We are at the second of four Tuesday sessions: Skills was on 6 October, today is Projects, then Village on 20 October and Deliverables on 27 October. Skills are the method; a project is the context the method works on. The goal is plain: by the end of this hour you can turn a piece of client work into a project the brain opens in seconds. Today is the context layer only; the team and the deliverable parts come in the next two sessions. Spend about a minute here, then move straight into the problem that motivates the whole thing.
A chat ends and the context dies with it. The same project lives scattered across email threads, SharePoint folders, sticky notes and people's heads, so "where are we on this client?" means reconstructing from four places. Show that: you cannot answer without paging through mail and asking a colleague. This is the cost the project folder removes — one place, always reachable. Spend forty seconds here; it justifies the rest of the talk.
Ask the room to picture a real client and the question "where are we on this account?". In a fresh chat the brain has nothing: you retype the client's situation, the contacts, where the search stands. In a project the brain loads the folder, the current state, and the last diary entries. Use one concrete example: last week's candidate search, whose notes now live in your head or in mail. Thirty to sixty seconds. Do not reveal the mechanics yet — that is what the next slides do.
WESP is the memory model. Working memory is the chat in front of you; Episodic is the diary, what happened and when; Semantic is the wiki, what things are; Procedural is skills, how to do things. A project is a semantic container in the wiki. It links to episodic memory through diary entries tagged with the project, and to procedural memory through the skills used on it. Draw three boxes plus the chat. Keep it to ninety seconds: the audience only needs the map, not the theory.
A project is a folder in the wiki. The path is wiki/projects, then a slug naming the client or the work, and inside it an index.md. That file is mandatory — the brain does not treat the folder as a project without it. Everything else in the folder is optional: separate files for deep dives, contacts, budgets. The index is the entry point the brain reads first, which is why it stays short. Show the path on the slide and read it aloud. One minute.
Three short slides. First the definition: a project is a text file that points to many other files. Then what it points to: the diary for what happened, the wiki for what we know, skills for how we work, and external links. Pause here: the next slide opens an example project.
Plain words, no file formats. A project holds four things: a short description, key info with a few KPIs, the do and don’t for working on it, and the current state. Each part lights up on the example as you name it. Ask the room which of the four they could write today for one of their own clients. One sentence on rules: company-wide rules (tone, privacy, what the brain may do) are not set per project and a project cannot override them; they apply to the whole brain, so leave the detail for the rules part.
That is why a project matters: when the brain does not do something you want, or you want it consistent at something specific for that project and not in general, you ask it to remember it in the project. Three places can hold a correction. The project: for this project only, the common case. A skill: for a way of working across projects. The boot, the base rules: very rarely, and with care. Correcting the answer is fine for the moment, but asking the brain to correct the project or the skill fixes it for next time. Keep this short today: it opens a bigger conversation about how to correct the brain, which we take up later in the course.
The most important idea of the hour. The project carries the current state: one line, today’s truth, replaced when things change. It does not carry the history. The history lives in the diary: every entry about this client is linked to the project, so you can rebuild the whole story at any time by reading those entries in order. Show the five dated entries one by one. Message to leave: keep the project short and up to date, let the diary be long and complete.
Four places hold four kinds of content, and they do not overlap. The index is the summary and the one-line current state. The diary holds dated events — deploys, decisions, calls — each tagged with the project. Issues and changelog go in issues.md or log.md inside the project folder. Quotes and budgets get their own separate files. When in doubt, ask: is it dated? diary. Is it a list of problems? issues. Is it a summary? index. About 30 seconds, walk the table row by row.
The dry rule: the brain is a notebook, not an archive. Big files stay where they are — an Excel tracker, a SharePoint folder, a document. The project only records that they exist and where to find them. Never copy the file into the brain, never paste a whole spreadsheet. Example: "Offer tracker: Excel on SharePoint, link." The brain points, the source of truth stays in one place. About 20 seconds.
Five stages, left to right, one line each. Experiment: an idea or a test, and it may die with no blame. Beta: it runs on real cases and a person checks every result. Active: trusted, part of the daily work. On hold: paused, nobody waiting, it can reopen. Archive: closed. The stage is a label on the project: you set it, or ask the brain to change it ("put Acme on hold"). The stage is private to your own project: it helps you and the brain, nobody else reads it. Ask the room where their own clients sit today.
One idea, said clearly: a project is your private notebook. It is yours and your brain’s. It is not how the team sees the work. To share there is email, and the Village with shared channels, tasks and decisions, which is the next session on 20 October. Artifacts (pages, sheets and reports for clients and candidates) come on 27 October. Do not go deeper today: the point is to set the expectation that nothing in the project is visible to colleagues unless it is sent or shared.
Walk through what happens when someone types /project. The brain does not start empty: it reads the short index first, so it knows the client, the phase and the current state before opening any deep file — that is the zero-token-waste rule in practice. Example: "/project acme-logistics" immediately shows the role search, the phase and what not to do, without loading the whole wiki. Invite one person to try it on a real client name and read the output aloud. Allow two minutes for this and the first reactions.
A chat should serve one target. When you open it, name the project (an existing one or a new one) and say what this chat is for: "Acme, send the shortlist to Maria today". Work until that target is reached. Then start a new chat: fresh context, shorter conversations are faster and cheaper, and each chat is easy to find later because it has one purpose. One chat per project is allowed, but one chat per target is the habit to build.
A different angle on the same project: follow one question through the brain. You ask where you are on Acme. The brain opens the project and reads the current state, then the latest diary entries, then the latest mail, and goes to the external source only if it needs the detail. The answer is built from what the project links to, without you retelling anything. Ask the room to tell you which of those four stops they would have to do by hand today.
Set up the exercise. Ten minutes, one real client each — not a fake. They type /project plus the client name, and the brain interviews them: a few questions, one at a time, then it shows a draft and writes the project once they approve. Tell them to answer from what they actually know; an empty field is fine, phase and state can be filled later. Example: one consultant used a live client and had a working index before the break. Timebox ten minutes, then a short show-and-tell.
Close with a check you can do in two minutes. After the project is written, open it and tick six things: a short summary, key info with the KPIs that matter, do and don’t, a phase, a one-line current state, and a history made of linked diary entries. Most people will already have projects: tell them not to start over. Open the existing project and ask the brain what is missing; it compares the project with this list and only asks about the gaps. Give them five minutes to do it on one of their own clients.
Close the loop and bridge to the next session. A project gives the brain the context of a mission; on 20 October, Village shows how a team follows that mission together, with checklists and deadlines, and on 27 October Deliverables turns it into documents for clients and candidates. Invite questions now and keep them concrete; if someone raises a point that part 3 will cover, note it and park it. Example: answer one question about phases, then point forward to Skills. Two minutes for questions, then hand over to part 3.