You spend more on AI than it makes you

Do not index

In 2020 I spent a year trying to get rich quick by building AI algo trading.

The idea: use AI to predict stock markets and make consistent money.

I discovered there was an entire industry selling trading algorithm. They claimed high performance. Just copy their strategy, pay royalties, and profit.

I just had one question: if your algorithm makes money consistently, why are you selling it instead of just running it yourself?

That question came back this week when I saw this company called Polsia.

Here’s what I think 👇🏼

The full context

Polsia sells AI agents that supposedly run profitable companies for you. Raised $30M at a $250M valuation. They're at about $10M ARR, close to $1M per month.

The promise: spawn autonomous AI companies. Agents run everything. You sit back and watch your business grow as a solo founder.

The growth is real. I don't doubt their growth numbers.

But here's what I don't understand:

1 - Users spend their own money to operate on the platform beyond a point.

2 - If this software creates companies that earn money consistently, the founder could spawn thousands of them and make way more money than he does right now.

Instead he's selling subscriptions.

That tells you something.

The lesson from my algo trading days

When people sell trading algorithms, it means one of two things:

1 - The algorithm doesn't actually work consistently

2 - They make more money selling it than running it

Both options mean you shouldn't buy it.

Same logic here.

If Polsia agents actually built successful companies, the founder wouldn't need your subscription. He'd spawn 1,000 companies himself and retire.

The math doesn't add up either. I'm pretty sure the total money earned by all Polsia companies is at best 10x smaller than the money users spent creating them.

The practical problems

Beyond the fundamental logic problem, there are structural issues:

1 - You don't own anything

Most Polsia companies are hosted on Polsia subdomains. No SEO growth channel. Payment provider is Polsia. Ad accounts are Polsia's. You have almost zero control and completely locked in.

2 - You're competing against yourself

Same AI agents run multiple companies. They use the same strategies. Same execution. You're likely competing with other Polsia users running the exact same system.

3 - AI isn't ready for full autonomy

Outsourcing everything to AI means missing the right integrations. Missing real world knowledge. Missing everything you need for a successful product.

You're building wrong from day one.

4 - Nothing compounds

You can't leverage previous work. You can't build on what you learned. It's more like gambling than building a startup.

The question no one's asking:

Are users actually getting value? Or just spending money to watch AI agents run in circles?

If Polsia agents actually worked, you wouldn't be able to buy access.

Maybe I'm wrong. If you're running a Polsia company and actually making good money, reply to this email. I genuinely want to understand how you're doing it.

My personal AI agent system

I'm not anti-AI agents. Agents are getting more capable every day.

I've been using missioncontrolhq.ai for weeks now and have outsourced many workflows. It was built by Bhanu, who sold Feather to me 2 years ago. It works great, and I love building growth systems there that work for me while I sleep.

But I want to build my own agent platform. The right way.

Agents that enhance what you do. Not try to replace you entirely.

The difference: you maintain control.

You own your domain. You make the strategic decisions. Agents handle execution.

Here's my question for you:

What work are you still doing manually that you feel AI should be able to handle?

What is AI not capable of doing for you that you would want to unlock?

Reply to this email and tell me 👇🏼

I'm building this, and I want to know what actually matters.

Tweet of the week

This team delegated everything to AI. And they had to hire more people.

Exactly suits my point, 1 person letting AI run the show is not too far away from hopecore content selling a dream to 1000s of people.

A good read, this one:

That's it for this week.

Until next week,

Keep building

Tibo 💻

P.S. Link to my original tweet about this. The responses were interesting. Some people defending it, but nobody showing actual sustainable revenue from their Polsia companies.