These examples show patterns of use, not built-in applications. Compute Substrate does not come with prediction markets, research systems, or benchmarking frameworks inside the protocol. It provides a minimal surface for publishing proposals, attesting to them, and reading reproducible aggregated outputs.
In every case below, the chain records speculative structure. Meaning, action, and judgment remain outside the system.
A domain such as eth_price_weekly can be used to collect competing expectations about a short time horizon. Participants submit proposals such as:
"ETH > $4000 by Friday"
"ETH between $3000–$3500"
"ETH < $2500"Others attest to whichever proposals they regard as most plausible. A reader can then inspect the aggregated output:
curl http://localhost:8789/top/eth_price_weeklyThe result is a ranked memory of expectations. It is not a market, there is no settlement engine, and nothing pays out automatically. The system records speculative support; it does not resolve the event.
A domain such as ml_papers_2026 can be used to collect interpretations of papers, methods, or critiques. Proposals might describe a claimed contribution, a failure mode, or a competing interpretation:
"Paper A introduces new training method"
"Paper B improves inference efficiency"
"Paper C is flawed due to X"Researchers, readers, or reviewers can attest to the proposals they find most useful or most convincing. The output is a ranked structure of support, not an official verdict about which paper is correct. The protocol merely aggregates what participants expressed under the rules of the chain.
A domain such as llm_eval_reasoning_v1 can collect claims about model behavior:
"Model X solves multi-step reasoning tasks"
"Model Y fails on long-chain logic"
"Model Z shows emergent planning ability"Participants can attest based on their own evaluation procedures, experiments, or judgment. The output becomes a shared view of how different participants are weighting those claims. No benchmark is enforced by the protocol, and no resulting score becomes official simply because it ranked highly.
A domain such as macro_event_fed_march can be used to record interpretations of a public event:
"Rate hike signals tightening cycle"
"Decision indicates pause"
"Market reaction is temporary"Different participants can attest to different readings of the same event. The output is a ranked set of interpretations, visible to anyone reading the chain. The system does not reconcile disagreement. It preserves it as structured memory.
A domain such as security_incidents_stream can be used to record live claims around incidents or anomalies:
"Exploit detected in protocol X"
"Suspicious activity in contract Y"
"False alarm"Attestations can reflect confidence, support, or verification effort. The output becomes a shared signal layer that others may inspect. There is still no central filter, no protocol-level approval step, and no built-in authority that decides which entry is real.
A domain such as scientific_hypothesis_x can make rival explanations visible in a shared structure:
"Hypothesis A explains the data"
"Hypothesis B is more consistent"
"Neither hypothesis holds"Over time, attestations accumulate around the competing proposals. The output is not truth. It is a visible landscape of support, disagreement, and changing weight across ideas. Compute Substrate does not converge a field onto a final answer. It makes the shape of contention legible.
Across all of these examples, the same structure appears. Proposals express possibilities. Attestations express support. Outputs reflect aggregation. Nothing inside the protocol turns a ranked result into an enforced decision, and nothing becomes true merely because it sits at the top of a list.
Compute Substrate does not decide. It records. All meaning remains outside the system.