Research
in motion.
Field definitions, technical positions, formal primitives, benchmark agendas and experimental results from NEO Labs.
Trajectory Intelligence for Autonomous Systems
The object of study, boundary conditions and irreducible capabilities of intelligence operating across time, state and consequence.
Read the research objectA Trajectory Is Not a Training Sample
What continual-learning pipelines discard when production traces become training objects.
Read the research objectFrom Identity to Effect
Federated attestation across agent identity, authority, execution evidence and independently verified consequence.
Read the research objectThe Right Granularity Is Claim-Dependent
Adaptive trajectory evaluation without evidence loss at stage boundaries, milestones or early stops.
Read the research objectFrom Memory Write to Governed Belief Commit
Reliability, provenance, applicability and durable-state consequences in autonomous memory systems.
Read the research objectThe Runtime Is Part of the Policy
How interfaces, defaults, control surfaces and render boundaries shape autonomous outcomes.
Read the research objectThe Harness Is Not the World
Agent operating systems, self-improving runtimes and the unresolved boundary between task success and consequence truth.
Read the research objectClaims should survive
contact with evidence.
NEO Labs distinguishes original contribution from prior art, hypotheses from results, and structural integrity from independently verified truth.
Read the live frontier