Because this company has done it 20 times before and made them all profitable. Evernote. Meetup. WeTransfer. Remini.

This is the story of Bending Spoons.

A startup repair factory that buys companies that are dying, revives them, then holds them forever.

They’ve been profitable since day one. 1% annual employee turnover. Didn’t raise venture capital until they were making $100 million a year. All powered by an unusual playbook.

Subscribe now

I break down the full story here

Watch on YouTube

Five takeaways from Bending Spoons’ playbook:

1. Finding product-market fit is the hard part. Execution is repeatable.

Bending Spoons’ first startup crashed. They raised a million dollars, went to market, and flopped. Founder Luca Ferrari called it “arrogant to think we knew what the market would want.”

So they stopped trying to predict what people wanted, and started buying what already had product market fit. Their first acquisition was a broken keyboard app for $15,000. They fixed it. Turned a profit. Did it again.

Most founders romanticize building from scratch. Bending Spoons decided the hardest part of startups isn’t building. It’s knowing what to build. They let others take that risk, and bet they can execute better.

2. If you’re buying to hold forever, you can afford to pay more.

Most bidders model a 3-7 year exit. Bending Spoons models lifetime ownership. They never sell. Luca calls it “25% private equity, 75% tech company.”

This changes the math. When you’re capturing decades of compounding value, you can outbid everyone focused on a quick flip. That’s why Bending Spoons has never lost an acquisition bid.

3. Standardize your tools. Customize your treatment.

When Bending Spoons acquires a company, a strike team of elite operators arrives and examines four things: organizational sprawl, inefficient costs, gross margin leaks, and under-monetization.The diagnosis is usually similar, but the treatment plan is always custom.

What makes it scale is that Bending Spoons has built 50 proprietary tools to power the repair process. Strike teams plug newly acquired companies into this platform. Within days, they know what’s broken and how to fix it. Once it’s profitable, the team moves onto the next opportunity, and the tools travel with them.

4. Talent density beats talent volume.

In 2025, Bending Spoons had 800,000 applicants. They hired 250. That’s a 0.03% acceptance rate, which is harder to get into than Harvard. What’s interesting is how they keep these elite operators. There are no bonuses or equity grants. Luca believes those incentives “make things transactional.”

Instead, raises are tied to impact, and employees can convert part of their salary into stock options at a discount. The result: 1% annual turnover. Most tech companies lose 15-20% of their people every year.

A small team of exceptional people, armed with the right tools, will outperform a large team without them.

5. Debt preserves control over long-term growth. Equity can cost it.

Bending Spoons didn’t raise venture capital until they were nearly 10 years in. They bootstrapped for five years, reinvesting profits from each acquisition into the next.

When they needed capital for bigger deals, they took out loans. Banks trusted their cash flows because they’ve always been profitable.

Today they’ve raised roughly $3 billion in debt for acquisitions and $700 million in equity for operations.

Most startups raise equity early, take on zero debt, and hope they grow fast enough before the money runs out. Bending Spoons flipped the model. They scaled without giving away much control.

Share

Hiten

=)

Transcript

Hiten: On the fifth floor, a man with glasses pops champagne. His company is now worth $1 billion. Phones around the room light up. Messages stream in from investors. “Congrats on building a rocket ship.”

Ten years later, the same man is working from home. Scrolling Reddit: “It’s so slow, it’s unusable.” “There’s so many bugs.” “I wish I switched to Notion earlier.”

He opens another tab to check financials, but Evernote hasn’t made a profit for years. He starts writing his resignation. Then an email hits his screen. It’s marked urgent. An acquisition offer from a company based in Milan.

No one on his team has heard of them, but the number is 50% higher than any other offer they’ve gotten. He accepts.

The takeover begins immediately. An Italian man fires most of the team. By noon, a new crew arrives. Europeans, fresh out of college.

Within two years, Evernote is profitable again.

But this isn’t the work of private equity. It isn’t a buy and flip, and it isn’t a one-off. It’s been done 20 times in 10 years, all by one company, across apps just like Evernote. Dying but not dead.

This is how Bending Spoons revives zombie apps at scale. By building a startup repair factory that’s been profitable since day one.

Luca Ferrari: “I remember we were making half a million a year. Now we’re making a billion.”

Hiten: But first, you quickly need to know why they don’t start from scratch. Because it unlocks the company’s entire playbook.

For Luca Ferrari and his co-founders, the insight came after their first startup failed in 2013. They raised a million dollars, went to market, and flopped.

Luca Ferrari: “We crashed and burned by being arrogant in thinking we knew what the market would want.”

Hiten: So they took the money that was left and founded Bending Spoons in 2013. Only this time, they didn’t build.

They bought a broken keyboard app for roughly $10,000, made it better, and turned a profit – used that money to buy a second app and bring it back to life, then did it over and over again.

Luca Ferrari: “We let others seek product-market fit, and then we acquire their company and try to make it even better.”

Hiten: This strategy powers the repair factory, and in 2025, it was valued at $11 billion.

So how did they go from fixing one zombie app to reviving 20 of them? What you’re about to see is Bending Spoons’ playbook.

First, let’s start with a surprising fact about the business model.

Luca Ferrari: “You know what? We have never lost a bid before.”

Hiten: How is this possible?

It’s because they see a different future than everyone else. Other buyers model an exit in three to seven years. Bending Spoons models lifetime ownership.

Luca Ferrari: “We acquired to hold forever. We never sell.”

Hiten: Luca calls this 25% private equity, 75% tech company – going deep on changes to product, engineering, and design. Decades of compounding improvements.

Luca Ferrari: “I think we’re a hybrid. It’s almost like private equity had a baby with Google.”

Hiten: This is why they sometimes pay double the next offer.

So how do they know what to buy? Acquisitions are guided by three signals.

• One: Software people already love with a strong brand and loyal user base.

• Two: Management is stuck and either can’t or won’t revamp the company.

• Three: A clear path to profitability based on changes they would make.

Luca Ferrari: “We buy businesses where almost all the value lies in existing customers or users.”

Hiten: In other words, they execute better on what already works.

The next question is: how do they do that?

Well, once they acquire a company, it goes on the operating table. Bending Spoons sends in a strike team like an emergency medical unit. They examine the patient for four things: organizational sprawl, inefficient costs, gross margin leaks, and under-monetization.

The diagnosis is similar across acquisitions, but the treatment is almost always custom.

Luca Ferrari: “It could be rewrite the software, re-architect the cloud infrastructure, launch lots of features, redesign the UI, optimize monetization and marketing, rebuild big chunks—sometimes the entirety of the organization.”

Hiten: Here’s three examples:

• At Remini, they rewrote every line of code and added features.

• At Meetup, they fixed 400 bugs and cut bloat.

• At WeTransfer, they reworked pricing and limited free usage.

Two things are consistent though. They improve the product for paying users, then raise prices on the core offering.

Luca Ferrari: “We’ve really focused on what these customers painfully needed.”

Hiten: The result? Retention always goes up.

But that still doesn’t explain how they make decisions at scale or move so fast.

What’s interesting is Bending Spoons doesn’t build consumer products from scratch, but they do build tools for reviving zombies.

Luca Ferrari: “More than 50 proprietary technologies, some of which could easily be sold as a B2B business—maybe worth hundreds of millions.”

Not all of them are public knowledge, but here’s what I found in my research.

Before they acquire:

• Lumen converts raw data into clean insights

• Crystal finds targets

After they acquire, three tools power the transformation:

• Pico answers: Will users pay for this?

• Minerva automates the big decisions – what to build, what to cut, and what to charge

• Pantheon gives every app access to the same AI infrastructure

Strike teams simply plug newly acquired companies into this platform. Within days, they know exactly what’s broken and how to fix it. Once it’s profitable, the strike team moves on and the tools travel with them.

Luca Ferrari: “It’s so easy to allocate people to where they can create the biggest impact. That makes a ton of difference.”

Hiten: The thing is though, tools don’t revive zombie apps on their own. Elite operators do.

Talent mobility is important, but talent density is the key to the entire repair factory.

Luca Ferrari: “We think about the jobs we offer as our most important product.”

Hiten: In 2025, 800,000 people applied. They hired 250. That’s a 0.03% acceptance rate – tougher to get into than Harvard.

Here’s how they find operators worth hiring. Custom AI screens applications for potential. It also filters out people it thinks won’t handle the intense culture. Next, candidates face assessments – called harder than Google, Amazon, or Meta. It’s testing three things: talent, motivation, and problem-solving skills.

New hires start in Milan and get a document called “Controversial Principles.” Then they spend three months soaking up the culture before joining a strike team.

Luca Ferrari: “You learn faster than anywhere else. Most of our general managers run businesses—on average $50 to $100 million in revenue.”

Hiten: The compensation makes it worth their while. Bending Spoons has been described as “a little Goldman Sachs in Italy,” paying London and New York salaries in Europe.

But what’s interesting is there are no bonuses or equity grants.

Luca Ferrari: “Those kind of incentives tend to hinder relationships. They tend to make things more transactional. It’s more difficult to have a proper problem-solving session where all we’re thinking about is: how do we win together?”

Hiten: So how do they keep their elite operators? Raises are tied to impact only. There’s also a unique system for employees to build meaningful stakes.

Luca Ferrari: “Anyone can convert a portion of their compensation, up to a certain limit, to stock options at a steep discount.”

Hiten: The result of this whole thing is pretty impressive: 1% annual turnover. Most companies lose 10 to 20% of their people every year. With a team of just 700, that’s Bending Spoons’ real moat. Talent density without churn.

But none of this works without funding. Which brings me to the final page in Bending Spoons’ unusual playbook: they didn’t raise venture capital until they were nearly 10 years in.

Luca Ferrari: “We really wanted to build this with a multi-decade view, and we felt that it was quite dangerous to relinquish control so early.”

Hiten: They bootstrapped for five years, reinvesting profits from each acquisition into the next. When they were ready to go after bigger deals, they took out loans. Remember, they’ve been profitable since day one, and banks trusted their cash flows.

By the time they raised their first equity round, Bending Spoons was making $100 million in annual revenue. And investors aren’t limited to VCs. There’s a bunch of celebrities who have also bought in, including The Weeknd, Ryan Reynolds, and Andre Agassi.

Most startups raise equity early, take on zero debt, and hope they grow fast enough before the money runs out. Bending Spoons flips that model.

Today they’ve raised roughly $3 billion in debt for acquisitions and $700 million in equity for operations – all without giving away much control.

Luca Ferrari: “We can reinvest in making our platform more powerful. So build better proprietary technologies, better access to talent, and go after new, bigger acquisitions.”

Hiten: That’s how Bending Spoons revives zombie apps at scale. And according to analysts, it’s a strong match for current markets.

There’s a growing number of dead unicorns and companies trading below their IPO. Insiders say Bending Spoons has 5,000 companies on their acquisition list.

2025 was their biggest year yet—over $3 billion invested in Vimeo, AOL, and Eventbrite alone.

The deals are getting bigger, but the thesis is still the same.

Subscribe now

At its peak, Terra Luna’s market cap was $40 billion.

Then, in May 2022, it collapsed to $0.

Subscribe now

I break down the full story here

Watch on YouTube

The story is a cautionary tale about crypto, or fraud, or hubris. But there’s a more uncomfortable truth buried underneath: Do Kwon ran a similar playbook to Steve Jobs.

Building a religion around your product isn’t new. Jobs did it. Elon does it. When it works, you get customers who tattoo your logo on their body and defend you in public for free.

The playbook is remarkably consistent:

Create an enemy. Jobs had IBM. Elon has legacy automakers. Do Kwon had the entire traditional finance system.

Build identity, not just utility. Nobody buys Apple or Tesla products because it’s the most cost-effective option. They buy it because of what it says about them. Terra Luna worked the same way. Everyday investors were joining a movement.

Make believers feel like insiders. Early Apple adopters felt like rebels against corporate conformity. Tesla owners felt like co-conspirators on a mission to save the planet. UST and Luna holders felt like they were playing a role in the future of money. That feeling of secret knowledge is intoxicating. It also blinds people to red flags.

Do Kwon executed this playbook almost perfectly. But the problem wasn’t the tactics: it was what they were built on top of.

Here’s where Terra Luna diverges from Apple.

When the iPhone launched, it delivered.

Do Kwon built his religion on a story that couldn’t survive contact with reality. His “algorithmic stablecoin” promised the holy grail of crypto: a fully decentralized stable currency you could actually use.

Turns out, the entire project was built on fraud. Faked transaction data. Secret payoffs to prop up the price. Stolen Bitcoin reserves.

So when investors started panic-selling, there was nothing to stop the collapse.

The founder lesson isn’t “don’t build a movement.” Movements work. They create moats and compounding advantages.

The lesson is simpler, and darker: movements amplify whatever’s underneath them.

If your product delivers real value, a religious following will defend you through rough patches and evangelize you into new markets. Apple fans forgave the hockey puck mouse. Tesla owners moved past panel gaps. Airbnb hosts fought regulations city by city.

But if your product doesn’t deliver the value promised, the community you built will turn on you.

Build movements. Just make sure the thing underneath them is real.

Share

Hiten

=)

Transcript

This is Nationals Park. Inside is the Terra Club. It’s a private lounge full of politicians and bankers. In the corner, a Korean man is sitting in a black hoodie. This man has just signed a $40 million deal to sponsor the home team. His nickname: the King of the Lunatics.

Two blocks away, a college student is booking flights home. He scrolls past Apple and Google Pay at checkout. Instead, the kid taps UST and pays with crypto.

5,000 miles away in Montenegro, Interpol agents are putting together a case. The target has been on the run for nearly a year, but the hunt is almost over. They know that tomorrow he’s on a flight to Dubai.

This is the rise and fall of Do Kwon, the man who built a $40 billion religion around a cryptocurrency that collapsed in a week.

In January 2025, he was on trial for his role in the crash and pleaded not guilty. But in August he did a U-turn and pleaded guilty to two fraud charges.

How did he convince millions to believe in his product, and what were the lies that brought it all down?

Before we go any further, you need to understand what Do Kwon promised everyday investors.

This is Chad. He’s in corporate making a decent salary. He’s just been denied a mortgage. Chad’s been looking into crypto as a way to earn more, but he’s skeptical. The news is full of meme coins. An image of a clock just sold for $52 million.

But one night scrolling Twitter, Chad sees something less crazy. It’s a thread from Do Kwon: “Banks lend your money to make profits. This is the system keeping you poor. Terra changes that.”

The thread explains how it works. Two tokens: UST, a stablecoin, and Luna, a volatile one. The price of Luna adjusts to keep UST stable. No banks, no government.

Chad skips ahead. All he wants to know is: can you actually use this? He goes on Reddit to find the answer. Turns out you can use UST to buy food, Apple shares, and even book flights. He sees that the private lounge inside Nationals Stadium was renamed the Terra Club.

Chad knows he should be nervous. This is his entire savings. But he pulls the trigger and transfers $100,000 into UST. His worries don’t last long. By the end of the month, he’s up 20%. A few more months and he can finally afford that house.

But here’s the thing — Chad wasn’t alone. From 2021 to 2022, the amount invested in UST grew to $18 billion. Millions had bought into Do Kwon’s promise: a stable cryptocurrency you could actually use, free from banks and government control.

So now you understand what made Terra so appealing. But a vision alone doesn’t get millions of people to go all in. What you’re about to see are the tactics Do Kwon used to turn Terra into a must-have for people like Chad. These are the same tactics Steve Jobs used at Apple, Elon Musk at Tesla, and Brian Chesky at Airbnb. The difference? Their products actually worked.

This is how Do Kwon built a religion around his product. His playbook was made up of four tactics.

First, he sold a story. Banks and governments were the enemy, and the hero could be anyone. If you bought UST. Do Kwon proved it. People were using UST to buy groceries in Korea on an app called Chai. In America, they were using it to trade stocks 24/7. This narrative helped Terra raise $200 million in venture capital.

Next, Do Kwon gave his followers an identity: Lunatics. And he showed them how to act — naming his daughter Luna, mocking critics in public with lines like “Enjoy being poor,” and even betting one critic $11 million that Terra would keep growing. Believers felt superior. Skeptics looked small. Eventually, the Lunatics did all of this for him.

The next step was to reward his followers. In March 2021, he offered 20% returns to anyone who held UST, 40 times what you could earn in a savings account. People asked, “How is this possible?” His answer: “It’s like dividends. As long as Terra grows, everyone wins.”

Once people had a stake, they became evangelists. And soon they were needed. In May 2021, Terra faced its first real test: UST became volatile. Critics said it would collapse, but the system stabilized. Do Kwon called it a triumph of decentralization. Lunatics spread the story everywhere. To them, it proved the system worked and their 20% was safe.

Finally, Do Kwon built social proof. In April 2022, he raised money and bought $3.5 billion of Bitcoin. The message: UST is backed by mainstream crypto. But the most powerful validation came from Wall Street and the MLB. Mike Novogratz tweeted a viral photo of his Terra tattoo. The Nationals announced a $40 million sponsorship. Every high-profile endorsement made doubters think, “What am I missing?”

By May 1st, 2022, the four tactics had worked. Millions of people held UST and Luna. Not because they fully understood how it worked, but because Do Kwon had created a sense of FOMO and social pressure. If you spoke out, you either looked dumb or against the future of money.

But here’s what Jobs, Musk, and Chesky understood that Do Kwon didn’t: You can create a connection, build empathy, and scale a narrative quickly. But eventually the fundamentals catch up.

So the product better work. But Terra’s didn’t. In less than one week, it all came crashing down.

It’s 6:00 PM in a small, crowded airport terminal. Do Kwon hears his boarding call to Dubai. He closes his laptop and walks toward the gate. Behind him, two plainclothes officers get up and follow.

Right as Do Kwon reaches for his boarding pass, the Interpol agents detain him. Do Kwon is on the floor in handcuffs. His fake passport skids across the tiles.

The question is: why is he being arrested?

On the 9th of May, 2022, both UST and Luna collapsed. Investors panicked and there was a massive sell-off. Terra’s market cap fell from $40 billion to zero.

As investigators dug in, they found something far worse than a failed cryptocurrency: a web of lies. And one by one, the SEC and DOJ exposed them.

Lie #1: Some of the use cases weren’t real. Remember the Korean payment app called Chai? The data was faked by Do Kwon. And the 24/7 stock market? Bots made it look like people were trading.

Lie #2: Demand wasn’t organic. Remember those returns Do Kwon compared to dividends? Investigators found Terra was paying for those 20% returns with runway, not revenue. It was subsidized growth just to get more people holding UST.

Lie #3: Do Kwon potentially stole from the Bitcoin reserve. After the crash, most of it disappeared. Investigators still can’t trace where it ended up.

Lie #4: He knew the two-token system would fail. It turns out Do Kwon had tried this exact stablecoin design before under the alias “Rick Sanchez.” But that project collapsed in under two months. Yet he still built and scaled Terra anyway.

Lie #5: When UST went volatile in 2021, the system didn’t fix itself like Do Kwon said. Instead, he made it look like it did — secretly paying a trading firm $100 million to buy UST. It wasn’t self-stabilizing and it wasn’t decentralized.

So when investigators pieced it all together, Do Kwon’s story fell apart. The Lunatics quickly turned their backs. Terra subreddits went from worshiping Do Kwon to hating him.

That’s why in August 2025, Do Kwon changed his mind. He pled guilty to fraud and conspiracy, admitting to misleading investors while leading Terraform Labs.

On December 11th, he was sentenced to 15 years in prison.

And Chad, who was up 20%, already planning for that house, never got his money back.

Subscribe now

Not on DraftKings. Not on FanDuel. On Kalshi — a federally regulated startup that moves more money than the entire Vegas Strip ($1B a month).

It looks like a casino. It acts like a casino. But Kalshi doesn’t have a gambling license. And that’s not the most unusual thing about it: Most startups in legal gray areas build first and fight regulators later (Uber, Airbnb, Robinhood).

Kalshi did the opposite. They spent years asking for permission before launching. Sixty-five lawyers said it was impossible. Then, in November 2020, they became the first federally approved “prediction market” in US history. Five years later, they just raised $1B at an $11B valuation.

This is how Tarek Mansour built Kalshi by optimizing for defensibility over speed — and why that bet is now being tested.

Subscribe now

I break down the full story here

Watch on YouTube

Five takeaways from Kalshi’s playbook

1. They convinced regulators when everyone said no

Sixty-five lawyers told the founders their idea was illegal. But Kalshi found a gap: the Commodity Futures Trading Commission (CFTC) had authority to define “gaming” and “wagering,” but never did. It was a legal loophole that went unnoticed for nearly a century. Over thirty meetings, they made the case that prediction markets aren’t wagering and instead measure uncertainty — similar to how traders hedge oil or gold. The CFTC agreed and awarded Kalshi a federal license in November 2020. By doing the hard thing first, Kalshi unlocked a new market (legally).

2. Why going slow created their moat

Federal approval gave Kalshi something competitors can’t replicate: legitimacy. By optimizing for defensibility over speed, Kalshi overtook Polymarket to become the dominant player in the space. As co-founder Tarek Mansour framed his early strategy: “If we want this thing to be big, we have to play by the book.” In October 2025, total trading volume passed $11B.

3. How they scaled by suing their own regulator

Kalshi’s growth isn’t linear; it’s shaped like a hockey stick. The inflection point came after the company sued their own regulator, made the same case, then won approval for political event contracts. For the first time in 100 years, it became legal to bet on elections in the US — just in time for the 2024 Presidential election. Trading volume has grown 200x since.

4. How Kalshi makes money

It’s easy to assume Kalshi operates like a casino, but legally it doesn’t. There’s no “house” taking the opposite side of your bet. Every contract matches two traders (yes and no). Kalshi then takes a small cut on every transaction, deposit, and withdrawal. The odds are set by the market, not the company. That’s why, by October 2025, Kalshi was generating roughly $13 million in monthly revenue entirely from taxing activity, not from pricing risk. And at a federal level, everything is above board (for now). By building alongside regulators, they essentially helped develop the compliance framework for their own market.

5. Why their defensibility strategy is now being tested

By October 2025, roughly 80% of trades were on sports. This is a problem for Kalshi, because states tax sports betting. Ohio, Nevada, Massachusetts, and New York are trying to stop Kalshi in their state (they want their cut). The more states that join in, the more scrutiny they face. Kalshi knows the stakes: they need their federal license to operate. That’s why Kalshi has stacked its board with former CFTC officials and gambling lobbyists (even Donald Trump Jr. as an advisor). If their license comes under threat, it’s a safe bet they will leverage those connections to protect it.

Share

Hiten

=)

Transcript

It’s 10 p.m. on a dark, rainy Wednesday night.

In a small apartment in New York, South Park plays on the TV.

A kid in his twenties sits at a desk, his MacBook wired into a cable box.

Picking up the same episode, only 28 seconds faster.

That head start is about to make him real money.

He’s running a custom script.

Lines of code, reading the show’s captions in real time.

Online, there’s a platform letting people bet on what’s going to be said in the episode:

“Trump.” “Elon.” “Tylenol.”

As soon as one of those words appears, the code buys low, before anyone on streaming can react.

Seconds later, the prices jump.

He’s already out, cashing in on the lag between cable and the cloud.

But this isn’t some underground gambling site.

It’s a federally regulated startup backed by Wall Street and Silicon Valley.

Every month, a billion dollars flows through this platform – more than the entire Vegas Strip.

It looks like a casino.

It acts like a casino.

But Kalshi doesn’t have a gambling license.

It has something far more powerful.

In this video, I’ll break down:

• How Kalshi makes money whether you win or lose

• The legal loophole that keeps them alive even as states try to ban them

• And the one thing that could make their house of cards collapse

But before we go further, you need to understand how a bet on Kalshi works.

Every “bet” is called an “event contract.”

It’s a yes-or-no position on something that might happen.

Like the New York City election or the Pro Football champion.

Each contract trades between 1 and 99 cents.

The price is the crowd’s odds.

For each option, Kalshi matches a trader who thinks “yes” with another who thinks “no.”

When new information hits – a poll, a touchdown, a tweet – prices move.

Then, when the event resolves, all of the money staked by the losing side funds the winners.

To make this concrete, let’s follow a single bet.

You’re betting that a South Park character says the word “Trump” in an episode.

You put down $1,000 when the price is 1 cent – 1% odds.

That gets you 100,000 “yes” contracts.

If “Trump” is said, each contract pays you a dollar.

And you walk away with a hundred grand.

If not, you lose it all.

That’s an extreme case – but it has happened.

Like when a punter put $500 on the next pope and made more than fifty grand.

Most of the time, prices land somewhere in the middle.

Say you bought in at 70 cents, and the odds drop to 40.

You could lock in a $430 loss by selling, or hold and hope it swings back.

That’s the whole machine.

A live market where people trade yes/no contracts.

And the outcome decides who gets paid.

Now that you know how Kalshi works, the question is: how do they actually make money?

Every month, roughly a billion dollars moves across the platform.

But trade volume isn’t the same as revenue.

Instead, Kalshi takes a small cut every time money changes hands – just like an online marketplace.

They charge a sliding fee on every transaction: 1 to 3.5%, depending on trade size.

There’s fixed fees as well: 2% on deposits, and a $2 withdrawal fee.

Win, lose, or just trade again – Kalshi earns every time money moves.

It’s a tax on activity, not on risk.

By my estimates, Kalshi’s pulling in about $13 million a month – roughly $160 million a year.

When this video was recorded, total trading volume passed $11 billion.

This earning mechanism is what separates Kalshi from traditional gambling.

In a casino, the house sets the odds and bakes in a tiny markup.

Every spin, roll or hand earns a few cents.

One loss doesn’t matter.

Over thousands of rounds, those cents stack into guaranteed profit.

Kalshi doesn’t do that.

There’s no house, so they have no stake in the outcome.

Remember, it’s just people trading against each other.

Knowing how Kalshi makes money differently to a casino is going to help you understand my next point:

Which is why they’re allowed to operate.

When Kalshi started, 65 lawyers told the founders the same thing:

*“*This idea isn’t new, and it’s illegal.”

So they took an unusual path for a startup.

As founder Tareck Mansoor put it:

“For this thing to be big, we need to do it by the book.”

In mid-2019, Kalshi raised a small seed round and began building.

A team of roughly ten worked on shipping the product.

But Kalshi’s breakthrough didn’t come from tech, it came from the law.

Buried in the Commodity Exchange Act is a 90-year-old clause,

giving the Commodity and Futures Trading Commission, the CFTC, power to define “gaming” and “wagering.”

But the law never actually defined those words.

That gap was Kalshi’s opening.

Over the next year, the team met more than thirty times with the CFTC.

Each meeting was part negotiation, part education – showing how Kalshi’s contracts weren’t wagers, but predictions.

They leaned on two core arguments:

One: Similar to how traders hedge gold and oil, each contract is a tool to measure uncertainty in real-world events.

Two: Kalshi has no stake in the result.

And they came armed.

Not just a demo, but binders of research.

Proposed compliance frameworks, risk models, even mock audits.

The kind of prep you’d expect from Goldman Sachs, not a ten-person startup.

Against all odds, the CFTC said yes.

In November 2020, Kalshi became the first federally approved “prediction market” in U.S. history.

But that alone doesn’t explain how they went mainstream.

Because after getting their license, platform growth slowed.

The founders knew they needed a breakout moment.

And they saw one: the US presidential election in 2024.

But they had a problem.

Kalshi never got permission to list political event contracts – meaning people couldn’t put money on who would be the next President.

So they sued their own regulator.

Kalshi made the same case from before.

Only this time, the CFTC said no.

So Kalshi took the case to federal court, and won.

For the first time in nearly a century, Americans could legally trade on political events.

The election hype put them in the spotlight, but it was just the start.

They’ve expanded beyond politics – adding sports markets and parlays.

Since October 2024 to now, trading volume has surged 200×.

But not everybody is a fan.

Several states have tried to ban Kalshi.

But all those lawsuits have failed so far.

In the U.S., both federal and state governments make laws.

When they clash, federal law wins.

It’s called preemption.

Kalshi’s founder summed it up simply:

“State laws don’t apply to us.”

But Kalshi is aware of its strategic vulnerability.

One license is keeping their business alive.

If enough states push back, Kalshi’s house of cards might not stand forever.

This brings me to their next move…

How Kalshi is protecting their license.

Since they got it, they’ve quietly built a powerful network of political and financial allies.

Brian Quinntenz, a former CFTC Commissioner, joined their board in 2021.

In 2025, they added Donald Trump Jr. as a strategic advisor.

And they brought in Sara Slane.

One of the key figures in legalizing sports betting across America.

It’s not all one-way traffic either.

Several former Kalshi employees have taken roles inside the U.S. government.

And their legal counsel, Josh Sterling, is now being vetted by the White House to lead the CFTC itself.

This could give Kalshi a direct line inside Washington.

On the financial side, Kalshi is backed by three Wall Street titans.

Charles Schwab, Henry Kravis, and Peng Zhao – whose firms collectively manage trillions.

All of this – the connections, the capital, the influence – fortifies one thing.

Their federal license.

If that license ever came under threat, it’s a safe bet Kalshi could use those connections to defend it.

Which might soon be necessary.

Remember how Kalshi is leaning hard into sports?

When they first applied for their federal license, they deliberately stayed away from sports-related contracts.

Now, in 2025, those markets make up nearly 80% of all money traded on the platform.

That’s a problem.

States that allow sports betting collect taxes from it.

For example, New York pulled in more than $1 billion in sports-betting tax revenue last year.

Guess how much Kalshi paid?

Zero.

And if there’s one thing governments hate, it’s not getting their cut.

That’s exactly why Ohio is now trying to ban Kalshi.

Ohio is the second-largest sports-betting market in the U.S.

The more states that try to tax Kalshi,

the more scrutiny it faces.

But as long as that federal license stands...

Kalshi is a casino the government can’t shut down.

The same economics that made Spotify possible are now making AI music inevitable.

It’s a system built for scale: cheap distribution, passive listening, and algorithms that reward time spent over taste.

In 2024, that machine finally worked. Spotify turned a profit for the first time since launch.

This tends to surprise people. How can a platform with 700 million users be allergic to making money?

The answer lies in Spotify’s licensing agreements. Roughly 70% of its revenue goes to rights holders: record labels, publishers, and distributors.

The very artists who make Spotify popular are the reason its margins stay razor thin.

So while Taylor’s 26 billion streams made her Spotify’s most-played artist last year, they also made her the most expensive.

And that’s the problem. Because Spotify’s business works better when real artists aren’t involved.

AI filler, fake artists, and stream farms aren’t bugs in the system. They’re features of a model that rewards quantity over quality, and cost-cutting over creativity.

A popular piano loop with no copyright is sometimes worth more to Spotify than a chart-topping hit.

Put simply: Spotify isn’t optimizing for better music, it’s optimizing for cheaper music.

When your business depends on paying for other people’s IP, your next innovation will likely find a way to replace it.

In Spotify’s case, replacement gets easier every day. Most people can’t tell (or don’t care) whether a song was made by a person or a machine.

If it sounds good, costs less, and keeps users listening... what incentive does Spotify really have to police AI content?

Hiten

=)

Subscribe now

I break down the full story here

Watch on YouTube

Video chapters

(00:00) Why Taylor Swift isn’t profitable (for Spotify)

(00:43) Daniel Ek’s roots in piracy

(01:41) How Spotify saved the music industry

(02:04) Spotify’s surprising economics

(03:20) Why Spotify is perfect for generic content

(04:22) Bots, stream farms, and royalty scams

(05:39) The fake artist with more streams than Abba

(07:30) Major vs. indie artist takedowns

(09:26) Why Spotify doesn’t want to police AI content

(10:07) The interesting case of Velvet Sundown

(11:14) Why you should expect more AI music

Share

Transcript

Taylor Swift was Spotify’s most-streamed artist last year with 26 billion plays.

That’s roughly 100 billion minutes of listening.

But across Spotify, users listened to 2.5 trillion minutes in total.

Taylor was just 4%.

So where’s the other 96% coming from?

That question stuck with me, and it got personal.

A while back I was in a café, heard a song I liked, and tried to Shazam it.

But nothing came up.

It got me thinking: Who made it?

That moment sent me down a rabbit hole:

Reddit threads, old news reports, stories about fake artists, and Bulgarian stream farms.

But the most interesting thing I uncovered is how Spotify might be complicit.

Because ideally, they don’t want to pay artists at all.

Before we go any further, first we need to break down how the music industry ended up here.

As you know, Spotify provides a giant online library of music that anyone can access for a monthly fee.

It ended the days of buying CDs or single songs on iTunes.

But what most people don’t know is the steps it took to get there.

This is Daniel Ek, the founder of Spotify.

He was born in Sweden, arguably the home of internet piracy.

Daniel even said this on a Freakonomics interview:

“I don’t think I would have been that interested in music if it weren’t for piracy, to be honest... We couldn’t afford all the records that I wanted.”

Before starting Spotify, he built his reputation as a teen web designer, eventually selling an online advertising company for $1 million.

He even had a stint as the CEO of uTorrent – one of the world’s most popular downloading tools, often associated with pirated content.

Then in 2008, he took on an even bigger challenge: creating a cheaper alternative to music piracy.

His idea... Spotify.

The narrative was heroic: democratize music, and hand power back to artists.

His timing couldn’t have been better.

The music industry was at rock bottom in 2008.

Their revenues had halved since the nineties.

They were desperate.

And they reluctantly agreed to license music to Spotify, a deal that would reshape the entire industry.

Fast forward to today and Spotify is a household brand name.

But the interesting thing is they have a problem.

It’s their surprising economics.

Despite 700 million users and $16 billion in annual revenue, Spotify survives on razor thin margins.

In fact, 2024 was the company’s first full profitable year since launching.

Why? Because Spotify only keeps about 30% of the money it makes.

The other 70% goes to music rights owners: Record labels, publishers, and distributors.

Every Taylor Swift or Drake stream costs Spotify more than it makes back.

At scale, those hit songs squeeze their margins.

It adds up when you’re paying out stream after stream, billions of times.

Spotify reports paying roughly $60 billion to the music industry since launching.

This model made sense when Spotify was prioritizing growth over profit.

But those days are over.

Investors are now pushing Spotify to make more money.

Raising prices more than they already have isn’t the answer.

Spotify is scared of user backlash.

The obvious way forward then is cutting licensing costs.

Analysts from Goldman Sachs have already modeled Spotify’s business.

They concluded margins would widen if more listening shifts away from major label music toward in-house, cheaper content.

In other words, Spotify doesn’t make more when you listen to better music, it makes more when you listen to cheaper music.

So if margins are the problem, here’s why AI is the answer, and why Spotify’s system is built for it.

It comes down to three key conditions.

First: the economics.

It’s a pay-per-stream model, but not all streams cost the same.

Remember, major label tracks command full royalty rates.

Meanwhile, music from smaller indie artists can be licensed for a fraction of the cost, sometimes even a one-time fee.

Every stream shifted away from Taylor Swift and toward low cost filler is better for Spotify’s bottom line.

Second: user behavior.

Most people don’t listen to music the way they used to.

You don’t sit down with an album studying the lyrics.

You put on music while you’re cooking, driving, or working.

Most listening today is passive, meaning music has become background noise.

That makes Spotify ripe for generic content.

And third: the algorithm.

Spotify doesn’t care if music is made by a human.

They care if you keep listening.

If a chill piano beat keeps you on the app, it gets pushed to your feed or placed on an official playlist, no matter who made it.

Put those three together and you get a system that rewards quantity over quality.

And it didn’t take long for outsiders to figure out how to exploit it.

Take the case of a Bulgarian man who created a massive stream farm.

He uploaded hundreds of short 30 second tracks, then used bots to play his music nonstop.

His playlists, Soulful Music and Music From The Heart, even broke into Spotify’s top 50 by revenue, despite only a couple thousand followers.

By the time he was caught, he had made over a million dollars.

What’s wild though is it was all completely legal.

He just gamed Spotify’s system.

But this had nothing on Michael Smith’s operation.

This guy combined AI generated music and stream farming.

He used AI tools to create over 500,000 tracks.

Then set up bot farms to stream them billions of times.

He made $10 million in royalties.

His scheme was so sophisticated, it flew under the radar for years.

But in 2024, he was indicted for wire fraud, conspiracy and money laundering.

The DOJ is still working the case.

And we don’t know whether what he did in regard to gaming Spotify’s system is actually illegal.

These two cases prove something important.

Spotify’s system, and its three key conditions, doesn’t just allow gaming, it rewards it.

Here’s the irony.

Every fake stream saves Spotify money compared to a real one.

And if outsiders can pull this off, imagine the incentive for insiders.

Here’s where the story takes a darker turn.

At first, it sounded like a conspiracy theory.

On Reddit and music forums, users noticed strange patterns.

Artists with no social media presence, no fan base, no touring history.

Yet somehow they showed up again and again on Spotify’s curated playlists.

Speculation turned into evidence when a Swedish newspaper launched an investigation into Yohan Raw.

Don’t worry if you haven’t heard of him, because you’ve probably heard his music.

He released nearly 3000 tracks under 650 fake names and racked up 15 billion streams.

That’s more than Ah-bah or Elton John.

You may have seen his music on official playlists, like Stress Relief and Peaceful Piano.

Just not under his actual name.

And Raw wasn’t the only one.

The newspaper traced more than 800 fake artists back to Firefly Entertainment.

A Swedish label with close ties to a former Spotify exec.

The reporting showed that roughly 20 real people were behind at least 500 of those aliases.

There were also questions about whether the label benefited from Spotify recommendations or playlist placements…

Firefly’s revenues grew tenfold between 2017 and 2020.

The controversy didn’t stop there.

Multiple investigations suggested Spotify struck reduced royalty deals with a number of labels and artists.

Here’s how it worked.

Instead of the standard payout, the 70/30 split, some artists accepted a quarter of the usual rate.

This was in exchange for spinning up fake artist accounts, and pumping out cookie cutter background music.

They were also promised spots on playlists, like Deep Focus or Ambient Chill.

These arrangements allowed Spotify to fill its music library with low-effort filler every week to keep users streaming.

While there’s no concrete proof, the incentives make a compelling case.

It’s not the in-house content those analysts from Goldman Sachs once imagined.

But every time these tracks replace a real artist, on an official playlist, Spotify makes more money, while musicians lose income.

Now, you might be thinking, surely Spotify has rules against this, right?

Well, not necessarily.

Spotify’s official line is clear.

They don’t ban AI generated content.

All tracks are treated equally.

And in terms of fake artists, their response is that “stage names” have been common practice in the industry for decades.

What they do ban is deceptive content.

But what counts as deceptive?

That seems to depend on who’s being deceived.

Take the 2023 viral song Heart On My Sleeve, which was apparently a new track by Drake and The Weekend.

It exploded across social media.

15 million views on TikTok and over 600,000 streams on Spotify in just one week.

It was even submitted for a Grammy nomination for best rap song.

But it was eventually disqualified for lacking “human authorship.”

Because it turned out the song was created by someone called Ghostwriter677.

They used AI tools to clone the voices, and probably make the beat too.

But when Drake’s team got involved, the song was yanked from every platform within days.

Now compare that to Emily Portman, a British folk singer.

In August, 2025, she woke up to fans congratulating her on a great new album.

The problem was she hadn’t released one.

Someone had uploaded an AI generated record called Orca to her official Spotify page.

It was complete with convincing writing credits and song titles.

Even Emily herself admitted the AI had done a scary job imitating her delivery and songwriting style.

She even said: “I’ll never be able to sing that perfectly in tune”.

But despite reporting the issue immediately, it took Spotify three weeks to remove the album.

This was long enough to rack up thousands of streams and mislead her actual fans.

And within days of taking it down, another fake album appeared in its place.

The message was clear.

If you’re Drake, the takedown is swift.

If you’re an indie artist, you’re on your own.

This imbalance has the whole industry on edge.

Universal Music Group and others are lobbying hard for AI takedowns.

They back legislation like Tennessee’s Elvis Act, which makes unauthorized voice cloning illegal.

And more than 200 artists, including Billie eye-lish, Stevie Wonder, Nicki Minaj, and Katie Perry signed an open letter demanding more protection.

The letter stated: “The use of AI to replace human artistry with machine generated content dilutes the royalty pool that is meant to reward human creativity”.

But here’s the fundamental tension.

Spotify’s business model makes policing AI content nearly impossible.

Every extra dollar spent on detection, every human hour spent reviewing flagged content, every real artist stream they protect.

It all works against Spotify’s bottom line.

Essentially, labels are asking Spotify to act against its own business model.

And history suggests which way Spotify will lean.

But while the industry fights back, the technology isn’t waiting.

It’s actually getting better.

There’s been an explosion of AI music startups like Soo-no and MusicLM.

Think of these like the pickaxes in the Gold Rush.

People are using them to flood music platforms with AI generated tracks.

We don’t know the exact number for Spotify.

But a French streaming service called Dee-zer recently said 20,000 AI tracks are uploaded to its platform every single day.

None have been more impressive though than a band called Velvet Sundown.

In June, 2025, they released an album to critical acclaim and genuine fan enthusiasm.

It got over a million streams within weeks.

Two more albums dropped in just a few months.

Music bloggers praised the vintage vibe, calling it an authentic 70s rock revival.

But then came the reveal.

The entire album was AI generated using Suno.

Every vocal, every guitar riff, every drum fill.

Suddenly the same songs were dismissed as soulless AI.

The interesting thing here is that most people can’t tell the difference.

And people like it until they know it’s machine generated.

So this raises a burning question.

If it sounds good, people like it, and it’s cheaper to distribute…

What incentive does Spotify really have to remove AI content?

This brings us full circle to Spotify’s founder, the man who swooped in to solve music piracy with Spotify.

Artists once praised him.

Now they say he’s created something worse.

If AI music becomes infinite, the long tail of smaller independent artists may simply vanish.

Spotify isn’t asking whether AI music is coming.

They’re asking how much of it can they afford to put in front of you.

Because the economics point in only one direction.

The tech keeps getting better.

Spotify wants to make their platform as sticky as possible.

And users want recommendations that match their mood.

I hope it’s clear by now that Daniel never set out to destroy human creativity.

But he might have built the perfect machine to do exactly that.

And this time there might not be a heroic narrative for Spotify to hide behind.

Because Taylor Swift may be their biggest star, but the real winner might be AI.

So the next time you hear a track you can’t Shazam, don’t be surprised like I was.

It might not be a person at all.

Subscribe now

Early-stage AI tools like Showrunner are already generating episodes South Park episodes that rack up 80 million views, and original shows like Exit Valley. No studios, no budgets, no permission.

The founder's vision for the future is even bolder: creative control over the stories you actually care about. Imagine remixing Star Wars after Disney's takeover, or finally settling the Game of Thrones debate by rewriting the final season yourself – all with a single prompt from your couch.

This isn't to say Netflix is about to collapse. With 250 million subscribers and strong profits, it's still the clear leader. But the dynamics fit a familiar pattern of previous disruption cycles, where incumbents get stuck protecting short-term profits, while scrappy outsiders bet on future shifts in consumer behavior. At first, those behaviors look niche, and then they compound. Fast.

This cycle has valuable lessons for anyone building products today.

Share

Four lessons from Showrunner's approach

1. Blockbuster was rich. It didn’t matter.
   

   Blockbuster was pulling in billions when Netflix showed up. Executives assumed that kind of dominance would protect them. It didn’t. Netflix is profitable today, but those numbers don’t shield it from a shift in user behavior. A healthy balance sheet isn’t a moat when technology makes something users want suddenly possible.
   

2. Scrappy often wins (if the shift is big enough).
   

   Netflix's first streaming product was clunky: a tiny catalog, Internet Explorer-only, and buggy. But it didn’t matter. The shift from discs → internet streaming was so powerful that “good enough” won.
   Showrunner’s AI episodes feel rough today, but they don’t need to be polished. They just need to be credible enough to hint at the future and prove people want it.
   

3. User behavior = moat.
   

   Netflix built its empire on passive watching: sit back, scroll, hit play. AI flips that model. Showrunner’s alpha testing revealed users don’t just consume; they play. People wanted to cast themselves (and their friends) in the show. That’s a completely different kind of engagement that streaming platforms aren’t built for. It’s also why Amazon has already invested in Showrunner, and why Disney is in licensing talks.
   

4. What made you strong can make you slow.
   

   Blockbuster couldn’t give up on late fees. Netflix can’t easily walk away from studio deals and billion-dollar productions. The very things investors reward (like subscriber growth and new TV shows) make it harder to take risky bets on what consumers might want. That’s the innovator’s dilemma in action: your strengths turn into shackles. And by the time the giants react, the outsiders have already scaled.

Netflix vs. Blockbuster was one of the clearest examples of this cycle. AI vs. streaming could be next. Still, even Showrunner’s founder admits his vision isn’t guaranteed. People might ultimately prefer passive consumption.

But that uncertainty is exactly the point. The question isn’t whether AI replaces Netflix. It’s whether a new way of consuming content compounds faster than incumbents expect.

The future always looks scrappy at first. The winners are the ones who see past the rough edges and get to work. Paul Graham put it best: “Live in the future, then build what’s missing.”

Hiten

=)

Subscribe now

I break down the full story here

Watch on YouTube

Video Chapters

(00:38) How Blockbuster shaped Friday nights

(01:11) Netflix’s unlikely origin story

(01:44) Betting it all on the internet

(02:35) Inside the streaming wars

(03:12) The rise of streaming fatigue

(03:39) When AI wrote nine South Park episodes

(04:32) Meet Showrunner: TV made by AI

(05:22) Casting you and your friends in your own show

(05:52) Remixing Star Wars

(06:25) Why Amazon has already invested

(06:46) Why Netflix can’t build this in-house

(07:34) Changing Game of Thrones’ ending

(07:58) Could AI replace Netflix?

Transcript

AI could replace Netflix the same way Netflix replaced video rentals. Here's how it works. While that's generating, let's break down the last time. Hollywood ignored a game-changing technology like this. It's a cleanest example of the innovator's dilemma when big companies go for short-term profits over long-term shifts.

A concept explained perfectly by Steve Jobs in this clip

It's 1998. Friday nights had a ritual. You drive to the blockbuster store, walk those blue and yellow aisles lined with videotapes, grab some overpriced candy and rent a movie for the weekend.

At its peak, blockbuster was making $6 billion a year with 800 million of that coming from late fees. That's how much money they made off people forgetting to return. Shrek Blockbuster looked invincible. It was drunk on its own dominance, fat on profits, and blind to change until one scrappy startup deleted their entire business model, all because of a single late fee.

In 1998, Reed Hastings got hit with a $40 charge for Titanic. He was pissed, but it gave him an idea. What if you could rent DVDs by mail with no late fees at all? He called it Netflix. At the time, fewer than 10% of American households owned a DVD player, but Reid bet on adoption. By 2004, nearly 70% of homes had one. was thriving. Then in 2007, he pivoted.

Internet speeds were getting faster. He said, forget disks. We're going all in on streaming.

The first version was Scrappy. A tiny catalog that only worked on Internet Explorer, but streaming caught fire fast. By the time Blockbuster launched their own service, it was too late. Within three years, Netflix had 20 million subscribers and Blockbuster was bankrupt.

But here's the irony. In 2000, Reed flew to Blockbuster's HQ and offered to sell Netflix to them for $50 million. Executives laughed him out of the room. That's the innovator's dilemma in action.

Blockbuster was so focused on running physical stores and making money from late rentals. They couldn't imagine a world without them.

What if I told you that today Netflix is the new blockbuster and we're going to see history repeat itself despite being profitable With 250 million subscribers worldwide, Netflix is locked in a spend to win arms race with Disney, Amazon, and Apple. The industry now burns over a hundred billion dollars a year on content to win the streaming wars.

This spend is colliding with a clear signal. Churn is climbing in the us. Nearly 40% of people cancel at least one service every six months.

Some are calling it streaming fatigue. Almost half of consumers say they overpay, they're frustrated. Too many subscriptions, weaker content and more ads. The gold rush is over. Investors reward, short-term subscriber growth. Not risky innovation, but just like in Blockbuster's day, the disruption won't come from another giant.

It'll come from a scrappy outsider.

Because what if people don't just wanna watch shows? They wanna make them instead. That's the bet behind showrunner ai. Now you might be thinking, this is just AI swap, it can't replace an actual TV show. But what if I told you it actually had in 2023 showrunner's, early model produced nine South Park episodes. These were unauthorized. The original creators had nothing to do with them. Here's an example. Clip

Reactions were mixed. Some were impressed by how good they were. Others thought they were soulless. But the point is, people watched, they racked up over 80 million views online. No studio, no budget, no permission. Just AI

What's cool about Showrunner now is you can actually make your own shows and in a few seconds I'll show you how. In 2024, they launched an original called Exit Valley. It's a satire of Silicon Valley featuring AI versions of Sam Altman, Elon Musk. Here's how it works. You write a short, prompt, cast the characters, pick a scene and generate the episode.

Let's check in on our prompt from earlier.

I thought that was pretty good. And other clips from Exit Valley have also gone viral. There's already a hundred thousand people on the wait list for showrunner's official platform launch. They're also working on several more original shows, But what's more interesting is user behavior. In early testing, one thing stood out. People didn't just wanna spin up new storylines. They wanted to create a character and cast themselves in the show, not just watching, starring. That's a completely different consumer behavior than what Netflix was built for.

Showrunner's founder Edward Saachi believes the tech is going to improve quickly. And yes, he's the son of the advertising legend behind Saachi and Saachi.

He's also a Emmy-winning storyteller who worked on VR at Oculus. His vision is for a future where TV is playable, people create original shows, cast their entire friend group, and remix the IP they already love.

Imagine opening Disney Plus and clicking remix Star Wars. You could rewrite canon or explore thousands of fan made storylines after the controversial Disney takeover. You can bet fans would jump on that. imagine never having to argue over how a story should have ended. You could just settle the debate in cinematic quality.

Think of Game of Thrones. Nearly 2 million fans signed a petition demanding a remake of the final season.

Showrunner's vision is why Amazon has already invested and Disney is in licensing talks, not because they believe in it, but because it's too dangerous to ignore, maybe even as a hedge against streaming itself.

One important note, showrunner has some limitations. Episodes are animated. Shorts, one-off stories, and the comedy often falls flat.

It is scrappy, but remember, so was Netflix's first streaming service

At this point, you're probably wondering, couldn't Netflix just build this in-house? In theory, yes, but here's the innovator's dilemma in action. Netflix is tied to billion dollar productions and studio deals. It's the same way Blockbuster was tied to retail leases. That makes pivoting much harder. Instead, Netflix is using AI for visuals, dubbing and feed recommendations Useful but not transformative.

Netflix is handcuffed. It's focused on safe bets that increase its stock price. Apple, Amazon, and Disney are breathing down its neck, taking their eye off the ball. To build new tech could cost them market share, but if showrunner takes off, Netflix might find themselves in the same situation. They put blockbuster in one where a new technology unlocks an entirely new way to consume and interact with content. That's the impossible choice. Protect profits today or build the tech that disrupts you tomorrow.

The future might look like this. It's Friday night. Instead of scrolling Netflix for 20 minutes, you can describe the exact show you want and it's made on the spot. Don't like the ending, remake it. Want 10 more episodes? Get them instantly. Or better yet, your group chat lights up. Check out this Game of Thrones episode I made where John Snow becomes king. Suddenly your couch is a studio and your group chat is the TV network.

Blockbuster laughed at Netflix, then vanished. Streamers now risk laughing at AI generated content. It won't be Netflix or Disney that decides the future. It'll be people like you. How AI could replace Netflix. Not with one hit show, but with millions of stories all at once. And just like Blockbuster, by the time they see the threat, it could already be too late.

Share

Waymo looks like the winner in self-driving.
It’s fully autonomous. Ultra-precise. Barely crashes.

But Waymo’s success is an illusion.
And Tesla is set to leapfrog them.
Not by building the best tech, but by deploying it the smartest.

Waymo is expanding city by city.
Its cars are packed with sensors. Each road is pre-mapped.
It’s safe, but slow and expensive.

Tesla is moving much faster.

They’ve already shipped over 500,000 semi-autonomous cars globally.
Those vehicles are collecting billions of real-world miles, training a system that learns on the fly.
No lidar.
No pre-mapping.
Just cameras, neural nets, and fast iteration.
Tesla sends out new software every few weeks.

The goal?
One update that turns millions of Teslas into robotaxis overnight.
Elon says they’re 2–3 years from making that happen.

But the deeper story isn’t just about Tesla’s strategy.
It’s the pattern — the same playbook Elon’s used for rockets, electric vehicles, AI, and now robotaxis.

Four founder lessons from Elon’s approach:

1. Solve problems in the right order: Elon always seems to ask: What’s the one constraint that, if solved, makes everything else easier? With self-driving, Elon didn’t start with full autonomy. He started by solving the real blocker: distribution. Deploy semi-autonomous cars first, then farm real-world training data at scale to improve the product.

2. Find the non-obvious solution to the obvious problem: If everyone agrees on an approach, Elon will almost always do the opposite. At SpaceX, the big barrier was cost. Most people assumed the solution was better spacecraft. But Elon bet on reusable rockets instead — and changed the economics of space entirely.

3. Build for tomorrow’s constraints: Elon doesn’t care about being right today. He cares about being right in the end. With EVs, he didn’t listen to early skeptics and wait for battery breakthroughs or charging networks. He built an electric sports car people wanted — then let infrastructure follow demand.

4. Turn users into advocates: This is a similar play to Airbnb and Uber: Build something people love and profit from, and they’ll fight to protect it. Elon’s version? He wants your Tesla to become your side hustle, earning money as a robotaxi while you sleep. Owners won't just want autonomy — they'll lobby for it.

Tesla’s robotaxi plan might be Elon’s biggest bet yet, and the one that proves the pattern works at global scale.

I break down the full strategy here:

Watch on YouTube

Chapters

(00:35) Why Waymo seems unstoppable

(01:46) Tesla’s dropout advantage

(02:38) How Tesla hacked the liability problem

(03:20) The four-move checkmate to autonomy

(04:28) Your Tesla’s secret side hustle

(05:38) Elon’s genius failure formula

Pledge your support

Transcript

Tesla has got a massive problem with its self-driving tech, but no one is talking about what happens when they fix this issue because Waymo success is an illusion. And Tesla is on track to leapfrog them with a strategy to deploy a fleet of Robotaxis in every city around the world, all within the next three years made possible by Elon's playbook. That's continuously defied the odds and skeptics.

First, let me introduce you to who everyone thinks is the market leader. Imagine you're in downtown San Francisco, you open an app, tap a button, and a car glides up. No driver, just empty seats waiting for you. That's Waymo. It started inside Google back in 2009. Today it's also live in cities like Austin and Phoenix.

Each car is a rolling super sensor, 29 cameras plus lidar and radar. It fires millions of laser pulses every second to build a 3D model of the street. It can see shapes, judge distances, and track movement faster than any human.

But the real magic isn't the hardware, it's the prem mapping. Before a single passenger steps in Waymo vans have already driven those same roads hundreds of times, recording every traffic light, curb, and road marking. When a Waymo robot taxi hits the road, it's not just reacting, it's cross-checking live data against a blueprint. It's like deja vu for robots. The results are hard to argue with. Waymo has only been at fault in one crash since launch. It's now handling 250,000 rides per week on paper. Waymo looks like it's already won the race, but the question isn't does it work? It's how fast can they take this global? That's where things get interesting.

In the opposite corner, there's Tesla, no lidar, no radar. Just nine cameras worth a thousand dollars. Instead of custom mapping cities,

Tesla's cars, figure it out on the fly. It's like dropping someone in Rome and telling them to drive. Tesla calls it vision-based ai.

It's a system relying only on cameras and neural nets to interpret the world. Unlike Waymo's cautious rollout, Tesla ships over there updates to users every few weeks. All while collecting billions of miles of real world footage.

At first glance, Waymo's Tech looks superior. Cameras have less rain than lidar. They struggle in rain fog or at night. To be fair, this isn't apples to apples. Waymo's are fully driverless. Teslas are still semi-autonomous, but that's the point. Waymo's a valedictorian who studied too hard. Tesla's a dropout who's already making money.

So how is Tesla getting away with it? The answer is liability hacking. Since Tesla's system is still semi-autonomous, it's classified as driver assistance.

That means the human behind the wheel is still legally responsible. If something goes wrong, it's your fault. Most of the time it's not Tesla. If Tesla had gone the Waymo route, they'd be on the hook for every crash. But by offloading liability to the driver with partial autonomy, Tesla gets to avoid approvals, delays, and lawsuits that's not just clever legal work. It's what unlocks Tesla's biggest advantage in ai, real world data at scale.

Waymo's launching like NASA, slow, precise, and expensive. Tesla's doing it like SpaceX, ship fast, then learn from failure.. Here's Tesla's four step strategy.

One, deploy semi-autonomous vehicles. There are already 500,000 on the road. Two, collect billions of miles of real world data. Three, train the AI on edge cases.

Lab testing. Can't replicate four. Flip the switch instantly upgrading the entire fleet overnight via software update. That's 6 million Teslas turned into fully autonomous vehicles.

Right now Tesla is deep into steps two and three, but step four won't happen until they hit an AI safety threshold. This is where governments agree. Tesla's AI is statistically safer than a human driver. Tesla estimates it'll take about 6 billion miles of data to clear that bar.

This year, a model Y delivered itself to a customer. The robotaxis went live in Austin, already covering more ground than Waymo's service. And their semi-autonomous fleet has logged 3.6 billion miles at this pace. They're two to three years away from flipping the switch and leaving everyone else behind.

But will the public trust Tesla's technology?

Elon's end game isn't owning a fleet. He wants you and your car to become the fleet. Tap a button and your Tesla joins. A pool of robo taxis making money while you sleep. Think Uber, but you don't have to drive the car.

It could turn every Tesla into an appreciating asset, justifying higher margins, pricing power, and long-term valuation..

This is Elon's master stroke. When autonomy arrives, millions of Tesla owners will have money riding on it. They'll advocate for regulatory approval, maybe not out of belief, but out of profit.. It's the same playbook. Uber and Airbnb used. Build something people love and earn from, then regulators will have little choice, but to legalize it, wall Street gets it. Tesla's $800 billion. Valuation assumes autonomy works. If it fails, that value disappears. But investors aren't just betting on autonomy, they're betting on a story they've already seen.

Tesla's Robotaxis strategy follows a pattern Elon has used before. Look at SpaceX critic said he couldn't build better rockets than nasa, but that was never the point. He made space exploration cheaper by creating reusable rockets.

After years of blowing them up and learning. SpaceX now handles 80% of the world's payload. It's the same with electric cars. Elon didn't care that batteries were weak or charging took too long. He bet that if he built a car, people want it, the infrastructure would follow demand. That's why Tesla started with a hundred thousand dollars sports car, not a budget sedan.

Elon's pattern is that his companies look like expensive failures until they suddenly work. His Robo Taxii strategy is the same play. Waymo might beat Tesla in precision, but Tesla's betting it can beat everyone in distribution. There are four things founders can learn from Elon's approach. One, solve problems in the right order. What's the constraint that if solved, makes everything else easier? Two, find a non-obvious solution to the obvious problem.

When everyone agrees on an approach, Elon does the opposite. Three, build for tomorrow's constraints. Elon doesn't care about being right today. He cares about being right in the end. Four, turn customers into advocates. When they love what you build, they'll fight for it. But most importantly, sometimes you just have to lock in, block out the noise and build.

Consensus creates comfort while conviction builds empires.

Share

Hiten
=)

Hiten Shah is free today. But if you enjoyed this post, you can tell Hiten Shah that their writing is valuable by pledging a future subscription. You won't be charged unless they enable payments.

Pledge your support

But in 2025, a leaked memo exposed the company’s true AI strategy: automate education. The backlash was instant. Reddit turned. TikToks went viral. Loyal users walked away from thousand-day streaks.

This is the inside story of Duolingo's unraveling: how a brand built on personality and playfulness lost user trust overnight, alienated its community, and put years of brand equity at risk all while selling a different story to investors.

Watch on YouTube

Chapters

(0:00) Why Duolingo users started quitting

(0:42) AI as a teaching tool or a threat to quality?

(1:12) When Duolingo went all-in on AI

(2:07) The memo that kickstarted Duolingo’s downfall

(2:57) From viral love to viral hate

(3:17) Duolingo’s off-brand response

(3:52) The scramble to win users back

(4:12) What every founder (and brand) can learn

Subscribe now

Transcript

In May, 2025, Duolingo users started quitting. I quit Duolingo. We are watching the live downfall of Duolingo.

Tiktoks went viral. The Reddit community turned on them all because of a memo posted by the CEO AI is really positive for our business.

In this video, I'll break down what happens when a consumer brand tells one story to users and another to investor.

How that tension backfired for Duolingo and what not to do during a PR crisis. Guys, Duolingo’s downfall is here.

You already know this, but Duolingo was one of the most loved brands in tech since 2012. They've worked to create a cult following the green owl, the memes, the passive aggressive streak reminders.

It felt chaotic. Funny, personal, but internally, the company was moving in a different direction. By 2020. Duolingo was integrating AI first with Bird Brain, then GPT-3, but they kept it quiet in public.

They still leaned on the owl, their human experts, and a tone that made the app feel alive. This was a contradiction on the outside playfulness, community, personality on the inside, automation scale, efficiency.

It raised a tricky question for Duolingo. How do they scale using AI without breaking the brand? While selling a story investors wanted to hear by 2023, the shift was harder to Miss.

Duolingo launched Max Duolingo Max powered by GPT-4. It offered AI role play instant explanations and personalized feedback inside the company.

Some called it Duolingo ChatGPT moment. The stock tripled within a year and the messaging change. Its tagline became AI and Education make a great duo. Duolingo worked with OpenAI for years. The same year, the New Yorker profiled, CEO Luis von Ahn. He said replacing teachers is a worthwhile trade off.

Then came the rollout in April, 2025. Duolingo released 148 new courses built with generative AI. von Ahn said our first a hundred courses took about 12 years. The next 150 took just one. The internal roadmap shifted towards full stack automation from lesson content to gamified features with AI led characters, but users started to suspect a drop in quality.

Recently, things have gotten so bad. Then came the memo. In May, 2025, von Ahn on wrote a post that made the company's direction clear, automate everything. It was the first time Duolingo had said it out loud. One line stood out, humans won't get us there. The memo detailed what AI first meant In practice, headcount would only grow if AI couldn't do the job.

New hires and existing teams would be evaluated on how well they used AI. More jobs would be cut, support would be handled by chatbots. None of this was a pivot. It was a strategy, months in motion. It was just never shared. Publicly, the goal was to scale personalized education, but to many users, it felt like the soul had been stripped out.

Alan, are you spending on infrastructure, content? Where's the money going? I'm demanding Answers from the CEO. Duolingo set out to optimize the backend, but what they were really changing was the front end, their brand, and it didn't go unnoticed. The response was instant. Reddit threads called the new content AI slop full of mistakes on TikTok.

Users said the app felt like a robot pretending to care. Videos tagged Duolingo AI racked up over 10 million views in days, most of them negative. Even LinkedIn piled on. Some users walked away from streets they'd kept for over a thousand days. After users turned, Duolingo pulled back. The company known for chaos and humor went quiet.

First, they deleted comments. Next, they temporarily wiped their TikTok and Instagram feeds yours of brand voice gone. Then von Ahn went on a podcast and added fuel to the fire. He said, I'm not sure that there's anything that computers can't really teach you. I think speculating schools would mostly exist for childcare.

While AI handled the teaching to users, it sounded like the company had stopped listening. So Duolingo resumed posting only this time with carefully worded lines. Community feedback is valued and humans are still in the loop. Then at the end of May von Ahn walks some of it back. Our employees are what make Duolingo so amazing.

So we're gonna continue having employees, and not only that, we're actually gonna be hiring more employees, but he also didn't reverse course. I see AI as a tool to accelerate what we do at the same or better level of quality. If anything, his comments suggested he still believed the trade-offs were worth it.

This is a case study in nailing the right messaging to the right audience. I know you're upset. I know people are upset. I wanna make this right. That matters even more when different audiences care about different things. In duo's case, that was public market investors versus everyday people trying to learn a new language.

For your company, it might be investors versus users or employees versus customers. The same principles apply. Know exactly who you're talking to and what matters most to them.

Share

Hiten

=)

Subscribe now

Subscribe on YouTube

Timestamps

• (00:00) What’s going to come up on my call

• (01:15) 5 things to know about Tim & Saasco

• (01:33) Focusing on value, not price

• (03:39) Rethinking Saasco’s positioning

• (07:19) The art of product sequencing

• (10:21) Finding the right design partners

• (13:10) Aligning business goals with customer needs

• (16:04) Next steps for Tim & Saasco

• (17:35) The key takeaway

Where to find Tim

• Twitter/X

• Website

Subscribe now

Subscribe on YouTube

Where to find Part 1 & 2:

• Cofounder Scenes From Breaking Bad (Part 1)

• Distribution Scenes From Breaking Bad (Part 2)

The pivotal scenes:

1. Early founder mistake (Season 2, Episode 1)

2. Founder ego challenges (Season 5, Episode 6)

3. Knowing when to sell (Season 5, Episode 6)

4. Consider your ultimate goal (Season 5, Episode 6)

Episode timestamps:

• (00:00) Coming up on today's episode

• (00:35) How much is enough

• (01:35) Founder ego challenges

• (02:47) Know when to sell

• (04:08) Walter’s real motivation

Subscribe now

Subscribe on YouTube

Where to find Part 1:

• The Business of Breaking Bad (Part 1)

The pivotal scenes:

1. Sales & Distribution(Season 1, Episode 6)

2. Co-founder Disagreements (Season 2, Episode 11)

3. The Best Product Doesn’t Always Win (Season 2, Episode 11)

4. The Investor/Founder Relationship (Season 3, Episode 5)

Episode timestamps:

• (00:00) Coming up on today's episode

• (00:31) Building a distribution engine

• (01:50) Navigating co-founder disagreements

• (03:37) One common founder trope

• (04:23) Founders always forget this

• (05:50) Finding partners

• (06:55) What investors want from founders

Subscribe now

Subscribe on YouTube

Scenes

1. The Cancer Diagnosis (Season 1, Episode 1)

2. Walter Wants to Cook (Season 1, Episode 1)

3. Walter Shows Jesse the Cooking Equipment (Season 1, Episode 1)

4. Walter and Jesse’s First Cook (Season 1, Episode 1)

5. Walter and Jesse Argue About Business (Season 1, Episode 6)

Timestamps

• (00:00) Coming up on today's episode

• (00:26) Walter’s cancer diagnosis

• (01:29) Walter and Jesse team up

• (02:59) Alignment is key

• (04:20) The Batman and Robin framework

• (05:53) Building a quality product

• (07:05) Product market fit is a “vibe”

• (08:14) Setting cofounder expectations

• (10:07) The hack to finding a cofounder

Subscribe now

Subscribe on YouTube

The pitches:

1. Squatty Potty (Season 6, Episode 9)

2. Jax Sheets (Season 12, Episode 13)

3. Haven Lock (Season 10, Episode 18)

4. Scrub Daddy (Season 4, Episode 7)

5. Tree-T-Pee (Season 5, Episode 7)

Episode timestamps:

• (00:00) Coming up on today's episode

• (00:33) Squatty Potty

• (03:36) Jax Sheets

• (05:34) Haven Lock

• (08:56) Scrub Daddy

• (11:42) Tree-T-Pee

Subscribe now

Subscribe on YouTube

The scenes from Succession:

• Celebration (Season 1, Episode 1)

• Vaulter (Season 2, Episode 2)

Episode timestamps:

• (00:00) Coming up on today’s episode

• (00:15) Show intro

• (00:31) The founder's objection to the deal terms

• (01:53) What acquirers don’t do

• (02:40) Unrealistic founder dialogue

• (03:35) Negotiations shouldn’t be personal

• (04:36) Pitching the acquisition to stakeholders

• (05:40) Unexpected challenges post-acquisition

• (06:52) Gutting the company post-acquisition

• (08:41) Breaking promises

• (09:38) The importance of empathy in layoffs

Subscribe now

Subscribe on YouTube

The pitches:

• Hilton (Season 3, Episode 9)

• Heinz Ketchup (Season 6, Episode 4)

• Heinz Beans (Season 5, Episode 6)

• Lucky Strike (Season 1, Episode 1)

• Kodak Carousel (Season 1, Episode 13)

• Jaguar (Season 5, Episode 11)

• Belle Jolie (Season 1, Episode 8)

• Kellogg's (Unfrosted)

Episode timestamps:

• (00:00) Coming up on today’s episode

• (00:20) Show intro

• (00:36) The Hilton pitch

• (01:53) The Heinz Ketchup pitch

• (03:09) The Heinz Beans pitch

• (04:26) The Lucky Strike pitch

• (06:30) The Kodak Carousel pitch

• (08:34) The Jaguar pitch

• (09:53) The Belle Jolie pitch

• (10:52) The Kellogg’s pitch

• (12:28) That’s a wrap!

This video is not intended to be a hit piece or to drag the founders, employees, or anyone else involved with Humane.

Instead, I aim to uncover valuable insights we can all apply to future challenges when building our own companies.

Subscribe now

Subscribe on YouTube

Referenced

• Bethany Bongiorno (co-founder)

• Imran Chaudhri (co-founder)

• Humane (Website)

• The potential buyout (Verge)

Timestamps

• (00:00) Coming up on the show

• (00:30) Everything you need to know about Humane AI in 30s

• (01:25) Humane AI’s biggest failure

• (02:48) The one thing a company can’t lose

• (04:20) What I would do as CEO

• (06:09) What’s going on inside Humane AI

• (08:52) The only way out of this

In this episode, we discuss:

• Cody biggest challenges and my targeted advice

• The common mistake self-funded founders make

• Which market SwellAI should win first

• How Cody can differentiate SwellAI

Subscribe now

Subscribe on YouTube

Timestamps

• (00:00) Episode trailer

• (00:50) Cody’s Pitch for SwellAI

• (03:12) More engineers and money ≠ growth

• (04:19) Cody’s biggest challenge

• (06:38) Advice on differentiation, opportunity, and growth

• (15:15) Key takeaways

• (16:15) That’s a wrap!

Subscribe now

Subscribe on YouTube

The pitches:

• Jian Yang’s hot dog app

• DoorDash’s YC application

• DoorDash’s YC demo day

• Doorbot’s Shark Tank pitch

Episode timestamps:

• (00:00) Show intro

• (00:10) Silicon Valley’s hot dog app

• (01:54) DoorDash’s YC application

• (02:38) DoorDash’s YC demo day

• (06:29) Doorbot’s Shark Tank pitch

• (07:44) That’s a wrap!

I discuss:

• Finding the right hook

• What information will make investors say yes

• When to focus on the product versus the problem

• How each founder should refine their pitch

Subscribe now

Subscribe on YouTube

Today’s founders:

• Jacob Peters (superpower.com)

• Luke Yarnton (therave.co)

• Valentin Haarscher (easop.com)

Episode timestamps:

• (00:00) Show intro

• (00:31) Jacob, founder of Superpower (Longevity tech)

• (02:50) Luke, founder of The Rave (20% MOM growth)

• (06:29) Valentin, founder of Easop (Acquired by Remote)

• (09:52) That’s a wrap!

We discuss:

• What not to say

• How to talk about traction

• Telling a compelling story in under 60s

• Avoiding investor confusion

• Communicating your unique value

Subscribe now

Subscribe on YouTube

Today’s founders:

• Austin: Howie (howie.ai)

• Omar: UnInbox (uninbox.com)

• Luke: Pixels (pixels.xyz)

Episode timestamps:

• (00:00) Show intro

• (00:33) Luke, founder of Pixels (400k DAU)

• (05:35) Austin, founder of Howie (AI scheduling assistant)

• (10:56) Omar, founder of UnInbox (open-source email)

• (15:26) Reflecting on today’s pitches

• (16:14) That’s a wrap!

• Big mistakes and how to avoid them

• Why “hiring slow, firing fast” is wrong

• The problem with offshore talent

• Building remote versus in-person

• Ways to hire top talent

• Tips that will help you scale

Subscribe now

Subscribe on YouTube

Episode timestamps:

• (00:00) Show intro

• (00:20) Rapid Fire Round

• (00:28) Q1: How we we keep morale high after a pivot?

• (01:28) Q2: How much financial info should we share with employees?

• (02:31) Q3: What’s the biggest hiring mistake you’ve ever made?

• (03:00) Q4: What’s the biggest firing mistake you’ve ever made?

• (03:52) Q5: Do you have a golden rule for hiring?

• (04:48) Q6: How do I lower an employee’s pay?

• (05:23) Q7: What’s your favorite interview question?

• (06:06) Q8: Thoughts on poaching employees?

• (08:04) Hot Takes

• (08:18) You should hire slow and fire fast

• (08:45) In-person startups > remote ones

• (10:10) Work trials are a great way to evaluate job candidates

• (11:33) Hiring offshore talent is a great way to build a team

• (12:10) That's a wrap!