WEEKLY CHALLENGES / COMPLETE HISTORY

Practice,
kept in motion.

Previous challenge sets from the homepage. Use these when you want a different constraint, prompt style, or weekly focus.

28archived prompts
ready to run again

THE BACK CATALOG

Previous challenges

Latest set →
BUILD 15 MIN

Archived Mar 3, 2026

Reliability Audit

Choose one AI workflow you trust. Test whether it still holds up when inputs are messy and incomplete.

  1. Run the workflow on 5 real-world inputs, not polished prompts.
  2. Document the top 3 failure modes and note which are silent.
  3. Implement one guardrail this week: validation, fallback, or human review.
Use Audit Worksheet →
LIVE 30 MIN

Archived Mar 3, 2026

Repeatable Meal Challenge

Build one diabetic-friendly meal you can repeat without decision fatigue during a busy week.

  1. Pick one recipe with simple ingredients and realistic prep time.
  2. Cook it once and note prep friction, portions, and satiety.
  3. Set a second cook day and improve one step for speed or consistency.
Run Meal Challenge →
CREATE 20 MIN

Archived Mar 3, 2026

Creative Consistency Sprint

Create one small piece this week with a fixed constraint and a fixed finish line.

  1. Choose one constraint: tempo, word count, instrument, or format.
  2. Ship a rough version in one session with no polish pass.
  3. Write three lines on what worked and what to repeat next week.
Run Sprint Protocol →
DISCOVER 25 MIN

Archived Mar 3, 2026

Open-Domain Stress Test

Test an AI assistant on one ambiguous research question that requires synthesis, not a single fact.

  1. Define a question with unclear boundaries and multiple possible sources.
  2. Require citations for every key claim in the response.
  3. Rate the answer for evidence quality, not writing fluency.
Use Team Worksheet →
BUILD 25 MIN

Archived Feb 24, 2026

Retrieval Reality Check

Test whether one RAG workflow retrieves evidence that is useful, current, and specific enough to support its answer.

  1. Choose three questions that require different source documents to answer well.
  2. Inspect the retrieved passages before reading the generated response.
  3. Fix one retrieval failure involving ranking, chunking, metadata, or source freshness.
Audit Retrieval →
LIVE 20 MIN

Archived Feb 24, 2026

Spillover Circuit Breaker

Interrupt one recurring pattern where stress from one part of life contaminates another.

  1. Name the situation that reliably depletes your emotional battery.
  2. Identify the first person or activity that receives the spillover.
  3. Design a ten-minute transition ritual and use it three times this week.
Build Circuit Breaker →
CREATE 30 MIN

Archived Feb 24, 2026

Eight-Bar Escape

Turn one polished loop into a complete arrangement by designing contrast before adding detail.

  1. Duplicate the loop into a rough beginning, middle, and ending.
  2. Remove at least one element from each section to create contrast.
  3. Bounce the full structure before changing any sound design.
Finish the Arrangement →
DISCOVER 20 MIN

Archived Feb 24, 2026

Agent or Software Decision

Determine whether one proposed AI agent is solving genuine uncertainty or avoiding straightforward engineering.

  1. Write the workflow as a sequence of inputs, decisions, and outputs.
  2. Mark every step where the correct behavior is already deterministic.
  3. Replace one unnecessary model decision with code, rules, or validation.
Run Decision Test →
BUILD 25 MIN

Archived Feb 17, 2026

Authorization Boundary Map

Make one sensitive action path explicit from identity through policy to the protected resource.

  1. Choose one action that changes data, permissions, money, or system state.
  2. Map the principal, delegated rights, policy checks, and final resource.
  3. Add one missing denial test or audit event at the weakest boundary.
Map Authorization →
LIVE 25 MIN

Archived Feb 17, 2026

Weekday Meal Friction Audit

Redesign one weekday meal around repeatability, stable energy, and the reality of a busy schedule.

  1. List every decision and preparation step required by the current meal.
  2. Remove or pre-stage the two steps most likely to break on a tired day.
  3. Repeat the revised meal twice and record energy, hunger, and effort.
Reduce Meal Friction →
CREATE 25 MIN

Archived Feb 17, 2026

Creative Inheritance Map

Make the influences behind one piece of work visible, then recombine them with greater intention.

  1. Name three works, people, or traditions shaping the piece.
  2. Write down the specific technique or judgment inherited from each.
  3. Combine two influences in a way neither source used directly.
Map Your Influences →
DISCOVER 20 MIN

Archived Feb 17, 2026

Stack Retention Review

Evaluate one tool in your stack by evidence of durable value rather than familiarity or hype.

  1. Choose one paid tool, framework, or service that has not been reviewed recently.
  2. List the outcomes it enabled and the maintenance or switching costs it created.
  3. Decide to retain, replace, reduce, or remove it and document the reason.
Review the Stack →
BUILD 25 MIN

Archived Feb 10, 2026

Configuration Drift Drill

Find one environment difference that could turn an ordinary deployment into an incident.

  1. Compare one production configuration against staging or its declared source.
  2. Identify a value, secret, feature flag, or dependency that has drifted.
  3. Codify the correction and add a check that detects the drift automatically.
Run Drift Drill →
LIVE 20 MIN

Archived Feb 10, 2026

Burnout Cost Ledger

Translate one unsustainable work pattern into its real cost to energy, quality, and future capacity.

  1. Choose one recurring behavior that looks productive but leaves you depleted.
  2. Estimate its weekly cost in recovery time, mistakes, and displaced priorities.
  3. Redesign the workload, boundary, or expectation that keeps recreating it.
Build the Ledger →
CREATE 25 MIN

Archived Feb 10, 2026

Imperfect Shot Study

Create a more believable image by replacing polished perfection with specific photographic imperfection.

  1. Choose one image prompt that currently feels synthetic or overproduced.
  2. Add lens, exposure, framing, texture, and environmental imperfections.
  3. Generate a small comparison set and identify which flaw increased realism.
Run the Shot Study →
DISCOVER 20 MIN

Archived Feb 10, 2026

Governance Failure Map

Trace one security risk back to ownership, incentives, and enforcement rather than stopping at the missing control.

  1. Choose one recurring exception, ignored alert, or unresolved security task.
  2. Identify who owns the decision, who bears the cost, and who can defer action.
  3. Change one escalation rule, ownership boundary, or review cadence.
Map the Governance Gap →
BUILD 30 MIN

Archived Feb 3, 2026

Semantic Search Prototype

Prototype one intent-based code search that exact text matching cannot answer reliably.

  1. Write five natural-language queries developers actually ask about the codebase.
  2. Create a small local index over one bounded directory or component family.
  3. Compare semantic results with grep and document where each method wins.
Prototype Search →
LIVE 20 MIN

Archived Feb 3, 2026

Emotional Regulation Loop

Treat one emotional trigger as a system with inputs, state changes, feedback, and recovery paths.

  1. Describe the trigger, body signal, interpretation, and usual response.
  2. Identify the earliest point where a different action remains possible.
  3. Precommit one small intervention and test it during the next occurrence.
Design the Loop →
CREATE 30 MIN

Archived Feb 3, 2026

Character Consistency Sheet

Build a visual reference grid that makes one generated character stable across pose, angle, and expression.

  1. Define immutable facial, wardrobe, proportion, and material details.
  2. Generate front, profile, three-quarter, full-body, and expression views.
  3. Use the strongest grid as the reference for one new scene.
Build the Grid →
DISCOVER 20 MIN

Archived Feb 3, 2026

Latency Tolerance Test

Measure where user trust actually begins to decay instead of relying on stated patience.

  1. Choose one workflow with a noticeable wait and define three latency bands.
  2. Observe behavior at each band using completion, abandonment, or retry signals.
  3. Improve feedback, progressive results, or processing time at the first trust break.
Test Real Tolerance →
BUILD 20 MIN

Archived Jan 27, 2026

Ticket as Contract

Rewrite one implementation ticket so architecture intent and operational boundaries survive the handoff.

  1. Choose a ticket with ambiguous scope, dependencies, or acceptance criteria.
  2. Add the invariant, constraints, failure behavior, and ownership boundary.
  3. Ask an implementer what decision is still missing before work begins.
Rewrite the Ticket →
LIVE 25 MIN

Archived Jan 27, 2026

Life Architecture Review

Check whether your current ambition is building the life and relationships you actually want to inhabit.

  1. Write the future your current calendar and spending are constructing.
  2. Name one important relationship or value that receives leftover capacity.
  3. Reallocate one recurring block of time toward what success is meant to serve.
Review the Architecture →
CREATE 25 MIN

Archived Jan 27, 2026

Compounding Craft Calendar

Design a sustainable week of creative and technical repetitions that can compound without requiring perfect motivation.

  1. Choose one music, writing, or technical skill worth twelve weeks of attention.
  2. Schedule three small repetitions and one feedback session each week.
  3. Define the minimum version that still counts on a low-capacity day.
Build the Calendar →
DISCOVER 20 MIN

Archived Jan 27, 2026

Assumption Kill Test

Replace one confident business-plan narrative with a test that could prove the underlying assumption wrong.

  1. Choose the claim most responsible for revenue, adoption, or cost projections.
  2. Write the observable evidence that would contradict the claim.
  3. Run the cheapest test that can produce that evidence this week.
Test the Assumption →
BUILD 25 MIN

Archived Jan 20, 2026

Identity Trust Graph

Map one service path around explicit identities and permissions rather than assumed network trust.

  1. Choose one request that crosses users, devices, workloads, or services.
  2. Draw every identity, credential, delegation, and resource in the path.
  3. Remove one implicit trust edge or add verification where identity is unclear.
Draw the Trust Graph →
LIVE 20 MIN

Archived Jan 20, 2026

Modification Friction Audit

Notice where your instinct to customize creates agency and where it creates unnecessary cost for you or someone else.

  1. List three things you changed recently instead of using as designed.
  2. For each, name the gained control and the added time or relationship cost.
  3. Choose one system to accept, one to adapt, and one to rebuild deliberately.
Audit the Modifications →
CREATE 20 MIN

Archived Jan 20, 2026

Compression Without Distortion

Explain one complex idea simply without removing the relationship that makes it true.

  1. Write the idea in full, including its important dependencies and exceptions.
  2. Rewrite it in three sentences: claim, mechanism, and boundary condition.
  3. Give it to one reader and ask what relationship still feels unclear.
Run the Compression →
DISCOVER 25 MIN

Archived Jan 20, 2026

Resistance Stack Interview

Separate technical risk, identity threat, and incentive misalignment in one stalled AI initiative.

  1. Interview one person resisting the change without trying to persuade them.
  2. Sort each concern into technical, identity, incentive, or coordination layers.
  3. Address the strongest nontechnical layer before proposing another tool or demo.
Interview the Resistance →