Forecast Evidence Swarm · NEAR HACKATHON · synthetic data
Would you pay for a forecast that turned out wrong?
You pay for research you can check, not for luck. Two recorded runs, shown separately: an AI forecast (8 Nov 2026) and a paid job on NEAR testnet (9 Nov 2026).
Recorded AI forecast · Synthetic data
…
- Evidence before the cutoff
- …
- Model forecast
- …
- Range estimated by the model
- …
- How the model got the number
…(Laplace smoothing)- Outcome
- … · The less likely outcome occurred. The delivery was still valid.
…
No TEE attestation of the model run in this version (attestation: null).
NEAR testnet receipt · Job 0 · Scripted worker
- Reward
- …
- Status on chain
- ✓ …
- Publication consent
- …
Accepted delivery hash
…
This job used the scripted analyst (75% YES), not the recorded AI run.
View acceptance transaction ↗Check the delivery yourself
This checks whether the report matches the bytes accepted on NEAR.
…
Hash computed in your browser
…