I have deleted Twitter five times. Each time, I came back. The intervals between deletion and return have varied — three weeks in 2018, six months in 2020, two weeks in 2022, four months in 2024. The pattern has been consistent enough that I am no longer surprised by it. The surprise, when I come back, is that I had forgotten how bad it was. The relief, when I delete it again, is that I had forgotten how good it was to not have it.
This time, in August 2026, I deleted it for the sixth time. I have no plans to come back. I have made this statement before. I am making it again. The difference, this time, is structural. I have done something different.
This is the cycle, what I have learned from it, and what I did this time that I have not done before.
The cycle, in five iterations
2018: The first deletion
I joined Twitter in 2011 because I was at a conference where the speakers were tweeting during their own talks, and I felt excluded from the conversation. I deleted it the first time in 2018 because I noticed I was checking it during every commercial break while watching TV. The behavior was obvious. The fix was deletion.
I came back in three weeks because I had a work event where Twitter was the coordination channel. The deletion was impulsive. The return was necessary. The lesson I drew was that Twitter was a useful tool and the overuse was the problem, not the tool.
2020: The political deletion
I deleted it again during the 2020 US election because the platform was making me physically anxious. The return, six months later, was during a period when I had more time and the political content had moderated. I came back because I missed the people I followed, not the platform itself.
2022: The productivity deletion
I deleted it in 2022 because I had read Cal Newport's "Digital Minimalism" and decided that the platform was incompatible with deep work. I came back two weeks later because I had not done anything productive with the recovered time, and the failure to translate deletion into productivity felt like a reason to return.
2024: The boredom deletion
I deleted it in 2024 because I had stopped finding it entertaining. The platform had become angrier, the timeline more cluttered, the people I followed less interesting. I came back four months later because I had a minor public event that required a Twitter presence, and the return was easier than the workaround.
2026: The structural deletion
The 2026 deletion is the one I have more confidence in. The reason is what I did differently.
What I did differently this time
The pattern in the first four deletions was that I deleted the app and the account, and then either reactivated the account or created a new one when I needed Twitter for a specific reason. The structural problem was that I had not actually changed anything in my life that made Twitter unnecessary. The deletion was a behavior change, and behavior changes without structural changes usually fail.
What I did differently:
I identified the specific uses I had for Twitter. Not "I use Twitter" but specifically: which accounts I followed, what I got from them, what conversations I participated in, what value the platform added.
I found replacements for each use. For professional updates, RSS readers and Substack. For conversation with specific people, email and Signal. For news, traditional news sources. For entertainment, books and podcasts.
I told the people I cared about that I was leaving. Most people did not notice. A few people moved to other channels. Two people said they were considering the same thing.
I deleted the app, the account, and the email recovery option. This is the change that matters. The earlier deletions all left the option to reactivate. This deletion removes the option. The cost of return is now significantly higher than the cost of staying away.
I made a written commitment to a friend. Not a public post. A private email to a friend saying I had deleted it and would not be reactivating. The social accountability is the lever I had not pulled in previous iterations.
The combination of these changes is what makes this one different. The previous deletions were decisions. This one is a decision plus a structure that supports it.
What I miss
I will not pretend the deletion has been costless. There are things I miss.
The professional serendipity. Twitter was where I learned about half of the things I would not have learned about otherwise. Job opportunities. Conference announcements. Articles I would not have found. Book releases. This serendipity was genuine and valuable.
The conversations with people I do not see often. I had a small network of people I had been following for ten years, and the replies under their posts were a kind of ambient social connection. The replacement (email, Signal) is more effortful and less ambient.
The sense of being in a conversation. The platform made me feel like I was part of a larger professional conversation. The replacement is more curated and less chaotic. It is also less fun in some ways.
The novel information density. Twitter was the highest signal-to-noise platform I used for certain kinds of professional information. RSS comes close for technical content. For cultural and political information, nothing else has the same density.
These costs are real. They are also why the previous deletions failed — the value I was getting was real, and the deletion was giving up real value. What I had not done in previous iterations was honest about what was actually valuable versus what was habitual.
What I do not miss
More things than I expected.
The anger. The platform in 2026 was angrier than it was in 2018. The political content was a constant low-grade irritant. The replacement is calmer.
The performance anxiety. Posting on Twitter required me to think about how a post would be received, who would disagree, how to phrase things to minimize argument. The performance cost was real and constant. The deletion removed the performance.
The doomscrolling. The reflex to check the app during low moments — waiting in line, in a meeting that ran short, between tasks — was the largest single behavioral cost. The deletion removed the reflex entirely.
The comparison. Twitter was where I encountered the curated lives of people I knew and did not know. The comparison was mostly unconscious. The deletion has made it conscious in the opposite direction — I now notice when I am comparing and recognize the comparison as the problem, not the people being compared to.
The sleep impact. I checked the app within ten minutes of waking and within ten minutes of going to sleep. The first half of the day was shaped by what I had seen on Twitter before I had even brushed my teeth. The deletion has changed the morning routine in ways I underestimated.
What I learned about social media in general
The Twitter cycle taught me things that apply to other platforms.
The first ten days are the hardest. Deleting a habit-forming platform produces withdrawal-like symptoms. Irritability. Boredom. The reflex to check the phone. These pass in ten to fourteen days. The previous deletions failed because I relapsed before the withdrawal was over.
The replacement matters more than the deletion. Deleting without replacing leaves a vacuum that the platform will rush to fill. The replacements do not have to be perfect. They have to be present.
The structural changes are what make the deletion stick. Behavior changes without structural changes are willpower, and willpower fails. The 2026 deletion will succeed because I have removed the easy path back.
The loss is real but the gain is also real. The platforms provide value. The platforms also extract value. The honest accounting has to include both. My accounting in previous iterations was incomplete.
The honest summary
I do not know if this one will stick. I have been wrong five times before about exactly this question. What I know is that the structural changes make this iteration different from the previous ones, and the cost of return is high enough that I would need a real reason to come back.
If I do come back, I will write about it. If I do not, I will also write about that, eventually. For now, the platform is gone, the morning is quieter, the time I would have spent on it is being spent on other things, and the Sunday evening ritual is the only ritual I have kept.
The lesson, if there is one, is that deleting a platform is easy. Deleting it in a way that sticks is hard. The hard version requires identifying what you actually used it for, replacing those uses, removing the easy path back, and creating accountability. The easy version fails. The hard version might not.
Most of life is this. The pattern is what to look for.