What Social Media Did to the Definition of Success

What this argues. Not that social media is poison and everyone should delete it. The evidence for population-level harm is genuinely contested, and I get into that below. The argument is narrower: these platforms quietly changed what counts as success, from something you can assess privately to something that has to be witnessed to be real. That shift is doing more damage than the screen time itself.

Comparison was always here. The sample changed.

People have measured themselves against each other for as long as there have been people. Festinger wrote up social comparison theory in the 1950s, and the basic finding is unremarkable: in the absence of an objective standard, we work out how we are doing by looking at others. It is not a defect. It is how you calibrate anything when you have no ruler.

What changed is the sample you calibrate against. For almost all of human history the comparison set was the village: a few hundred people, observed in full, including their bad days. What you see now is a set assembled by a ranking system optimising for engagement, which means it is not a sample of people at all. It is a sample of moments, selected because they performed well.

The result is a broken sense of base rates. Scrolling gives you the strong impression that you are seeing the middle of a distribution when you are looking at its right tail, repeatedly, in a way no amount of knowing better seems to correct. You cannot reason your way out of a prior that gets updated three hundred times a day.

The asymmetry that makes it worse

Comparison would be survivable if it ran on equivalent information. It does not. You experience your own life from the inside, with every hesitation, every unglamorous stretch, every thing you got wrong. You experience other people's lives as a finished cut, produced with the specific intent of looking good.

So the comparison is not you against them. It is your unedited interior against their edited exterior, and that is a contest with a predetermined outcome. The most reliable way to feel behind is to compare a process with a highlight.

Turning regard into a number

The more consequential move was quantifying approval. Before, social regard was real but fuzzy. You knew roughly where you stood and the estimate was noisy enough to be survivable. Now it arrives as an integer, attached to each thing you make, visible to everyone, and comparable across time.

Numbers invite optimisation. This is the part engineers should recognise immediately, because it is Goodhart's law: when a measure becomes a target it stops being a good measure. Point a metric at a system and the system reorganises itself around the metric, including when the system is a person and the metric is other people's attention.

Applied to a self, that reorganisation looks like drift. You start making things you would have made anyway and posting them. Then you start noticing which ones do well. Then, without any deliberate decision, you are choosing projects partly on how they will read. The preference has been rewritten by the feedback signal, and from the inside it still feels like taste.

Success becomes whatever is legible

The damage is not that people post. It is that a generation learned to treat the unwitnessed parts of a life as though they did not happen, which is most of a life.

Here is the part I think matters most, and it gets discussed least.

Plenty of real accomplishment is illegible. Becoming competent at something difficult mostly looks like a long stretch of unremarkable effort with no visible milestone. Being reliable over years does not produce a moment. Getting slowly better at a craft, staying with a hard problem past the point it stopped being fun, quietly supporting people, changing your mind about something important: none of it photographs.

A public and quantified definition of success cannot see any of that, so it systematically undervalues it. What it can see is announcements. Offers, launches, admissions, numbers, awards, anything with a date and a headline. Those become the visible surface of a career, and then, by a short and unexamined step, they become the definition of one.

The cost is not that people chase visible milestones. Some of those milestones are worth having. The cost is that the invisible parts start to feel like they do not count, which means the years when you are actually getting good at something register as failure to yourself while they are happening.

Self-image under observation

There is a difference between doing something and being seen doing it, and the second one changes the first. Research on self-objectification describes people taking an observer's view of themselves, evaluating from the outside in. Platforms built on posting make that stance the default, because to post is to consider in advance how a thing will look to strangers.

Once that habit is established it does not switch off when you close the app. You catch yourself narrating ordinary experience as content. Your sense of who you are stops being something you have and becomes something you maintain, and maintenance is exhausting in a way that having is not.

Where I have to be careful

The honest position is that the population-level evidence is messier than either side of this argument usually admits. Orben and Przybylski's specification-curve work found that the association between digital technology use and adolescent wellbeing is small, comparable in size to effects nobody worries about. Other researchers, Haidt and Twenge most prominently, read the same period and argue for something much larger and specifically tied to smartphones and social platforms. That disagreement is live, the causal direction is hard to establish, and anyone telling you it is settled is telling you about themselves.

It also is not uniform. For people isolated by geography, illness or identity, these platforms are the reason they have a community at all, and that is not a small thing to weigh against a diffuse cost. Plenty of people use them lightly and are fine.

My argument does not depend on the strong version being true. Even if the wellbeing effect is as small as the most cautious estimate, the redefinition of success is a separate phenomenon and it is much easier to observe. You can see it in how people describe their own work.

The only thing that has helped me

Not a detox, which I have never managed for long. What helped was writing down what I would consider a good year if nobody found out about any of it. It is an uncomfortable exercise, mostly because the first list you produce is obviously performative and you have to throw it away and do it again.

What survives the second pass tends to be specific and unimpressive. Understand a thing I currently do not understand. Finish something I said I would finish. Be someone a couple of particular people can rely on. Get measurably better at one skill in a way I can test myself rather than announce.

That list is a private metric, and a private metric is much harder to game, because the only person available to fool is you. It does not make the comparison instinct go away. It does give you something to check against when the feed tells you that everyone your age is further along, which is a claim the feed is structurally incapable of evaluating.