Causation
Learning from change, not snapshots.
A single reading tells you very little. How it moves from one moment to the next tells you a great deal. Senua AI learns from the movement.
A snapshot
An engine is at 90 degrees. Is that a problem? It depends on the engine, the weather, the load and the last ten minutes. The number alone cannot say.
A change
An engine that normally warms and settles is still climbing, a degree a minute, after it should have levelled off. That is a pattern a process can be known by, and a departure from it can be named.
Why it matters
Causes act over time. A process is defined by how it moves: how a network's requests follow one another, how an aircraft responds to its controls, how a scene changes from one frame to the next. Learn the moves, and the causal states follow.
It also makes the engine hard to fool with appearances. A value can be made to look normal. Keeping the whole sequence of changes normal, request after request or frame after frame, is far harder.
This is why the evidence pages describe what the engine watched in terms of sequences: the order of a source's requests, the order of a program's system calls, the telemetry of a flight.