release brings together trustless on-chain compute and real-time visualization, making advanced surgical robotics training accessible to anyone.
What’s New in 2.5
Upgraded RL Training Pipeline
A more stable, responsive RL engine optimized specifically for high-precision surgical tasks.
Real-Time Visualization
Watch your robotic agents learn live.
Smooth rendering, high-resolution views, and interactive inspection — all powered by the IC.
Phasma Robotics — New Version Demo is Live (on $ICP)
We’re pleased to announce that the new Phasma Robotics version demo is now live.
This release introduces a platform for training surgical robotics reinforcement learning (RL) models with real-time visualization. The system allows users to observe model behavior dynamically as training progresses, offering a clearer understanding of robotic decision-making in simulated surgical environments.
This is currently a demo version, designed to showcase the architecture, training workflow, and visualization capabilities of the platform.
The goal is to build infrastructure that supports:
Training of surgical robotics RL models
Interactive real-time 3D visualization
Foundations for future autonomous surgical systems
The leaderboard now tracks the top 10 runs ranked by reward score
It updates live during playback, so you can watch your current run climb the rankings in real time — no refresh needed.
What’s New in 2.5
Upgraded RL Training Pipeline
Real-Time Visualization
Who It’s For
Try it here
Explore the demo:
