An educational crypto, finance, and AI vocabulary puzzle about blockchain communication, credit analysis, dependable agents, and computing capacity.
Cross-Chain Messaging
These terms cover message delivery, remote verification, proof data, and duplicate-message safety.
RelayerLight ClientMessage ProofReplay Protection
- Relayer: A relayer observes one blockchain and submits messages or proofs to another network.
- Light Client: A light client checks selected blockchain data without downloading and executing the entire chain history.
- Message Proof: A message proof provides cryptographic evidence that a cross-chain message was included or authorized.
- Replay Protection: Replay protection prevents the same signed message or transaction from being executed more than once.
Credit Risk Basics
These terms describe risk premiums, default likelihood, expected loss severity, and value recovered.
Credit SpreadDefault ProbabilityLoss Given DefaultRecovery Rate
- Credit Spread: A credit spread is the extra yield a borrower pays above a lower-risk benchmark.
- Default Probability: Default probability estimates the likelihood that a borrower will fail to meet an obligation.
- Loss Given Default: Loss given default estimates the share of an exposure that remains lost after recoveries.
- Recovery Rate: Recovery rate is the portion of a defaulted claim that creditors ultimately recover.
Agent Reliability
These terms cover structured tools, controlled retries, escalation to people, and traceable actions.
- Tool Schema: A tool schema defines the fields and types an AI agent must use when calling an external tool.
- Retry Policy: A retry policy controls when and how an agent repeats a failed request without looping indefinitely.
- Human Handoff: A human handoff transfers an uncertain or sensitive task from an AI agent to a person.
- Audit Trail: An audit trail records an agent's inputs, decisions, tool calls, and outputs for later review.
Compute Economics
These terms describe accelerator usage, facility efficiency, guaranteed capacity, and interruptible servers.
GPU HourPower Usage EffectivenessCapacity ReservationSpot Instance
- GPU Hour: A GPU hour measures the use of one graphics processor for one hour of computing work.
- Power Usage Effectiveness: Power usage effectiveness compares total facility energy with the energy delivered to computing equipment.
- Capacity Reservation: A capacity reservation secures computing resources in advance for a workload that needs predictable availability.
- Spot Instance: A spot instance is discounted cloud capacity that the provider may reclaim when demand changes.