The Light Switch Test

It's 2026, end of the day, and I just want to accomplish a very simple task: turn on my TV and watch the next episode of a show. To do that I still need to perform three clicks with my Apple TV remote control. And that extra friction is there, every time. To contrast, I open my Music app, select an album or a song, hit play and it just plays. Same decade, same company crafting the software (Apple vs Apple!). Why does friction exist in TV and not in music?

What's utility? Computers, and thus software, are interfaces to assist people in getting from point A to point B. The higher the utility, the less the interface should get in the way. Intent → interface → realization. Maximum utility and the lowest possible friction is what interfaces should aspire to provide. The old and boring light switch is a prime example of utility. Dependable and simple to operate.

Now lets add the business layer. We now have intent → business → interface → realization. When the business dictates the experience, utility is lost.

Take the same light switch example and make it smart. You may get new features, and have to pay for a subscription. Perhaps you get more utility from it like better control of colors and dimming and scheduling. Then the company that owns it fails to make profit and needs to shut down. Now you're left with utility again. A boring and simple on and off toggle, which again, just works.

Why can you just play a song with one click and need at least three clicks to play a TV show (select show → skip recap → skip intro)? Because Apple Music doesn't care about which songs you play. Apple Music earns the same whichever song you play, because of the way music licensing works, so completing your intent is a profitable outcome. A show can spend $100–500k on its title sequence (FX, licensed music) and wants that investment watched, not skipped — so redirecting your intent is the profitable outcome. An interesting fact is that Netflix's skip intro button gets pressed 136 million times a day1, signaling awareness without providing a better default experience.

Having spent half a decade working on Meta Quest OS, I now have a better grasp of why friction exists: to move business metrics and satisfy internal competition among teams. Every surface is someone's team OKR. Friends are usually surprised when I mention that designers at Meta favor shipping solutions that are good for users, but often face incentive roadblocks. And they are surprised again when I point out that the utility of Meta apps as communication tools (WhatsApp\IG) is still there, despite users getting distracted by what is designed to steal their core intent: send a message to a friend or a business.

In the 90's, waiting was the computer's default state. In 2026, forcing people to wait is a business decision.

Imagine you physically turn a light switch off and a popup is displayed: are you sure you want to turn your light off? (No joke, it has shipped before.) Will LLM software save us all? I don't know, yet, but I have a sense (and some early experience) that it may. If anything it is going to equip curious minds to cook more software for themselves, without a business standing between intent and realization.

Footnotes

  1. https://about.netflix.com/en/news/looking-back-on-the-origin-of-skip-intro-five-years-later