AI是場財富排位賽,你站好隊了嗎？

The other day I was scrolling Twitter and saw indie developer Peter Levels post a tweet.

He said everyone around him is scrambling to make money, buying up assets — stocks, ETFs, gold, real estate — all with one goal: be well-positioned before AGI arrives.

Honestly, I usually don't find this kind of talk convincing, because it tends to feel like fear-mongering.

But that day, after reading it, I quietly hit that like button.

Because I genuinely agree with him.

AI is driving a massive wealth transfer, and those who see it clearly are already racing to stake their position.

I know someone's going to say, "See? See?"

"You're just spreading anxiety too, aren't you?"

But those "don't worry, just focus on yourself, everything will be fine" videos — I actually think that's a form of dishonesty.

It's like the early days of the Industrial Revolution — telling workers about to be replaced by machines, "Don't worry, just relax" — nobody could actually relax.

When the wave hits, you only have two choices: face it head-on and figure out where you stand, or bury your head in the sand and pretend nothing is happening.

And this race — I believe you can't sit it out, whether you like it or not.

So as an investor, how do you position yourself?

Today I want to talk through the four dimensions of investing in the age of AI.

But before we dive in — I want to take 30 seconds to thank this channel's long-time partner, Interactive Brokers.

IB is an award-winning, globally recognized brokerage and the platform I personally use to invest in US stocks.

I love it for mainly two reasons: first, trading costs are extremely low — commission-free in many regions, or so minimal it barely matters.

Second, its FX conversion rates are very competitive. I buy US stocks from Canada, so I'm constantly converting CAD to USD.

Compared to Canadian brokers charging 1.5%+ in conversion fees, IB's FX fees are in a completely different league.

That alone is enough reason to choose them.

IB has plenty of other perks too — like it pays interest on cash in qualifying accounts, its options trading is also very powerful, and very cost-effective.

I highly recommend clicking the link in the comments to check out Interactive Brokers.

When you use my link, I earn a small commission — so it's also a great way to support the channel.

Alright, now back to the video.

To understand what to invest in, you first need to understand what this race is actually about.

Long-time viewers know I've been talking about ideas like this for a while.

Last year I made a video called All in AI — why you should start Vibe Coding right now.

I put forward one core thesis: the golden age of the "tool person" is being ended by AI.

What does that mean?

At its core, there are only two ways to build wealth: one is labor — your physical effort, brainpower, time — the other is capital — assets that earn money for you.

But the vast majority of people rely on the former.

We're essentially "tool people," tied to some skill or tool, trading it for income.

And that path used to work, because no matter how powerful machines got, they always needed people to operate, maintain, and connect them. The finer the division of labor, the more demand there was for tool people.

But AI is changing the rules.

AI is fundamentally different from past tech revolutions. The steam engine, electricity, the internet — all partial automation.

Machines handled part of the work, but couldn't function without humans.

AI is the starting point toward full automation.

In the digital realm, it can already handle knowledge work end-to-end.

And in the physical world, the rise of humanoid robots means physical labor may now be on the chopping block too.

Half a year later, these predictions are being validated one by one.

The most direct evidence: this past February, after Anthropic dropped Claude's new capabilities, global software and services stocks lost nearly a trillion dollars in market cap in a single week.

What does that tell us?

The market finally got it: the model itself is the app.

Every model upgrade directly eats into the functionality of the application layer.

Those SaaS companies that seemed to have moats — they're just the tool layer. And the tool layer is collapsing.

And the tool people dependent on that tool layer?

Their situation can only get harder.

Layoffs at Silicon Valley giants have never stopped, and every single business owner is thinking about how to use AI to cut costs and boost efficiency.

That's the underlying logic behind Peter Levels' tweet — once AGI arrives, capital won't need labor anymore.

Your physical strength, your brainpower — none of it will ticket you into the wealth game.

At that point, the only thing keeping you at the table is the assets you own.

Where you stand right now is your starting position entering that new world.

That's why everyone is running this race.

But hold on — didn't Elon Musk say this?

[laughs] After AGI, abundance for everyone — no need to work, everyone gets Universal Basic Income, just go do what you love.

Honestly, I want to believe that too.

Who wouldn't want to live in that world?

But ask yourself honestly: do you actually think that's going to happen?

About the future of AGI, the only thing we know for sure is that nobody knows for sure.

And that uncertainty is exactly what's driving so much anxiety.

So what do we do?

I want to borrow something someone figured out over 300 years ago — Pascal's Wager.

Pascal faced the ultimate question: does God exist? And his answer was: you don't need to know the answer.

You just need to figure out which choice costs you the least.

Believe — if God exists, you go to heaven.

If he doesn't, you lose nothing. Don't believe — if he exists, you're in big trouble.

So regardless of the truth, believing is always the better bet.

Now apply that to AI. You have two choices: position yourself now, or do nothing and wait to see.

If the future locks in a hierarchy and you've built up assets, you win.

If the future is universal abundance, what you built beforehand didn't hurt you — no loss at all.

But if you do nothing, and that utopia turns out to be wishful thinking?

You're locked into a position you don't want to be in, with no way out.

So whether you're an optimist or a pessimist, the rational choice is the same: position yourself first.

But here's the question: how, specifically?

How do you stake that claim?

Since we're in this race, the most obvious move is to just buy AI stocks, right?

But there's a question most people haven't really thought through: which layer exactly are you buying?

A lot of companies you think are AI-related are actually the very ones being disrupted by AI.

There's an important concept here: AI-native.

Right now, there are two types of companies doing "AI."

Type one: bolting AI onto an old architecture.

Think Salesforce Notion QuickBooks Adobe.

They have massive user bases and years of accumulated product architecture.

In the past, those were core assets.

But in the AI era, those have become liabilities, because these structures represent an existing way of producing and organizing resources — one that's already outdated. But these companies can't let go, and can't afford to. So they patch and tinker, force-fitting AI on top.

That's not AI-native.

That's putting a motor on a horse-drawn carriage.

Type two: built for AI from day one.

The entire system architecture is designed around AI from the ground up.

AI is the engine. Everything else is the drivetrain.

The starting point isn't "how do we add AI to our product?" —

it's "AI is the product.

Everything else revolves around it."

The difference between these two is fundamental.

Take an example: you want to control a smart speaker at home.

Why do you have to open that clunky app? An AI assistant handles it directly — you don't even

app? An AI assistant handles it directly — you don't even need that app.

Or take taxes.

Tax season is coming up. All you want is an accurate tax return.

Why do you care what the software interface looks like?

AI does the math, fills it in, submits it. Done.

That's the AI-native mindset: outcome-driven.

You want the result, not the process.

All those menus, buttons, forms, workflows, interfaces — they existed to help humans operate a system.

But once AI replaces the operator, those things lose their reason to exist.

Now I'm not saying software or interfaces will disappear entirely.

Consumer-facing software and interfaces won't go away — people still want great visual experiences.

But the interfaces and software designed for human operators and producers — those are disappearing.

And what this reflects is a fundamental shift in how value is created.

When you view the market through this lens, you'll find that the vast majority of so-called "AI stocks" aren't AI-native at all.

They've just wrapped AI around old products.

Truly AI-native companies are just starting to emerge, and there are far fewer than you'd think.

Now, this doesn't mean every legacy app company will die.

Some will survive, but they'll go through an incredibly painful transformation.

And even if they make it, their scale probably won't be what you'd expect, because AI has reset them back to the starting line, competing neck-and-neck with AI-native startups.

So when you say "I want to invest in AI," the first question to ask yourself is: am I investing in AI itself, or in the old world being disrupted by AI?

But that's only half the story.

What we've discussed is mostly the software side.

There's something you can't quite see yet, but whose impact will be even greater: physical-world AI.

There's a question a lot of people have that no one's really answered clearly: AI is so powerful now — has productivity actually improved?

Is it showing up in GDP?

Andrej Karpathy, in his Dwarkesh Podcast interview, asked the same question.

He builds AI himself, and even he's asking: can AI push GDP growth to 10%?

My take: the real productivity unlock won't come from digital AI — it'll come from physical-world AI.

Right now, all software is fundamentally virtual.

It has value, but ultimately it has to translate into action in the physical world.

Software improves decision-making, optimizes processes — but GDP ultimately comes from physical-world production: building houses, roads, manufacturing goods, moving freight.

When AI can truly move atoms — that's when it truly unlocks productivity.

So what is physical-world AI?

Physical entities embedded with an AI brain that can execute tasks in the real world.

Self-driving vehicles, creating value 24/7.

Humanoid robots that can use any tool or machine designed for humans.

The factories that build them.

And the compute, energy, and infrastructure powering it all.

When masses of AI robots are working in factories and construction sites, delivering services in the physical world, even running computers to operate digital-world AI — that's when the GDP growth curve jumps to another dimension entirely.

So as an investor, keep your eye firmly on embodied AI.

AI moving from bits to atoms. Remember what I said at the beginning?

The endgame of this race is assets that grow on their own, no longer dependent on human labor.

The day embodied AI becomes widespread, that endgame arrives. Factories run themselves, fleets drive themselves, robots do the work.

Whoever owns those assets is positioned.

And physical AI has one more key feature: you can't easily replicate them.

Copying code is free. But copying a factory, a vehicle, a robot — that's not so easy in the physical world.

And that brings us to the third dimension: doing things that are hard to scale.

This is less about investing strategy and more about what it means to be human in the age of AI.

AI is replacing the tool layer, replacing the tool people.

So where does human value come from?

When AI releases one new feature, an entire group of professionals can lose their jobs.

Years of expertise you've built up — looked at through a results lens, it can suddenly lose much of its meaning.

Let me share my own example.

[laughs] People used to call me a "knowledge creator."

But starting last year, I felt like purely spreading knowledge didn't have much reason to continue.

Because AI already does it better than I do.

I've been using NotebookLM a lot lately.

It can generate tutorials, podcasts, even videos, from hundreds of sources.

Knowledge has entered the era of personalization.

ByteDance recently released SeedDance 2.0.

The editing skills I'd always been proud of, my understanding of visual storytelling — I always thought those weren't so easy to replace.

But when AI is specifically trained on those things, you realize it does it way better than me.

I've also been Vibe Coding my own software.

And now, AI building software has reached a "say it and it's done" level.

The other day someone used AI to clone a very well-known piece of software, added new features, open-sourced it all — free for everyone to use.

That gave me a clear realization: in the future, most software might become like writing articles or making videos — it's free. It's not a product; it's a traffic channel.

To borrow a line from The Three-Body Problem: "Photography is gone. Editing is gone.

Programming is gone.

The thing you were good at — it's gone."

When your primary value was in providing tools, operating tools, or being the tool itself, once those aren't needed anymore, where does your value come from?

I believe your real moat is the things you have that can't easily be scaled.

And here's the key concept: arbitrage.

You're doing something hard to scale, but inside it you're using cutting-edge AI to automate, cut costs, and boost output. That gap is where huge arbitrage lives.

Say you have a physical business.

AI hasn't eaten into it yet. Hard to immediately replace.

It might not look "sexy" — hard to scale. But that's exactly your moat.

Use AI to automate the back end, cut out the tool layer.

While your competitors haven't realized this yet, you've already got the edge.

So what kinds of things are hard to scale?

First: genuine human relationships.

AI can help you "handle" ten thousand people at once, but that's not a relationship — that's customer service.

Real trust and community require your actual presence.

Second: businesses that deal with the physical world.

Running a hardware company, a brick-and-mortar business, a physical space — naturally constrained by physics.

But you can use AI inside them to automate operations, cut costs, boost efficiency. That's the arbitrage.

Third: your experience, judgment, and personal brand.

AI can generate infinite content, but it doesn't have a real life.

The mistakes you've made, the intuition extracted from failure, your personality — those form the irreplaceable parts of you.

So in the AI era, lean into things that are hard to scale, and use AI inside them to arbitrage.

Because the easily scalable territory already belongs to AI.

And that leads to a very practical question: how do you actually pull off that arbitrage?

Which brings me to the fourth dimension: a thousand hours.

Regular viewers know I've talked about the 10,000-hour rule from Malcolm Gladwell's Outliers — that number is definitely exaggerated, but the core idea is right: if you want to truly master something, you need to put in enough time.

In the age of AI, you don't need ten thousand hours.

But you need at least a thousand hours of genuinely using and studying AI.

To pull off that arbitrage, you have to truly understand AI. Not just reading articles, scrolling through some tweets.

Actually spending thousands of hours working alongside it, knowing what it can do, what it can't, where the edges are.

And those thousand hours can also answer two questions a lot of people have: "I don't know what to invest in" and "I don't have capital — how do I even invest?"

Let's talk about no capital first. The endgame is assets doing the talking.

But that endgame isn't today.

Today we're still in a window. Your intelligence,

labor, and creativity can still help you build capital.

But the key is to use AI to amplify your output.

The reason this window is short is that the space for earning income through pure labor keeps shrinking.

But before that window closes, AI is precisely the strongest lever you have.

So the first step isn't thinking about which stock to buy — it's to use AI to build your first pot of capital outside the market.

Now, about not knowing what to invest in: legendary investor Peter Lynch, in his book One Up on Wall Street, said your best investment ideas come from your experience as a user.

When you genuinely use a product, you feel things others can't.

And AI is something you have to be immersed in to truly understand.

Those of us who live and breathe AI, scrolling Twitter every day, often feel like "a day in AI is a year in the real world" — like the whole world already knows how powerful AI is.

But when you step back into real life, you realize most people around you are barely paying attention.

We actually live in a very small bubble.

That's why some say AGI is already here — it's just not evenly distributed.

But that unevenness — that information gap — is your opportunity. And this opportunity isn't just about arbitrage and investing.

Those thousand hours also give you something even more important: conviction.

Conviction is crucial in investing.

It determines whether you can hold on when others are panicking, whether you can stay the course when tempted by wrong moves.

Like with my Tesla investment — I got FSD early, use it every single day.

I've watched this car transform day by day.

That kind of conviction can't be given to you by someone else — it comes from your own lived experience.

This past half year going all-in on AI, my biggest takeaway is: you have to actually use it for your mindset to truly change.

You go from someone being pushed forward to becoming an AI native.

But a lot of people past 30 are reluctant to learn new things, don't want to start from zero.

But when it comes to AI, the unwillingness to start over — that is the real risk.

So spending a thousand hours on AI is something everyone can do.

The barrier to entry is extremely low; the payoff is extremely high.

The only barrier is mental: resistance, laziness, and dismissiveness.

Finally, on investing in your AI practice, two more points.

First: don't be cheap when it comes to spending on AI.

A lot of people think using free ChatGPT means they're "using AI" — they're not. Subscribe to the best models,

they're not. Subscribe to the best models, use the most cutting-edge productivity tools.

Spend a few extra dollars a month, and you'll actually know where AI's true capabilities and limits are.

That's not spending — that's the most direct investment in yourself.

Second: seriously integrate AI into your actual work.

Don't just use it as a search engine, and don't just chase the shiny new thing, playing with flashy demos every day.

What you should really do is experiment seriously — figure out how it can genuinely replace your work, cut your costs, boost your output.

That's the real meaning of a thousand hours.

I can promise you: once you actually try it this way, your entire perspective on AI will fundamentally shift.

So that's the fourth dimension I think every investor needs to consider in the age of AI.

Before you think about which stock to buy, invest in your thousand hours first.

This is probably the most important investment you can make.

It builds your capital outside the market, builds your information edge, and builds your conviction.

Those thousand hours are your entry ticket into this race.

Alright, that's everything for today's video.

Hope it gave you something to think about. Thanks for watching, and I'll see you next time.

Take care!