IDInternals Decoded
All series
Series · 10 of 10 parts · Beginner

AI, From Zero

The ground floor: what a language model actually is, why it works, and why it fails, explained for someone starting from nothing.

What you’ll understand

By the end you can follow any AI conversation, read model announcements critically, and know exactly what is and isn't magic.

Episodes
01What an LLM Actually Is (and Why Predicting Words Works)It predicts the next word. That one trick, at scale, is doing everything you've seen.12 min02Tokens: The Currency of AIWhy models see 'strawberry' as three pieces, why pricing is per token, and why that matters to you.12 min03Attention, Gently: How Models Decide What MattersThe mechanism behind the magic, explained with zero math.14 min04Embeddings: Meaning as NumbersHow machines measure similarity, and why it powers search, recommendations, and memory.12 min05Training vs Inference: Why Building Costs Millions and Asking Costs CentsTwo completely different phases of a model's life, constantly confused.10 min06Context Windows: The Model's Working MemoryWhat fits, what falls out, and why long conversations get weird.9 min07Why Models Make Things UpHallucination isn't lying. It's the same mechanism that makes models work at all.10 min08Open vs Closed Models: What 'Open Source AI' Really MeansWeights, data, licenses: the three things people conflate when they say open source.10 min09Fine-Tuning, Explained: Teaching an Old Model New TricksWhen it helps, when RAG beats it, and why most teams never need it.14 min10How to Actually Choose a ModelBenchmarks, price, latency, context: a working decision framework instead of hype.11 min
The Newsletter

Keep up with AI. One email a week.

One thoughtful email each week. Unsubscribe whenever you like.