Getting an answer is easy. Getting it to stay up is the job.

Both renew every month and you cancel either one yourself, any time. A class you finish stays yours — the recordings, the exercises and the written curriculum for it remain in your login after the subscription stops. Classes can also be taken one at a time, at their own prices — this path runs from $100 to $500 a class.

Tagged by philosophy, this path is 32 make · 22 think · 14 own · 3 secure. A capability can carry more than one tag, so those add up to more than 41.

1. 01

   Using Large Language Models

   • generalise a reasoning prompt pattern
   • produce a reusable system prompt
   • operate a model as a first pass editor
   • +2 more

   Prove itThe pattern applied to three tasks with paired outputs against a plain prompt, plus one task where it made the output worse.

   • 6 clips · 10.7 h
   • 55 exercises

   Adds 5 — after this class you can do 5 of the 41 things on this path
2. 02

   Control AI Spending

   • segment text into tokens and cost it
   • run an open model locally

   Prove itA token count for a real document, with the resulting cost and the fraction of the model's context window it occupies.

   • 2 clips · 3.1 h
   • 12 exercises
   • Syllabus
   • next session 12 Sep
   • $100 on its own

   Adds 2 — after this class you can do 7 of the 41 things on this path
3. 03

   AI Alignment

   • verify a rubric against independent graders
   • measure whether a fine tune changed behaviour
   • falsify a prompt with a benchmark
   • +5 more

   Prove itTwo rubric versions, the adversarial inputs that forced the revision, and inter-grader agreement measured before and after.

   • 1 exercise
   • Syllabus
   • next session 25 Sep
   • $300 on its own

   Adds 8 — after this class you can do 15 of the 41 things on this path
4. 04

   Context Engineering

   • design a recursive summariser for oversized documents
   • segment text into tokens and cost it
   • configure inference hyperparameters

   Prove itA run over a document several times the context window, a fact list written beforehand, and the recall score of the output against it.

   • 27 clips · 35.7 h
   • 329 exercises
   • Guided app
   • next session 16 Sep
   • $400 on its own

   Adds 1 — after this class you can do 16 of the 41 things on this path
5. 05

   Claude Model Context Protocol

   • generalise a tool interface for model use
   • execute a tool call round trip by hand
   • locate a tool in an open tool hub

   Prove itThe tool definition, a transcript of a model calling it correctly from the description alone, and a transcript of it recovering from a bad call.

   • 3 clips · 5.2 h
   • Syllabus

   Adds 3 — after this class you can do 19 of the 41 things on this path
6. 06

   Intro to Agents

   • construct an agent that uses tools
   • constrain model output to a schema
   • produce a reusable system prompt
   • +2 more

   Prove itA run log showing tool selection the author did not script, a benchmark over held-out tasks, and a case where it correctly declined to continue.

   • 42 clips · 56.7 h
   • 32 exercises
   • Syllabus
   • next session 15 Sep
   • $350 on its own

   Adds 1 — after this class you can do 20 of the 41 things on this path
7. 07

   Prompt Engineering

   • synthesise a metalanguage for a problem domain
   • falsify a prompt with a benchmark
   • parameterise a prompt template
   • +3 more

   Prove itThe notation's grammar, a work produced in it that exceeds the context window, and a consistency check a reader runs to catch contradictions.

   • 58 clips · 63.5 h
   • 142 exercises
   • Syllabus

   Adds 1 — after this class you can do 21 of the 41 things on this path
8. 08

   RAG & Memory

   • synthesise an agent that carries notes across its own runs
   • construct a retrieval system
   • transform a corpus into an embedded index

   Prove itPaired runs of the same multi-step task with the note store kept and cleared, where the cleared run repeats a step the noted run skips, plus the note text the agent wrote and the later turn that cites it.

   Adds 3 — after this class you can do 24 of the 41 things on this path
9. 09

   Advanced Retrieval Augmented Generation

   • construct a retrieval system
   • transform a corpus into an embedded index

   Prove itA running system, a benchmark of retrieval quality over held-out questions, and citations traceable to source documents.

   • 5 clips · 6.8 h
   • 7 exercises

   Goes deeper — after this class you can do 24 of the 41 things on this path
10. 10

    Production Agent Engineering

    • synthesise a self extending agent behind a review gate
    • select among agent architectures
    • reconcile agents from different toolchains into one run
    • +3 more

    Prove itA diff the agent authored to its own prompt or toolset, the benchmark run that accepted it, and a logged rejection with the reason recorded.

    • 19 clips · 22.3 h
    • next session 17 Sep
    • $200 on its own

    Adds 6 — after this class you can do 30 of the 41 things on this path
11. 11

    Agentic SDLC

    • synthesise an agent that carries notes across its own runs
    • synthesise an accountability regime for an unattended agent
    • design a review gate that catches what the doer cannot see
    • +4 more

    Prove itPaired runs of the same multi-step task with the note store kept and cleared, where the cleared run repeats a step the noted run skips, plus the note text the agent wrote and the later turn that cites it.

    • 17 clips · 26.5 h
    • 46 exercises
    • Companion
    • next session 7 Sep
    • $500 on its own

    Adds 6 — after this class you can do 36 of the 41 things on this path
12. 12

    Agentic AI Security: Securing What You Build

    • verify model written code against its specification
    • characterise an applications injection surface
    • produce a list of where untrusted input enters a system

    Prove itThe diff, a review note per changed hunk, and a test run that is red on the parent commit and green on the child.

    • 1 exercise
    • Reference

    Goes deeper — after this class you can do 36 of the 41 things on this path
13. 13

    Keep It Running

    • verify you would know your system broke before a user tells you
    • characterise what one user costs you to serve
    • characterise what breaks when you cut the connection
    • +2 more

    Prove itA real failure you learned about from your own alerting with the timestamps to prove it, and one you learned about from a user, with what you added so that one cannot happen the same way twice.

    Adds 5 — after this class you can do 41 of the 41 things on this path