building business in AI era

Building is cheap

Building is getting easier. Cheaper, faster, more commoditized.

Back in the days it was a moat. To create a software product it required a team of experienced developers and months of work.

I remember I did a job interview for a senior software engineer around three years ago.

I asked him about his latest project experience. He told me he migrated a backend app from Ruby to Golang. It took him around six months and a lot of solo work.

Let’s say he got paid around $7k/month at the time. So that’s $42k and six months spent to migrate an application from one programming language to another.

Today, give the same task to the latest Claude model and you’ll get it done in the afternoon.

Andrej Karpathy famously said that “new hottest programming language is English,” and there is a lot of truth in that.

We moved one layer of abstraction higher.

Programming languages give us a tool that’s easy to understand, so we don’t have to write instructions to computer in binary or assembly.

A compiler or interpreter converts this higher-abstraction-level code to machine code.

What happened is that now we have a layer that converts English to a high-level programming language.

We just added one more abstraction layer on top with LLMs.

Skill set that matters

World changes rapidly, but some things stay the same.

Building a business has fundamental elements that didn’t change with AI.

That’s good news and bad news for every aspiring entrepreneur.

I remember I spent a lot of my free time learning skills that were not really related to my core expertise - backend development with Go.

Since like 2019 I was obsessed with learning about marketing & sales, design, content creation, writing, human psychology, etc.

Skills that not directly related to “being a good software engineer”.

I knew that going too deep on the tech side (learning more languages and frameworks, staying on track with all the latest updates) had low ROI for me.

It’s optimising for an employee path, but the path itself didn’t inspire me.

I had that gut feeling that I need to become a generalist and build expertise across a broad range of topics. I was curious how business makes money and how every element of it works. Not only the tech side.

Usually, that’s what’s important to have to become an entrepreneur.

Business is selling a product or service for a profit. That’s it.

So, that means that you need to know how to build a product (or get it for a cheap price somewhere that you can sell it later for a premium) or provide a service.

You need to know how to find customers. Then you need to understand how to sell it.

Also, you need to find a way to build that system in a predictable, repeatable way, that also has net positive unit economics and generates profit for the company.

And usually, it requires human traits and skills that AI doesn’t change:

• You have to take risks as an entrepreneur and be okay with uncertainty
• You have to be creative enough to find solutions to questions without clear answers
• You have to be good at distribution and sales
• You have to have taste and judgement
• You have to be resilient enough to continue pushing after failure

Those are mostly human traits that AI can’t replace. You can be a great builder. And AI amplifies your existing skillset.

But does it mean you can succeed in business?

Not really. If you lack the skillset and risk-taker mentality, you’ll probably fail and continue to work for a salary.

It’s not bad, it’s the way the economy works. Not everybody has to own a company.

You just have to be realistic with yourself and understand what it really takes to win.

And then ask yourself a question: “Am I even willing to pay this price?”