Every six months, someone proposes a new fix for social media.
Sometimes it is age verification. Sometimes it is content moderation. Sometimes it is breaking up the companies. Sometimes it is banning certain platforms entirely. Sometimes it is teaching media literacy in schools. Sometimes it is requiring algorithm transparency. Sometimes it is fining the platforms for harm done to users.
None of the fixes have worked. Some have made specific harms slightly better. The general condition — polarization, identity collapse, emotional volatility, performative outrage, the steady degradation of public discourse — has continued unabated, and in many measurable ways has gotten worse.
The fixes are not working because they are aimed at the wrong level. This chapter is about what level the actual problem is at.
What people think the problem is
The standard analysis of social media goes something like this. Platforms are designed to maximize engagement. Engagement is maximized by content that provokes strong emotional reactions. The strongest emotional reactions are produced by outrage, fear, and tribal identification. Therefore, platforms amplify outrage, fear, and tribal content. This makes users angry, divided, and unhappy. Solution: regulate the platforms.
This analysis is not wrong. The platforms do work this way. The amplification does happen. The harms are real.
What the analysis misses is why the amplification works in the first place. The platforms cannot make users feel outraged about something the users do not already have a felt response to. The platforms cannot make users tribal if there is no tribe in the user's identity for the platform to activate. The platforms are amplifying something. The thing they are amplifying is not produced by the platforms. It is brought to the platforms by the users.
What the actual problem is
The actual problem is that most modern users come to social media with externally regulated identities — identities that depend on the responses of other people for their stability. Approval calms them. Disagreement destabilizes them. Visibility regulates them. The platforms did not create this condition. The platforms found a population already in this condition and built a business model around amplifying the regulatory loop.
A user with an internally regulated identity uses social media differently. They post things they think are worth saying. They engage with things they find interesting. They are not pulled by the engagement metrics, because the engagement metrics are not what is regulating them. When they encounter content designed to provoke them, they notice the design and decline to react. When they encounter people who disagree with them, they consider the disagreement on its merits and respond or move on. They are unbothered by being unliked, because being liked was not the source of their stability to begin with.
Such users exist. They are rare. The platforms do not destabilize them, because they were not destabilizable to begin with.
The majority of users, however, did not arrive at the platforms with internally regulated identities. They arrived already partly dependent on external validation for their sense of self. The platforms then accelerated that dependency by offering a constant stream of validation and the constant possibility of its withdrawal. The dependency was preexisting. The platforms exploited it. They did not create it.
Why regulation cannot fix it
If the underlying issue is externally regulated identity, then regulating the platforms does not address the issue. It just changes which external regulator is in charge.
You can break up the platforms. The dependency moves to whatever new platforms emerge.
You can require content moderation. The dependency now sits in front of moderated content, but the user is still organized around external validation; the moderators have just become another external authority deciding what they are allowed to feel.
You can ban algorithms. The dependency now sits in front of chronological feeds, but the user is still scrolling for validation; the algorithm has just become slower at producing it.
You can teach media literacy. The user now knows the platform is manipulating them, and the dependency is unchanged because the dependency is not at the level of knowledge. It is at the level of nervous system regulation.
Each proposed fix replaces one external regulator with another. None of them addresses the question of whether the user is internally regulated. As long as the user is externally regulated, any platform — moderated, unmoderated, algorithmic, chronological, broken up, intact — will function as a regulatory prosthetic, and the harms of that prosthetic will reappear in slightly different form.
The category error
This is the category error at the heart of every social media reform proposal.
The proposals treat the symptoms as if they were the disease. They assume that if the platform behavior is fixed, the user behavior will follow. The reverse is closer to the truth. The user behavior produces the demand that the platform behavior meets. Fix the platform without changing the user, and the user will recreate the same dynamics on whatever surface is available.
This does not mean platforms are blameless. They are not. The amplification they perform is genuinely harmful. The business model is genuinely corrosive. Specific design choices have made specific harms worse. Where the platforms can be held accountable for specific harms, they should be.
It does mean that platform reform alone will not produce the outcomes people are hoping for. The thing the reformers want — a population that can hold complex conversations, tolerate disagreement, maintain stable identity across contexts, and not be manipulated by engagement metrics — is not deliverable through platform regulation. It requires the underlying condition to change. The platforms can be made less harmful. They cannot be made benign for a population that is not internally regulated to begin with.
What this looks like in practice
If you watch carefully, you can see the category error operating in real time.
A community is in an uproar about something on social media. The uproar is treated as a problem to be solved by the platform. The platform makes a change. The uproar continues, redirected at the next thing.
A politician proposes regulating social media to protect children. The regulation passes. Children are harmed by social media at slightly different rates, on slightly different platforms, in slightly different ways. The general harm continues.
An expert publishes a book about how the algorithms are damaging us. The book is widely read. Nothing changes, except that some readers now have a vocabulary for being damaged.
In each case, the proposed solution operates at the level of the platform. In each case, the underlying issue is unchanged. The damaged population produces the demand. The platforms meet the demand. The damage continues.
What would actually help
The interventions that would actually help are not platform interventions. They are interventions at the level of the user's underlying regulation.
Reducing the amount of childhood time spent in environments where identity is performed for an audience would help.
Restoring the kind of family, neighborhood, and community structures that used to do regulatory work without social media would help.
Slowing children's exposure to platforms designed for adults until their identities are more developed would help.
Building cultural permission for adults to disconnect, to be unwitnessed, to exist without producing content, would help.
Reducing the economic and professional pressures that make personal branding feel necessary would help.
None of these are platform regulations. All of them are at the level of how people live, what conditions they live under, and what kinds of regulation are available to them other than the regulation the platforms offer.
These are the harder fixes. They are at the level of culture, of household, of community, of the texture of daily life. They do not produce headlines. They do not produce political wins. They do not produce easy enforcement mechanisms.
They are also the only fixes that would actually work.
What this chapter is for
This chapter is not arguing against social media regulation. Where the platforms can be held accountable for specific harms, they should be.
It is arguing that platform regulation will not deliver what its advocates are hoping for, because the platforms are amplifying an underlying condition that the regulation does not touch.
If you want to see less harm in your own life, the move is not to wait for the platforms to be fixed. The move is to work on the regulation that you bring to the platforms. Use them less. Use them differently. Notice when they are activating you. Notice what is being activated. Notice what would have to be different in the rest of your life for the activation not to find anything to hold onto.
The platforms are not going away. The regulation, if it comes at all, will come slowly and will not address the underlying issue. The thing you can change is what you bring with you when you log on.
That is not a satisfying answer. It is also the only one that has ever worked.