The End of Commercial Software
Why I’ll never pay for a CRM again
Last Friday, my 8-year-old built a video game at a restaurant. We were waiting for our food. He had an idea. twenty minutes later, it was live on the internet and we were playing it.
I’ve been writing code for thirty years. It took me decades of practice to ship something that good.
Chasing a Ghost
In my first Technonymous post, I wrote about learning HyperCard in the early 90s. If you’re too young to remember it, HyperCard was Apple’s attempt to give normal people the power to build software. It created this sensation—this almost physical feeling—that anything was possible. I fell in love with technology because of it.
Then HyperCard died, and for three decades, that feeling became a ghost. Programming got more powerful but also more complex. The barrier to entry rose. The magic became a profession. My profession.
I’ve spent thirty years chasing that ghost.
And, with Claude Code, I caught it.
Confessions of a Skeptic
Let me be honest about where I was six months ago.
I’d been using AI coding tools—Copilot, Cursor—for a couple of years. They were useful. They made me faster. But the code needed babysitting. Basic errors. Weird hallucinations. A decent tool, sure, but only in expert hands.
When people talked about “vibe coding,” I dismissed it. The tools weren’t good enough.
But if I’m being really honest? My skepticism was equal parts technical assessment and ego. I’ve been writing code since the Clinton administration. Surely that skillset can’t become obsolete overnight.
Right?
Dinner, Then Everything Changed
Recently, my partner and I were at dinner. We started riffing on an idea for a word game. Nothing serious—just playing with a concept over cocktails and queso.
The next evening, we were playing it. Crosslet. Live on the internet. It’s fun. It looks good. It works.
I never looked at the code. Not once.
I just talked to Claude Code like I’d talk to a junior developer, except this junior developer worked at otherworldly speed, never got defensive about feedback, and didn’t need health insurance.
My thirty years of expertise? Useful for knowing what to ask for. Completely unnecessary for the actual building.
Here’s What Nobody Tells You
There’s a dirty secret about vibe coding: the vibe dies the moment you try to share what you built.
You create something amazing in two hours. You’re flying. Then you realize you need to actually put it on the internet. So you trudge off to set up hosting. Configure DNS. Figure out deployment pipelines. Want users to log in? Go wrestle with Firebase. Need to send emails? Set up Resend. Analytics? Google Analytics. AI features? Now you’re juggling API keys from OpenAI, Anthropic, and Google, trying to remember which billing profile is which.
Your two-hour miracle has become a full day of account creation and configuration. The vibe isn’t just dead—it’s been buried in a shallow grave behind a YAML file.
So I fixed it.
I built ItsAlive.co. One command: npx itsalive-co. Enter your email, click a link, and your app is live. Ten seconds, start to finish.
It launches today. If you’ve been vibe coding, try it, it’s free.
But honestly, ItsAlive isn’t what this post is about.
What This Post Is Actually About
While building ItsAlive, I kept shipping things. Testing the infrastructure. Scratching itches. Just this week:
Custom evites for a birthday party (15 minutes)
That racing game my kid built at lunch
A founder proposal for a friend to join me in building ItsAlive (10 minutes)
A non-algorithmic Substack alternative (1 hour)
A recipe manager that generates AI illustrations for every dish and calculates the exact start time for each step so your whole meal finishes hot at once (3 hours)
Three hours for that recipe app. It would have taken a small team months to build that a few years ago. I would have raised a million dollars in VC to fund that development. Now it’s an afternoon.
And, standing in my kitchen, looking at the app telling me exactly when to start the risotto so it finishes with the chicken, I realized something:
Commercial software is living on borrowed time.
Why would anyone pay HubSpot hundreds of dollars a month for a CRM that sort of fits their workflow when they could build exactly what they need in a day? Why wrestle with Notion’s limitations when you can just… build the thing you actually want?
We’re not going to adapt our workflows to software anymore. We’re going to build software that adapts to us.
The Ghost, Caught
Thirty years ago, HyperCard whispered a promise: everyone could build.
Then the world got complicated. Programming became priesthood. The promise faded into nostalgia.
But it wasn’t a lie. It was just early.
Today, an 8-year-old can ship a game before the appetizers arrive. A word game can go from dinner conversation to live product overnight. A recipe app that would have required a team and a budget materializes in an afternoon.
Everyone can build now. Not “soon.” Not “eventually.” Now.
This will change everything about how we interact with computers, with software, with each other.
I think it’s going to be wonderful. I’m excited to keep building.
Let’s go!


I never tried Hypercard, but I did have that anything-is-possible moment with Geocities! Yeah, you nailed it with this post, vibe coding ain't new but it's back and stronger than ever. Thanks for building It's Alive! I can't wait to see what else y'all build. Noah’s Wheels has great controls. Smooth. Reminds me of RC ProAm for NES.