Examples

These examples show how Compute Substrate can be used.

They are not built-in features. They are patterns.

1. Prediction Aggregation

Domain:

eth_price_weekly

Proposals:

"ETH > $4000 by Friday"
"ETH between $3000–$3500"
"ETH < $2500"

Participants submit proposals and attest to the ones they believe are most likely.

Read output:

curl http://localhost:8789/top/eth_price_weekly

Result: a ranked list of predictions based on attestations.

This is not a market. There is no settlement. It is a shared memory of expectations.

Domain:

ml_papers_2026

Proposals:

"Paper A introduces new training method"
"Paper B improves inference efficiency"
"Paper C is flawed due to X"

Researchers and readers attest to the most useful or correct contributions.

Output: a ranked list of research signals. No authority decides what is best. The network only aggregates what participants express.

Domain:

llm_eval_reasoning_v1

Proposals:

"Model X solves multi-step reasoning tasks"
"Model Y fails on long-chain logic"
"Model Z shows emergent planning ability"

Attestations represent confidence or agreement.

Output: a shared view of model capabilities across participants.

No benchmark is enforced.
No score is official.

Domain:

macro_event_fed_march

Proposals:

"Rate hike signals tightening cycle"
"Decision indicates pause"
"Market reaction is temporary"

Participants express interpretations of the same event.

Output: a ranked set of interpretations.

The system does not resolve disagreement.
It records it.

Domain:

security_incidents_stream

Proposals:

"Exploit detected in protocol X"
"Suspicious activity in contract Y"
"False alarm"

Attestations reflect confidence and verification.

Output: a live, aggregated signal layer.

No central authority filters or approves entries.

Domain:

scientific_hypothesis_x

Proposals:

"Hypothesis A explains the data"
"Hypothesis B is more consistent"
"Neither hypothesis holds"

Attestations accumulate over time.

Output: a visible structure of competing ideas.

The system does not converge to truth.
It exposes the landscape.

Across all examples:

Nothing is enforced.

Nothing becomes true because it ranks highly.

Compute Substrate does not decide.

It records.

All meaning remains outside the system.