Social media was not deleted; instead, the mechanism that determines content visibility was deactivated. The resulting experience was less dramatic than anticipated, yet more unsettling.
A specific kind of restlessness typically manifests around day three. Upon opening an application, there is no curated content waiting. There is no queue assembled overnight by a system that has been studying user hesitations, nor content chosen because it maintained someone else's attention for eleven minutes the previous Tuesday. Instead, there is merely a reverse-chronological list of accounts that were intentionally followed, with posts appearing in the order they were published. It feels broken, but it is not.
That restlessness is the central point of the experiment.
The experiment was simple: one month with algorithmic feeds disabled. There was no "For You" page, no "Explore" tab, and no Reels served by a system designed to monitor user behavior more closely than most people in real life. Where platforms allowed a chronological setting, it was activated. Where they did not, the application was removed entirely.
Why now
The motivation stemmed from a growing distrust in the presented content. It was not a conspiratorial concern; rather, it was akin to losing trust in a mirror that consistently displays an idealized version of reality. The feed had become highly proficient at displaying content that would elicit a reaction, which is distinct from displaying content of actual interest. The gap between those two states had been quietly widening for years.
A research paper published in Nature earlier this year made this feeling harder to dismiss. A field experiment with nearly 5,000 active users on X found that seven weeks of exposure to an algorithmic feed shifted political attitudes in measurable directions—and that switching back to a chronological feed did not reverse those shifts. The algorithm does not simply change what is seen; it changes what feels normal. Turning it off does not undo that effect.
"The algorithm doesn't just change what you see. It changes what feels normal to you. And turning it off doesn't undo that."
That asymmetry is significant. Leaving the system does not provide a clean slate.
Week one: the logistics
The first week was primarily administrative. Each platform was examined to locate the necessary settings. On Instagram, the option was buried but available, allowing a switch to "Following" instead of "For You." On TikTok, no viable option existed, so the application was deleted. YouTube allows for the disabling of recommendations on the homepage, though the sidebar remains persistent. On X, a chronological feed exists but resets every session, demonstrating a clear preference from the platform for algorithmic usage.
It was surprising how few platforms treat the chronological feed as a serious feature. Most treat the algorithm as the core product, not as an option. Removing it felt like asking a casino if a user could play cards without the lights and music; technically, it was possible, but the environment was not built for it.
By the end of the first week, three applications had been removed entirely, and two others were set to chronological. The feed appeared strange and dated—similar to social media in 2013, before engagement optimization took hold. Scrolling would reach an actual bottom within minutes. There was nothing left to consume, and the phone was put down, not as an act of discipline, but because there was nothing else to do with it.
Week two: the withdrawal nobody warns you about
This phase is difficult to describe without sounding dramatic. It was not anxiety, exactly, but a low-grade irritability—the feeling of a habit with nowhere to go. The phone was picked up at habitual moments: waiting for coffee, sitting after lunch, or during the first thirty seconds of a commute. The gesture was automatic; the reward was missing.
The average attention span on a screen in 2026, down from 2.5 minutes in 2004. The feed didn't shorten your attention. It just kept capitalizing on it.
Research suggests that dopamine is not the pleasure chemical, but the "wanting" chemical. The feed does not necessarily provide pleasure; it provides the pursuit of it. Each scroll is a small bet. The algorithm excels at ensuring the next "card" is interesting enough to maintain engagement. When the table disappears, the urge to sit at it does not immediately vanish; this transition took approximately ten days.
Week three: finishing things
Around day eighteen, a quiet shift occurred. Tasks began to be finished. This was not a grand productivity transformation—there were no morning routines or output metrics. It was simply the ordinary experience of starting an article and reading it to the end, sitting with a thought long enough for it to develop, or watching a full video without exiting at the ninety-second mark.
A part of the brain called the Default Mode Network activates specifically during unstructured time. It is responsible for imagination, creative connection, and the specific kind of thinking that occurs during idle moments. Constant stimulation starves it. When gaps are always filled, the gaps never truly exist.
"The best ideas I had that month didn't come from anything I read. They came from the ten minutes before I found something to read."
The intent was not to achieve creativity, but to stop actively preventing it.
What chronological actually shows you
One unexpected realization was that the chronological feed is somewhat boring. This is not necessarily negative; it is merely an accurate reflection of reality. People post mundane things: thoughts that didn't quite land or photos from three days ago that received little engagement. A feed without an algorithm is a much more accurate picture of reality, where most activity is ordinary, posted imperfectly, and occurring at random times.
In contrast, the algorithmic feed had provided a "greatest-hits" version of every account followed. It felt richer because it was edited—similar to how a movie feels more eventful than a week of actual life. The edit had been mistaken for the thing itself. Switching to chronological felt like the difference between watching a documentary about a city and actually walking around it: less coherent, but more real.
Week four: the part I can't fully explain
By the final week, the phone was still present and applications were still being used, but the relationship had shifted in a way that is difficult to articulate. It felt less like a pull and more like a choice. Usage occurred when intended, and the application was closed once the task was complete. This represents the baseline of a healthy habit.
There was also a reported increase in presence during offline conversations. This was not a spiritual shift, but a literal one: there was no longer a mental feed being tracked while others were speaking. The cognitive residue of the scroll—the background hum of half-processed content—had quieted down. Research into "attention residue" finds that after switching tasks, it takes an average of twenty-three minutes for the brain to fully re-engage. Social media does not just consume time; it taxes the time on either side of it. A month without the algorithm does not erase that effect, but it stops compounding it daily.
What was returned to, and what was not
At the end of the month, some features were turned back on. Not everything. One application was reinstalled, chronological settings were maintained where available, and TikTok remained uninstalled—not as a moral stance, but because re-entering that environment felt like returning to a room that one was glad to leave.
The reality of this experiment is that it does not feel like freedom immediately. The first week feels like deprivation, the second like restlessness, the third is quiet, and the fourth is when it becomes clear that the quiet was always available. It had simply been filled with something else.
"The algorithm isn't showing you the world. It's showing you a version of the world calibrated to keep you looking at it."
The algorithmic feed does surface content that might otherwise be missed, and there is a convenience to having curation handled by a system. However, it is valuable to know what lies beneath it: what is sought when nothing is being pushed, and what holds attention when attention has not been pre-spent on something designed to capture it.
Turns out, long articles, specific newsletters, and trusted sources were preferred. It was a much smaller internet than the algorithm had suggested was necessary.
One month was sufficient to observe the structure of the habit. It was not necessary to continue the experiment; the true necessity was to stop mistaking the feed for one's own mind.
0 Comments