May 6, 2026

Building Used to Be the Hard Part

Two years ago, when I first started making apps, if you had any idea for any type of app or software, the hardest part was actually building it. AI was starting to get there, but it was a lot of copying and pasting, to say the least. Copying, pasting, building, looking at the error, going back, copying, pasting, looking at the error, going back. Now, with tools like Claude Code or cursor, that issue is pretty much basically gone. Honestly, most ideas died right there at that first step of copy and pasting. Most people didn’t really want to go through that hardship, which is totally understandable.

That gate is basically gone now.

With the tools that are provided today, anybody can ship anything. It might not be the greatest in that one thing, and it might not be the most polished, and it might have purple gradients everywhere. A person who actually cares, knows generally a good amount of stuff about the problem that they’re solving and who’s willing to put in the time to learn the tools that are available can now build things that would have required an entire department of a Fortune 500 company to build.

It’s pretty interesting because the bottleneck has kind of moved. It’s not really “can you build it.” It’s more “are there people who want this and are you obsessive enough to scale with this app or software that you’re building?” Most of the time, people will kind of stop before they’re promoting the product. They build something, they use it, maybe a few friends use it, and then they stop there, but pushing for the next step of sales is where a cool idea turns into a startup and a startup becomes a scaled company.

This new wave also definitely changes something about the ideology of figuring out what is worth pursuing when building. It was expensive, not even expensive financially, but expensive mentally. You had to be really confident about the market that you’re going to go into. Most people do a very large market evaluation before they spend tens of thousands of dollars on an MVP for a platform, typically. Now I can open a quick Claude Code tab, create an entire new website in 10 minutes, and then go find the people who want to use it later. I don’t know if that’s good or bad. I probably say not the greatest thing, but I guess if you have an MVP while you’re seeking validation for the software itself, then that’s probably good. Especially if it only took you a few days.

The thing that hasn’t really changed is the obsession requirement. Tools like Claude Code and Cursor don’t really care about if you fix the problem or not. They don’t make you go pitch to 100 shops in 2 days, and they don’t make you roll out of bed when you’ve been up till 7am working. That part is still completely on you, the founders who will actually get it done are the ones who are obsessive and cannot stop thinking about the problem that they are solving.