Domains
Robotics: measure the simulator first.
Robots learn in simulation, but few can say where a simulator is wrong, or by how much. The Senua AI Robotics Core is the software a robot is developed and run on, and it measures that gap.
Four parts, one language
World, the simulator
Contact and friction, with sensors and motors that speak the real device's wire format. The same inputs give the same bytes every run.
Body, on the robot
Device drivers, a real-time layer and a 1 kHz reflex loop, with a stop latch and force limits below the mind.
Mind, the causal engine
It reads the robot's sensors as streams, learns their causal states, and turns each moment into a decision with a confidence.
Measure, the gap instrument
It reads two sources of the same motion and reports whether they differ, where, and by how much, and holds a physics engine to the right answers.
The first deep dive: SuperDex
Measure held Meta's open physics engine for robot hands to the laws of physics: twelve departures from the right answer, each traced to its cause and corrected by seven opt-in patches with five of SuperDex's own settings. Every rule was written before its data existed. The same work applies to any physics engine or robotics framework. The deep dive and the open record.
We do not use models. No LLMs. No SLMs. Of any kind. The engine learns the structure of a signal from the signal itself: no GPU, no data centre and no AI token costs, ever. It runs on low-end CPUs: the whole engine has run on a Raspberry Pi Zero 2 W. Next it will read a recorded machine against its own simulation. If you build robots or maintain a simulator, talk to us.