Cognitive Barbell: The Automated Core Architect
automate the safe low-yield core
Chapter VIII
Cards marked with a teaser are in the Vault. Unlock everything — $97, one time.
automate the safe low-yield core
design asymmetric convex bets
eliminate the draining middle tasks
frame premium sponsor/partner deals
configure an autopilot RAG loop
Filter a week of inputs against your situation — 3–5 decision-changing signals, 1–2 anti-signals, one blind-spot input
Turn one signal into 3–4 three-year scenarios with a cross-scenario read (robust position vs speculative bet)
Test current commitments against scenarios — keep/reconsider positions, gaps to fill, 3–5 concrete 30-day actions
Run the full Signal Scan → Scenario Map → Position Stress-Test loop in one chat, no infrastructure
Builds a quality rubric from strong and weak examples
Builds business, audience, voice, decision, and quality context blocks
Eight-question self-review of your own process
Finds the simplest sufficient tool setup for a task
Converts recurring thinking into an SOP or playbook
Synthesizes any section across four research lanes into a direct-use practice
Extracts standards, principles, assumptions, and tradeoffs into a reusable block
Seven-stage frame-to-artifact thinking pipeline
Eight-question pressure test that refuses to agree by default
Six-point weekly thinking log designed to fire as a scheduled agent
Labels each obstacle structural vs narrative and finds the binding constraint
Exposure map, immediate actions, and a binding one-page data policy
Sorts your AI usage into Surface, Structured, System tiers and plans promotion
Four exercises defining data and surveillance boundaries before pressure arrives
recombine elements for novelty
escape the most-likely sources
collide unrelated domains
force creativity via constraints
invert the problem
force divergent, non-generic output
fracture into multiple perspectives
autopsy the default assumptions
pre-mortem a plan
talk through a problem out loud
find gaps in your understanding
Socratic interrogation of your thinking
steel-man the opposing view
audit a belief against evidence
map second-order consequences
map your thinking in NotebookLM
excavate + rate your assumptions
pressure-test a decision
mirror/challenge/objection — don't fix
extract your writing voice fingerprint
spar with AI on an idea
check how original an idea is
a daily creative warm-up
who captures the value? does it compound?
Prompt 2: The Ownership Audit Pattern: Structured Decomposition Not all output is equal. Some of what you’re building belongs to you. Some of it disappears the moment you stop. This prompt helps you tell the difference. <role> You are a business strategist who specializes in helping solopreneurs and creators build assets they own rather than effort that evaporates. </role> <context> Everything I'm currently working on or producing: [LIST ALL PROJECTS, TASKS, CONTENT, SERVICES] My primary income sources: [LIST] Things I'm building that could generate income without my direct time: [LIST OR "NONE YET"] </context> <task> Audit everything I'm working on and classify each item by who actually benefits and whether it builds something I own. </task> <methodology> FOR EACH item on my list: 1. Who captures the value: me or someone else? 2. Does this build an asset (something that works without me) or does it require my time every time? 3. Does this compound over time or reset to zero? 4. Am I the only one who could do this, or am I doing it because it's available to me? </methodology> <output_format> ## Ownership Audit ### Builds Assets I Own (keep and expand) [Items with rationale] ### Generates Value for Others First (evaluate) [Items with honest assessment] ### Resets to Zero Every Time (reduce or eliminate) [Items with alternatives] ### Where AI Capacity Should Be Redirected [Top 3 recommendations] </output_format>
expand on purpose, not by default
Prompt 3: The Intentional Expansion Plan Pattern: Goal-Forward Design Once you know what you’re building, this prompt turns AI-recovered time into a deliberate plan rather than a list that grows on its own. <role> You are an execution strategist who helps self-directed professionals convert recovered time into meaningful progress on things they actually own. </role> <context> My top goal for the next 90 days: [ONE GOAL ONLY] Time I currently waste or spend on low-ownership work per week: [ESTIMATE IN HOURS] AI tools I use that save me time: [LIST] </context> <task> Design a 90-day intentional expansion plan that redirects AI-recovered time toward my stated goal. Every hour AI frees up has a pre-assigned destination before it gets filled with something else. </task> <output_format> ## 90-Day Intentional Expansion Plan ### Weekly Time Budget [Hours freed by AI, hours pre-assigned to goal, hours protected for rest] ### Weekly Actions That Directly Advance the Goal [Specific, not general] ### What to Say No To This Quarter [Three categories of work to decline or reduce] ### 30/60/90 Day Milestones [Concrete checkpoints] ### How to Know It's Working [Observable signals, not vanity metrics] </output_format> <constraints> One goal. Not three. If I list multiple goals, ask me to pick one for this quarter and park the others. </constraints>
make the week build assets
Prompt 5: The “Work for Me” Weekly Plan Pattern: Constraint-Based Scheduling This is your weekly planning prompt. It builds your schedule around your goals first, then fits everything else around them -- not the other way around. <role> You are a scheduling strategist who builds weekly plans for self-directed workers where personal goals come before reactive tasks. </role> <context> My top priority this week (tied to my 90-day goal): [ONE THING] Non-negotiable personal commitments this week: [HEALTH, FAMILY, REST, ETC.] Tasks I know are coming that serve someone else's needs: [LIST] Hours I want to work this week (hard cap): [NUMBER] </context> <task> Build a weekly plan where my top priority gets protected time first, personal commitments are treated as non-negotiable, and reactive work fills only what's left. </task> <output_format> ## Weekly Plan ### Protected Time (goal work, immovable) [Days and times blocked for priority work] ### Non-Negotiable Personal Time [Scheduled, not aspirational] ### Reactive Work Window [When you handle requests, messages, and tasks for others] ### What Doesn't Fit This Week [Explicitly parked until next week or later] ### This Week's Definition of Done [How you'll know the week was successful] </output_format> <constraints> If my reactive task list is longer than my protected time, flag it. The plan should reflect what I value, not just what's being asked of me. </constraints>
check compounding over time
Prompt 7: The Compound Life Check-In Pattern: Ritual Design This quarterly prompt steps back from the work entirely and asks the bigger question: is what you’re building with AI actually moving you toward the life you want? <role> You are a life and business advisor who helps driven people stay connected to why they started building in the first place. </role> <context> What I was doing 12 months ago: [WORK, INCOME, DAILY LIFE] What I'm doing now: [SAME CATEGORIES] What I hoped AI would change about my life: [HONEST] What has actually changed: [HONEST] What I want my life to look like 12 months from now: [SPECIFIC] </context> <task> Conduct a quarterly compound life check-in. Measure not just productivity, but whether my AI-augmented work is actually moving me closer to the life I want to be living. </task> <output_format> ## Quarterly Compound Life Check-In ### What's Actually Compounding in My Favor [Real progress, not activity] ### What's Keeping Me Busy Without Moving Me Forward [Honest assessment] ### The Gap Between What I Said I Wanted and What I'm Building Toward [Direct feedback] ### One Adjustment That Would Make the Next Quarter More Aligned [Specific and actionable] ### What "More" Still Needs to Mean for Me [Revisit and update your definition] </output_format> <constraints> Don't let me off the hook with vague answers. If my responses suggest I'm building for the wrong reasons or toward someone else's version of success, say so directly. </constraints> Prompting Tips Tip 1: Start with Prompt 1, then let it lead Prompt 1 (”Define Your More”) is the foundation. Every other prompt builds on what you discover there. If you skip it and jump straight to scheduling or income mapping, you’re optimizing for a target you haven’t actually set. Do Prompt 1 once, save the output, and reference it every time you run the others. Tip 2: Run these inside a Claude Project or persistent chat These prompts compound when they talk to each other. Store your “Define Your More” output, your Ownership Audit results, and your 90-day plan in one place. When you run Prompt 5 (weekly planning) or Prompt 7 (quarterly check-in), paste in the earlier outputs as context. The AI stops giving you generic advice and starts giving you advice that fits your actual situation. Tip 3: Honest inputs only Every one of these prompts has a context block that asks you to be specific and honest. The more you tell it what you want it to hear, the more it will tell you what you want to hear. List the tasks you’re embarrassed are still on your plate. Name the income that belongs to someone else. Admit the hours. That’s where the real output lives. Tip 4: Use Prompt 7 every 90 days, not just when something feels wrong The Compound Life Check-In is most useful before you feel off track. Run it at the end of every quarter as a scheduled ritual, not an emergency response. The goal is to catch drift early, when a single adjustment fixes it, rather than after six months of building in the wrong direction.
define what 'more' should mean
map income by ownership
sustainable output system
map where AI touches your work + failure modes
stress-test worst-case AI error impact
scan contracts for AI liability
disclose AI use to clients
document AI-assisted decisions
why B beats A, dimension by dimension
expose gaps vs expert judgment fast
articulate why something is good
build a quality vocabulary
identify markers of quality
what no competitor will sacrifice
copy that doesn't read like an ad
map deep audience empathy
test an idea against Ogilvy's 5 questions
reverse-engineer a creative brief
collide constraints for new ideas
think as 5 disruptive personas
surface limiting assumptions
diagnose why you're stuck
stop premature exits from ideas
write the failure autopsy first
what structures are you missing?
the decision you're afraid to make
map business vulnerability
does this expand or contract options?
define genuine security
what breaks without you?
are you acting from scarcity?
attack your preferred choice
when each option would be right
imagine failure, work backward
steel-man every option
clarify the choice before consulting
cost/differentiation/focus positioning
train rapid OODA decision-making
strategy = what you sacrifice
cascade strategic choices
architect a durable advantage
decision framework from an 'already-decided' future
reverse-engineer the breakthrough that solved it
break a claim into checkable parts
cross-check sources + credentials
detect bias in a source
systematic verification workflow
spot the bias operating in your thinking
form your view before reading conclusions
measure your cognitive independence
resist conformity pressure
trace where your beliefs came from
weekly intelligence gathering (Observe)
synthesize observations (Orient)
prioritize content/actions (Decide)
act fast on decisions (Act)
continuous improvement of the loop
avoid disaster by asking how it fails
surface hidden costs of a decision
stress-test assumptions to a killer question
predict who/what will oppose a plan
plan an exit before you commit
biomimicry/analogical mapping → actionable insights
reverse-engineer strategy via narrative prototyping
blend a Core × Lens concept into breakthrough ideas
strip brand/jargon → concrete product specs
deconstruct assumptions → rebuild radical strategies
guerrilla strategy where constraints are the idea
red-team every hidden assumption
build novel solutions from first principles only
break the default analogy, invent new models
Socratic questions → fundamental truths
push past symptoms to root cause
lean 90-day plan to validate a hypothesis
find evidence that challenges your idea
non-obvious 2nd/3rd-order consequences
imagine the failure, then work backward
skeptical expert panel red-teams your idea
break a creative block with 5 unconventional starts
map 10-year industry trends + disruption opportunities
work backward from failure (few-shot)
question assumptions down to fundamentals
apply + synthesize 3-5 models on one challenge
Bayesian scenario planning with probabilities
game-theory 3-round response simulation
design decisions that preserve flexibility
map components→leverage points→consequences
sort commitments aligned/neutral/misaligned + yes/no screen
screen commitments via energy/opportunity-cost/regret/excitement
use AI to sharpen your thinking, not replace it
analyze a problem as 3 different experts
make AI steel-man the opposition before validating you
persistent setting so AI always pushes back
Find the gaps in a handwritten explanation — 3 places wrong/oversimplified, the fuller answer, a source to study
Connect sources to a hand-drawn mind map — nodes to link, branches to add, one assumption the sources contradict
Extend handwritten notes — 3 highest-potential ideas, one next question + one assumption to pressure-test each, no new prose
Translate concepts from an unrelated field into your problem — 3 structural parallels and the mechanism, not the metaphor
Generate 10 anti-optimization practices sized 15–90 min, screen-free, un-automatable, at least 3 slightly silly
find the hidden theme connecting 3 sources
delegate multi-variable project planning
learn a topic via guided questions, no direct answers
analyze a problem via first-principles/second-order/inversion
AI finds the top-5 flaws in your plan