Infosoul Laboratories · AI for the Layman
Syllabus
Five classes, numbered like a catalogue and booked one at a time, for people who are not engineers. The heart of the set is the history, the philosophical questions and the social and political implications; the technology is taught only as far as those need it, and then, in 301 and 401, as far as running it yourself needs it. Each is listed the way a programme coordinator needs it: what the room will be able to do afterwards, how long it runs, what everyone leaves with, what the room needs, and what it costs. About eleven contact hours for the set, which is roughly one continuing-education unit. The glossary gives our plain words beside the catalogue words.
- Presenter — Brock Satter, Director of Infosoul Laboratories. BA student in science, technology and culture at the University of Minnesota, focusing on information technology, biotechnology and artificial intelligence, with the anthropologist Harvey Sarles, the department’s former chair, as senior project adviser.
- Access — all ages. No technical background assumed, and none required. Large-print guide on request.
- Where — your room, or online on Zoom or your own platform. Hybrid on request.
- Paperwork — invoice on booking, W-9 on request, net 30. Purchase orders welcome. Three weeks' lead time.
What this is not. Colleges now sell certificates in generative and agentic AI: twelve to eighteen weeks of Python, frameworks and rented services, at about three thousand dollars a seat. They teach building on somebody else's machine, and they teach it well. They do not address what these classes do: explaining the thing plainly, checking it, and running it in your own building, on your own laptop, under your own control. If a member of your staff wants the building track, we will say so and point them at the right programme. This ladder is the other thing.
AI 101 · The talk
The AI Age, For the Layman
50 minutes for a class or lunch slot, 35 and 15 for questions · 75 minutes as an evening, under the title Superintelligence for the People, 50 and 25
- Outcomes
- Place the machines in seventy years of history, so that "it is all new" no longer works on you.
- Ask the questions the room came with, properly: does it think, is it creative, is anything home, is it alive, and what would follow if it were. Know the main answers people have given, from Turing to the "stochastic parrot", why each is still open, though not in the same way; what the machine shows about our own thinking, and whether meaning lives in language itself; and where the presenter stands.
- Say what one of these machines actually is, in one sentence, and why that neither settles the questions nor excuses you from them.
- Name its two failures, flattery and invention, and the habit that protects against both.
- Tell what you are dealing with in the inbox, on a website or in a call, with four questions and a privacy ladder: who you are talking to, where it goes, who else can read it, and whether it will teach the machine.
- Take three discoveries of this moment, the protein atlas, the enzyme found in viral DNA, and the claimed proof about the equations of flowing water, and follow their implications for medicine, patents, security, universities, nations, and the locks every bank relies on: the implications almost nobody has put to the people they affect.
- Say who paid for it, what they want back, what it is doing to work and to the truth, why the largest warehouse may not be the only road to the frontier, and what "whose machine" means for your organisation.
- Say, in their own words, what some of the people who own the machines believe about democracy, people and machines, and why this course holds that people, not owners, decide.
- The room does
- Asks the questions. At some point the room watches the machine get one wrong, live, on a question the room chose.
- Leaves with
- The one-page guide, for every seat. It may be photocopied freely.
- Room and tech
- Any size, seminar room to hall. A screen with HDMI. Internet for the live demonstration.
- Contact hours
- 1 to 1¼
- Price
- $500 community and congregations · $1,500 campus, library and professional
AI 201 · Workshop
Ask It Yourself
90 minutes, hands-on · best after 101, not required
- Outcomes
- Put a question from your own work to a model in a form that gets a usable answer: who you are, what it is for, who reads it, how long, what to include and leave out.
- Recognise the sound of a confident wrong answer.
- Check the one claim that would cost you something if false, and make that a habit.
- Know which questions it can invent an answer to (names, numbers, dates, quotations, citations) and which it cannot.
- The room does
- Everyone brings a real question they already know the answer to and watches what comes back. Then a question they do not know, and checks it.
- Leaves with
- A checking habit, and the guide.
- Room and tech
- Up to 30 with laptops or phones, 60 in pairs. Internet needed.
- Contact hours
- 1½
- Price
- $750 · $2,000
AI 202 · Workshop
Make It Yourself
90 minutes, the generative one · built in advance from three of your own documents
- Outcomes
- Produce something real for your organisation with a model: a newsletter, a flyer with a picture from an open image model, a grant paragraph, an announcement in two languages and in a spoken voice.
- Work from your own documents as the source, so the model is grounded in your material and not the pattern. In catalogue terms: retrieval over private data; in ours, a second brain that answers back.
- Use models from the open-source galleries, Hugging Face and the Ollama library, that you can download and keep, and choose them safely: from the laboratories’ own accounts, in formats that hold only numbers.
- Know what to give a speaker to check before anything is printed or sent.
- The room does
- Makes a finished piece, each person, on the organisation's real material, and reads it as the recipient would.
- Leaves with
- A finished piece, and the method.
- Room and tech
- Up to 25 with laptops. Internet needed. Three documents from the host, two weeks ahead.
- Contact hours
- 1½
- Price
- $750 · $2,000
AI 301 · Build
Run Your Own
Half day · the half-day the book's last chapter is about
- Outcomes
- Install a free, open-source model on the organisation's own laptop and use it with the network cable out.
- Choose a model for the machine you have: what the size numbers mean and what memory they want, and which of the open models, Llama, Gemma, Mistral, DeepSeek, Qwen, Kimi, is the right one today.
- Understand why a small model asked well, with room to check its own work, reaches on many tasks what the largest reached a few months ago: method travels, and it fits on a table.
- Know what a local model is good for on day one (summarising, rewriting, translating for a speaker to check, comparing two documents) and what it is not.
- Dictate to your own machine: an open speech model (Whisper) behind a key, turning speech into text that stays on your computer.
- Let it look things up: a local model can search the web, which makes a small model more capable as well as more current; know that your search words leave the machine when it does.
- Go beyond the chat window: an open vision model that reads a photograph or a scanned page, a voice that reads text aloud (Piper), and, where the machine is strong enough, an open image model that draws. The room sees what each needs and what each gets wrong.
- Say plainly what you give up by running your own and what you keep: slower, a little dumber, yours.
- The room does
- Installs it, runs it, proves nothing leaves the room, tries a better model, puts the day's minutes through it, and dictates a letter to it with the cable out.
- Leaves with
- A running machine, and a written setup sheet for the next one.
- Watch next
- Five YouTube channels for running AI at home. Videos date fast: check when one was made before buying anything it recommends.
- Matt Williams — the gentlest start: he helped build Ollama and explains it plainly, with privacy in mind.
- Techno Tim — self-hosting at home, calmly explained.
- Alex Ziskind — which laptop or Mac runs what, tested against the cloud on the same job.
- Digital Spaceport — building a dedicated home AI machine: what to buy and what to skip.
- Fahd Mirza — installs each new open model on a home machine as it comes out.
- Room and tech
- Up to 12, each with a laptop from the last five years and its charger. Power strips and internet for the download.
- Contact hours
- 3½
- Price
- $1,200 · $3,500
AI 401 · Agents
Give It Hands
Half day, for staff and committees · best after 301
- Outcomes
- Explain, in plain words, an agent (a script running in the machine's window, with permission to act), the socket (MCP, the standard plug that lets it reach your files and tools) and the foreman (orchestration: several of them working a job).
- Tell apart the free agents people are installing this year: Pi, OpenClaw and Hermes Agent, and what each is for.
- Connect the model you run to one of your own folders, read-only, and ask it questions answered from your own records.
- Understand hidden instructions (prompt injection) and why an agent that can send, pay or delete needs a person between the decision and the act.
- Write, as a group, the organisation's policy: what it may touch, who checks its work, what it may never do, whose machine it stands on.
- The room does
- Builds a working read-only agent on the organisation's own machine, tries to trick it, and writes the policy before leaving.
- Leaves with
- The agent, running, and the written policy.
- Room and tech
- Up to 12, each with a laptop that has done 301 or an afternoon's equivalent. Internet needed.
- Contact hours
- 3½
- Price
- $1,200 · $3,500
For teachers
The book behind these classes, The AI Age, For the Layman, is a primer in the history of science and technology: how these machines emerged, whose work made them possible, who controls them, and what they change in how we understand knowledge and ourselves, with every claim sourced in the notes. It moves from history to practice, so a class can test the competing accounts on a machine it can run. It reads well beside Hans Moravec’s Mind Children in an introductory course.
The set
101, 201, 202 and 301 in order across a month, for a library, congregation or local; 401 when the room is ready. About eleven contact hours. Quoted, less than the sum of the parts. Every participant who completes the set receives a certificate of completion. Continuing-education credit through a partner college on request. Any class can be given as a staff edition, for a team's training day, with a policy conversation added.
A room that cannot pay can host it for costs. Say what the budget carries.
Book a class
The glossary
The one-sheet