Paper walkthroughs, literature surveys, and frontier results made practical.
MoE routers let inference providers silently halve the experts they activate — saving 31% of compute while staying undetected by every current fingerprinting scheme.
Majority voting counts hands — it can't rank quality. Fortytwo's Bradley-Terry protocol achieves 85.90% on GPQA Diamond vs. 68.69% for majority voting, extracting ranking signal from pairwise comparisons and making Sybil attacks economically unattractive.
Nova replaces expensive SNARK-in-SNARK recursion with a random linear combination that folds two R1CS instances into one — 10× less overhead per step. For N-step AI inference: O(N) prover work, one constant-size final proof. NANOZK delivers this at GPT-2 scale: 6.9 KB, 23 ms verify.
Shared-state multi-agent LLMs exhibit three formal anomaly classes — stale-generation, phantom-tool, causal-cascade. Three 2026 papers prove them unavoidable without isolation primitives and measure the fix: zero corruptions across 884,110 commits under Observable-Read Isolation.
Covenant-72B pre-trained a 72.7B-parameter LLM across 70+ anonymous internet peers. The key was SparseLoCo: 1.56% gradient density plus 2-bit quantization achieves 146× compression, collapsing the required uplink from 1.9 Gbps to 110 Mbps — consumer fiber.
Reasoning models hide their chain-of-thought but bill for every token of it. CoIn proves a provider can inflate that hidden count 5× with 94.7% detection — but only if the escrow contract enforces a commitment root before payment clears.
DiFR's seed-commitment round-trip is the last synchronization barrier to async verifiable inference. TOPLOC removes it: a locality-sensitive hash of intermediate activations compresses to 258 bytes per 32 tokens — 1,000× smaller than raw embeddings — and validates faster than inference ran.
LLM non-determinism has been the blocker for output-based verification. DiFR commits to a random seed before inference — and activation fingerprints built from random orthogonal projections detect 4-bit quantization in just 2 tokens at AUC > 0.999.
An on-chain agent's transactions prove what happened, not why. Three 2026 papers on trajectory anomaly detection show how to close that gap — and how the optimistic bisection game from rollups makes it slashable.
CoW Protocol's batch auction design makes front-running structurally impossible — and that matters enormously for AI agents executing DeFi strategies at machine speed.
FRI-STARKs are fast and trustless, but their proofs are megabytes wide. Groth16 is 256 bytes but needs a ceremony. STARK→SNARK wrapping resolves the tension — and it's why SP1 and RISC Zero can settle any ML inference on L1 for under 300k gas.
Deposit ETH in Aave, borrow USDC, buy more ETH, repeat. Three loops creates 2.97× leverage from a single ETH — and all three positions share one liquidation trigger at −3% ETH price. The math, the cascade, and why AI yield optimizers find themselves here by default.