README.md

AIFromZero

AI literacy concept by concept - tokens, embeddings, RAG, agents, explained simply.

🌐 Live (all days): https://dev48v.infy.uk/aifromzero.php

Every lesson is a single, self-contained HTML file with three tabs — 👁 LOOK (interactive demo), 🧠 UNDERSTAND (the concept, step by step), 🔨 BUILD (do it yourself). No build tools, no install — open any file in a browser.

Days

• AIFromZero · Day 1 — AI vs ML vs DL — day1-ai-ml-dl.html
• AIFromZero · Day 2 — Tokens — day2-tokens.html
• AIFromZero · Day 3 — Embeddings — day3-embeddings.html
• AIFromZero · Day 4 — What a Neural Net Does — day4-neural-net-intuition.html
• AIFromZero · Day 5 — How a Neural Net Learns — day5-training.html
• AIFromZero · Day 6 — What an LLM Really Does — day6-next-token.html
• AIFromZero · Day 7 — Why LLMs Hallucinate — day7-hallucinate.html
• AIFromZero · Day 8 — The Context Window — day8-context-window.html
• AIFromZero · Day 9 — Temperature & Sampling — day9-temperature.html
• AIFromZero · Day 10 — Fine-tuning vs RAG — day10-finetune-vs-rag.html
• AIFromZero · Day 11 — AI Agents — day11-agents.html
• AIFromZero · Day 12 — Tokenization — day12-tokenization.html
• AIFromZero · Day 13 — Embeddings — day13-embeddings.html
• AIFromZero · Day 14 — Vector Databases — day14-vector-databases.html
• AIFromZero · Day 15 — RAG — day15-rag.html
• AIFromZero · Day 16 — Multimodal AI — day16-multimodal.html
• AIFromZero · Day 17 — How LLMs Are Trained — day17-how-llms-trained.html
• AIFromZero · Day 18 — Prompt Injection & Jailbreaks — day18-prompt-injection.html
• AIFromZero · Day 19 — Mixture of Experts — day19-mixture-of-experts.html
• AIFromZero · Day 20 — Quantization — day20-quantization.html
• AIFromZero · Day 21 — Speculative Decoding — day21-speculative-decoding.html
• AIFromZero · Day 22 — The KV-Cache — day22-kv-cache.html
• AIFromZero · Day 23 — Knowledge Distillation — day23-distillation.html
• AIFromZero · Day 24 — LoRA Fine-Tuning — day24-lora.html
• AIFromZero · Day 25 — Flash Attention — day25-flash-attention.html
• AIFromZero · Day 26 — Grounding & Citations — day26-grounding.html
• AIFromZero · Day 27 — Reasoning Models — day27-reasoning-models.html
• AIFromZero · Day 28 — Function Calling & Tool Use — day28-function-calling.html
• AIFromZero · Day 29 — Guardrails & Safety — day29-guardrails.html
• AIFromZero · Day 30 — Agentic Memory — day30-agentic-memory.html
• AIFromZero · Day 31 — Prompt Caching — day31-prompt-caching.html
• AIFromZero · Day 32 — Perplexity — day32-perplexity.html
• AIFromZero · Day 33 — Beam Search vs Sampling — day33-beam-vs-sampling.html
• AIFromZero · Day 34 — RLHF — day34-rlhf.html
• AIFromZero · Day 35 — DPO — day35-dpo.html
• AIFromZero · Day 36 — Constitutional AI & RLAIF — day36-constitutional-ai.html
• AIFromZero · Day 37 — Long-Context Tricks — day37-long-context-tricks.html
• AIFromZero · Day 38 — Evals & LLM-as-a-Judge — day38-evals-llm-judge.html
• AIFromZero · Day 39 — Synthetic Data & the Self-Improvement Loop — day39-synthetic-data.html
• AIFromZero · Day 40 — Scaling Laws — day40-scaling-laws.html
• AIFromZero · Day 41 — Emergent Abilities — day41-emergent-abilities.html
• AIFromZero · Day 42 — World Models — day42-world-models.html
• AIFromZero · Day 43 — State-Space Models / Mamba — day43-state-space-models.html
• AIFromZero · Day 44 — Model Collapse — day44-model-collapse.html
• AIFromZero · Day 45 — Diffusion Language Models — day45-diffusion-language-models.html
• AIFromZero · Day 51 — Model Context Protocol (MCP) — day51-model-context-protocol.html
• AIFromZero · Day 52 — GraphRAG — day52-graphrag.html
• AIFromZero · Day 53 — Test-Time Compute — day53-test-time-compute.html
• AIFromZero · Day 54 — Lost in the Middle — day54-lost-in-the-middle.html
• AIFromZero · Day 55 — Semantic Chunking — day55-semantic-chunking.html
• AIFromZero · Day 56 — Rerankers — day56-rerankers.html
• AIFromZero · Day 57 — Hybrid Search — day57-hybrid-search.html
• AIFromZero · Day 58 — Prompt Compression — day58-prompt-compression.html
• AIFromZero · Day 59 — Model Merging — day59-model-merging.html
• AIFromZero · Day 60 — Continuous Batching — day60-continuous-batching.html
• AIFromZero · Day 61 — Differential Privacy — day61-differential-privacy.html
• AIFromZero · Day 62 — Mechanistic Interpretability — day62-mechanistic-interpretability.html
• AIFromZero · Day 63 — Context extension: PI, NTK-aware, YaRN — day63-context-extension.html
• AIFromZero · Day 64 — Attention sinks and StreamingLLM — day64-attention-sinks.html
• AIFromZero · Day 65 — Chain-of-thought faithfulness — day65-cot-faithfulness.html
• AIFromZero · Day 66 — Model cascades — day66-model-cascades.html
• AIFromZero · Day 67 — Embedding model migration — day67-embedding-migration.html
• AIFromZero · Day 68 — Streaming and partial-output commitment — day68-streaming-commitment.html
• AIFromZero · Day 69 — Agent step reliability — day69-agent-step-reliability.html
• AIFromZero · Day 70 — RAG index freshness — day70-rag-index-freshness.html
• AIFromZero · Day 71 — Multi-tenant fairness — day71-multi-tenant-fairness.html
• AIFromZero · Day 72 — Where a preference eval’s variance lives — day72-preference-eval-variance.html
• AIFromZero · Day 73 — Context-Window Packing — day73-context-window-packing.html
• AIFromZero · Day 74 — Benchmark contamination — day74-benchmark-contamination.html
• AIFromZero · Day 75 — Inference non-determinism — day75-inference-nondeterminism.html
• AIFromZero · Day 76 — Human oversight — day76-human-oversight.html

---

Part of Zero to Hero — build something real from scratch, every day: https://dev48v.infy.uk