Home/Concepts

// the encyclopedia

The Encyclopedia

25 AI concepts. 4 sections each. Organized by category. Written for practitioners who need to know what's under the hood.

// foundations

Foundations

perplexity
How language models measure uncertainty and why it reshapes search ranking
burstiness
The statistical property that separates topical depth from keyword stuffing
tokenization
How models break language into computable units and why token economics matter
context window
The finite memory that determines what a model can see and process at once

// architecture

Architecture

transformer
The architecture behind every modern LLM and search model
attention
The mechanism that lets models decide what matters in context
hidden state
The internal representations that drive model reasoning and drift
embeddings
Vector representations that power semantic search and similarity

// generation

Generation

temperature
The knob that controls randomness vs. determinism in model output
inference
The runtime process where trained models produce actual outputs
chain of thought
Prompting models to show their work for better reasoning output
prompt engineering
The craft of structuring inputs to get reliable, useful model outputs

// learning

Learning

fine-tuning
Adapting pre-trained models to specific domains and tasks
zero-shot
When models perform tasks they were never explicitly trained on
few-shot
Teaching models new patterns with just a handful of examples
alignment
Training models to follow human intent and behave predictably

// applications

Applications

RAG
Retrieval-augmented generation: how AI answers get grounded in real sources
semantic search
Search by meaning, not keywords. The paradigm shift that changed everything.
knowledge graph
Structured entity relationships that power contextual understanding
entity recognition
How models identify and classify named entities in unstructured text
grounding
Tethering model outputs to verifiable external sources
hallucination
When models generate confident nonsense, and what practitioners need to know

// frontier

Frontier

agents
Autonomous AI systems that plan, execute, and adapt across tool chains
tool use
Models that call APIs, run code, and interact with external systems
distributed authority
How trust and credibility signals propagate across decentralized networks

// go deeper

Concepts Are Just the Start

Discuss implementations, ask questions, and connect with practitioners in the free community.

Join the Skool Group →