AI is splitting the economy in half.

On one side: companies with so much specific knowledge about their specific customer that AI makes them more powerful. More sticky. More profitable.

Every day that passes, they get harder to compete with.

On the other side: companies that were selling "good enough." Fast enough. Cheap enough. Convenient enough.

AI is making "good enough" basically free. And free beats "good enough" every single time.

There’s one word that separates the winners from the losers: Personalization.

Allow me to explain.

To start, let’s look at the graveyard.

Because you need to see the bodies to understand what’s happening… and how to make a lot of money from it.

The Bodies On the Floor

Chegg.

Remember Chegg? Homework help.

$14.7 billion market cap in 2021. Today? $88 million.

That's not a typo. $14.7 billion to $83 million. A company that was once worth more than Dropbox is now worth less than a mid-sized car dealership.

Why? College students started asking ChatGPT.

Chegg's entire business model was: we have the answers and you don't. Then OpenAI showed up and said: we also have the answers, and also you can get them for basically free.

Game over.

Shutterstock. Down 87%. Getty Images and Shutterstock announced a $3.7 billion merger. Two companies that AI is actively destroying have decided the solution is to become one company that AI is actively destroying.

Why? Because making images is basically free.

Fiverr. The freelance marketplace. Down 97% from its peak. The CEO cut 25% of the workforce and declared the company "AI-first."

Here's what's actually interesting…

Chegg. Shutterstock. Fiverr.

All three were sitting on personalization gold mines. And all three chose to be vending machines instead.

Chegg watched students fail the same problems for years and never built a personalized tutor. Shutterstock tracked what converted for whom across billions of placements and never built a personalized brand brain. Fiverr accumulated millions of real working relationships—who hired whom, what worked, what didn't—and never built a matching engine.

Now the survivors.

Because this is actually the good news.

Unbothered. Focused. Flourishing.

Spotify. Not a streaming company anymore. A personalization company that happens to stream music.

AI users on Spotify spend 140 minutes a day on the platform. Non-AI users spend 99. That's 42% more engagement.

Every song you skip, every playlist you ignore, every artist you replay at 2am—that's data. Spotify has years of your data. Millions of people's data.

Building a competing product is possible.

Building your Spotify? Impossible.

Toast. Full year 2025 revenue: $6.15 billion, up 24% year-over-year. Toast ended 2025 with approximately 164,000 restaurant locations globally—adding a record 30,000 net new locations across the full year.

Why is Toast winning? Because Toast knows restaurants. Not "businesses." Restaurants. Kitchen display systems. Table-turn timing. Tip pooling laws in 47 states.

And it gets smarter every day because it's embedded in 164,000 locations watching everything.

ServiceTitan. ServiceTitan knows contractors better than anyone.

Technician dispatch routing. Job costing by trade type. Financing structures for a $12,000 HVAC replacement that a homeowner wasn't expecting to need.

Revenue up 24%. Gross dollar retention above 95%. Net dollar retention above 110%. That last number is the whole story. The average customer spends more every year than the year before.

Try building an AI that replaces ServiceTitan for a plumbing contractor. You'd need to know every local permit requirement, every supplier relationship, every technician routing optimization, every financing structure they offer customers. You'd need years of that contractor's specific data.

The stock has fallen below its IPO price—which is probably an opportunity.

The Biggest Opportunity

Here's the thesis in one sentence:

Personalization is now worth more than it has ever been worth in the history of capitalism.

If AI can do what you do, you're in trouble. If AI makes what you do better—because what you do requires context, history, relationships, and domain knowledge that can't be replicated—you're in the best position you've ever been in.

And this isn't just about big companies.

The same shift is happening at the individual level. And the numbers are insane.

AI prompt costs dropped 96% between early 2024 and 2025.

A complete solo operator tech stack—AI assistant, code editor, app builder, design tools, analytics—runs $87 to $230 a month. A traditional five-person team doing equivalent work costs $300,000+ annually.

The real story isn't just about which companies win and lose.

It's about who gets to play.

For the first time in history, an individual with deep knowledge in a specific domain can build, distribute, and operate a real business without VCs, without a huge team, without an MBA, without a computer science degree, and without asking anyone's permission.

That's the even bigger story.

And that’s why I’m putting together the AI agent masterclass.

Early beta testers already saw 1.0.

The full tutorial is coming next week for Altucher Confidential.