Now It's Apps' Turn to Get Cheap

Now It’s Apps’ Turn to Get Cheap
Once the foundational tools are good enough, technique and training outperform hardware. — Seth Godin, “You don’t need a better camera”
Seth wrote that about cameras this week. The gist: past a certain point, a better one won’t make you better. New snow tires beat a new car for winter driving.
It got me thinking about something bigger. The tools don’t just get good enough — they get cheap enough, over and over, until the only expensive thing left is the idea.
Take blogs. In 1995, having one was a project. A server, or money for hosting. Some HTML, maybe a web server to configure. And even then you were shouting into an empty room — no feed, no algorithm to find you. Real money, real knowledge, real time.
Ten years later, none of that was true. You signed up for Blogger, picked a theme, and wrote. Free. Today you can stand up a blog on your own domain in the time it takes to make coffee.
I keep noticing this pattern. The cost of making a kind of thing starts impossibly high, then collapses.
The same arc, over and over
- Words. A blog went from servers-and-HTML to a free signup in about a decade. Now anyone can publish under their own domain in seconds, for almost nothing.
- Music. Recording once meant a studio and a label. Then a laptop and a cheap interface made a bedroom good enough.
- Video. Cameras, editing suites, broadcast deals — then YouTube and a phone in your pocket.
- Graphics, then video again — this time with AI. What used to need a designer, or a film crew, now needs a prompt.
Each time, the same thing happens. The expensive parts — the gear, the skill, the distribution — fall away one by one. What’s left is taste and the idea: knowing which thing is worth making, and being the one to think of it. Those don’t get cheaper. When everyone can make the thing, the only edge left is picking the right thing and making it well.
And each time, a wave of people who could never have made that thing before suddenly can. With something like Claude Code, the distance from “I wish this existed” to a thing that runs is an afternoon — you describe it in plain words, it writes the code.
The other half: getting found
But making was only ever half of it. The other half is getting someone to look.
Before social media, that was pure word of mouth. Seth tells a story from the days before a Share button existed: people would print out his posts and leave them on a colleague’s desk with a note — “you have to read this.” They faxed them between offices. They forwarded the email to the whole team. Not to promote Seth — to look smart themselves. A good idea was a thing you passed along, because passing it along made you look sharp.
And he still works that way. Seth doesn’t really promote the blog. He just shows up — every single day, for years — and writes something that might be useful to someone else. The showing up is the whole strategy.
Apps are next
Software was the holdout — two costs in one: writing the code, and running it. Both are caving now, just not in step. Running went first; cloud platforms ate most of that pain over the last decade. Writing’s going now, with AI doing the typing. And this week I felt the running side fall even further.
The making I’ve written about before: I build my apps with Claude Code and barely touch the code by hand. What got me this week was the running. So I ran an experiment — moved two apps, GribGrab and MarrowMinded, from Render to Railway, just to feel how it goes now.
I didn’t really do it. I told Claude Code to, and it did — figured out the build, the start command, wired up the database. One command, a couple of minutes, a live URL. Turbo fast.
Not all smooth. The first deploy landed on top of my database and wiped it, because the tool defaults to whatever service you touched last. One flag fixes it. Every platform has an edge like that; the trick is finding it before it finds you.
Money? Roughly a wash — a few dollars either way.
So price didn’t decide it. Time did. The gap between “works on my laptop” and “live on the internet” used to be a chore. Now it’s one prompt. And when shipping is that cheap, you ship more — even the small things you’d never have bothered with.
What this is really about
So it’s not the tools. It’s the direction. Shipping an app used to be a wall. Now it’s a button — same as blogging, music, video, images before it. The expensive parts fall away, and you’re left with the only two questions that ever mattered: do you have something worth making, and can you get anyone to see it?
Takeaways
- Let an AI agent write the code. You keep the taste and the decisions; it handles the typing.
- Pick a platform that ships in one command. I tried both Render and Railway and I’m happy with each — the point isn’t which one, it’s that the hop from “works locally” to “it’s live” is now a single step.
- Every tool has one sharp edge — find it early. Mine was Railway deploying on top of my database and wiping it. Trip over it once, write it down, move on.
- Ship the small thing. When deploying is free and instant, build the tiny app you’d never have set up a server for. The point isn’t the app. It’s that you finally can.
What have you been sitting on because shipping it felt like too much? The wall’s gone — go make something. → gribgrab.com
Thanks for reading! If you'd like to share your thoughts send me an email.