A daily crypto, finance, and AI vocabulary puzzle covering identity primitives, central-bank tools, model training, and distributed compute.
Onchain Identity
These words describe identifiers and credentials that can connect accounts with verifiable claims.
Wallet AddressDecentralized IdentifierAttestationSoulbound Token
- Wallet Address: A wallet address is a public identifier used to receive assets or interact with applications on a blockchain.
- Decentralized Identifier: A decentralized identifier, or DID, is an identifier designed to be created and controlled without relying on one central registry.
- Attestation: An attestation is a verifiable statement that an issuer makes about an account, credential, or event.
- Soulbound Token: A soulbound token is a non-transferable token used to represent a credential, membership, or reputation signal.
Monetary Policy
These words explain policy rates, balance-sheet actions, and inflation-adjusted returns.
Policy RateOpen Market OperationsQuantitative TighteningReal Yield
- Policy Rate: A policy rate is a benchmark interest rate a central bank sets to influence borrowing costs and economic activity.
- Open Market Operations: Open market operations are purchases or sales of securities used by a central bank to manage liquidity and short-term rates.
- Quantitative Tightening: Quantitative tightening reduces a central bank's asset holdings, which can withdraw liquidity from the financial system.
- Real Yield: Real yield measures an investment's return after accounting for expected or realized inflation.
Training Pipelines
These words describe training data, optimization steps, and saved model states.
Training CorpusMini-BatchGradient DescentCheckpoint
- Training Corpus: A training corpus is the collection of text, images, audio, or other examples used to train a model.
- Mini-Batch: A mini-batch is a subset of training examples processed together during one optimization step.
- Gradient Descent: Gradient descent updates model parameters in a direction intended to reduce the training loss.
- Checkpoint: A checkpoint stores model parameters and related training state so work can be resumed or evaluated later.
Decentralized Compute
These words describe marketplaces, providers, routing, and verification for distributed AI compute.
- Compute Marketplace: A compute marketplace connects buyers of processing capacity with infrastructure providers that have available resources.
- Compute Provider: A compute provider supplies GPUs, CPUs, storage, or networking resources to run workloads.
- Inference Network: An inference network routes model requests to distributed workers that produce outputs.
- Verifiable Compute: Verifiable compute uses proofs or validation methods to help others check that a computation was performed correctly.