A daily crypto, finance, and AI vocabulary grouping puzzle for beginners.
Stablecoin Liquidity Management
These concepts describe liquidity planning and redemption operations for stablecoin systems without suggesting any investment action.
- Reserve Ladder: A reserve ladder spreads reserve assets across different maturity dates so cash availability can be planned over time.
- Cash Redemption Buffer: A cash redemption buffer is readily available money set aside to help process token redemptions or withdrawals.
- Settlement Window: A settlement window is the expected period during which a payment, redemption, or asset transfer is completed.
- Redemption Timetable: A redemption timetable sets expectations for when redemption requests are accepted, processed, and completed.
ETF Creation Mechanics
These terms explain how exchange-traded fund plumbing connects share supply, market baskets, and underlying asset value.
Share Creation BasketPrimary Market BasketIn-Kind RedemptionIndicative NAV
- Share Creation Basket: A share creation basket is the set of assets delivered to a fund when new ETF shares are created.
- Primary Market Basket: A primary market basket is the asset basket used for direct ETF creation or redemption activity with the fund.
- In-Kind Redemption: An in-kind redemption exchanges ETF shares for a basket of underlying assets rather than for cash.
- Indicative NAV: Indicative NAV is an estimated intraday value of a fund's underlying holdings, often used as a reference for ETF pricing.
Agent Evaluation Ops
These concepts help teams test whether an AI agent uses tools correctly, completes tasks, and stays within expected boundaries.
Tool-Call TraceTask Success RateRegression EvalSafety Rubric
- Tool-Call Trace: A tool-call trace records which tools an AI agent used, what inputs were sent, and what outputs came back during a task.
- Task Success Rate: Task success rate measures how often a model or agent completes a defined task according to evaluation criteria.
- Regression Eval: A regression eval checks whether a model or agent has lost expected behavior after a prompt, model, or system change.
- Safety Rubric: A safety rubric is a scoring guide used to judge whether an AI system follows expected safety and policy constraints.
Verifiable AI Data
These ideas describe ways to record where data came from, how claims are checked, and how external facts can reach applications.
- Training Data Provenance: Training data provenance records where training data came from, how it was collected, and what transformations were applied.
- Content Credential: A content credential is metadata that helps describe the origin, editing history, or authenticity claims attached to digital content.
- ZK Attestation: A ZK attestation is a verifiable claim that can prove something about data or computation while revealing limited underlying information.
- Data Oracle Feed: A data oracle feed supplies external facts or measurements to software systems, including smart contracts or AI workflows.