From a single token,
to a working agent.
A cinematic, five-chapter course on how generative AI actually works — from tokenization and attention to training, multimodality, RAG and agents. Every idea is animated, narrated and built from first principles.
One payment · lifetime access · no subscription
Foundations — What Generative AI Actually Is
Prediction, not magic: what these models are, how they differ from every AI that came before, and why 2022 changed.
Inside the Model — Tokens, Meaning, and Attention
Text becomes tokens, tokens become geometry, attention decides what matters — and the Transformer falls out.
How Models Learn — and Why They Lie
Pretraining, fine-tuning and RLHF; the sampling dials you actually turn; and the real mechanism behind hallucination.
Beyond Text — Images, Sound, and Everything Else
Diffusion from pure noise, multimodal models that see and hear, and generation across code, speech, music and video.
Building Real Systems — Prompting, RAG, Agents, Production
Prompting that survives contact with reality, retrieval as memory, tool-calling agents, and the cost/latency/eval/safety bill that comes due in production.
What you'll actually understand
- What a generative model actually is — prediction, not magic — and why 2022 was the turn
- Tokens, embeddings and the geometry of meaning: how text becomes something a machine can reason over
- Attention, from first principles, and how the Transformer falls out of it
- Pretraining, fine-tuning and RLHF — the three stages, and what each one really buys
- Temperature, top-p and top-k: the sampling dials you actually turn, shown on a live distribution
- The real mechanism behind hallucination — and why 'it lies' is the wrong mental model
- Diffusion: how an image resolves out of pure noise, step by step
- Multimodality — one shared space of meaning for text, images and sound
- Prompting that survives contact with reality, and RAG as memory
- Tool-calling agents, and the cost / latency / eval / safety bill that comes due in production