Resume One Stalled AI Task
Recover one real decision, finish one small part and leave a restart note. The prompts fill in from your task's name.
Open the member guide →Practical resources
Keep the thing that helps. Leave the rest.
Earlier free resources, Standard member companions, and expanded Premium guides are collected here. The full weekly Brief stays public.
Member companions
Standard and above · Learning Hub
Make the Graphic You Actually Need
A visual-brief builder, actual image examples, prompts and a review checklist.
Open the member guide →GPT-6 Astra: Use It or Save It?
Choose the jobs that earn OpenAI's strongest model, protect your allowance, and run one fair comparison before you make Astra your default.
Built for the August 31 through September 6 brief after Astra and Fable 5.1 raised the same practical question: which work actually needs the stronger model?
See the member guideExpanded in Build ModeBuild Mode → Model decisions
The full guide compares GPT-6 Astra and Claude Fable 5.1 across common kinds of work, plan usage, effort choices, and a same-job test using your own project.
Free resources
8 tools you can use now
The Safe AI Handoff Card
Give AI one bounded job, keep consequential actions with you, and make sure the handoff removes an old step instead of adding another checkpoint.
Built for the August 17 through 23 brief after AI tools reached further into messages, files, browsers, and agent actions.
Write the safe handoffThe Computer History Safe Start Guide
Recover work, notice patterns, and find repeatable workflows without opening your whole computer at once.
Built for the August 10 through 16 brief after Computer History found useful patterns in Joyce's own work within its first day.
Run the safe-start testThe AI Stack Snapshot
See which AI tools are earning their keep, which need one defined test, and which possible pause or cancellation deserves a closer look.
Built for the August 3 through 9 brief from the live member signal around subscription creep and tool-choice hesitation.
Run the free snapshotExpanded in the Builder LibraryPremium + VIP → Your AI Stack
The full skill creates a reusable stack file and routes future work only to tools, Skills, Projects, and connections you already have.
The Findable Quick Setup
Recover one useful thing you already made in ChatGPT, give it a reliable home, and avoid rebuilding it next month.
Originally shared as the FINDABLE comment giveaway from the July 19 brief.
Use the free setupExpanded in Build ModeBuild Mode → Workflows
The expanded Build Mode version adds Claude's different search behavior and a fuller system for keeping important work in files you own.
The Ten-Minute Quarter Audit
Compare what you planned with what you actually completed, count the useful work that appeared along the way, and decide what deserves to continue.
Originally shared as the QUARTER comment giveaway from the July 1 brief.
Run the free auditExpanded in Build ModeBuild Mode → Workflows
The expanded guided version lives in Build Mode for members who want to work through the review as a fuller operating exercise.
The Outage-Proof Context File
Build a portable context file that lets you keep working when a platform changes, a thread disappears, or you move to another tool.
Originally shared as the CONTEXT giveaway from the June 15 brief.
Build the free fileThe AI Summary Engine Audit
See how an AI search system could describe your homepage or offer page, then fix the parts that hide what you do and who it is for.
Originally shared as the AUDIT comment giveaway from the May 24 brief.
Run the free auditClaude Fable 5: When It Is Worth Using
Decide whether one difficult job justifies the model's cost, access, time, and data-retention tradeoff before you begin.
Rebuilt from the Lab's Fable 5 training and checked against Anthropic's current plan, prompting, safety, and retention documentation.
Choose and test the modelBuild Mode
Expanded and member-only resources
From Recovered Decision to Reviewed Deliverable
A practice packet, ready-to-copy prompts, an answer key and a review worksheet for one deliverable.
Open the member guide →One Source, Three Finished Assets
One source, an editable handout, a short deck, a graphic and a way to check the results.
Open the member guide →GPT-6 Astra vs. Claude Fable 5.1
A plain-language guide for choosing which model should do the job, what each may use from your plan, and whether the difference earned the cost.
Members: open the expanded guide →Which Model, Which Effort
A visual guide for choosing the model and effort level that match the job without burning through your usage limit.
Open in Build ModeThe Match Kit
Turn a useful AI update into the right tool, workflow, and smallest next move for the work already on your plate.
See Build Mode