When building is cheap

For most of the history of software, the hard part of starting a company was building the thing. You needed engineers, months of runway, and usually someone else’s money to pay for both. Most ideas died at that stage, before a single customer had the chance to ignore them.

That stage is mostly gone. With the AI tools available now, one person can ship in a weekend what a small team used to ship in a quarter. I see it in my own companies and in almost every pitch I read. The prototypes are good. Some are better than products that raised real money a few years ago.

So the useful question is what’s still hard. Whatever is still hard is where the value goes.

Getting chosen

When anyone can build, products stop being scarce. What stays scarce is attention, and trust once you have it. A market full of capable products doesn’t reward the best one. It rewards the one people heard about from someone they believe.

Naval has a framework I keep coming back to. In How to Get Rich he describes code and media as permissionless leverage. You don’t need anyone’s approval to write software or to publish, and both keep working while you sleep. For most of the last twenty years, code was the expensive one. Media was treated as marketing, something you bought once the product was done.

I think AI has flipped that. Code is close to free now. Media hasn’t got any easier. If anything it’s harder, because the same tools that let anyone build an app let anyone flood a feed. The one form of leverage that still takes years to build is the one most founders leave until last.

What having both taught me

I didn’t plan it this way, but I ended up with both. I spent years in digital marketing, growing audiences for other people’s products, before I built my own. Today around 200,000 people follow what I write on Instagram, and Profit AI, the investing assistant I founded, is used by more than 200,000 traders. Each made the other easier.

An audience doesn’t rescue a bad product. People leave just as fast, maybe faster, because they trusted you and you let them down. What it changes is day one. You don’t launch into silence. You launch to people who already know how you think and will tell you quickly if you got it wrong. That feedback is worth more than any launch budget I’ve seen.

How I invest now

It has also changed the first question I ask founders. I used to want to know whether they could build it. That question has almost stopped being useful. Now I ask how their first thousand customers will hear about it, and why those people would believe them.

A founder with a real answer is in a different position from one who plans to work it out after the round. The answer might be a community, a newsletter, a reputation in a small industry, or an audience they grew themselves. The first founder can test ideas for free. The second pays for every lesson.

If you’re building now

Start writing before you start the company. Not as a marketing plan, but because publishing is the fastest way to learn what people care about. The comments are free research, and they arrive faster than any survey.

Put your name on your work. A company brand gets rebuilt every time you start over. Your name comes with you. Mine is on the companies, the investments and this site, so each new project starts with some trust it didn’t have to earn from zero.

Be specific about who it’s for. Generic products are the easiest thing to build with AI and the hardest thing to sell. The more precisely you can describe one customer, the more that customer recognises themselves in what you make.

And stay put. Trust compounds, but only if you stay in one place long enough for it to add up. Naval’s version is to play long-term games with long-term people. I’d add: under the same name.

None of this means the product matters less. It matters more, because the bar has gone up for everyone. It just isn’t the moat anymore. The moat is that when you ship something, people are already listening.

I’ll write about this at more length here than Instagram allows. If you’re building something and want to compare notes, email me at zak@zakosman.com.