Crypto & Blockchain · Sun, Nov 1, 2026

Bitcoin, Ethereum, stablecoins, DeFi, smart contracts, scaling and consensus.

RSS feed
arXiv cs.AI · cs.CR · cs.DC

A Compute Center Anyone Can Join: A Peer-to-Peer System for AI Inference

In briefA Compute Center Anyone Can Join: A Peer-to-Peer System for AI Inference

SummaryThe paper proposes the Open Inference Market (OIM), a peer-to-peer system in which individuals and small institutions contribute idle GPUs to AI inference. Its core mechanism, earnings as collateral, has workers deliver results without knowing whether they will be audited, with unsettled earnings held as collateral during a challenge window and forfeited entirely if cheating is confirmed. In a simplified model with example parameters, about 0.3 replays per worker per day make cheating unprofitable for a rational worker.

Yimo Lin, Zhen Zhang

arXiv q-fin.TR · cs.CE

AIDA: An AI Outage Registry with Evidence from On-Chain MEV

In briefAIDA: An AI Outage Registry Tracking Failures and On-Chain MEV Activity

SummaryThis paper introduces the AI Incident and Disruption Archive (AIDA), a registry of 5,868 deduplicated outage incidents across 30 AI providers from 2025 to 2026, with foundation-model incidents verified against primary sources. The authors pair it with verified maximal extractable value (MEV) transaction data on Ethereum and Base. They find that active arbitrage addresses quadrupled from late 2024 to late 2026 while sandwich attackers declined, and execution stayed concentrated among the top fifty addresses. During AI outages, the release of new trading code slowed, while already-deployed bots continued trading.

Wenbin Wu

arXiv cs.CR

Attack Patterns in Cross-chain Bridges: A Trust-based Empirical Analysis

In briefAttack Patterns in Cross-chain Bridges: A Trust-based Empirical Analysis

SummaryThis paper empirically studies 48 real-world cross-chain bridge attacks from January 2022 to August 2026. It proposes six attack patterns, each defined by the trust assumption the attacker violates, and maps them to five functional domains of bridges. On the primary loss basis, Proof Forgery and Credential Seizure account for 86.2% of losses, and the Verification and Governance and Operations domains hold 93.9%.

Ishraq Samiha, Yasir Rahman, Rahma Mukta