I Sell AI for a Living. Take the Phone Out of Your House.

I sell AI skills for a living. I'm about to tell you to take the phone out of your house.

And not because I beat the thing. Because I lost to it — every night, for years — until I stopped trying to win and started removing it instead.

I'm the last person who should be telling you this. That's exactly why it's worth hearing.

A Bluetooth rotary phone did $789K last year

Cat Goetze is 29, a Stanford grad, and last year her company Physical Phones sold nearly eight hundred thousand dollars' worth of landline phones. Bluetooth handsets. A wall mount. A rotary dial you spin with your finger. She got the idea in 2023 after spending the pandemic trying to claw back her own screen time.

She's not alone. There's dumb.co — a $25 flip phone that forwards your number so you can ditch the smartphone for a few hours. There's the Flow Alarm Clock, a brick a teenager built so you have to physically stand up to kill your alarm instead of scrolling in bed.

The smart read on My First Million was that phone addiction is the biggest business opportunity of the decade, and the moat on these products isn't technology. It's the anti-positioning. The 1% who are exhausted by the giant.

I think that read is half right. The products sell because they're a decision someone already made for you. Buy the rotary phone, and you've outsourced the willpower. But the rotary phone is one frozen choice. What you actually need is the rule underneath it — the thing that decides, every time, what gets to come inside.

I'm the addict, not the guru

Here's who's talking.

I am the guy with the phone in his hand at 11pm. I've got a two-month-old asleep in the next room and a wife who runs our entire family on a shared calendar — and I'm the variable that doesn't sync. I show up to half of what she planned because the other half got eaten by YouTube Shorts I told myself I'd close after one.

And it's not even the phone, not really. Some nights it's the laptop — me starting one more business with AI at 2am, a thing I do not need to build, calling it work because there's a terminal open. That's the one I'm actually ashamed of. Junk is easy to call junk. The loop that wears the costume of productivity is the one that takes the most and asks for the least guilt in return.

The most profitable thing I do all week is fix washing machines in my garage. Hands dirty, no notifications, real money. The least profitable thing I do is whatever the feed convinces me is urgent. The gap between those two facts is the whole problem, and I've known it for a long time, and knowing it changed nothing.

So when I tell you to take the phone out of your house, understand: this is not advice from a man with discipline. It's a confession from one without it.

Willpower is a tax. A filter is a purchase.

One thing finally moved me.

Willpower is a tax you pay every day, and the price goes up when you're tired, stressed, or two months into no sleep. You will miss a payment. You always do. Discipline that depends on you being your best self at 11pm is not discipline. It's a coin flip you've decided to feel guilty about losing.

A filter is different. A filter is paid once. You make the hard call when you're clear-headed and rested, you build the rule, and then the rule holds the line on the nights you can't. The rotary phone works not because Cat Goetze's customers are strong, but because they removed the option. You can't doomscroll a phone that doesn't scroll.

Removal beats resistance. Every time.

The filter, stated

So here's the rule I'm trying to live by. One question, asked before any new piece of technology gets to enter the house:

What does this cost me around the people I love — and am I willing to pay it at the dinner table?

Default answer: no.

Not "no" forever. "No" until it earns a yes. New app, new device, new feed — it doesn't get the benefit of the doubt just because it's clever or free or everyone has it. It has to argue its way in past the people in the room with me. Most things can't. The smartphone, if it showed up new tomorrow and I ran it through that question honestly, would not survive it. Neither would most of what's on it.

The filter doesn't care whether the loop looks like waste or looks like work. Shorts at 11pm and a fresh repo at 2am are the same machine.

That's the move the dumb-phone companies are selling, packaged as a product. You don't have to buy the product. You can just run the question.

This is the same muscle as the work

This is why an AI guy gets to write this without being a hypocrite.

I've written before that the best work is defined by what you refuse to include — the Rick Rubin "reducer" idea, selection as the real skill. I've written about shipping 26 products and finally winning only when I cut down to one. And on the store where I sell these skills, I pruned the catalog from 61 down to 16, because nobody drowning in AI tools needs a longer list. They need someone to tell them what to ignore.

The analog filter is that exact same muscle, just pointed the other direction. AI is the most powerful production tool I've ever held. It makes infinite. The filter is the intake valve — it decides what gets in. Same discernment, two directions. The reason I'm any good with the tools is the same reason I'm trying to keep them out of my living room: I've spent two years learning what to leave out.

Subtraction in the work. Subtraction in the house. It's one skill.

What I'd actually do

Three things, in order of how much they cost you and how much they give back:

The phone sleeps outside the bedroom. Buy a $10 alarm clock. That single move kills the 11pm scroll and the 6am one, the two worst windows in the day. Hardest first step, biggest return.
One analog block a day, on the calendar, where she can see it. For me it's the garage and a washer. For you it's whatever pays or whatever heals — hands busy, phone in another room. Put it on the shared calendar so it's real to the people counting on you.
Run the question on the next new thing. Before the next app or device comes home: what does this cost me around the people I love? Default no. Make it earn the yes.

None of this requires you to be strong. That's the point. I'm not. I just got tired of losing to a rectangle, so I took it out of the room.

If you want the same discernment pointed at your AI tools — the 16 things worth keeping instead of the 61 that aren't — that's the whole idea behind what I build at apexskills.dev. But you don't need me for the important version of this. You need a $10 alarm clock and the willingness to say no to something clever.

Subtraction is the whole game. Start with the thing in your hand.