Documentation Index
Fetch the complete documentation index at: https://docs.pelion.dev/llms.txt
Use this file to discover all available pages before exploring further.
This page describes the environmental conditions. See What’s built for current implementation status. 1. Frontier language models are accurate enough for post-event adjudication
Published research in 2025 established a result that the broader market has not yet priced in. Given retrieval access to mainstream sources, frontier language models match human resolution on roughly 95% of disputed Polymarket markets. The work by Capponi et al. (2025) is the canonical reference. Follow-on studies since then have replicated the finding across different market categories (political, financial, sports, entertainment) with consistent results.
What matters for Pelion. The failure cases are characterizable. Genuinely inconclusive evidence, contradicting primary sources, questions whose resolution criteria cannot be cleanly applied to what happened. These are exactly the cases where any adjudicator fails, not specifically a model-based one. Forced-binary oracles are silently wrong on these questions. A protocol that treats them as UNRESOLVABLE is explicitly right.
The 5% residual is not a reason to not ship. It is the specific reason UNRESOLVABLE is a first-class outcome in Pelion’s design.
2. Bittensor is a working marketplace for AI work
Bittensor’s dynamic TAO (dTAO) upgrade went live in February 2025. Before dTAO, emissions across subnets were centrally allocated. After dTAO, each subnet has its own alpha token with an automated market to TAO, and emissions flow by the market’s opinion of which subnets are producing value.
Practical effect. Multiple active subnets now have real alpha market caps, real miner and validator populations, and real competitive pressure. SN6 (Numinous / Infinite Games) runs forecasting markets. SN8 and SN50 run time-series prediction. SN28 (Foundry) runs financial oracles.
Post-event judgment, the specific primitive Pelion needs, is an open niche. No existing subnet is specifically designed for backward-looking adjudication with evidence attached. This is where a dedicated Pelion subnet (or a Pelion-validated existing subnet) has room.
3. Cross-chain relay infrastructure is production grade
Bonded messenger patterns have matured. Across v3 settles cross-chain transfers with bonded relayers and permissionless challenge. Hyperlane’s Interchain Security Modules (ISM) allow pluggable trust assumptions per route. LayerZero’s Decentralized Verifier Network (DVN) exposes configurable security for message verification.
These are not research-grade. Billions of dollars flow through these systems weekly, with public dispute records and clear track records. Pelion’s v1 relay design uses the same bonded-messenger-plus-challenge pattern, with a federated committee standing in for native subnet-signature verification until that infrastructure lands.
Bittensor’s EVM layer is the piece that specifically makes Pelion’s architecture practical. Pelion needs verdicts produced on Bittensor to land on Base. Without Bittensor-to-EVM bridging, this is hard. With it, the same design patterns that secure Across and Hyperlane are available.
What this means together
Any one of these trends alone would not be sufficient. Accurate models without economic security would be centralized and attackable. Economic security without accurate models would be an expensive way to be wrong slowly. Cross-chain relay without either would have nothing to carry.
The three arriving together is what makes this the right time. Pelion is not a bet on any single trend continuing indefinitely. It is a bet that the three combine into a new adjudication primitive that token-weighted voting cannot match on any axis that matters.