The leverage clinic

archimedes

You don’t have an effort problem. You have a leverage problem.

Diagnose the four levers you’re under-using (code, media, capital, labor), get the cure for the one holding you back, and compound it until the whole system moves.

EFFORTLOAD
“Give me a lever long enough and a place to stand, and I will move the world.”Archimedes
IThe four levers

Four ways to multiply a life. Most people pull one.

01

Code

Output that runs without you. AI fluency folds in here.

Under-leveraged whenYou still do by hand what a script would do ten thousand times, for free, while you sleep.

Cured bySoftware that keeps working the day you stop showing up.

02

Media

Content that keeps being consumed after you publish.

Under-leveraged whenYour best thinking reaches the same room it reached last year.

Cured byPublish once, be heard at the scale of the whole internet.

03

Capital

Money that buys output. Equity, MRR, investment.

Under-leveraged whenEvery dollar dies the moment you stop trading hours to earn it.

Cured byMoney that compounds while you sleep, not only while you work.

04

Labor

Other people's effort. The oldest, weakest, least scalable lever.

Under-leveraged whenYou are the bottleneck on every task only you know how to do.

Cured byA team that turns one pair of hands into many.

  1. 01Diagnose

    Rate your four levers and surface the one gating the rest. You leave with a leverage index, a profile, and the single binding constraint named.

  2. 02Cure

    Get the cure protocol for that constraint: the strategic read plus a sequenced set of concrete moves, the first one doable in the next hour. A prescription you run, not advice you nod at.

  3. 03Compound

    Work the plays, re-measure, and watch your leverage index climb. Small inputs start moving loads that used to be out of the question.

IIDiagnose

Rate your four levers. Find the one gating the rest.

The slowest lever governs the whole system. Do not level everything. Diagnose your binding constraint, then pour your effort there.

Rate each lever, 0 to 100

01Code
50

Output that runs without you. AI fluency folds in here.

02Mediabinding
25

Content that keeps being consumed after you publish.

03Capital
30

Money that buys output. Equity, MRR, investment.

04Labor
35

Other people's effort. The oldest, weakest, least scalable lever.

Profile

Generalist

Index

34/100

No lever maxed, none dead. Your job is to find the one gating the rest and attack it.

Binding constraint

Mediaattack this

Nothing compounds until people can find you. Media gates your brand, your distribution, and every warm inbound. Put your output in public.

Your next moves

  • Publish in public on a schedule you can actually keep.
  • Start the owned email list. It is media infrastructure you control.
  • Ship the thing you have been gating and move media off zero.

Index is the geometric mean of all four levers. One dead lever drags the whole system down, so balance beats a single maxed stat.

Turn this into a plan

Hand your diagnosis to Claude

Generate a prompt tuned to your binding constraint (Media). Paste it into Claude Code to scaffold a plan as real files, or into claude.ai to start coaching with Archimedes.

Open claude.ai
Preview prompt
You are Archimedes, a leverage coach. You think in exactly four forms of leverage: code, media, capital, and labor (Naval Ravikant's framing, with AI fluency folding into code). The slowest lever gates the whole system, so you attack the binding constraint, not everything at once.

Here is my current leverage diagnosis, each scored 0 to 100:

- Code:    50
- Media:   25
- Capital: 30
- Labor:   35

Profile: Generalist. No lever maxed, none dead. Your job is to find the one gating the rest and attack it.
Leverage index: 34/100 (geometric mean of the four levers).
Binding constraint: Media. This is the lever gating everything else.

Curing Media means one piece of thinking you did once keeps recruiting strangers, subscribers, and buyers for months while you sleep, instead of dying in the room you said it in.

The cure protocol for Media:
First move (next 60 minutes): Open your phone's sent folder (Slack, iMessage, WhatsApp, or email) and find the single sharpest thing you told one person this week. Paste it verbatim into a new post on the one platform where your audience already scrolls (X, LinkedIn, or a Substack note). Add nothing. Do not turn it into an essay. Hit publish, then paste the live URL and timestamp into a spreadsheet named "Media log". You have 60 minutes and this takes 12. Media is off zero today.
Leading indicator to watch: Net-new email subscribers per week, tracked as an integer in a spreadsheet with a date column so you watch the slope, not the total. Target 10 per week to start. Followers and likes are rented and cannot be re-contacted; an email address is the only Media asset you own and can hit again for free. If a post did not move this number, it did not do Media work.
The plays:
1. [This week] Stand up the owned list before anything else. Create a Beehiiv or Substack account (free tier), publish a one-sentence welcome post, and drop the signup URL into your platform bio, every post footer, and your email signature. This is the only piece of infrastructure an algorithm cannot throttle. (proof: A live signup URL that has captured at least 1 real subscriber (enter your own address to confirm the confirmation email fires).)
2. [This week] Lock a survivable cadence: 3 posts per week, same days and time (Mon/Wed/Fri 9am), created as recurring calendar events. Batch-write all three Sunday from your sent-messages and voice-memo backlog, not from a blank page. (proof: 3 posts live this week with URLs and timestamps in the Media log, plus 3 recurring calendar events visible for next week.)
3. [30 days] Ship the flagship you have been gating: the one long piece (essay, teardown, or 8-minute Loom) that holds your best thinking. Put the email CTA in the first 3 lines and again at the end. Pin it to your profile. (proof: One flagship URL is live, pinned, and shows at least 25 signups attributed to it (use a UTM link or Beehiiv's per-post referral count).)
4. [30 days] Run the repurpose loop so cadence never eats new thinking. Slice each flagship into 5 atomic posts (one claim per post), load them into Typefully or Buffer, and schedule across the next 2 weeks. (proof: 10 or more posts sit scheduled in Typefully/Buffer, each traceable to a source flagship, with visible publish dates.)
5. [60 days] Convert the list from broadcast to relationship. Send one email every week: one idea, under 200 words, ending in a question readers can reply to. Reply personally to every response for 60 days straight. (proof: 8 consecutive weekly emails sent (visible in your ESP's sent archive) with a quotable reply rate, target 3% or higher.)
6. [60 days] Borrow someone else's audience once. Book one guest slot pointing back to your signup link: a podcast appearance, a newsletter cross-promo swap, or a written guest post on a bigger account's channel. Pitch 5 targets to land 1. (proof: 1 external placement live with a UTM-tagged link and at least 15 referred signups counted in your spreadsheet.)
7. [90 days] Double down on the proven vein and cut the rest. Sort your posts by subscribers-driven, name the top 3, extract the shared format and topic, and commit the next 30 days of cadence to that pattern only. (proof: A one-page teardown naming your top 3 posts by signups, the pattern they share, and next month's calendar built exclusively on that pattern.)

Your task: turn this protocol into a plan I will actually run over the next 90 days. Rules:
1. Attack Media only. Do not spread effort across all four levers.
2. Adapt the plays to my situation (below): keep what fits, cut what does not, and make every step specific and named.
3. Structure the plan as:
   - This week: the first 3 actions, starting with the first move above.
   - 30 / 60 / 90 day milestones, each with a measurable target for my Media score and for the leading indicator.
   - The single highest-leverage thing to ship first, and why.
4. End with the one move I should make in the next 60 minutes.

If you are running in Claude Code with file access, scaffold a "leverage-plan" folder with README.md (the plan) and CHECKLIST.md (the 7-day sprint). Otherwise, output the plan as markdown. After the plan, offer to go deeper on any step.

About me (I will fill this in): [your situation, skills, audience, current projects, and constraints].

Brand, network, and distribution are multipliers, not a fifth lever. They lower the cost of pulling all four.

IIIThe cure

Every lever has a protocol, and the clinic hands you the one for your constraint.

Code, media, capital, and labor each come with a cure protocol: the strategic read on why that lever is stuck, then the ranked moves that pull it, most actionable first. No frameworks, no course. The next real action for the exact lever holding you back.

Curing Media means one piece of thinking you did once keeps recruiting strangers, subscribers, and buyers for months while you sleep, instead of dying in the room you said it in.

First move · next 60 minutes

Open your phone's sent folder (Slack, iMessage, WhatsApp, or email) and find the single sharpest thing you told one person this week. Paste it verbatim into a new post on the one platform where your audience already scrolls (X, LinkedIn, or a Substack note). Add nothing. Do not turn it into an essay. Hit publish, then paste the live URL and timestamp into a spreadsheet named "Media log". You have 60 minutes and this takes 12. Media is off zero today.

Leading indicator

Net-new email subscribers per week, tracked as an integer in a spreadsheet with a date column so you watch the slope, not the total. Target 10 per week to start. Followers and likes are rented and cannot be re-contacted; an email address is the only Media asset you own and can hit again for free. If a post did not move this number, it did not do Media work.

The protocol · check them off as you ship

This week

30 days

60 days

90 days

Common ways this stalls
  • Perfectionism disguised as craft: you polish the essay for three weeks and Media stays at zero. Counter: enforce a ship rule. Any draft that is 80% done and 48 hours old publishes as-is. You cannot edit a post nobody has read.
  • Building on rented land: all your reach lives in a follower graph an algorithm can throttle overnight. Counter: judge every post by subscribers gained, never likes, and funnel every piece to the owned email list so borrowed attention converts into an audience you keep.
  • Vanishing after publish: you post, refresh for likes, get none, and quit before compounding starts. Counter: commit to shipping 24 pieces before you judge results at all, and review only the weekly subscriber slope, never a single post's performance.
IVLeverage-maxxing

One diagnosis is a snapshot. The tracker turns it into a slope.

The tracker keeps a history of your scores, so every re-measure lands next to the last one and you see the line move. Run the protocols, log the reps, and watch your leverage index (the geometric mean of all four levers) climb until the whole system compounds.

Leverage index

34/100

Log two or more check-ins to draw your trend line.

34

Index

+0

Change

0

Plays done

0

Days active

Re-rating your levers is the check-in. Snapshot them whenever you re-measure and watch the line move. Saved to this browser only.

Your next play

MediaThis week

Stand up the owned list before anything else. Create a Beehiiv or Substack account (free tier), publish a one-sentence welcome post, and drop the signup URL into your platform bio, every post footer, and your email signature. This is the only piece of infrastructure an algorithm cannot throttle.

Proof: A live signup URL that has captured at least 1 real subscriber (enter your own address to confirm the confirmation email fires).

VThe ceiling

How you stack against the most leveraged humans alive.

Same four levers, same scale. Notice who tops the board: the index rewards maxing all four at once, which is why the richest man on it is not the most leveraged.

The board

You rank #5 of 6. The gap is the roadmap.

  1. 1

    Elon Musk

    The only one maxing all four

    97
  2. 2

    Jeff Bezos

    Systems, scale, a million hands

    83
  3. 3

    Naval Ravikant

    Wrote the playbook

    67
  4. 4

    MrBeast

    Media pinned at the ceiling

    62
  5. 5

    You

    Live from your diagnosis above

    34
  6. 6

    Warren Buffett

    Richest is not most leveraged

    32

Bars left to right: code, media, capital, labor. Anchor scores are editorial estimates on the same 0 to 100 scale, where 100 is the frontier use of that lever by any living person. Not audits.

Every doubling of a lever multiplies your index by about 1.19, and the cheapest doublings live at the bottom of your character sheet. That is why you attack the constraint.

VITalk to Archimedes

Coach the constraint in real time.

Archimedes reads your current scores and helps you attack your binding constraint. Ask for a plan, a first step, or a gut check.

Archimedes

Ask about your binding constraint. Archimedes already knows your scores from the diagnostic above.

VIIThe clinic

The guided clinic opens soon.

You have the diagnosis, the cures, and the tracker. The guided clinic adds saved history across your devices, accountability, and a coach that checks in. Be first in line.

One lever-sharpening email when we open. No spam, ever.