I replaced our smart TV with an old VHS player. My kids watch less TV, fight less, and don't throw tantrums.

VHS vs. Smart TV: What One Family's Swap Reveals About AI's Friction Problem

By · Published November 8, 2026 · Updated November 8, 2026

We did something that sounded almost like a joke at first. We unplugged the smart TV, carried it out of the living room, and hooked up an old VHS player in its place. No apps. No recommendations. No autoplay. Just a tape, a play button, and a screen that shows one thing until someone decides to stop it.

The result was not what we expected. Our kids watch less TV. They fight less. The tantrums when the screen goes off, the ones that had become a normal part of the evening, mostly disappeared. Nobody had to install a parental control app. Nobody had to set a timer. The machine itself did the work, simply by being less helpful.

That small home experiment is a useful lens for one of the biggest questions in tech right now: as AI makes every product smoother, faster, and more personalised, what are we losing? And what should builders of AI systems do about it?

What Actually Happened When the Smart TV Left the Room

The interesting part is not that the kids watched less TV. It is why they watched less TV. The old VHS setup changed the shape of the experience in three ways.

1. The content ran out

A tape ends. That is it. There is no next episode queued up, no "because you watched" row, no countdown telling you the next thing starts in eight seconds. The end of a tape is a natural stopping point, and it arrives without anyone having to make a decision or win an argument. When the tape ends, the room gets quiet, and the kids find something else to do.

2. Choosing took real effort

Picking a tape means walking to the shelf, reading a label, sliding it in, and waiting. That tiny bit of work filters out the idle, bored, "I don't know what I want" scroll. When watching requires a small commitment up front, the default is not watching. With a smart TV, the default is always watching, because the next thing is already loaded.

3. There was no algorithm working against the parent

This is the big one. A smart TV is not a neutral screen. It is a recommendation engine with a display attached, and its job is to keep the session going. Every design choice, autoplay, previews, personalised rows, the seamless jump from one show to the next, points the same direction. A parent is competing against a system that never gets tired and never stops suggesting.

The VHS player has no such agenda. It plays what you put in it and then it stops.

The Hidden Cost of Frictionless Technology

Strip away the nostalgia and this is a story about friction, and about what happens when a product removes all of it.

For the last decade, the dominant design goal in consumer tech has been to reduce effort. Fewer taps. Fewer choices. Fewer decisions. Recommendation systems were built to do the choosing for you. AI is now pushing that logic to its limit: models that predict what you want before you ask, agents that complete tasks without your input, interfaces that adapt so smoothly you stop noticing them at all.

That is genuinely valuable in many contexts. But the VHS experiment shows the flip side. When you remove every stopping point, every decision, and every moment of waiting, you also remove the moments where a person would normally pause and ask, "Do I actually want to keep going?"

A system optimised for engagement will fill those pauses for you. A system optimised for your wellbeing would leave them in.

The tantrums in our house were never really about turning the TV off. They were about the mismatch between a child's nervous system and a machine designed never to let the ride end. Remove the machine's agenda, and the conflict mostly evaporates. The fight was never between the kid and the parent. It was between the kid and the algorithm.

Friction Is a Feature, Not a Bug

This reframes how we should think about AI product design. Friction has been treated as pure cost, something to measure and eliminate. But friction does real work. It creates reflection. It creates intention. It creates natural endpoints.

Consider how this plays out across a few areas:

What This Means for the Future of AI

As AI moves from a tool you open to an ambient layer that surrounds you, the friction question becomes urgent. A chatbot you visit is one thing. An assistant embedded in your phone, your car, your home, and your kids' devices is something else entirely. When the intelligence is everywhere, there is no shelf to walk to. There is no tape that ends.

That points to three shifts we should expect, and push for.

Stopping cues become a design category

Just as "dark mode" and "accessibility" went from afterthoughts to standard requirements, stopping cues will become a recognised part of product design. Natural endpoints. Session summaries. "You've been here a while" moments. AI makes these easy to build and easy to personalise, a system can learn what a good stopping point looks like for a given person and surface it at the right moment.

Engagement stops being the only scoreboard

Most consumer AI is measured on time spent, sessions, and retention. Those metrics reward the smart TV model. Expect growing pressure to measure something else: did the user get what they came for and leave satisfied? Did they regret the session afterwards? Those are harder to measure, but AI is actually good at inferring them, which means there is no excuse for not trying.

Agency becomes the premium feature

As more of the digital world becomes automated on our behalf, the ability to stay in the driver's seat becomes valuable. Products that let you see what the AI is doing, override it easily, and set your own boundaries will command trust. "Frictionless" will still sell, but "respectful" may be what people actually pay for once they have been burned by the alternative.

Practical Implications for Businesses

If you build or buy AI products, the VHS lesson has direct commercial implications.

For product teams: Audit your AI features for runaway loops. Where does your system create an endpoint, and where does it quietly remove one? If your roadmap is entirely about reducing effort, you are optimising on one axis and ignoring the other. Add friction deliberately where the stakes are high, spending, sharing, publishing, anything irreversible.

For marketers and growth teams: Short-term engagement gains from aggressive recommendation can create long-term brand damage. The family that rips out the smart TV is a customer you have lost permanently. Trust compounds; addiction erodes it.

For enterprise AI: The same principle applies inside companies. An AI agent that drafts, sends, and files without a checkpoint is not efficient, it is a liability. Human-in-the-loop is often framed as a technical safeguard. It is also a friction feature, and it belongs exactly where the consequences are large.

For parents and consumers: You do not need to go full VHS. But the experiment shows that the environment matters more than willpower. If a device is designed to keep going, you will lose that fight eventually. Change the device, not just the rules.

Actionable Insights

The Bigger Picture

The most important technology of the next few years will not be judged on how much it can do. It will be judged on how well it knows when to stop.

AI is getting very good at removing effort. That is a gift when the effort was wasted, the boring form, the repetitive task, the search through ten tabs. It is a trap when the effort was the point, the pause before a purchase, the decision to end a session, the moment a child learns to entertain themselves.

Our old VHS player is not smart. It cannot predict what anyone wants to watch. It has no personalisation engine and no engagement metrics. But it does one thing that our smart TV never did: it lets the evening end. And in a world where every system is optimised to keep us engaged, that turns out to be the most valuable feature of all.

TLDR: Replacing a smart TV with an old VHS player cut screen time, fighting, and tantrums, not because the tape was better, but because it removed the frictionless, never-ending loop that recommendation systems are built to create. For AI, the lesson is that friction is a feature: stopping cues, finite recommendations, human checkpoints, and wellbeing metrics matter as much as personalisation and speed. The next wave of AI products will be judged less on what they can do and more on how well they know when to stop.