AI Concepts
AI concepts, explained simply
A free visual library of 133 AI concepts — one infographic and a short, plain-language read each. Understand any idea in the AI stack in a single sitting.
AI Foundations
What is AI (Artificial Intelligence)?
How AI works and its types — narrow, general, super.
What is Machine Learning?
Supervised, unsupervised and reinforcement learning.
What is Deep Learning?
Multi-layer neural networks that learn features.
What is a Neural Network?
Layers of connected neurons that learn patterns.
What is a Transformer?
The attention-based architecture behind modern AI.
Inference
Turning a trained model into predictions.
Batch Inference
Process many inputs at once for efficiency.
Language Models
Large Language Model (LLM) — The Basics
Massive data → training → understanding → generation.
What is a Large Language Model (LLM)?
Inside an LLM — tokenization, embeddings, transformer.
GPT (Generative Pre-trained Transformer)
How GPT predicts text, one token at a time.
GPT-4
OpenAI's multimodal model — text and images.
ChatGPT
From your question to the answer, step by step.
Claude
Anthropic's helpful, harmless and honest AI assistant.
PaLM / Bard
Google's conversational AI powered by PaLM.
Llama 2
Meta's open-source large language model.
What is Generative AI?
Models that learn patterns and create new content.
Image & Generative Models
GAN (Generative Adversarial Network)
Generator vs. discriminator, competing to improve.
VAE (Variational Autoencoder)
Encode to a latent space, sample, then decode.
Diffusion Model
Add noise, then learn to reverse it into new data.
DALL·E
OpenAI's model that creates images from text.
Stable Diffusion
Open-source text-to-image diffusion model.
Tokens & Embeddings
Token
The basic unit of text AI models read and generate.
Tokenization
Breaking text into tokens the model can read.
Tokenizer
Breaks text into tokens the model can understand.
Token Embeddings
Turning tokens into meaningful vectors.
Embedding
Turning words and data into numbers AI understands.
Context Window
How much text an AI can consider at once.
Cosine Similarity
Measuring similarity by the angle between vectors.
Prompting
Prompt
The input you give an AI — and how to write better ones.
Prompt Engineering
The art of giving AI the right instructions.
System Prompt
Sets the AI's behavior, personality and rules.
Few-Shot Prompting
Give a few examples so the AI learns the pattern.
Zero-Shot Prompting
Ask the AI to do a task with no examples.
Chain-of-Thought Prompting
Make the AI reason step by step.
RAG & Knowledge
RAG (Retrieval-Augmented Generation)
Ground LLM answers in external, up-to-date knowledge.
Retrieval
Finding relevant information from a knowledge source.
Knowledge Base
A centralized store an AI uses to answer questions.
Vector Database
Stores and searches embeddings by similarity.
Hallucination
When AI confidently makes things up.
Agents, Tools & Protocols
AI Agent
An autonomous system that perceives, reasons and acts.
LangChain
A framework for building LLM-powered applications.
MCP (Model Context Protocol)
The universal adapter connecting AI to tools and data.
WebMCP (Web Model Context Protocol)
MCP for the web — AI in the browser using online tools.
Fine Tuning
Adapt a pre-trained model to your specific data and task.
Fine-Tuning (Deep Dive)
The full fine-tuning workflow and hyperparameters.
Loop Engineering
Feedback loops that make AI outputs improve over time.
API
How software systems talk and exchange data.
GitHub Copilot
Your AI pair programmer inside the IDE.
Production AI Stack (2026)
1 · The Big Picture
The complete end-to-end production AI stack (2026).
2 · User Request Lifecycle
How one request flows through the whole system.
3 · LangChain Ecosystem
How LangChain, LangGraph, LangServe and LangSmith fit together.
4 · LangGraph Orchestration
Stateful, multi-step agent workflows.
5 · AWS Bedrock Integration
How LangChain talks to models through Amazon Bedrock.
6 · Amazon AgentCore
The enterprise runtime for LangChain/LangGraph agents.
7 · Memory & State Management
The four memory layers that make agents reliable.
8 · Tools & MCP Integration
Give agents real actions and data via MCP.
9 · Observability & Monitoring
Tracing, metrics and logs across the agent stack.
10 · Identity, Policy & Security
Fine-grained, context-aware authorization with Cedar.
11 · Enterprise RAG Pipeline
From question to grounded, cited answer.
11 · Enterprise RAG Pipeline (v2)
Alternate layout of the enterprise RAG pipeline.
12 · Chunking Strategies
Chunk quality decides retrieval quality.
13 · Retrieval Pipeline
From user query to the most relevant chunks.
14 · Prompt Engineering Architecture
Context + instruction + memory = better answers.
15 · Streaming Architecture
Real-time streaming for better UX and lower latency.
16 · Multi-Agent Patterns
Supervisor, parallel, sequential and debate patterns.
17 · FastAPI Architecture
The API gateway for your AI agent platform.
18 · Bedrock AgentCore Architecture
Build, run and operate production agents at scale.
19 · LangChain + AgentCore Integration
Developer experience plus enterprise-grade runtime.
20 · Enterprise RAG + Agent System
Production-ready RAG plus agents with LangChain + AgentCore.
21 · LangGraph + AgentCore Multi-Agent System
Intelligent agent teams at enterprise scale.
22 · Enterprise AI Agent Platform Blueprint
Build, run, observe, secure and scale.
23 · AI Agent Platform (LangChain + AgentCore)
Design, build, run, observe, secure, scale.
24 · AI Agent Platform Architecture
End-to-end enterprise agent platform architecture.
25 · Complete Enterprise AI Architecture
The full production-ready reference architecture.
26 · Writing Better Tools for AI Agents
What makes a tool reliable and production-ready.
27 · Agent Skills
Composable capabilities that make agents modular.
28 · Single vs Multi-Agent Systems
Choose the simplest architecture that solves the problem.
29 · Agent Design Patterns
Sequential, parallel, routing, supervisor, collaborative.
30 · Choosing the Right Agent Architecture
A decision framework for pattern, stack and services.
31 · Agent Failure Modes
Common failures and how to recover automatically.
32 · Context Engineering
The right information, in the right form, at the right time.
33 · Production Checklist
Go from POC to production to scale, safely.
Agent Brain Deep-Dive
Section 1 · Inside the Agent Brain
How AI agents break down complex problems.
Section 2 · Reasoning Patterns
ReAct, Plan-Execute, Reflection, Tree-of-Thoughts, Self-Consistency.
Section 3 · Planning Graphs Instead of Prompts
Stateful graphs that plan, decide and act.
Section 4 · Memory in AI Agents
Working, long-term, semantic and episodic memory.
Section 5 · Reasoning Engine
Think, observe, reflect and refine until confident.
Section 1.5 · How Agents Decide What To Do Next
The observe → decide → act → learn loop.
RAG & Retrieval (Enterprise)
Where Hallucinations Come From
Failures across the query, retrieval and model layers.
RAG Retrieval Optimization Layer
Retrieve the right context with high recall and relevance.
RAG: Retrieval Re-ranking
Re-rank retrieved docs to keep only the most relevant.
RAG Optimization 1 · Hybrid Search
Keyword + semantic search for higher recall.
RAG Optimization 2 · Re-ranking
Find the most relevant docs before the LLM.
RAG Optimization 3 · Context Pruning
Remove noise; keep only what matters.
RAG Optimization 4 · Semantic Cache
Stop asking the same question twice.
Generation Guardrails
Make the LLM stick to the truth and cite sources.
Enterprise Hallucination Prevention Pipeline
Verify every answer with layered checks.
Optimized Enterprise RAG — Final Architecture
All optimizations working together.
15 Techniques for a Fast, Cheap RAG System
The complete enterprise RAG optimization checklist.
AWS Bedrock AgentCore
Insurance Support Agent on AgentCore
Reference architecture on AWS Bedrock AgentCore.
Why Strands Framework for AgentCore
Strands vs building agents yourself.
Deploy to AgentCore Runtime
Packaging and shipping an agent to AgentCore.
What Happens When a User Runs a Query
End-to-end AgentCore flow and how charges apply.
Lifecycle of an Enterprise AI Agent
Development to production on AgentCore.
What AgentCore Manages For You
The ~15 systems AgentCore replaces.
Inside Runtime (Internal Brain)
Step-by-step execution inside AgentCore Runtime.
The Runtime Decision Engine
How the agent decides what to use and why.
Enterprise Explainers
Enterprise AI Agent Architecture
What an enterprise AI agent actually consists of.
Where Do Tokens Come From?
The anatomy of an LLM request and why bills explode.
Agent Execution Lifecycle
What happens after a user sends a prompt.
One Interface, Every Model
Orchestrate any LLM through one unified integration layer.
Claude Interpretability (J-Space)
Inside Claude's Brain
What happens before Claude speaks — the J-Space workspace.
What Is J-Space?
A silent internal workspace inside Claude's neural network.
Global Workspace Theory (GWT)
The theory that inspired J-Space.
How Claude Thinks Silently
Claude thinks first, reasons, then speaks.
How Anthropic Reads Claude's Thoughts
The J-Lens (Jacobian Lens) tool explained.
J-Space vs Chain of Thought
Two layers of thinking: hidden vs visible.
Reading Claude's Thoughts — The Pipeline
Neural activity → J-Lens → words/concepts.
Why This Changes AI Engineering Forever
From blind trust to informed engineering.
Other Key Results From the Paper
Additional insights on J-Space and interpretability.
Global Workspace Theory (in Humans)
A clean single-panel view of GWT.
Top 20 Cheat Sheets
Top 20 Bottlenecks in LLMs
From prompt to production — fixing what breaks LLMs.
Top 20 AI Agent Bottlenecks
How enterprises solve agent bottlenecks.
Top 20 Token & Cost Bottlenecks
Optimize token usage and control cost.
Top 20 Python Libraries for AI Engineers
Essential libraries and the problems they solve.
Top 20 RAG System Bottlenecks
Fixing what breaks RAG in production.
YC AI Startup Bets
The Infrastructure Layer (Picks & Shovels)
YC bet: the tools and infra every AI company needs.
AI-Native Firms (Be The Firm)
YC bet: startups that use AI to become the whole company.
AI That Acts in the Real World
YC bet: agents, robots and systems in the physical world.
Intelligence That Costs Pennies
YC bet: making intelligence radically cheaper.
