From neurons and tokens to autonomous agents — a no-fluff, no-jargon guide to modern AI, large language models, and the frameworks shaping the next decade.
AI is moving fast and most explanations are either dry textbooks or marketing fluff. This is neither. Just clear ideas, real diagrams, and the stuff you actually need to know.
Five rabbit holes. Read in order, or jump wherever your brain wants to go.
What AI actually is. Neural networks, training, the difference between AI / ML / DL — explained without the math wall.
Read it →Large Language Models like GPT, Claude, and Gemini. Tokens, embeddings, attention, context windows. With an interactive tokenizer.
Read it →Models that think, plan, use tools, and act on your behalf. The ReAct loop, memory, multi-agent systems.
Read it →LangChain, LlamaIndex, Claude Agent SDK, CrewAI, AutoGen, and more. What each is good at — and when to skip it.
Read it →Every weird acronym defined in plain English. RAG, MoE, RLHF, MCP — searchable, no scrolling required.
Read it →This site is hand-coded, no trackers, no ads, no fluff. Just HTML, CSS, and a tiny bit of JavaScript.
About →This site has interactive bits. Don't just read — click around.
Type any sentence and watch it get sliced into tokens — the way an LLM actually sees text.
Try it →Watch the Think → Act → Observe loop animate through each step. The pattern behind every modern AI agent.
See it →An animated diagram of neurons "firing" as they pass signals through a tiny network.
Watch it →The 70-year version. From Turing's question to autonomous agents.
Alan Turing asks "Can machines think?" and proposes the imitation game — the original AI benchmark.
The Dartmouth workshop gives the field its name. Optimism is sky-high. (It will not last.)
Rumelhart, Hinton & Williams revive neural networks with a way to actually train them efficiently.
A deep convolutional net crushes the ImageNet challenge. The deep learning era begins for real.
Google publishes the Transformer paper. Every modern LLM — GPT, Claude, Gemini, Llama — descends from it.
OpenAI ships a chatbot. It hits 100M users in two months. AI becomes a household word.
Tool-using, multi-step AI agents go mainstream. Claude, GPT, and open models all gain agentic abilities.
You're learning AI in the most exciting moment in its history. Pick a topic above and dive in.
Artificial Intelligence is the field of building systems that can do things normally requiring human thought — recognizing images, understanding language, making decisions, learning from experience.
Modern AI is powered by machine learning: instead of writing rules by hand, we let the system learn patterns from huge amounts of data.
A Large Language Model is a neural network trained on a massive pile of text. It learns to predict the next word — and from that simple objective, it picks up grammar, facts, reasoning patterns, and more.
Examples: GPT-4, Claude, Gemini, Llama, Mistral.
An AI agent is an LLM hooked up to tools and a loop. It can decide what to do, take an action (like searching the web, running code, or sending an email), see the result, and decide what to do next.
Think of it as the LLM with hands.
For the first time in history, we have machines that can read, write, reason, and take actions in the digital world. The difference between someone who understands this and someone who doesn't will get bigger every year.
Better to learn it loud than miss it quiet.