Questions with truth labels
Claims remain distinguishable from proposals, measurements, reproductions, failures and inconclusive results.
Animus is an AI Institute for secure systems, human-controlled intelligence, reproducible research, applied engineering and open learning. Commercial projects fund real work; real work returns to the Institute as evidence and curriculum.

The Institute does not separate theory from operating reality. A useful contribution can be an experiment, benchmark, bug fix, review, negative result, reproducible artifact or reviewed production change.
Claims remain distinguishable from proposals, measurements, reproductions, failures and inconclusive results.
Partner problems become bounded engagements with explicit scope, evidence, security boundaries and production acceptance criteria.
Project programsLink, Datalab, Amity and Coder represent different maturity levels and authority domains.
Systems portfolioGamification may visualize real progress, but reputation, payments or XP never create security authority or scientific truth.
Adventure modelPersonal Cell state, private Research and security-sensitive capability remain separate trust zones. Public media can explain; models can calculate; only explicit authorized paths can mutate valuable state.
Identity, memory, preferences and private knowledge do not become public because a page can render them.
Production capability is stated only when implementation, release and runtime evidence support it.
Negative and inconclusive results are useful institutional output when they improve the next model.
Start an applied project, study the systems, or follow the first-party production and media archive.