Applied AI · Unification of Physics

Applied AI for the Unification of Physics.

We are encoding the experiments that pin down the laws of physics into bds421's temporal knowledge graph, then letting you question them directly. One interactive case every month, every answer traceable to its source.

rho/tkg · open sourcesigma/tkgd · Cypher + Tylanew case every month
01 · The idea

Physics, encoded as facts you can question.

A century of experiments, in one graph. Each landmark result becomes a time-indexed, sourced fact in bds421's temporal knowledge graph. You can then ask what the evidence says, and get an answer that is retrieved and verified, not generated and hoped for. The same engineering behind our consulting work, pointed at the open questions of physics.

Facts, not vibes

Experiments become facts

Every published result is stored as a fact with a source and a window of validity in time. rho/tkg holds the graph.

Two engines

Ask what is, and what follows

Cypher answers what is in the graph. Tyla reasons over rules to derive what follows from it, with interval arithmetic over time. sigma/tkgd.

Provenance

Every answer shows its working

Each claim carries the source it came from and the time it was valid. Verified by construction, so the answer you read is the answer you can check.

Monthly

One experiment a month

A new interactive case lands every month, each one wired into the same growing graph. The archive only gets deeper.

02 · The graph

Ask the graph.

The end goal of this project: a chat that answers physics questions straight from the graph, running Cypher and Tyla queries over sigma/tkgd and returning every answer with its provenance. Here is a preview of how that conversation feels.

Ask the graph demo · scripted
you
Why didn't the gamma-ray speed depend on the pion's velocity?
the graph
Because the second postulate of special relativity holds: light moves at c in vacuum no matter how fast its source moves. In the 1964 test, neutral pions travelling at β = 0.99975 emitted gamma rays whose measured speed stayed at c, not c(1 + β). The graph stores this as a measurement event linked to the postulate it tests.
Alväger, Farley, Kjellman, Wallin · Phys. Lett. 12, 260 (1964) matched in sigma/tkgd · Cypher

This preview is scripted. The live system does not generate answers, it retrieves them: each claim is a stored fact with a source and a valid-time interval, and the engine returns only what the graph entails, with full provenance. If a fact is not in the graph, it says so rather than inventing one. That is the same neuro-symbolic approach behind datAInsights.

03 · Examples

One experiment a month.

Each case is a small, interactive instrument: change the inputs, watch the physics, read the verdict. Then ask the graph about it. New experiments are added at the top.