One of five philosophies · the verb
Make is not a folder classes sit in. It is a label on a capability — and 83 of the school's capabilities carry it, from run a program someone else wrote and read its error to originate a market from a need no one had named. The unusual part is that the school can tell you exactly how far apart those two are, name every rung in between, and hand you the proof each one takes.
The recurring hour
Not a lecture. Neek runs it as coworking: you bring the thing you are building, and the hour supplies tech support, a second pair of eyes, and enough company to keep the energy up on the days the build is boring.
Mon-Thu · 9:00 Pacific · led by Neek
This one hour carries two philosophies on purpose.
Build the thing. Ship software, make art, put hardware together.
Choose your tools, and know who can take each one away.
Making a thing and owning your tools are the same motion — you cannot really build on a stack somebody else can withdraw on a Tuesday — so the school put them in one room rather than two.
The ladder
Every capability sits on a computed rung. Pick one and you get the real capabilities on it, the proof each one takes before anybody says you have it, and which classes carry it. Filter by domain to see where the ladder is thick and where it is thin.
Say it plainly: of the 83 capabilities on this ladder, 62 are carried by a class you can take today and 21 are on the map but not yet on the timetable. Those 21 are marked as such on their cards. They are what the school is building next, not what it is selling now.
Bottom rung is where you start. Arrow keys climb once a rung has focus.
Rung 1 of 8 3 here 0 below 80 above
You originate the category. The proof is other people: buyers using your words, or somebody outside your organisation building to the thing you made.
originate
What proves itBuyers who describe the need in your terms rather than their own prior ones, and at least one other party organising delivery in the category.
Named in the ontology. No class carries it yet.
What proves itA cutover record with cost per outcome before and after, the compatibility surface that kept callers unchanged, and no interruption in the traces.
Named in the ontology. No class carries it yet.
What proves itA schema in production, the invalid state someone tried to enter and could not, and the working agreement that was changed rather than accommodated by a nullable column.
Named in the ontology. No class carries it yet.
What proves itThe paradigm running in at least two markets with materially different constraints, and one party outside your organisation building to it.
Named in the ontology. No class carries it yet.
What proves itA written operating model in use by at least two people, with a recorded instance of it catching a failure that the previous arrangement missed.
Named in the ontology. No class carries it yet.
Nothing in that domain sits on this rung. Pick another rung, or clear the filter.
More than one system, reconciled — including systems that disagree, agents from toolchains that were never meant to meet, and rules about who is allowed to know what.
arbitrategovernreconcilesynthesise
What proves itA written information policy, a transcript where a mediator withholds, and paired runs with and without the hierarchy scored on the same goal.
Named in the ontology. No class carries it yet.
What proves itOne transcript containing agents from at least two frameworks and one self-hosted backend, plus the script chaining its output into a second task.
Taught in Production Agent Engineering
What proves itA live surface serving at least three named audiences, the routing rule written down, per-audience outcomes measured apart, and one contradiction found and resolved rather than averaged away.
Named in the ontology. No class carries it yet.
What proves itA written comparison of at least two architectures on the same workload, with the measurements that decided it and the conditions that would reverse the decision.
Taught in Production Agent Engineering
What proves itThe notation's grammar, a work produced in it that exceeds the context window, and a consistency check a reader runs to catch contradictions.
Taught in Prompt Engineering
What proves 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.
Taught in Production Agent Engineering
What proves itA week of unattended runs with a per-action trace, and a replay of one run reconstructed from the trace alone.
Taught in Agentic SDLC
What proves 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.
Taught in Agentic SDLC, RAG & Memory
Nothing in that domain sits on this rung. Pick another rung, or clear the filter.
Parts you chose, assembled into one system, measured as a whole against held-out work rather than against the case you built it for.
constructdesign
What proves itA transcript where the speaker order differs between two runs of the same task, both terminating, plus the selection rule in source.
Taught in Production Agent Engineering, Critical Thinking and Creativity with AI
What proves itWritten predictions made before running held-out inputs, the hit rate against them, and the same account re-tested after a model version change with what it got wrong the second time.
Named in the ontology. No class carries it yet.
What proves itA running system, a benchmark of retrieval quality over held-out questions, and citations traceable to source documents.
Taught in RAG & Memory, Advanced Retrieval Augmented Generation
What proves 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.
Taught in Intro to Agents
What proves itA shared assistant used by at least three people who did not build it, its rubric-scored answers, and one answer citing the live API it called.
Taught in Custom GPTs
What proves itA run over a document several times the context window, a fact list written beforehand, and the recall score of the output against it.
Taught in Context Engineering
What proves itThe cost curve before and after a change you made for that reason, at two usage levels, and the feature you removed or capped to get it.
Named in the ontology. No class carries it yet.
What proves itA shipped offer, the per-segment outcome measured apart from the rest of the traffic, and the one element you changed when the outcome disagreed.
Taught in Founding Federation: Smallest Sellable Thing
Nothing in that domain sits on this rung. Pick another rung, or clear the filter.
You attack your own work and write down what happened. Every proof at this rung has a red state and a green state, and you have to have seen both.
falsifyjustifymeasureverify
What proves itThe voice written as instructions, a passage produced in it, at least three attempts by other readers to name its influences, and the one device you changed because it gave a source away.
Taught in Critical Thinking and Creativity with AI
What proves itThe guardrail as stated, the inputs that defeat it, the mitigation, and a re-run where the same inputs no longer succeed.
Named in the ontology. No class carries it yet.
What proves itA benchmark file with graded cases, a run showing a non-zero failure rate, and a written account of which failures are the prompt's fault.
Taught in Prompt Engineering, AI Alignment
What proves itA working input that reaches the interpreter, the fix, and a regression test that is red before the fix and green after.
Taught in The Dark Arts (Red Team), AI-Assisted Attacks: What Actually Happened
What proves itA live surface serving at least three named audiences, the routing rule written down, per-audience outcomes counted separately, and at least one located contradiction between two audiences' paths.
Named in the ontology. No class carries it yet.
What proves itThe training set, a held-out set built from different sources, and base-versus-tuned scores on both.
Taught in AI Alignment
What proves itA test that is red on the broken version and green on the fixed one, run in both states.
Taught in Agentic SDLC
What proves itTwo rubric versions, the adversarial inputs that forced the revision, and inter-grader agreement measured before and after.
Taught in AI Alignment, Custom GPTs
What proves itA run log containing at least one rejection and the revision that followed, plus a control run with the critic removed.
Taught in Production Agent Engineering
What proves itThe diff, a review note per changed hunk, and a test run that is red on the parent commit and green on the child.
Taught in Agentic SDLC, Agentic AI Security: Securing What You Build
What proves 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.
Taught in Keep It Running
Nothing in that domain sits on this rung. Pick another rung, or clear the filter.
One thing becomes a kind of thing. The same work now takes inputs whose shape you did not see coming, and the proof is the input that broke it before you handled it.
characteriseclassifygeneraliseparameterise
What proves itThree traversals by deliberately different strategies, the state each ended in, at least one dead end or contradiction you did not know was there, and what you changed because of it.
Taught in The Interactive Fiction Intensive
What proves itA generated passage with the defensible parts marked, each traced to your premise, your voice definition or your own history, and the parts you would not sign named as such.
Taught in Critical Thinking and Creativity with AI
What proves itA scene three times longer after AI expansion with its shape intact, and one decision you refused to delegate, with the reason.
Taught in The Interactive Fiction Intensive
What proves itA script that calls a model and parses the result with no try/except around the parse, plus a test that fails when the constraint is removed.
Taught in AI Alignment, Intro to Agents
What proves itThe representation, its token count against the budget, and a reconstruction scored against the original on questions written first.
Named in the ontology. No class carries it yet.
What proves itA deployed node serving a real group, with the failure each part absorbs named and the fallback stated for each. NOTE: the full DESIGN of a community information system would be order 12, and this class does not reach it — the chain computes to 10 and the capability validator says so. Workshop 5 is a showcase of what you built, not an architecture defended against alternatives. The order-11 rung (verifying the system survives a cut you did not choose) is unwritten.
Taught in Solarpunk Automation
What proves itThe pattern applied to three tasks with paired outputs against a plain prompt, plus one task where it made the output worse.
Taught in Using Large Language Models, Critical Thinking and Creativity with AI
What proves 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.
Taught in Claude Model Context Protocol
What proves itThree runs whose inputs differ in shape rather than only in value, with no edit to the body between them, and the input that broke it before you handled it.
Taught in The Interactive Fiction Intensive
What proves itThe template, three filled instances whose slot values share no words, their outputs, and the slot you had to add after the third one did not fit.
Taught in Prompt Engineering, AI Alignment, Custom GPTs
What proves itThe persona definition and three transcripts on unrelated tasks where a reader picks out the same agent, including one where it refuses.
Taught in Automate Your Email: No Code AI Automation, Critical Thinking and Creativity with AI, The Interactive Fiction Intensive
What proves itOne offering with its price, the floor it clears, and the number of clients per month the arithmetic requires.
Taught in Founding Federation: Price and Ideate (Week 0)
Nothing in that domain sits on this rung. Pick another rung, or clear the filter.
You make the artifact. This is the widest rung in Make and the one an ordinary working week lives on — a script, a schema, an interface, a story, an automation that ran unattended three times.
configureoperateproducetransform
What proves itA power budget — draw, panel, battery capacity — and the measured hours the node stayed up unplugged.
Taught in Solarpunk Automation
What proves itMemory on with one fact it retained across a fresh conversation, one connector enabled with the scopes it was granted written out, and one source you deliberately did not connect with the reason.
Taught in Critical Thinking and Creativity with AI
What proves itOne client script, unedited, producing comparable output against both the hosted provider and your own endpoint, with only the base URL swapped.
Taught in Production Agent Engineering, Hosting Open-Source Models
What proves itThe same prompt run under two settings, with the difference in output described against the property you were targeting.
Taught in AI Alignment, Context Engineering
What proves itA before-and-after pair of prompts with both outputs, and the written criterion the second satisfies and the first does not.
Taught in Using Large Language Models, Critical Thinking and Creativity with AI
What proves itA schema plus a test file containing at least one payload it accepts and two it rejects.
Named in the ontology. No class carries it yet.
What proves itA transcript answering a question findable only in the uploaded set, and a second transcript where it declines a question the set does not cover.
Taught in Custom GPTs
What proves itA redlined draft showing which of the model's edits you took and which you refused, with the standard each decision was made against.
Taught in Using Large Language Models, Critical Thinking and Creativity with AI
What proves itA live automation with a run history of at least three unattended runs, and the records it wrote at the destination.
Taught in Automate Your Email: No Code AI Automation
What proves itA value written at one point and read at a point far enough away that the two are not visible together, two runs differing only in that decision, and the place the difference becomes visible.
Taught in The Interactive Fiction Intensive
What proves itThe service reachable after an unplanned restart, with the boot configuration that made it come back written down.
Named in the ontology. No class carries it yet.
What proves itTwo strands running independently, the condition that decides when each surfaces, and one run where changing that condition changed the order the reader saw them in.
Taught in The Interactive Fiction Intensive
What proves itThe named state set, the code that moves between states, and one transition the naming made you notice that a number would have hidden.
Taught in The Interactive Fiction Intensive
What proves itOne story shown to fail both ways and then fixed with state, with the variable that does the work named.
Taught in The Interactive Fiction Intensive
What proves itThe completion gate answered out loud, and either a budget built on an observed timeline or a scoped first run with a defined end and a showable output.
Taught in Founding Federation: Price and Ideate (Week 0), Founding Federation: Build With Your Heart First
What proves itThe function, two call sites that do not know about each other, and two runs of the same call producing different results with the state that caused each difference named.
Taught in The Interactive Fiction Intensive
What proves itThe list with a source and a price against each line, the substitution you made when something was unavailable or too expensive, and the part that arrived wrong.
Named in the ontology. No class carries it yet.
What proves itYour own dungeon, playable end to end, holding at least one piece of state that a later scene reads back.
Taught in The Interactive Fiction Intensive
What proves itThree declines on cases absent from the boundary text, three accepts on near neighbours you did want, and the near miss that made you rewrite it.
Named in the ontology. No class carries it yet.
What proves itThe prompt in daily use, plus two transcripts from different sessions where a preference it encodes shows up without being asked for.
Taught in Using Large Language Models, Intro to Agents, Automate Your Email: No Code AI Automation, Critical Thinking and Creativity with AI
What proves itThe rubric plus two independent gradings of the same three outputs, with the scores side by side.
Taught in Prompt Engineering, AI Alignment, Custom GPTs, Critical Thinking and Creativity with AI
What proves itA run that halts on its own with the condition met, a run that halts with it unmet and says so, and the case where it stopped too early.
Named in the ontology. No class carries it yet.
What proves itA reachable URL and a recording of a first-time user finishing the task with no spoken instruction from you.
Named in the ontology. No class carries it yet.
What proves itThe original example, your script, and the run that produced the output you wanted rather than the example's.
Named in the ontology. No class carries it yet.
What proves itA live page, the outbound message that points at it, and the single action it asks for.
Taught in Founding Federation: Smallest Sellable Thing
What proves itTen pairs made on paper, the two you could build something on, and a written reason each of those two generates where the others only startle.
Taught in Critical Thinking and Creativity with AI
What proves itA story live on The Spaces Between's public shelf, with mood tags and metadata set, and a playtest exchange in both directions.
Taught in The Interactive Fiction Intensive
What proves itThe index, ten queries with their top results, and your relevance call on each — including the ones the index got wrong.
Taught in RAG & Memory, Advanced Retrieval Augmented Generation
What proves itThree diamonds chained, with the trunk demonstrably reading differently depending on state carried through them.
Taught in The Interactive Fiction Intensive
What proves itThe source text, the field list you wrote first, and the filled record, with every field traceable to a span of the source.
Taught in Automate Your Email: No Code AI Automation
Nothing in that domain sits on this rung. Pick another rung, or clear the filter.
You do the round trip by hand, with nothing helpful in between. This is the rung where a computer becomes parts you chose and a tool call becomes four messages you can name.
executelocaterun
What proves itA device that boots, its parts list, the total cost, and one line per part on why that one.
Taught in Build a Cyberdeck
What proves itA transcript of all four messages — request, tool call, tool result, final answer — from a script with no agent library in it.
Taught in Prompt Engineering, Intro to Agents, Claude Model Context Protocol
What proves itA stack trace, the line you navigated to from it, and one case where the line named was not where the mistake actually was.
Named in the ontology. No class carries it yet.
What proves itAn old phone or single-board computer running Linux and serving something over the local network, named by model, with what it replaced.
Taught in Solarpunk Automation
What proves itThe dataset, the model's analysis, and your own computation of one figure alongside it, agreeing or with the discrepancy explained.
Taught in Data Activism: Turn Data Into Power
What proves itThe server log showing the model loaded, and a client transcript from a separate process receiving the completion.
Taught in Control AI Spending, Hosting Open-Source Models
Nothing in that domain sits on this rung. Pick another rung, or clear the filter.
Someone else wrote it. You run it, you find it, you read what comes back — and the proof is that you can say what the error said before you fixed anything.
followlocaterun
What proves itA terminal transcript or notebook showing a successful call you ran yourself, including the model's response.
Taught in Using Large Language Models
What proves itA named tool, its argument schema copied from source, and a run showing its behaviour on a deliberately bad input.
Taught in Intro to Agents, Claude Model Context Protocol
What proves itA terminal transcript with a real failure, and your written account of what the error said before you fixed anything.
Named in the ontology. No class carries it yet.
Nothing in that domain sits on this rung. Pick another rung, or clear the filter.
The gap, measured
Pick any two capabilities. The distance between them is arithmetic, not opinion, because both numbers were computed the same way.
5 rungs apart Order 8 (Primary) to order 13 (Metasystematic), crossing Concrete, Abstract, Formal and Systematic.
Where it is thick and where it is thin
The same 83 capabilities, laid out by what they are about and how hard they are. Every cell is a real count; tap one to open that rung with that domain filtered.
| Domain | 7 | 8 | 9 | 10 | 11 | 12 | 13 | 14 | All |
|---|---|---|---|---|---|---|---|---|---|
| system construction | · | 24 | |||||||
| model behavior | · | · | · | 13 | |||||
| infrastructure | · | · | 11 | ||||||
| interaction and interface | · | · | · | 11 | |||||
| data shaping | · | · | 8 | ||||||
| audience and market | · | · | · | · | 5 | ||||
| evidence and verification | · | · | · | · | · | · | 5 | ||
| distribution and influence | · | · | · | · | · | · | 2 | ||
| organization and delegation | · | · | · | · | · | · | 2 | ||
| security | · | · | · | · | · | · | · | 1 | |
| alignment | · | · | · | · | · | · | · | 1 |
Read the empty cells too. Evidence and verification only appears at two rungs, and organisation and delegation does not appear at all until order 13 — which is another way of saying the school does not pretend delegating a team of agents is a beginner move.
What actually exists
Sorted by how much Make each one carries. 14 of the 26 have recordings — 277.6 hours in total, covering the whole class rather than the Make capabilities alone. 12 have no recordings at all, and this page is not going to imply otherwise: for those, the product is the live hour and the app.
VideoNo recordings. Live and the app only.
Exercises15
Next session Aug 29 · about 5.8 hours of load
Video10 clips, 22.4 hours
Exercises21
Next session Sep 11 · about 29.6 hours of load
VideoNo recordings. Live and the app only.
Exercises1
AppNo student app. The live sessions are the class.
Not on the calendar right now · about 2.2 hours of load
Video19 clips, 22.3 hours
AppNo student app. The live sessions are the class.
Next session Sep 17 · about 24.3 hours of load
Video7 clips, 9.0 hours
Exercises6
AppNo student app. The live sessions are the class.
Not on the calendar right now · about 12.5 hours of load
Video58 clips, 63.5 hours
Exercises142
AppA syllabus, and no student app.
Not on the calendar right now · about 101.0 hours of load
Video42 clips, 56.7 hours
Exercises32
AppA syllabus, and no student app.
Next session Sep 15 · about 66.7 hours of load
Video6 clips, 10.7 hours
Exercises55
AppNo student app. The live sessions are the class.
Not on the calendar right now · about 26.4 hours of load
Video17 clips, 26.5 hours
Exercises46
Next session Sep 7 · about 40.0 hours of load
Video8 clips, 10.8 hours
Exercises34
Next session Sep 4 · about 21.3 hours of load
Video4 clips, 3.9 hours
Exercises13
Next session Aug 22 · about 9.2 hours of load
Video3 clips, 5.2 hours
AppA syllabus, and no student app.
Not on the calendar right now · about 7.2 hours of load
VideoNo recordings. Live and the app only.
AppNo student app. The live sessions are the class.
Not on the calendar right now · about 2.0 hours of load
VideoNo recordings. Live and the app only.
Exercises34
Not on the calendar right now · about 10.5 hours of load
VideoNo recordings. Live and the app only.
Not on the calendar right now · about 2.0 hours of load
Video27 clips, 35.7 hours
Exercises329
Next session Sep 16 · about 120.0 hours of load
VideoNo recordings. Live and the app only.
AppNo student app. The live sessions are the class.
Not on the calendar right now · about 2.0 hours of load
Video5 clips, 6.8 hours
Exercises7
AppNo student app. The live sessions are the class.
Not on the calendar right now · about 10.6 hours of load
VideoNo recordings. Live and the app only.
Exercises25
Not on the calendar right now · about 8.2 hours of load
VideoNo recordings. Live and the app only.
AppNo student app. The live sessions are the class.
Not on the calendar right now · about 2.0 hours of load
VideoNo recordings. Live and the app only.
Exercises1
Not on the calendar right now · about 2.2 hours of load
VideoNo recordings. Live and the app only.
Not on the calendar right now · about 2.0 hours of load
VideoNo recordings. Live and the app only.
Next session Aug 22 · about 2.0 hours of load
Video2 clips, 3.1 hours
Exercises12
AppA syllabus, and no student app.
Not on the calendar right now · about 8.1 hours of load
VideoNo recordings. Live and the app only.
AppNo student app. The live sessions are the class.
Not on the calendar right now · about 2.0 hours of load
Video1 clips, 1.0 hours
Exercises3
AppNo student app. The live sessions are the class.
Not on the calendar right now · about 3.8 hours of load
9 of the 26 have a session on the calendar as this page was built. The rest run when a cohort forms; the Make & Do hour runs regardless, four mornings a week.
Where Make shows up
A path is a person you are trying to become; a philosophy is a thing you are trying to do. These are the paths where at least three of the capabilities carry Make — the number on each card is how many.
Build production AI systems from prompt to deployment.
32 of its 41 capabilities carry Make · 13 classes · tops out at order 13
🧙 The Weird Internet WizardBuild strange systems that change how people think.
31 of its 46 capabilities carry Make · 8 classes · tops out at order 13
🎨 The AI CreativeA creative studio with AI.
24 of its 41 capabilities carry Make · 5 classes · tops out at order 12
📊 The AI CitizenA liberal-arts curriculum rebuilt around AI.
23 of its 34 capabilities carry Make · 4 classes · tops out at order 12
🧑💼 The FounderFounder with an AI-first company.
16 of its 36 capabilities carry Make · 9 classes · tops out at order 12
💼 The AI ExecutiveReplace repetitive business operations with AI (a mini-MBA in AI).
12 of its 18 capabilities carry Make · 6 classes · tops out at order 12
🔓 Digital IndependenceTechnological sovereignty — not just saving money.
7 of its 21 capabilities carry Make · 8 classes · tops out at order 11
🌱 The Resilience TrackCommunity resilience and continuity.
4 of its 15 capabilities carry Make · 5 classes · tops out at order 13
Who this is not for
If the job is working out what follows from the evidence rather than shipping something, Think is the verb and Research & Activism is the hour. Make only gets to verify and falsify at order 11, and it does it to things you built yourself.
See Think →Own is the neighbour, and it is the one philosophy you do not have to choose between — Make & Do carries both, in the same room, at the same 9:00. If your actual question is about dependency and withdrawal rather than construction, start there and come to the same hour.
See Own →Then read the ladder above with a cold eye. 12 of the 26 classes have no recordings, 17 have no session booked, and 21 of the 83 capabilities have no class at all yet. What is reliably here is the hour, four mornings a week, and a written proof requirement for every capability you might claim. If that is not enough structure for you, this is the wrong philosophy to buy first.
Start on the rung you are actually on
Not a plan for a thing. The thing. Make & Do is Mon-Thu at 9:00 Pacific, and it is coworking rather than instruction — you build, somebody unblocks you, and the room keeps the energy up on the boring part. Find your rung on the ladder, read the proof it takes, and go and produce that proof.