A daily crypto, finance, and AI vocabulary grouping puzzle for beginners.
ETF Market Access
These concepts explain how exchange-traded funds issue shares, work with market makers, and compare fund prices with underlying holdings.
Spot ETFCreation UnitAuthorized ParticipantNet Asset Value
- Spot ETF: A spot ETF is an exchange-traded fund designed to track direct exposure to an underlying asset rather than a futures contract or derivative strategy.
- Creation Unit: A creation unit is a large block of ETF shares that can be created or redeemed through the fund's primary-market process, usually by specialized institutions.
- Authorized Participant: An authorized participant is a financial institution allowed to exchange baskets of assets or cash with an ETF issuer to create or redeem ETF shares.
- Net Asset Value: Net asset value is the estimated value of a fund's underlying holdings minus liabilities, usually expressed on a per-share basis.
Stablecoin Risk Controls
These terms describe operational controls that help users evaluate whether a stablecoin can support redemptions and manage liquidity stress.
- Reserve Attestation: A reserve attestation is a third-party report that gives a point-in-time view of assets backing a stablecoin or similar claim.
- Redemption Queue: A redemption queue is the ordered process for handling requests to convert a token or fund share back into cash or underlying assets.
- Peg Band: A peg band is a range around a target price where small deviations are expected before stronger market or operational responses may appear.
- Liquidity Buffer: A liquidity buffer is a reserve of cash-like or easily sold assets held to meet redemptions, payments, or unexpected funding needs.
Model Ops and Safety
These concepts help teams monitor model behavior, control unsafe outputs, and keep humans involved in sensitive decisions.
- Context Window: A context window is the amount of input and generated text a model can consider at one time when producing a response.
- Guardrail: A guardrail is a policy, filter, prompt rule, or system check that helps keep model outputs within intended safety and quality boundaries.
- Model Drift: Model drift is a change in model performance or behavior as input patterns, data quality, user needs, or surrounding systems evolve.
- Human-in-the-Loop: Human-in-the-loop describes a workflow where people review, approve, correct, or override automated model decisions in important cases.
DePIN Compute Markets
These ideas describe how decentralized infrastructure networks can list compute supply, route workloads, and evaluate provider quality.
Compute MarketplaceGPU BenchmarkWorkload SchedulerProvider Reputation
- Compute Marketplace: A compute marketplace connects workload demand with available CPUs, GPUs, or specialized hardware supplied by different providers.
- GPU Benchmark: A GPU benchmark is a standardized test or metric used to compare graphics processor performance for training, inference, rendering, or other workloads.
- Workload Scheduler: A workload scheduler decides where and when compute jobs should run based on resource availability, requirements, cost, and priority.
- Provider Reputation: Provider reputation is a score or record of past behavior used to estimate whether an infrastructure supplier is reliable, responsive, and accurate.