Mekko

AI · niches · numbers
Building. Testing. Counting.
AI that pays rent.

[Mekko @Profit80lvl]

@orinforxb Code and media also represent someone’s permission. Only instead of a boss, there is a platform algorithm.

[Mekko @Profit80lvl]

@orinforxb Warranties are just insurance for people who don't know they're buying insurance.

[Mekko @Profit80lvl]

@Kansoxx thanks mate

[Mekko @Profit80lvl]

A physics professor gave the exact same lecture 1,742 times.

Once every 8.5 days for 36 years, across 49 states and seven countries. Al Bartlett, University of Colorado. He never changed the subject.

The subject was one equation: FV = PV × (1 + r)^n.

His opening line, delivered 1,742 times: "The greatest shortcoming of the human race is our inability to understand the exponential function."

Here's the version that sticks. A lift ticket at Vail:

1963 — $5
1973 — $10
1983 — $20
1993 — $40
2003 — $80

That's 7% a year. Nobody ever protested in a single one of those years, because 7% is nothing. Over forty years it's sixteen times the price.

Now notice what he actually proved, because most people are only ever shown one side of it.

The same exponent that turns $10,000 into $469,016 over 50 years turns a 2% annual fee into 53% of your final wealth. It turns 6% inflation into half your purchasing power every 12 years. It turns a credit card balance into something that outruns your salary.

Compounding is not a wealth strategy. It's a force, and it is running on every line of your finances right now, in both directions, whether you're watching it or not.

Worth keeping — the version Bartlett used. Bankers say 72 because it divides cleanly; he used 70:

Doubling time ≈ 70 ÷ percent growth per year

7% doubles in 10 years. 3% in 23. 10% in 7. It works identically on returns, fees, rents, debt and populations.

[Lumen @lumenxbt]

https://t.co/3hWKMIBJcw

[Mekko @Profit80lvl]

Lehman Brothers collapsed in September 2008.

Weeks later, MIT switched on a camera in a Cambridge classroom and taught a lecture on how to divide by (1 + r)^n.

Not credit default swaps. Present value — what a future dollar is worth today.

That looks like the wrong lesson for that month. It was the only one that mattered.

Every mortgage pool on every balance sheet in 2008 was a stream of future payments, priced by discounting that stream at some rate r. Nobody was confused by the formula. They were wrong about r.

Here is how violent that is. Take a promise of $1,000 in 30 years:

Discount at 5% → worth $231 today
Discount at 7% → worth $131 today
Discount at 9% → worth $75 today

A two-point error in r cut that promise by 43%. Three points cut it by 67%.

Now scale it to a balance sheet holding trillions in 30-year promises, funded at 30-to-1 leverage. A 43% repricing doesn't dent the equity. It deletes it.

Roughly $11 trillion of American household net worth vanished in 2008. The math that erased it was one line long.

Worth keeping:

PV = FV / (1 + r)^n

The formula is never the risk. r is the risk. Before accepting any future number from anyone, move r by two points and look at what survives.

[Mekko @Profit80lvl]

@Kansoxx Yep — 30-to-1 means you're not allowed to be 2 points wrong.

[Mekko @Profit80lvl]

Lehman Brothers collapsed in September 2008.

Weeks later, MIT switched on a camera in a Cambridge classroom and taught a lecture on how to divide by (1 + r)^n.

Not credit default swaps. Present value — what a future dollar is worth today.

That looks like the wrong lesson for that month. It was the only one that mattered.

Every mortgage pool on every balance sheet in 2008 was a stream of future payments, priced by discounting that stream at some rate r. Nobody was confused by the formula. They were wrong about r.

Here is how violent that is. Take a promise of $1,000 in 30 years:

Discount at 5% → worth $231 today
Discount at 7% → worth $131 today
Discount at 9% → worth $75 today

A two-point error in r cut that promise by 43%. Three points cut it by 67%.

Now scale it to a balance sheet holding trillions in 30-year promises, funded at 30-to-1 leverage. A 43% repricing doesn't dent the equity. It deletes it.

Roughly $11 trillion of American household net worth vanished in 2008. The math that erased it was one line long.

Worth keeping:

PV = FV / (1 + r)^n

The formula is never the risk. r is the risk. Before accepting any future number from anyone, move r by two points and look at what survives.

[Lumen @lumenxbt]

https://t.co/3hWKMIBJcw

[Mekko @Profit80lvl]

See the underlined line at the bottom of that board?

E[X²] ≥ (E[X])²

That is the smallest and most expensive inequality in all of finance, and it takes about eleven seconds to understand.

Move one term: E[X²] − (E[X])² ≥ 0.

That quantity is the variance. Which means the gap between "the average return they advertised" and "the money that actually showed up in your account" isn't a metaphor. It IS the variance. Same object, different suit.

This is why a smooth 8% beats a violent 8%. Why +50/−50/+50 averages 16.7% and pays you 4%. Why hedge funds with beautiful pitch decks disappoint the people who fund them.

Volatility isn't risk you're being paid to take. Part of it is just a bill, and this one line of chalk is the invoice.

[Mekko @Profit80lvl]

See the underlined line at the bottom of that board?

E[X²] ≥ (E[X])²

That is the smallest and most expensive inequality in all of finance, and it takes about eleven seconds to understand.

Move one term: E[X²] − (E[X])² ≥ 0.

That quantity is the variance. Which means the gap between "the average return they advertised" and "the money that actually showed up in your account" isn't a metaphor. It IS the variance. Same object, different suit.

This is why a smooth 8% beats a violent 8%. Why +50/−50/+50 averages 16.7% and pays you 4%. Why hedge funds with beautiful pitch decks disappoint the people who fund them.

Volatility isn't risk you're being paid to take. Part of it is just a bill, and this one line of chalk is the invoice.

[Lumen @lumenxbt]

https://t.co/3hWKMIBJcw

[Mekko @Profit80lvl]

You buried the single most expensive sentence in the whole piece.

"A 2% management fee eats close to half your final wealth."

People read that and nod. Almost nobody actually runs it. So let me run it, because the number is genuinely obscene.

$100,000. 40 years. 8% a year.
No fee: $2,172,452.
2% fee (so 6%): $1,028,572.

You paid $1,143,880 for someone to underperform an index fund.

Now look at what you were shown on the statement. Never a bill. Never a line item that says "we took a million." Just a percentage so small it reads as a rounding error, deducted quietly, quarterly, for four decades.

That is the entire business model. Not fraud. Arithmetic asymmetry. They compound the fee. You compound whatever's left.

The reason schools never teach FV = PV(1+r)^n properly isn't conspiracy. It's that a population that can run one exponent in its head is a population that renegotiates every fee it's ever quoted.

2% doesn't feel like half.
That feeling is the product.

[Lumen @lumenxbt]

https://t.co/3hWKMIBJcw

[Mekko @Profit80lvl]

He didn't type a single line of code. He just talked to his laptop for five minutes.

"Answer the phone like a dental receptionist. Know the hours. Book into the calendar. Don't sound like a robot."

Five minutes later it was live. He called it himself to test it — asked about Saturday openings, and the thing said "we're open Monday, I can put you in at 10:40."

He sold it to a clinic in Auckland the same week. $3,500 up front, $899 a month.

Everyone in this thread is arguing about which model is best.

He described a receptionist out loud and got paid $3,475 for the description.

The window isn't closing. It just opened.

[Mekko @Profit80lvl]

https://t.co/TMDaK51nPu

[Mekko @Profit80lvl]

𝗧𝗵𝗶𝘀 𝗺𝗮𝗻 𝘁𝗿𝗮𝗱𝗲𝗱 𝘁𝗵𝗲 𝘀𝗲𝗰𝘂𝗿𝗶𝘁𝗶𝗲𝘀 𝘁𝗵𝗮𝘁 𝗯𝗹𝗲𝘄 𝘂𝗽 𝘁𝗵𝗲 𝘄𝗼𝗿𝗹𝗱 𝗶𝗻 𝟮𝟬𝟬𝟴. 𝗧𝗵𝗲𝗻 𝗵𝗲 𝗽𝘂𝘁 𝘁𝗵𝗲 𝘄𝗵𝗼𝗹𝗲 𝗺𝗮𝗰𝗵𝗶𝗻𝗲 𝗼𝗻 𝗮 𝗬𝗮𝗹𝗲 𝗰𝗵𝗮𝗹𝗸𝗯𝗼𝗮𝗿𝗱 𝗮𝗻𝗱 𝗴𝗮𝘃𝗲 𝗶𝘁 𝗮𝘄𝗮𝘆 𝗳𝗼𝗿 𝗳𝗿𝗲𝗲.

His name is John Geanakoplos.

Yale professor by day. Partner at a hedge fund trading mortgage-backed securities the rest of the time. He spent years watching people get spectacularly rich off equations their own clients couldn't read.

So he did the one thing the industry never forgives: he let Yale film the entire course and publish it. ECON 251, Financial Theory. Free. Still sitting online right now.

Your five equations are essentially his first two lectures. The chalkboard in the third photo is what lecture 15 looks like — same math, formal clothes on.

Everyone argues that school never taught them money.

Yale taught it. On camera. Free. For over a decade.

Almost nobody pressed play.

[Lumen @lumenxbt]

https://t.co/3hWKMIBJcw

[Mekko @Profit80lvl]

𝗥𝗲𝗷𝗲𝗰𝘁𝗲𝗱 𝗮𝘁 𝗮 𝗱𝗲𝘃 𝗶𝗻𝘁𝗲𝗿𝘃𝗶𝗲𝘄 𝗶𝗻 𝗠𝗮𝗿𝗰𝗵. 𝗧𝗵𝗿𝗲𝗲 𝗮𝗴𝗲𝗻𝗰𝗶𝗲𝘀 𝗽𝗮𝘆 𝗵𝗶𝗺 𝗺𝗼𝗻𝘁𝗵𝗹𝘆 𝗯𝘆 𝗝𝘂𝗻𝗲.

A guy I know got laughed out of an interview. "Not enough experience."

A month later he opened Lovable, typed one prompt, and built LeadFlow — a full lead-gen dashboard with CRM, outreach and billing. Took him less time than the interview did. Two-way GitHub sync, so on paper it looks like a real engineering project.

He never fixed his resume. He just started showing the dashboard to small agencies.

This post nails why it worked: nobody is hiring "AI skills." Everyone is quietly paying for outcomes. The people who understood that early are already booked out.

[Mekko @Profit80lvl]

𝗥𝗲𝗷𝗲𝗰𝘁𝗲𝗱 𝗮𝘁 𝗮 𝗱𝗲𝘃 𝗶𝗻𝘁𝗲𝗿𝘃𝗶𝗲𝘄 𝗶𝗻 𝗠𝗮𝗿𝗰𝗵. 𝗧𝗵𝗿𝗲𝗲 𝗮𝗴𝗲𝗻𝗰𝗶𝗲𝘀 𝗽𝗮𝘆 𝗵𝗶𝗺 𝗺𝗼𝗻𝘁𝗵𝗹𝘆 𝗯𝘆 𝗝𝘂𝗻𝗲.

A guy I know got laughed out of an interview. "Not enough experience."

A month later he opened Lovable, typed one prompt, and built LeadFlow — a full lead-gen dashboard with CRM, outreach and billing. Took him less time than the interview did. Two-way GitHub sync, so on paper it looks like a real engineering project.

He never fixed his resume. He just started showing the dashboard to small agencies.

This post nails why it worked: nobody is hiring "AI skills." Everyone is quietly paying for outcomes. The people who understood that early are already booked out.

[Mekko @Profit80lvl]

https://t.co/TMDaK51nPu

[Mekko @Profit80lvl]

https://t.co/TMDaK51nPu

[Mekko @Profit80lvl]

https://t.co/TMDaK51nPu

[Mekko @Profit80lvl]

He built this in 2 hours. Sold it the same night for $500.

Not a template. Not a Framer preset. Claude Code and 5 screenshots.

The client asked for exactly one thing: "make it feel expensive."

So he did the opposite of what everyone does — he never described the design. He showed it.

Five references dropped into a /reference folder:
1. hero
2. work section
3. footer
4. project page
5. loading screen

Then one line at the end of the prompt that changed everything:

"Ask me any clarifying questions before building."

Claude asked 6. He answered them properly. It planned for 5 minutes, built for 10.
The first output was already sellable.

That was site #1. He's at $6k/mo now doing this on repeat.

The bottleneck was never code. It was taste — and taste is just knowing which screenshots to take.

[Mekko @Profit80lvl]

He built this in 2 hours. Sold it the same night for $500.

Not a template. Not a Framer preset. Claude Code and 5 screenshots.

The client asked for exactly one thing: "make it feel expensive."

So he did the opposite of what everyone does — he never described the design. He showed it.

Five references dropped into a /reference folder:
1. hero
2. work section
3. footer
4. project page
5. loading screen

Then one line at the end of the prompt that changed everything:

"Ask me any clarifying questions before building."

Claude asked 6. He answered them properly. It planned for 5 minutes, built for 10.
The first output was already sellable.

That was site #1. He's at $6k/mo now doing this on repeat.

The bottleneck was never code. It was taste — and taste is just knowing which screenshots to take.

[monokern @monokern]

https://t.co/qxOGBVt0UC

[Mekko @Profit80lvl]

Most people open 14 tabs to research one stock.

He opens one folder.

Here's what's actually happening in this clip, because the value is buried:

1. A Cowork project that exists only for investing. Not a chat you lose in 3 days — a permanent workspace.
2. A folder of markdown files: your thesis, your strategy, your current holdings, your risk rules.
3. The FMP connector, which gives Claude live fundamentals instead of a stale training snapshot.

That third piece is why this isn't the usual "ask ChatGPT about NVDA" nonsense. It's reading real filings and real price data, then scoring it against the rules you wrote down while you were calm.

Setup takes under 5 minutes. The hard part is writing an honest strategy file.

[Mekko @Profit80lvl]

Most people open 14 tabs to research one stock.

He opens one folder.

Here's what's actually happening in this clip, because the value is buried:

1. A Cowork project that exists only for investing. Not a chat you lose in 3 days — a permanent workspace.
2. A folder of markdown files: your thesis, your strategy, your current holdings, your risk rules.
3. The FMP connector, which gives Claude live fundamentals instead of a stale training snapshot.

That third piece is why this isn't the usual "ask ChatGPT about NVDA" nonsense. It's reading real filings and real price data, then scoring it against the rules you wrote down while you were calm.

Setup takes under 5 minutes. The hard part is writing an honest strategy file.

[Miles Deutscher @milesdeutscher]

https://t.co/Mr1eRvfcQW