Large Language Models
A public walking-audio course from LLM 101 to how modern models appear to reason — tokens, training, context, tools, and chain-of-thought style thinking. Concepts and judgement, not API tutorials.
-
Lesson 1 ~10 min
What an LLM is — 101
Explain what a large language model is in plain language and how it differs from a search engine or a rules chatbot.
-
Lesson 2 ~10 min
Tokens, vocabulary, and why wording matters
Describe tokens and why the same idea in different words can change model behaviour and cost.
-
Lesson 3 ~10 min
Training versus inference — how the model got its skills
Distinguish pre-training, fine-tuning / alignment, and inference in everyday terms.
-
Lesson 4 ~10 min
Context windows, memory, and multi-turn chat
Explain what the model can 'see' in a conversation and why it seems to forget.
-
Lesson 5 ~10 min
Hallucinations, grounding, and healthy scepticism
Explain why LLMs invent plausible falsehoods and how to reduce reliance on ungrounded answers.
-
Lesson 6 ~10 min
Prompting that works — instructions, examples, and structure
Use practical prompting patterns: clear role, constraints, examples, and output shape.
-
Lesson 7 ~10 min
Tools, agents, and when the model should not guess
Explain tool use and agent-style loops at concept level — and their failure modes.
-
Lesson 8 ~10 min
How an LLM seems to reason — chain of thought and 'thinking'
Explain what people mean by LLM reasoning, chain-of-thought, and extended thinking modes — without mysticism.
-
Lesson 9 ~10 min
Limits, safety, and choosing the right job for an LLM
Judge when an LLM is the right tool, what it is bad at, and how to use it responsibly at work.