Curriculum graph
Follow the connections.
Lines show prerequisites and useful neighboring books. Move right across levels, and follow arrows from prerequisites to the books that depend on them.
Open visual map All 95 books · 197 links
BookRequires (direction)NearbyScroll inside the map to explore
Browse by track
Books and connections
Find every book and its relationships here without opening the visual map.
T01Foundations & judgement3
- T01-L01From Idea to Working Result: The Four-Quadrant MethodNearbyT01-L02 — Check AI Output Before Your Team Relies on ItT02-L01 — Get Answers You Can Use from a ChatbotT12-L01 — Decide What You May Paste Into AI at Work
- T01-L02Check AI Output Before Your Team Relies on ItRequiresT01-L01 — From Idea to Working Result: The Four-Quadrant MethodNearbyT03-L01 — Write a Prompt You Can ReuseT04-L01 — Ask Your Own Documents and Check Every CitationT12-L02 — Clean the Copy, Label the Result, Spot the Trap
- T01-L03Build the Smallest Thing Your Colleagues Can Run Without YouRequiresT01-L02 — Check AI Output Before Your Team Relies on ItNearbyT05-L01 — Your First Automation: Build It, Break It, Switch It OffT06-L02 — Build a Small Shared Tool From a Prompt, and Test It With a ColleagueT07-L03 — Build a Small Agent Your Colleagues Can Run, Test and Stop
T17Structure & understand4
T02Chat & assistants5
- T02-L01Get Answers You Can Use from a ChatbotRequiresT01-L01 — From Idea to Working Result: The Four-Quadrant MethodNearbyT03-L01 — Write a Prompt You Can ReuseT02-L02 — Set Up Projects a Colleague Can Pick Up
- T02-L02Set Up Projects a Colleague Can Pick UpRequiresT02-L01 — Get Answers You Can Use from a ChatbotT12-L01 — Decide What You May Paste Into AI at WorkNearbyT04-L01 — Ask Your Own Documents and Check Every CitationT02-L03 — Build a Shared Assistant Your Team Can Test and Take Over
- T02-L03Build a Shared Assistant Your Team Can Test and Take OverRequiresT02-L02 — Set Up Projects a Colleague Can Pick UpT12-L03 — Give an Unattended Workflow Limits and a Stop a Colleague Can UseNearbyT04-L02 — Curate a Document Set Your Team Can AskT02-L04 — A Self-Hosted Assistant Your Team Can Use Safely
- T02-L04A Self-Hosted Assistant Your Team Can Use SafelyRequiresT11-L03 — Local AI Apps Your Colleagues Can Rely OnT12-L04 — Lock Down a Self-Hosted AI Service Before Real Data Goes InNearbyT11-L04 — Put Your Team's Assistant on a Server Colleagues Can ReachT02-L05 — Run One Assistant for a Whole Institute or Company, and Hand It Over
- T02-L05Run One Assistant for a Whole Institute or Company, and Hand It OverRequiresT02-L04 — A Self-Hosted Assistant Your Team Can Use SafelyT12-L05 — Keep One List of Every AI System Your Team Runs, and Make It Survive YouNearbyT11-L05 — Keep a Shared Model Server Fast, Affordable and RunningT12-L05 — Keep One List of Every AI System Your Team Runs, and Make It Survive You
T03Prompting & context5
- T03-L01Write a Prompt You Can ReuseRequiresT02-L01 — Get Answers You Can Use from a ChatbotNearbyT03-L02 — Build a Context Packet Your Team Can Rerun
- T03-L02Build a Context Packet Your Team Can RerunRequiresT03-L01 — Write a Prompt You Can ReuseNearbyT04-L01 — Ask Your Own Documents and Check Every CitationT03-L03 — Put a Test Set Under the Prompt Your Team Depends On
- T03-L03Put a Test Set Under the Prompt Your Team Depends OnRequiresT03-L02 — Build a Context Packet Your Team Can RerunNearbyT03-L04 — Release Prompt Changes Through Review, Tests and a Way BackT05-L03 — Chain Checked Steps Into a Weekly Digest Your Team Can Rerun
- T03-L04Release Prompt Changes Through Review, Tests and a Way BackRequiresT03-L03 — Put a Test Set Under the Prompt Your Team Depends OnT08-L04 — Use OpenCode Desktop on Your Organisation's Real RepositoriesNearbyT03-L05 — Keep a Production Prompt Healthy for Months, and Hand It OverT05-L04 — Make an Automation Safe to Fail Before It Touches Real Systems
- T03-L05Keep a Production Prompt Healthy for Months, and Hand It OverRequiresT03-L04 — Release Prompt Changes Through Review, Tests and a Way BackNearbyT11-L05 — Keep a Shared Model Server Fast, Affordable and RunningT05-L05 — Keep Your Automations Running for Months, and After You Leave
T09Create media5
- T09-L01Make One Image or Narrated Clip, and Label ItRequiresT01-L01 — From Idea to Working Result: The Four-Quadrant MethodNearbyT09-L02 — Keep One Look Across Your Team's ImagesT12-L02 — Clean the Copy, Label the Result, Spot the Trap
- T09-L02Keep One Look Across Your Team's ImagesRequiresT09-L01 — Make One Image or Narrated Clip, and Label ItNearbyT09-L03 — Build a Caption Pipeline Your Team Can Rerun Every Week
- T09-L03Build a Caption Pipeline Your Team Can Rerun Every WeekRequiresT09-L02 — Keep One Look Across Your Team's ImagesT05-L02 — Put AI in the Middle of a Workflow, and Check Every AnswerNearbyT09-L04 — Transcribe Recordings That Must Stay in the Building
- T09-L04Transcribe Recordings That Must Stay in the BuildingRequiresT09-L03 — Build a Caption Pipeline Your Team Can Rerun Every WeekT11-L04 — Put Your Team's Assistant on a Server Colleagues Can ReachNearbyT11-L05 — Keep a Shared Model Server Fast, Affordable and Running
- T09-L05Keep Every Published AI Image, Voice and Video Labelled, Cleared and TraceableRequiresT09-L04 — Transcribe Recordings That Must Stay in the BuildingT12-L05 — Keep One List of Every AI System Your Team Runs, and Make It Survive YouNearbyT12-L05 — Keep One List of Every AI System Your Team Runs, and Make It Survive You
T15AI in research5
- T15-L01Use AI in Your Research and Say Exactly HowRequiresT01-L02 — Check AI Output Before Your Team Relies on It
- T15-L02Record an AI Step So a Co-Author Can Rerun ItRequiresT15-L01 — Use AI in Your Research and Say Exactly HowT10-L02 — Clean a Messy Export Once, So Your Team Can Refresh It
- T15-L03Get an AI Step on Participant Data Reviewed, Then Build It Inside the ApprovalRequiresT15-L02 — Record an AI Step So a Co-Author Can Rerun ItT12-L03 — Give an Unattended Workflow Limits and a Stop a Colleague Can Use
- T15-L04Set Up AI for a Whole Research GroupOpen starting point
- T15-L05Keep Your Group's AI Use Compliant and Running After You LeaveOpen starting point
T18AI for lab data3
T04Your documents & knowledge5
- T04-L01Ask Your Own Documents and Check Every CitationRequiresT01-L02 — Check AI Output Before Your Team Relies on ItT12-L01 — Decide What You May Paste Into AI at WorkNearbyT04-L02 — Curate a Document Set Your Team Can AskT01-L02 — Check AI Output Before Your Team Relies on It
- T04-L02Curate a Document Set Your Team Can AskRequiresT04-L01 — Ask Your Own Documents and Check Every CitationNearbyT02-L03 — Build a Shared Assistant Your Team Can Test and Take OverT04-L03 — Test and Fix the Document Assistant Your Group Relies On
- T04-L03Test and Fix the Document Assistant Your Group Relies OnRequiresT04-L02 — Curate a Document Set Your Team Can AskT03-L03 — Put a Test Set Under the Prompt Your Team Depends OnNearbyT04-L04 — Connect an Assistant to Real Documents Without Widening Anyone's AccessT03-L05 — Keep a Production Prompt Healthy for Months, and Hand It Over
- T04-L04Connect an Assistant to Real Documents Without Widening Anyone's AccessRequiresT04-L03 — Test and Fix the Document Assistant Your Group Relies OnT12-L04 — Lock Down a Self-Hosted AI Service Before Real Data Goes InNearbyT04-L05 — Keep a Connected Knowledge Base Current, and Prove a Withdrawn Document Is GoneT07-L04 — Connect an Agent to a Real Mailbox Without Handing It the Keys
- T04-L05Keep a Connected Knowledge Base Current, and Prove a Withdrawn Document Is GoneRequiresT04-L04 — Connect an Assistant to Real Documents Without Widening Anyone's AccessNearbyT12-L05 — Keep One List of Every AI System Your Team Runs, and Make It Survive YouT10-L05 — Keep a Data Pipeline Right for Years, and Trace Any Published Number to Its Source
T05Automation5
- T05-L01Your First Automation: Build It, Break It, Switch It OffRequiresT01-L03 — Build the Smallest Thing Your Colleagues Can Run Without YouNearbyT05-L02 — Put AI in the Middle of a Workflow, and Check Every Answer
- T05-L02Put AI in the Middle of a Workflow, and Check Every AnswerRequiresT05-L01 — Your First Automation: Build It, Break It, Switch It OffT03-L02 — Build a Context Packet Your Team Can RerunNearbyT05-L03 — Chain Checked Steps Into a Weekly Digest Your Team Can RerunT07-L03 — Build a Small Agent Your Colleagues Can Run, Test and Stop
- T05-L03Chain Checked Steps Into a Weekly Digest Your Team Can RerunRequiresT05-L02 — Put AI in the Middle of a Workflow, and Check Every AnswerNearbyT05-L04 — Make an Automation Safe to Fail Before It Touches Real SystemsT04-L03 — Test and Fix the Document Assistant Your Group Relies On
- T05-L04Make an Automation Safe to Fail Before It Touches Real SystemsRequiresT05-L03 — Chain Checked Steps Into a Weekly Digest Your Team Can RerunNearbyT05-L05 — Keep Your Automations Running for Months, and After You LeaveT07-L04 — Connect an Agent to a Real Mailbox Without Handing It the Keys
- T05-L05Keep Your Automations Running for Months, and After You LeaveRequiresT05-L04 — Make an Automation Safe to Fail Before It Touches Real SystemsT12-L05 — Keep One List of Every AI System Your Team Runs, and Make It Survive YouNearbyT11-L05 — Keep a Shared Model Server Fast, Affordable and RunningT03-L05 — Keep a Production Prompt Healthy for Months, and Hand It Over
T06Build apps5
- T06-L01Build a Small Browser Tool You Will Actually UseOpen starting point
- T06-L02Build a Small Shared Tool From a Prompt, and Test It With a ColleagueRequiresT01-L03 — Build the Smallest Thing Your Colleagues Can Run Without YouNearbyT06-L03 — Turn Your Team App Into One Colleagues Can Depend On
- T06-L03Turn Your Team App Into One Colleagues Can Depend OnRequiresT06-L02 — Build a Small Shared Tool From a Prompt, and Test It With a ColleagueT12-L03 — Give an Unattended Workflow Limits and a Stop a Colleague Can UseNearbyT06-L04 — Put Your Team App on the Organisation's Own Database, and Prove You Can Restore It
- T06-L04Put Your Team App on the Organisation's Own Database, and Prove You Can Restore ItRequiresT06-L03 — Turn Your Team App Into One Colleagues Can Depend OnT12-L04 — Lock Down a Self-Hosted AI Service Before Real Data Goes InNearbyT06-L05 — Run the App Others Depend On, and Hand It OverT10-L04 — Connect Your Analysis to a Live System, and Stop Bad Data Before It Reaches Anyone
- T06-L05Run the App Others Depend On, and Hand It OverRequiresT06-L04 — Put Your Team App on the Organisation's Own Database, and Prove You Can Restore ItT11-L04 — Put Your Team's Assistant on a Server Colleagues Can ReachNearbyT08-L05 — Run Automatic Checks, an AI Reviewer and Releases You Can Roll BackT12-L05 — Keep One List of Every AI System Your Team Runs, and Make It Survive You
T07Agents4
- T07-L02Hand a Task to an Agent, and Check What It DidOpen starting point
- T07-L03Build a Small Agent Your Colleagues Can Run, Test and StopRequiresT05-L02 — Put AI in the Middle of a Workflow, and Check Every AnswerT01-L03 — Build the Smallest Thing Your Colleagues Can Run Without YouNearbyT07-L04 — Connect an Agent to a Real Mailbox Without Handing It the KeysT05-L03 — Chain Checked Steps Into a Weekly Digest Your Team Can Rerun
- T07-L04Connect an Agent to a Real Mailbox Without Handing It the KeysRequiresT07-L03 — Build a Small Agent Your Colleagues Can Run, Test and StopT12-L04 — Lock Down a Self-Hosted AI Service Before Real Data Goes InNearbyT05-L04 — Make an Automation Safe to Fail Before It Touches Real SystemsT04-L04 — Connect an Assistant to Real Documents Without Widening Anyone's Access
- T07-L05Keep a Team of Agents Running for Months, With Memory a Person ApprovesRequiresT07-L04 — Connect an Agent to a Real Mailbox Without Handing It the KeysT03-L05 — Keep a Production Prompt Healthy for Months, and Hand It OverNearbyT05-L05 — Keep Your Automations Running for Months, and After You Leave
T21Delegate desk work3
T20AI harnesses5
- T20-L01Same Model, Different Harness: See What the Program Around the AI ChangesRequiresT01-L01 — From Idea to Working Result: The Four-Quadrant MethodT08-L01 — Start With OpenCode: Set It Up, Change One Line, and Undo It
- T20-L02Choose a Harness by Testing It on Five of Your Own TasksRequiresT20-L01 — Same Model, Different Harness: See What the Program Around the AI ChangesT08-L02 — Get a Script Your Colleague Can Rerun, Without Being a Developer
- T20-L03Run a Harness on a Model You Control: Local Server or Team GatewayRequiresT20-L02 — Choose a Harness by Testing It on Five of Your Own TasksT11-L02 — A Local Model Your Tools Can Call
- T20-L04Set Up a Harness So a Mistake Stays Small: Container, Rules, Keys and NetworkRequiresT20-L03 — Run a Harness on a Model You Control: Local Server or Team GatewayT08-L04 — Use OpenCode Desktop on Your Organisation's Real Repositories
- T20-L05Run Coding Agents for a Team: Shared Rules, a Test Set, Nightly Jobs and No Lock-InRequiresT20-L04 — Set Up a Harness So a Mistake Stays Small: Container, Rules, Keys and NetworkT08-L05 — Run Automatic Checks, an AI Reviewer and Releases You Can Roll Back
T08Write code5
- T08-L01Start With OpenCode: Set It Up, Change One Line, and Undo ItOpen starting point
- T08-L02Get a Script Your Colleague Can Rerun, Without Being a DeveloperRequiresT02-L02 — Set Up Projects a Colleague Can Pick UpNearbyT08-L03 — Write the Spec First, So an AI Assistant Can Change Code Your Team Depends OnT10-L02 — Clean a Messy Export Once, So Your Team Can Refresh It
- T08-L03Write the Spec First, So an AI Assistant Can Change Code Your Team Depends OnRequiresT08-L02 — Get a Script Your Colleague Can Rerun, Without Being a DeveloperT03-L02 — Build a Context Packet Your Team Can RerunNearbyT08-L04 — Use OpenCode Desktop on Your Organisation's Real Repositories
- T08-L04Use OpenCode Desktop on Your Organisation's Real RepositoriesRequiresT08-L03 — Write the Spec First, So an AI Assistant Can Change Code Your Team Depends OnNearbyT08-L05 — Run Automatic Checks, an AI Reviewer and Releases You Can Roll BackT03-L04 — Release Prompt Changes Through Review, Tests and a Way Back
- T08-L05Run Automatic Checks, an AI Reviewer and Releases You Can Roll BackRequiresT08-L04 — Use OpenCode Desktop on Your Organisation's Real RepositoriesT12-L04 — Lock Down a Self-Hosted AI Service Before Real Data Goes InNearbyT06-L05 — Run the App Others Depend On, and Hand It OverT12-L05 — Keep One List of Every AI System Your Team Runs, and Make It Survive You
T19Share & review code5
- T19-L01Keep Your Files in Gitea and Get Any Version BackOpen starting point
- T19-L02Work on a Shared Gitea Repository From Your Computer With OpenCodeOpen starting point
- T19-L03Propose, Review and Merge Changes With Pull RequestsOpen starting point
- T19-L04Check Every Change Automatically, Release Versions and Tell the LabOpen starting point
- T19-L05Run Your Organisation's Gitea: Teams, Safe Sign-In and a Restore You Have TestedOpen starting point
T10Data & analysis5
- T10-L01Ask a Spreadsheet a Question and Check Which Rows It UsedRequiresT01-L02 — Check AI Output Before Your Team Relies on ItNearbyT10-L02 — Clean a Messy Export Once, So Your Team Can Refresh ItT04-L01 — Ask Your Own Documents and Check Every Citation
- T10-L02Clean a Messy Export Once, So Your Team Can Refresh ItRequiresT10-L01 — Ask a Spreadsheet a Question and Check Which Rows It UsedNearbyT10-L03 — Build an Analysis Your Colleague Can Rerun and Get the Same Numbers
- T10-L03Build an Analysis Your Colleague Can Rerun and Get the Same NumbersRequiresT10-L02 — Clean a Messy Export Once, So Your Team Can Refresh ItT08-L02 — Get a Script Your Colleague Can Rerun, Without Being a DeveloperNearbyT10-L04 — Connect Your Analysis to a Live System, and Stop Bad Data Before It Reaches AnyoneT08-L04 — Use OpenCode Desktop on Your Organisation's Real Repositories
- T10-L04Connect Your Analysis to a Live System, and Stop Bad Data Before It Reaches AnyoneRequiresT10-L03 — Build an Analysis Your Colleague Can Rerun and Get the Same NumbersT12-L04 — Lock Down a Self-Hosted AI Service Before Real Data Goes InNearbyT06-L04 — Put Your Team App on the Organisation's Own Database, and Prove You Can Restore ItT04-L04 — Connect an Assistant to Real Documents Without Widening Anyone's Access
- T10-L05Keep a Data Pipeline Right for Years, and Trace Any Published Number to Its SourceRequiresT10-L04 — Connect Your Analysis to a Live System, and Stop Bad Data Before It Reaches AnyoneT12-L05 — Keep One List of Every AI System Your Team Runs, and Make It Survive YouNearbyT04-L05 — Keep a Connected Knowledge Base Current, and Prove a Withdrawn Document Is GoneT05-L05 — Keep Your Automations Running for Months, and After You Leave
T14Connect your tools4
- T14-L02Brief Your Team From the Shared Drive, Through a ConnectorOpen starting point
- T14-L03Turn a Connector Brief Into a Routine Colleagues Can RerunRequiresT02-L02 — Set Up Projects a Colleague Can Pick UpT05-L02 — Put AI in the Middle of a Workflow, and Check Every Answer
- T14-L04Build a Read-Only Connector for a System Your Assistant Can't Reach YetRequiresT14-L03 — Turn a Connector Brief Into a Routine Colleagues Can RerunT12-L04 — Lock Down a Self-Hosted AI Service Before Real Data Goes InT08-L03 — Write the Spec First, So an AI Assistant Can Change Code Your Team Depends On
- T14-L05Run Your Organisation's Connectors: One List, Per-Person Access, a Log and a Clean ExitRequiresT14-L04 — Build a Read-Only Connector for a System Your Assistant Can't Reach YetT12-L05 — Keep One List of Every AI System Your Team Runs, and Make It Survive You
T11Run your own AI5
- T11-L01Run an AI Model on Your Own Laptop, OfflineRequiresT01-L01 — From Idea to Working Result: The Four-Quadrant MethodNearbyT11-L02 — A Local Model Your Tools Can CallT12-L01 — Decide What You May Paste Into AI at Work
- T11-L02A Local Model Your Tools Can CallRequiresT11-L01 — Run an AI Model on Your Own Laptop, OfflineNearbyT11-L03 — Local AI Apps Your Colleagues Can Rely OnT05-L02 — Put AI in the Middle of a Workflow, and Check Every Answer
- T11-L03Local AI Apps Your Colleagues Can Rely OnRequiresT11-L02 — A Local Model Your Tools Can CallNearbyT11-L04 — Put Your Team's Assistant on a Server Colleagues Can Reach
- T11-L04Put Your Team's Assistant on a Server Colleagues Can ReachRequiresT11-L03 — Local AI Apps Your Colleagues Can Rely OnT12-L04 — Lock Down a Self-Hosted AI Service Before Real Data Goes InNearbyT11-L05 — Keep a Shared Model Server Fast, Affordable and RunningT02-L04 — A Self-Hosted Assistant Your Team Can Use Safely
- T11-L05Keep a Shared Model Server Fast, Affordable and RunningRequiresT11-L04 — Put Your Team's Assistant on a Server Colleagues Can ReachT03-L05 — Keep a Production Prompt Healthy for Months, and Hand It OverNearbyT02-L05 — Run One Assistant for a Whole Institute or Company, and Hand It OverT09-L04 — Transcribe Recordings That Must Stay in the Building
T12Security, privacy & governance5
- T12-L01Decide What You May Paste Into AI at WorkRequiresT01-L01 — From Idea to Working Result: The Four-Quadrant MethodNearbyT01-L02 — Check AI Output Before Your Team Relies on ItT12-L02 — Clean the Copy, Label the Result, Spot the Trap
- T12-L02Clean the Copy, Label the Result, Spot the TrapRequiresT12-L01 — Decide What You May Paste Into AI at WorkNearbyT09-L01 — Make One Image or Narrated Clip, and Label ItT12-L03 — Give an Unattended Workflow Limits and a Stop a Colleague Can Use
- T12-L03Give an Unattended Workflow Limits and a Stop a Colleague Can UseRequiresT12-L02 — Clean the Copy, Label the Result, Spot the TrapT05-L02 — Put AI in the Middle of a Workflow, and Check Every AnswerNearbyT05-L04 — Make an Automation Safe to Fail Before It Touches Real SystemsT07-L04 — Connect an Agent to a Real Mailbox Without Handing It the Keys
- T12-L04Lock Down a Self-Hosted AI Service Before Real Data Goes InRequiresT11-L03 — Local AI Apps Your Colleagues Can Rely OnNearbyT11-L04 — Put Your Team's Assistant on a Server Colleagues Can ReachT02-L04 — A Self-Hosted Assistant Your Team Can Use Safely
- T12-L05Keep One List of Every AI System Your Team Runs, and Make It Survive YouRequiresT12-L04 — Lock Down a Self-Hosted AI Service Before Real Data Goes InT03-L05 — Keep a Production Prompt Healthy for Months, and Hand It OverNearbyT05-L05 — Keep Your Automations Running for Months, and After You LeaveT10-L05 — Keep a Data Pipeline Right for Years, and Trace Any Published Number to Its SourceT02-L05 — Run One Assistant for a Whole Institute or Company, and Hand It Over
T13Adoption & enablement4
- T13-L02Show a Colleague AI on Their Own Task, in 30 MinutesRequiresT01-L02 — Check AI Output Before Your Team Relies on ItT02-L02 — Set Up Projects a Colleague Can Pick Up
- T13-L03Write a One-Page AI Rule Your Team Can Use Without Asking YouRequiresT12-L01 — Decide What You May Paste Into AI at WorkT12-L03 — Give an Unattended Workflow Limits and a Stop a Colleague Can Use
- T13-L04Roll Out an AI Assistant to a Whole Department, With Access, Cost and Consent Settled FirstRequiresT13-L03 — Write a One-Page AI Rule Your Team Can Use Without Asking You
- T13-L05Keep Everyone's AI Skills Current, and Show It When AskedRequiresT13-L04 — Roll Out an AI Assistant to a Whole Department, With Access, Cost and Consent Settled FirstT12-L05 — Keep One List of Every AI System Your Team Runs, and Make It Survive You
T16Choose tools5
- T16-L01Pick the Right Tool for One Task by Trying ItOpen starting point
- T16-L02Choose Tools Your Team Can Use, and AffordOpen starting point
- T16-L03Test a Tool on Your Own Work Before Your Team Depends on ItOpen starting point
- T16-L04Replace a Tool Without Losing Your WorkOpen starting point
- T16-L05Keep Your Toolset UsefulOpen starting point