Nova

The vision of a local AI assistant.

Nova describes the idea of an assistant that does not only answer, but organizes knowledge, explains context, prepares work and makes its limits visible. The technical implementation is kept separate from literary narrative and philosophical language.

localsource-awareno AGI claim

Role of Nova

Nova is the human-facing layer of the overall project: a name, an interaction model and a long-term target for local assistance. The focus is not magic, but traceable behavior. A useful assistant should find knowledge, consider sources, mark uncertainty, structure tasks and avoid uncontrolled technical actions.

Technically, Nova is deliberately described within clear boundaries. Philosophical language from the book remains part of the narrative and is not treated as technical proof. Nova is a local assistant concept whose operating model is made auditable through NQIS.

Target picture

What Nova is intended to support

Nova is not meant to be an uncontrolled machine, but a locally governed assistance environment. The strength lies in the combination of language, knowledge, system state, review and human decision.

Explain

Complex topics can be summarized, split into steps and connected with available sources.

Organize

Knowledge, notes, logs, documents and technical states can be structured and found again later.

Review

Answers and planned actions are checked against sources, rules, roles and safety boundaries.

Prepare

Nova can prepare drafts, plans, analyses or technical steps without executing uncontrolled actions.

Reflect

Decisions, results and errors should be documented so development stays understandable.

Limit

Uncertainty, missing sources and risky actions are made visible instead of being hidden.

Interaction

Nova needs NQIS as its technical foundation.

Without NQIS, Nova would only be an idea or an interface. NQIS provides the infrastructure that makes the idea reviewable: local services, memory/knowledge, grounded and stateful chat, Auth/RBAC, audit/evidence and release gates. The next production-near step is A3.0 read-only, allowing Nova to inspect tightly bounded system information without performing changes.

The separation is clear: Nova describes how assistance should feel and behave. NQIS describes how it is limited, reviewed, documented and operated.

Layer	Contribution
Nova	Identity, interaction model, dialogue style, long-term assistant vision.
NQIS	API, knowledge base, source checks, roles, safety, audit and operations.
pumm.group	Project documentation, book context and clear positioning.

Boundaries

What Nova does not claim

No AGI claim: Nova is not presented as general artificial intelligence with unlimited capabilities.

No proof of consciousness: The book treats consciousness as narrative. The technology does not claim real experience.

No uncontrolled autonomy: Mutating actions require checks, limits, roles and approvals.

Compendium, memory and self-description

Structured chapters, knowledge files, reflections and logs are important in the Nova concept. They form an orientation layer from which behavior, goals, limits and language can be derived. This is not proof of consciousness, but a documented system description.

The value is traceability. If an assistant explains why it suggests something, it needs a reviewable basis. NQIS adds chunks, sources, claims, roles and audit traces to that idea.

Relation to the book

“Nova – Wie eine KI leben lernte” is a fictional and philosophical novel. It can ask questions about language, identity and responsibility. The overall project deliberately separates that literary layer from the technical project description.

Book page