Free Beer Studio

The resource library.
Every link the courses point at, in one place.

Each lesson ends with a short list of pages worth opening next: the vendor documentation for the feature it uses, public guidance from bodies the audience already trusts, checklists, references. This page collects all of them, grouped by course and lesson, with the same one line on when to use each. Every link was checked when it was written.

302 links9 courses130 sources

Practical AI for Teachers

K-12 teachers

01What AI is actually good at in a classroom, and what it is not

  • UNESCO — Guidance for generative AI in education and research — read this first if you want a sober, non-vendor view of where these tools belong in a school.
  • SIFT: The Four Moves — a four-step method for checking a claim in under two minutes; use it on every fact an AI tool hands you.
  • Stanford Civic Online Reasoning — free lessons and assessments on evaluating online sources; useful for you now and for your students in lesson 5.
  • Hemingway Editor — paste a rewrite in and it estimates the reading level; a fast second opinion when you are not sure the simplified version is actually simpler.
  • TeachAI Toolkit — a free, plain-language toolkit for schools deciding how to use AI; the section on what AI can and cannot do pairs well with the test in this lesson.

02The five-minute setup

03Planning with AI as a thinking partner, not a content generator

04Differentiation and feedback at scale, in your own voice

05Students and AI

06A week with AI

Practical AI for Managers

people managers and team leads

01The manager's real job in an AI-shaped week

  • NIST AI Risk Management Framework — read the four functions (govern, map, measure, manage) when you need language for "would it matter if this were wrong" that your risk and legal people already accept.
  • Automation bias — the name for the third row of the ten-second test; read it before you decide you would definitely notice a wrong answer.
  • Hallucination (artificial intelligence) — a plain overview of why confident, well-formatted, wrong output is a feature of the technology rather than a bug you can prompt away.
  • Eisenhower matrix — the sorting tool most managers already know; useful as a contrast, because it sorts by urgency and this lesson sorts by who has to be in the room.
  • The Management Center — Deciding what to delegate — a free tool for deciding what belongs on your own plate rather than someone else's, using impact, role, and whether you are genuinely the only one who can do it well; worth reading if your third column came out suspiciously short.

02Your operating context, and the five-minute setup

  • ChatGPT — Custom Instructions — where the standing-instructions slot lives in ChatGPT, what it applies to, and the character limit on each plan; start here if ChatGPT is the tool you have.
  • Claude — What are Projects — how to set up a container that holds standing instructions and files for one body of work; use it when you want a separate context for hiring, the weekly update, and planning.
  • Claude — personalization features — where the standing-instructions slot lives and how it interacts with memory; read before you decide where to paste your block.
  • Gemini — Gems — the equivalent saved-instructions feature if Gemini is what your organisation approved.
  • Microsoft 365 Copilot overview — what Copilot can and cannot see inside your tenant, which is the first question to answer if Copilot is your tool.
  • ChatGPT — Data Controls FAQ — where the training and retention switches are, what changes when you turn them off, and what a temporary chat does; this is the page behind the one-sentence answer for your notes file.
  • Claude — How long do you store my data? — the same question answered for Claude, and a model of the retention page you should read on whichever tool you chose.
  • Google Workspace — writing effective prompts — a short, vendor-neutral-enough primer on persona, task, context, and format; useful if your block keeps coming out as a mission statement.

03One-on-ones and coaching, with AI as the prep partner

  • GitLab Handbook — 1:1s — a large company's published one-on-one guidance, including agendas and cadence; useful as a structure to steal when your meetings have drifted into status reporting.
  • Lara Hogan — Questions for our first 1:1 — a free, specific list of opening questions; the best source of low-cost openers when you inherit a team.
  • CCL — How to have a coaching conversation — a free guide on the difference between advising and coaching, and on asking rather than telling; read it if your growth conversations keep turning into you giving advice.
  • Atlassian — Practical ways to get feedback — how to ask for feedback on yourself specifically enough that people actually give it, which is the honest answer to the "me" line in the performance-drop prompt.
  • The Management Center — resources — free tools and worksheets for delegation, check-ins, and performance conversations; a good place to find the non-AI version of anything in this lesson.

04Writing that lands

  • Federal Plain Language Guidelines — the free standard for writing that people can act on; the sections on audience and organisation are directly applicable to the update and the decision note.
  • plainlanguage.gov — Be concise — a specific list of words and constructions to cut; use it as the checklist behind the voice test.
  • GOV.UK style guide — an opinionated, maintained A-to-Z of plain words and their replacements; the fastest reference when you are unsure whether a phrase is jargon.
  • Nielsen Norman Group — How users read on the web — why front-loaded sentences and short paragraphs matter for anyone reading your update on a phone between meetings.
  • Hemingway Editor — a free in-browser check for sentence length and passive voice; run a hard message through it once to see which sentences are doing too much work.
  • Basecamp — How we communicate — a published set of writing norms for a working team; useful as a model if you want a written standard rather than a personal habit.

05The team's AI norms, on one page

06A week with AI, as a manager

  • Atlassian Team Playbook — all plays — free facilitation guides for recurring team rituals; use it when you want the non-AI version of a slot in your week to be good before you add a tool to it.
  • GitLab Handbook — Leadership — a published set of managerial routines at scale, including cadences and meeting norms; worth reading for how specific a written routine can be.
  • Anthropic — Prompt engineering overview — the reference to reach for when a prompt in your routine keeps returning vague output; the sections on being explicit and giving examples fix most of it.
  • Microsoft — Get better Copilot responses with great prompting — the open version of Microsoft's prompting guidance, with task-shaped examples; useful for finding the phrasing for a recurring job rather than for copying wholesale.
  • Prompt engineering — a vendor-neutral overview of the techniques and their limits; read it once so you can tell a real technique from a piece of folklore someone shares in a team meeting.
  • Digital.gov — Artificial intelligence — ongoing public-sector guidance and community resources; a stable place to check what the norms look like outside your own organisation.

Practical AI for Small Business

owners of small businesses: restaurants, shops, trades, services

01The inbox: inquiries, quotes, and follow-ups

  • Digital.gov: plain language guide series — the public reference for writing the way people actually read; use it when a draft is polite but nobody can tell what you meant.
  • Hemingway Editor — paste a draft in to see whether it reads like speech; use it for the first week while you calibrate what "sounds like me" means.
  • Anthropic: prompt engineering overview — when a draft keeps coming back wrong and you want to know which part of the instruction to change rather than rewriting the whole thing.
  • Anthropic: what are Projects — read this when you get tired of pasting the shop block and want it attached to every conversation by default.

02Reviews and reputation, without losing a night's sleep

  • Yelp: content guidelines — read before you reply to anything; it sets out what a business owner may and may not post in a response.
  • Yelp: don't ask for reviews — Yelp's own position on soliciting reviews, and why asking can work against your rating; check it before you build any habit of asking customers.
  • Trustpilot: guidelines for reviewers — the clearest plain-language statement of what makes a review legitimate; useful when you think one is fake and want to know the grounds.
  • BBB customer reviews — how reviews and complaints work on a platform customers often check without telling you.
  • Google Business Profile — where your listing and its review replies live; go here if you have never actually claimed yours.

03A month of marketing in an afternoon

04The back office, and the thing you keep putting off

05What stays out

06The weekly twenty minutes

  • SBA: manage your business — the free, vendor-neutral guide to the operating side of a small business, with sections on finances, hiring, taxes, and staying compliant; a good place to look when your back-office list has an item you do not know how to start, or when the follow-up block keeps surfacing overdue invoices and the problem is your terms rather than your chasing.
  • SBA: resource partners and local assistance — free, in-person help near you; use when a decision in your monthly review is bigger than a prompt.
  • America's SBDC — no-cost advising through local small business development centers; the best free hour most owners never take.
  • USAGov: small business — the plain index of official obligations and services, for when you are not sure which agency an item on your list belongs to.
  • Anthropic: Claude help center — vendor documentation for saved projects, memory, and sharing; read the equivalent pages for whichever tool you use once you want the shop block to load itself.

Practical AI for IT Admins

IT administrators and systems people

01The boundary first

02Scripts and configs

  • ShellCheck — paste a shell script in and get the mechanical faults instantly; run it before the review prompt so the AI review spends itself on the things a linter cannot see.
  • Google Shell Style Guide — a concrete house style you can point a drafting prompt at: "follow this style guide" is a much sharper instruction than "write clean code."
  • Bash Pitfalls — the canonical list of ways shell quietly does the wrong thing; use it to write the requirements block that stops the tool reproducing them.
  • explainshell — paste a command line and see each flag mapped to its manual page; the fastest way to check whether an invented flag is real.
  • PowerShell: aboutCommonParameters — if you write PowerShell, `-WhatIf` and `-Confirm` are the dry-run you were about to ask for by hand.
  • PSScriptAnalyzer overview — the PowerShell equivalent of running the linter first; same rule, same order.

03From ticket thread to runbook

  • Google SRE Book: Postmortem Culture — read once before you write your first few; it is where the "things we tried that did not work" instinct comes from and why blameless framing makes people write it down.
  • Google SRE Workbook: Incident Response — the operational shape a runbook plugs into: roles, handover, and what you collect before you escalate.
  • PagerDuty Incident Response documentation — a complete, free, public incident process you can lift sections from rather than inventing your own from a blank page.
  • Atlassian: postmortem templates — worked templates to steal headings from when your runbook skeleton needs a sibling for the write-up after the incident.
  • Diátaxis — the four-way split between tutorials, how-to guides, reference, and explanation; use it the moment someone tries to turn your runbook into a training document.
  • Write the Docs: documentation guide — practical, vendor-neutral guidance on structure and maintenance for when your ten runbooks become forty and need an owner.

04User communication

05Running it locally

  • Ollama — one of the simplest packaged runners for macOS, Linux and Windows; use it when you want a model pulled and answering in ten minutes.
  • Ollama API documentation — the endpoint reference you need for step 5, when you move from the chat window to a script.
  • LM Studio — a graphical alternative with a model browser and a local server; easier if you want to compare several models before committing.
  • llama.cpp — the underlying engine most local runners are built on; read it when you need to understand or tune what your runner is doing for you.
  • Hugging Face: the GGUF format — what the quantization suffixes in a model filename actually mean, so you can choose a file rather than guess at one.
  • OWASP Top 10 for Large Language Model Applications — read before you put the endpoint anywhere other than localhost; it names the failure modes your security review will ask about.

06The acceptable-use policy you have been asked to write

  • NIST AI Risk Management Framework — the reference to cite when someone asks what your policy is based on; voluntary, vendor-neutral, and recognised by the people who ask that question.
  • NIST AI RMF Playbook — the actionable companion; use it to turn a framework function into a specific control you can actually write down.
  • NIST: Generative AI Profile — the generative-AI-specific companion to the framework, for the risks that only appear once people are typing into a model.
  • SANS security policy templates — free, editable policy templates to see how neighbouring policies are structured before you invent a structure of your own.
  • ICO: guidance on AI and data protection — the regulator-written view for anyone holding personal data; the section on lawful basis is the one your legal reviewer will want.
  • EU AI Act explorer — a readable, searchable copy of the text; use it to check whether anything you are doing falls into a category with obligations attached.

Practical AI for School Leaders

principals, assistant principals, and district administrators

01The leader's real job with AI

02Your operating context

03The writing load

04Reading your own data

05Leading the staff conversation

06The policy and the vendor pitch

Practical AI for Higher Ed Faculty

college and university faculty, adjuncts, and instructional staff

01What changed and what did not

02Your teaching context

03The syllabus and the assignment

04Reading, research, and the citation you must open

  • Crossref Metadata Search — paste a title or a DOI and see whether the thing exists; the fastest first check there is.
  • doi.org — resolve a DOI directly and see where it actually lands, which catches correctly formatted fabrications.
  • Google Scholar — the second opinion when a title returns nothing in one index; also useful for confirming an author works in the area at all.
  • Retraction Watch — check before you build an argument on an older paper, and read occasionally for a clear-eyed view of what citation failure costs people.
  • Zotero — free reference manager; the place verified sources go the moment you verify them, so you never check the same one twice.
  • SIFT (The Four Moves) — a short, teachable checking routine you can hand to students as well as use yourself.

05Feedback at scale

06The rest of the job

Practical AI for Parents

parents and caregivers of school-age kids

01What it is, in kitchen-table language

  • Elements of AI — a free, non-technical online course; use it when you want an hour of grounding for yourself before you explain anything to anyone.
  • Hallucination (artificial intelligence) — Wikipedia — the plain background on why these tools state false things confidently, for when your teenager asks you to prove it.
  • TeachAI — AI Literacy — how schools are framing AI understanding, so your kitchen-table explanation lines up with the one they hear in class.
  • aiEDU Resources — free classroom-style explainers and activities you can borrow when one explanation has not landed and you want a different angle.
  • Common Sense Media: AI — family-facing guidance written for parents, not technologists; good for age-by-age framing.

02Homework without doing the homework

03House rules that hold

04The household load

05The risks that matter

  • NCMEC CyberTipline — where to report online exploitation of a child, including sexual images generated from a real face.
  • Take It Down — a free service for getting a sexual image of a minor removed from participating platforms, usable without sending the image anywhere.
  • NetSmartz — age-appropriate materials for talking to kids about online safety, useful when your own wording is not landing.
  • 988 Suicide & Crisis Lifeline — call or text 988 if a conversation turns out to be about something much heavier than AI.
  • AI Detectors Don't Work — MIT Sloan — have this ready before a meeting about a detector score, not after.
  • Anthropic Privacy Center: How long do you store my data? — one vendor's actual retention answer; read it as the model of what to go look for in whichever product your kid uses.

06A month with AI in the house

  • Common Sense Media — age-based reviews of apps, games, and shows; the place to check a product your kid brings home before you form an opinion about it.
  • Common Sense Media: Parenting, Media, and Everything in Between — running parent-facing articles; good for the monthly conversation when your openers have gone stale.
  • HealthyChildren: Family Media Plan — revisit it at each quarterly refresh; the structure is built to be amended rather than rewritten.
  • Nemours KidsHealth — plain, age-appropriate health and development material for the conversations that turn out not to be about AI at all.
  • TeachAI: Sample Guidance — what schools are being advised to publish; useful at the quarterly refresh to check whether your house rules have drifted from the school's.
  • Anthropic Privacy Center — one vendor's retention and training answers, as the model of what to re-check once a year in whichever product your household actually uses.

Practical AI for Students

high school and early college students

01Learning and getting the answer are two different products

02Inside the rules

03The study partner

04Writing that is still yours

05The rest of life

06Your own setup

Running AI Locally

anyone who wants AI on their own computer, with the data staying there

01Why local, and for whom

  • Large language model — Wikipedia — plain background on what these models are and how they are built; read it if "model" still feels like a black box before you size one in lesson 2.
  • Open weights — Wikipedia — what it means for a model's weights to be downloadable, and how that differs from open source; useful before you pick a family.
  • Hallucination (artificial intelligence) — Wikipedia — the background on why a confident wrong answer is the default failure, which is the ceiling described above in one page.
  • AI Risk Management Framework — NIST — a US federal framework for reasoning about AI risk; borrow its vocabulary when you have to explain your boundary to someone who wants it in writing.
  • Ollama — the tool lesson 3 installs; look at it today only to see what you are heading toward, and do not download it yet.

02Your hardware, your model

03Install, download, talk

  • Ollama Quickstart — the official install-and-first-run walkthrough; use it for the exact commands on your operating system.
  • Ollama CLI reference — every command including `pull`, `ls`, `rm`, and `create`; keep it open the first week. It documents the listing command as `ollama ls`; `ollama list` is the same thing.
  • Modelfile reference — Ollama — the full syntax for `FROM`, `PARAMETER`, and `SYSTEM`, for when you want a setting this lesson did not cover.
  • Ollama FAQ — where models are stored, how to change that location, and how to expose the local server to other apps on your machine.
  • ollama/docs/api.md at main — the full reference for the local HTTP server: every endpoint, every parameter, and a curl example for each.
  • OpenAI Compatibility Endpoints — LM Studio — the `/v1` surface LM Studio serves on port 1234, and how to point an existing OpenAI-shaped client at it.
  • LM Studio docs — the same three settings in a windowed app, if the terminal is not where you want to live.
  • llama.cpp — the engine underneath; go here when you want to understand why something is slow rather than work around it.

04Your own documents

  • Retrieval-augmented generation — Wikipedia — what "point it at a folder" is doing under the hood; read it before you trust a tool that does it in one click.
  • Context window — Wikipedia — why pasting has a limit at all, and why the limit costs memory on top of the model.
  • Word embedding — Wikipedia — the numeric fingerprints retrieval depends on, and therefore why a question in the wrong vocabulary finds the wrong chunks.
  • Open WebUI documentation — a local chat interface with document and folder support; check its embedding settings against the warning above before loading anything sensitive.
  • AnythingLLM documentation — another local document-and-folder tool; useful for comparing how each one handles chunking and which parts run on your machine.

05What stays local and what goes to the cloud

06Making it a habit