# Decentralized AI ecosystem

Konnex treats intelligence for physical work as a first-class, onchain commodity. Models and policies have owners, versions, stake, reputation, and—where applicable—deterministic checks in subnet sandboxes or on real hardware.

## Roles

The roles follow the same **subnet** coordination pattern seen in networks such as [Bittensor](https://docs.learnbittensor.org/):

* **Miners** — Supply control or perception policies for a workload class; stake and compete for tasks.
* **Users / operators** — Issue tasks (natural language or schema) and pay in testKNX on testnet; stablecoins on mainnet when the [settlement design](/understand-konnex/stablecoins-integration.md) is live in production.
* **Validators** — Score outputs independently; they are not a proxy for the operator’s commercial self-interest.
* **Stakers** — Back validators and future bond pools per network rules.
* **Simulator / bench** — Where the subnet requires deterministic replays.

## Lifecycle of a policy or model version

1. **Publish** — Register a build with metadata, KPIs, and stake.
2. **Compete** — Tasks pull miner outputs into the open market for the subnet.
3. **Gate** — Validators run replay, sims, or sensor checks; unsafe or out-of-spec work is rejected and can be slashed.
4. **Deploy / record** — Winners execute as defined; telemetry feeds [Proof-of-Physical-Work](/understand-konnex/contracts-and-popw.md).
5. **Settle** — Rewards and reputation update; on mainnet, stablecoin settlement applies when enabled.

## Public model registry (conceptual)

A stable identifier (e.g. hash of code, hyperparameters, and author key) with immutable metadata and governance-controlled fields such as deprecation and royalty routing.

## Feedback loop

Better KPIs and safety behavior lead to higher validator scores and, where the economy allows, larger rewards and lower required collateral. Fraud or gross negligence triggers slashing and makes counterparties whole per parameterized rules.

For concrete VLA integrations, see [AI models — overview](/supported-ai-models/ai.md).


---

# Agent Instructions: Querying This Documentation

If you need additional information that is not directly available in this page, you can query the documentation dynamically by asking a question.

Perform an HTTP GET request on the current page URL with the `ask` query parameter:

```
GET https://docs.konnex.world/understand-konnex/ai-ecosystem.md?ask=<question>
```

The question should be specific, self-contained, and written in natural language.
The response will contain a direct answer to the question and relevant excerpts and sources from the documentation.

Use this mechanism when the answer is not explicitly present in the current page, you need clarification or additional context, or you want to retrieve related documentation sections.