Why I gave up on software.

The day my kid bypassed Apple Screen Time using a tutorial I had personally watched a year earlier was the day I stopped trying to outsmart him and started thinking about the wall.

When my son was nine, he disabled Apple Screen Time on my phone in under four minutes using a YouTube tutorial. I know it took him four minutes because I watched him do it. He did not know I was watching.

I had set up Screen Time the previous evening with what I thought was a clever PIN. I had read the App Store reviews. I had told my wife "this should hold." It held for less than 18 hours.

I had been here before.

Three years earlier, the router's built-in parental controls had fallen the same way. The kid figured out the device's MAC address, spoofed a different one in his phone's settings, and the whitelist treated him as a new device with no restrictions.

Two years after that, we paid for Circle Home Plus. It worked for about a month. Then he found the hardware reset button, held it for ten seconds, and the device factory-reset itself into a state where every household device had unrestricted access. He plugged it back in upside-down so the LED faced the wall.

Last year it was Bark, then Eero parental controls, then a smart plug with a schedule, then a manual unplug-and-take-the-cable-into-our-bedroom routine.

The pattern was the kid, not the product.

I work in hardware. I have spent two decades in product development. So I know what it looks like when a product gets out-engineered by its environment.

What was happening in my house was not a failure of any specific product. It was a structural feature of the category. Every parental-control product I bought was a software promise, and every software promise was a target for a smart, bored, motivated user with more time than I had.

The kid was always going to win the software war. He cared more about it than I did. He had eight hours a day to lose to the next workaround. I had ten minutes between work and dinner.

What I did instead.

I walked into the basement, found the router, and unplugged it.

The fight stopped in five seconds. There was no app to negotiate with. There was no schedule to argue around. There was no MAC address to spoof. The WiFi was just gone, the same way the lights are gone when you turn off a light switch.

My own laptop also lost connection. My wife's Kindle stopped syncing. The Sonos lost the radio station. The Roomba paused. I noticed all of these. So did the kid. Suddenly the rule was the same for everyone, which made it the rule for everyone.

And then the next night, I went down to the basement again. And the night after. And by the third night I was standing in a damp basement at 8:01 PM thinking, this is the best parental-control product I have ever used and it is also a 40-foot round trip.

So I built the button.

WiFiCurfew is the bedside button I wished I had when I was making that trip to the basement. The battery-free press button sits on my nightstand next to the lamp. The receiver outlet sits in line with the router on the other side of the house. I press the button before the lamp goes off. The router loses power. I press it again on my way out of bed in the morning, and the whole house comes back online.

v1 is deliberately blunt. Whole-house only. Manual press only. No app, no schedule, no per-device control. No battery either, because the press itself powers the signal. We chose this shape because every "smart" feature I encountered in the parental-control category became a way to lose. The dumb thing has no attack surface.

Yes, your own devices go down too. That is the cost. Yes, the kid will eventually figure out the button is on your nightstand. That is the cost. We are not selling you a miracle. We are selling you the one tool that worked for us when nothing else did.

The version 2 conversation is yours.

As customers ask for them, we will decide which features get added to v2. Per-device control. Scheduling. A lockable cover. A pair of synced buttons (one in your bedroom, one in your partner's). Probably one of those, not all of them. The point of v1 is to be small enough to actually ship and dumb enough to actually work.