The reason your last detox lasted nine days is not weak resolve. It is that you deleted a dependency without checking what depended on it.
When you remove a service from an architecture, you first enumerate its consumers. Nobody rips out the notification service on a Friday and finds out on Monday which four systems silently broke. But that is precisely how most people quit Instagram.
Enumerate the consumers first
For each app you want to cut, write down every genuine function it serves:
- Connection. Specific people you would actually lose contact with — named, not "everyone."
- Information. A feed you rely on for something concrete: your industry, a hobby, local events.
- Regulation. What it does for your mood at 11pm. This one is real, and pretending it isn't is why detoxes fail.
- Identity. The part of you that only exists when posted.
The first two have obvious replacements. The last two do not, and they are the actual dependencies.
Replace the function, then cut the tool
- Connection → a recurring standing call, or a group chat you actually keep.
- Information → RSS, two newsletters, one forum. Pull, not push.
- Regulation → a pre-committed alternative that is available at 11pm with no friction. A book on the pillow. A specific podcast. Not "meditate more."
- Identity → somewhere you still publish, but on your own schedule. This blog started as exactly that replacement.
Cut the app only after the replacements have been running for a week. Order matters. Cutting first and replacing later is how you end up reinstalling at 11:40pm on day nine, which I have done, twice.
Friction is a design tool, not a moral test
If a full cut is the wrong call — and often it is — then design the friction deliberately instead:
| Technique | What it actually changes |
|---|---|
| Log out after each session | Adds ~20 seconds and one decision point |
| Web-only, no native app | Removes push, removes the home-screen cue |
| Greyscale display | Strips the color reward from the feed |
| App on a second device left in a drawer | Converts an impulse into a walk |
None of these rely on willpower in the moment, which is exactly why they work. Every one of them moves the decision earlier, to a moment when you were calm and thinking clearly.
The honest part
I still use two of these platforms. I am not writing from a cabin. The goal was never purity — it was making the usage chosen rather than defaulted into.
A tool you open deliberately three times a week is a tool. The same tool opened forty times a day without a reason is an environment, and environments design behavior far more reliably than intentions do.