What If the Computer Never Came Home?
A prequel. Two what-ifs before we begin.
Before I start building, I want to sit with two questions.
What if the personal computer never happened?
And what if AI, as we know it today, never happened?
Neither is far-fetched. Both came closer to not happening than we remember.
What if the computer never came home?
In the 1970s, the sensible view was that computers belonged in rooms.
Big machines. Owned by governments, banks, universities. You rented time on them. You typed into a terminal, and the real thinking happened somewhere else.
Imagine that model had simply continued.
We would still have computing. Probably a lot of it. But it would be a utility, like electricity or the phone line. You'd pay for access. You'd use what you were given.
There's a real glimpse of this world. France had Minitel, a national network of simple terminals connected to central services. People booked trains, checked bank accounts, chatted. It worked. It was popular. And almost everything ran on the network operator's terms.
In that world:
- You don't own your tools. You subscribe to them.
- You don't tinker. There's no garage, no hobbyist, no kid taking apart a machine to see how it works.
- The people who build software are the people who own the central computers.
What the personal computer really gave us wasn't speed.
It gave us ownership. A machine that was yours. That you could change. That served you, not a central office.
Spreadsheets, word processors, games, the open web. Most of what followed came from people who weren't asked permission.
What if AI never happened?
This one also came close.
AI went through long winters. In 1969, a famous critique showed early neural networks couldn't solve simple problems, and funding dried up. Again in the late 80s. For decades, "neural networks" was almost an embarrassing phrase in research.
A small group kept going anyway. Then, around 2012, graphics chips built for video games turned out to be perfect for training neural networks. The rest moved quickly.
So imagine it hadn't.
We'd still have search engines. Apps. Smartphones. More screens, more feeds, more notifications. The tools would keep pointing outward: more to see, more to buy, more to respond to.
Work would stay mostly manual. You'd still be the one gathering information, sorting it, summarizing it, writing it up. Every report. Every email.
And the attention economy would carry on unchallenged. Each app competing for one more minute of us.
In some ways, that world is calmer. Fewer unknowns. No questions about machines that sound like people.
But it's also more crowded. The noise keeps growing, and we keep filtering it ourselves.
What these two what-ifs teach me
When I put them side by side, a pattern shows up.
Every big technology seems to start central and then has a chance to become personal.
| Started as | Became | |
|---|---|---|
| Computing | Mainframes in rooms | A computer on every desk, then in every pocket |
| AI | A few shared models in data centers | ? |
That question mark is where we are now.
The personal computer didn't become personal by accident. Chips got cheap. Hobbyists got curious. Someone built a spreadsheet, and suddenly an ordinary person had a reason to own one.
The same ingredients are showing up in AI. Smaller models that run on a desk. Curious people building in their garages, or their home offices. And a search for the "spreadsheet moment": the one use that makes it obviously yours.
The question underneath
So for me, the real question isn't whether AI becomes personal.
It's who it belongs to, and what it's for.
If it stays central, AI could become the most powerful attention machine ever built. It would know us better than any feed and ask for more of us than any app.
If it becomes truly personal, owned, private, shaped by our own values, it might do something different.
It might give attention back.
There's a line from Ramana Maharshi on the front page of this site: "The world is nothing but a projection of the mind."
Our tools shape where the mind goes. For forty years, they mostly pointed it outward.
I'm curious what happens if we build one that helps point it home.
That's the experiment. Next, I'll explain how I got here and what I'm attempting.