First principles
SIX POSITIONSPhysics is not a feature
A system that can be configured to ignore conservation laws is not a physics-informed system. The constraint layer has to be non-optional or it is decoration.
Provenance before prediction
An engineer's first question about a number is where it came from. Any system that answers that question second will not be trusted with the first.
Accountability cannot be delegated to a model
Someone signs. Software can prepare the case, assemble the evidence and state its confidence, but a named engineer accepts the risk.
The ontology is the product
Interfaces are replaceable. A shared, governed vocabulary for engineering concepts compounds in value every year it stays coherent.
Disagreement is a signal, not an error
When the twin and the instrument disagree, the useful move is to publish the residual — not to retune until the gap disappears.
Infrastructure earns trust slowly
Nothing in a plant adopts quickly, and it should not. The design goal is to be verifiable at every step rather than impressive at the demo.
Essays
UPDATED MONTHLYThe Engineering Intelligence Manifesto
The founding argument for a constitutionally bounded intelligence layer over physical assets.
Why predictive maintenance stalled at the pilot
A decade of pilots produced accurate models and very few changed maintenance plans. The missing layer was never the algorithm.
The residual is the reading
On digital twins that are graded by how loudly they disagree with the plant rather than how closely they match it.
Naming things: the case for an engineering ontology
Four plants, four names for the same pump. Why canonical vocabulary is an engineering problem before it is a data problem.
The formal version, for citation
The manifesto is also published as an ungated whitepaper with a revision number, references and a downloadable PDF.