They Built the Swamp, Part Two: Why the Other Side Can't Build This

Part One left off with the Washington Times — a newspaper that absorbed somewhere between $1.7 and $2 billion in subsidies from the Moon family's Unification Church over its lifetime, running an annual tab of roughly $35 million, and existed primarily as a prestige project for a cult leader who wanted to shape Washington's conservative conversation. It was Ronald Reagan's preferred morning read. It was never a business.

That distinction — between media that works as a business and media that works as an influence operation indifferent to profit — turns out to be the key to understanding why the left has spent thirty years failing to build anything like the right's communications infrastructure. The failure was never primarily tactical. It was structural, rooted in a fundamental misreading of what the machine actually was and how it actually worked.

Three Models, Not One

The right-wing media ecosystem runs on three distinct economic models that tend to get collapsed into one, which is how you end up with the wrong diagnosis and the wrong prescription.

The first model is genuine profitability. Rush Limbaugh died worth an estimated $600 million. The Daily Wire, Ben Shapiro's operation, crossed $100 million in annual revenue in 2022 and finished 2023 with a profit and more than $200 million in top-line revenue. These are real businesses. The audience appetite for the product is genuine, the advertising and subscription revenue follows, and the model is self-sustaining. This is the layer most people think of when they think of conservative media.

The second model is the subsidized influence operation — the Washington Times version, in which someone with money and a political project decides the influence purchased by a money-losing property is worth more than the losses. The Koch network, which budgeted roughly $889 million for the 2016 election cycle, didn't fund profitable ventures. They funded think tanks, state-level advocacy organizations, legal training programs, and human capital pipelines — all of which produced political returns on investment rather than financial ones. The return wasn't revenue. It was a federal judiciary reshaped over decades, state legislatures flipped, regulatory frameworks dismantled. You don't build that by expecting it to pay for itself.

The third model is new, and it's the one that's hardest to see clearly because it doesn't look like media funding at all. That's Elon Musk's contribution, and we'll get to it.

The point is that if you look at the right's communications infrastructure and conclude that the lesson is "build profitable outrage media," you've identified one-third of the picture and missed the structural explanation for why the other two-thirds exists. The left has donors. It has never had donors who understood that the infrastructure investment — the thing that doesn't produce electoral results this cycle but reshapes the terrain for the next twenty years — is where the real leverage lives.

Why the Mirror Image Doesn't Work

The most visible attempt to build a liberal equivalent to conservative talk radio was Air America, which launched in April 2004 with genuine institutional backing, genuine talent — Al Franken, Rachel Maddow — and a grand theory of the case. It filed for its first bankruptcy in 2006. It was sold for $4.25 million. It shut down entirely in January 2010.

The conventional explanation is management dysfunction and undercapitalization, both of which are true. But they're symptoms, not causes. The deeper problem was that the format itself didn't map onto the audience it was trying to reach.

Conservative talk radio is architecturally authoritarian. One voice. No genuine dissent. The host defines reality; callers exist to validate or be dismissed. This is a cognitively comfortable environment for an audience that prefers strong leaders, clear in-group and out-group distinctions, and a consistent worldview that doesn't require constant re-evaluation. Political psychology research going back decades — Jonathan Haidt's work on moral foundations is probably the most accessible entry point — finds consistent differences between conservative and liberal audiences in their appetite for exactly these features. Conservative audiences tend to have higher tolerance for hierarchy and authority, higher sensitivity to perceived threat, and stronger preference for group cohesion and loyalty. The format feeds all of those preferences simultaneously.

Liberal audiences, broadly, are less comfortable with a single authoritative voice and more suspicious of their own side's propagandists. They're more likely to want their assumptions challenged, more tolerant of internal disagreement, and more likely to find three hours of validation exhausting rather than energizing. These are not purely virtues in a political combat context, but they are structural features that make the talk radio format a poor product fit.

Add the coalition fragmentation problem. A unified message requires a unified audience. Fox News tells a coherent story because its audience is relatively homogeneous in its identity and grievances. A Democratic equivalent would need to speak simultaneously to Black voters in Atlanta, Latino voters in Phoenix, college-educated suburban women in Philadelphia, and union households in Detroit — audiences with overlapping but genuinely distinct interests, cultural vocabularies, and communication styles. The frame that activates one can easily alienate another, and the party has never solved this tension, in part because the activist and media infrastructure rewards amplifying the tension rather than managing it.

There's also what I'd call the own-goal problem. Every time a left-leaning media figure goes off-message — lights up a Democratic politician, amplifies a factional dispute that lands badly outside the base, or uses vocabulary that resonates in progressive circles but sounds alienating to working-class voters — it immediately becomes opposition research material that well-funded conservative media can distribute at scale. The right's ecosystem enforces message discipline through audience expectation and financial incentive; voices that deviate from the orthodoxy lose their audiences. The left's equivalent actively rewards heterodoxy, which is admirable in an intellectual culture and counterproductive in an information war.

The Bannon Innovation

The right's media infrastructure was built to deliver a message. Steve Bannon's insight — and it was a genuine tactical innovation, whatever else you want to say about him — was that the infrastructure could be used for something entirely different: destroying the informational commons itself.

In 2018, Bannon told writer Michael Lewis, in an interview for Bloomberg: "The Democrats don't matter. The real opposition is the media. And the way to deal with them is to flood the zone with shit."

The logic, as analyst Jonathan Rauch later observed, is that this is not about persuasion. It's about disorientation. If you generate twenty simultaneous controversies in a week, the informational environment can only sustain a small number of them at high public attention. Each individual scandal or revelation that would have been career-ending in a previous era gets diluted by the next one before the first one fully registers. The fact-checking and correction machinery — which operates on the assumption that establishing a fact constrains the subsequent conversation — simply cannot run fast enough to matter.

This is asymmetric warfare against an opponent that still believes in the rules of the old game. The mainstream press operates on the premise that truth, once established, has weight. The firehose strategy defeats that premise by volume. By the time the correction runs, the original claim has already done its work, been shared by the amplification network, and been replaced by three new claims requiring correction. The press is playing checkers against an opponent who has flipped the board.

The Democrats have no equivalent, and the honest answer is that they probably shouldn't want a pure version of it — the corrosive effect on public epistemics isn't a feature you want to import just because it works tactically. But they've also failed to develop an effective counter-strategy, which is a different problem and a more solvable one.

The Third Model: Infrastructure Capture

Which brings us to Elon Musk, and why what he's doing is qualitatively different from every prior model.

Rupert Murdoch funded a media operation. The Koch brothers funded an influence infrastructure. Both are significant. Neither is what Musk is doing.

Musk owns the distribution infrastructure. Not a newspaper, not a network, not a think tank. The pipes. The printing press and the delivery trucks simultaneously, serving 250 million accounts, with complete discretionary control over whose content gets amplified and whose gets quietly deprioritized.

In July 2023, X launched a revenue-sharing program for content creators. The first recipients were, as the Washington Post documented in real time, high-profile right-wing influencers — Ian Miles Cheong, Benny Johnson, Ashley St. Clair — who posted screenshots of their payouts before the general announcement. The program requires payment for X Premium verification, runs on a proprietary algorithm, and is subject to "human review" for eligibility. Its mechanics are opaque by design.

But the direct payments are almost a secondary mechanism. The primary subsidy is algorithmic reach. When a preferred voice's post gets surfaced to millions of accounts that didn't seek it out, and a disfavored voice's post with identical engagement metrics reaches a fraction of that audience, the difference in reach is a subsidy worth more than any direct payment — and it appears on no campaign finance disclosure. Musk amplifies his own posts to an audience that would dwarf any single media outlet. He does the same, selectively, for voices he favors.

This is not a media business model. It's infrastructure capture. And it compounds in a way that prior models don't: the platform's reach makes it the de facto public square for political discourse, the algorithm shapes what that discourse looks like, and the owner controls the algorithm. The Washington Times influenced Washington's conservative breakfast conversation. X shapes the informational environment for a significant fraction of the politically engaged population in real time.

The left has no equivalent and no obvious path to one. You can't build a competing platform at that scale with donor money. You can regulate the existing one, if you can summon the political will — which requires winning elections — which requires winning the information environment — which the platform shapes. The circularity is not accidental.

The Sykes Problem, Revisited

Charlie Sykes left Milwaukee radio in December 2016. He's spent the years since as a prominent voice in the Never Trump ecosystem — The Bulwark, the podcast circuit, the Republican-in-exile role that MSNBC and similar outlets have found useful. He's been honest, in his books and interviews, about what he built and what it produced.

But there's a limit to the self-reckoning, and it's worth naming. Sykes spent 23 years building an audience whose defining characteristic was that it had been trained not to trust information from outside the ecosystem. He is now, from outside the ecosystem, trying to reach that audience. The irony writes itself, and he's aware of it.

What he's less focused on is the infrastructure question. His critique of what he built tends to center on the content — the distrust inoculation, the tribal validation, the gradual substitution of performance for principle. These are real. But the content existed within a structure — the Fairness Doctrine rollback, the AM radio economics, the vertical integration with Fox, the think tank scaffolding, the subsidized outlets, and now the algorithmic amplification — that no individual voice, however honest, can dismantle by podcasting into it.

The machine was built deliberately over forty years. It was funded by people who understood they were making infrastructure investments with long time horizons and non-financial returns. It was built by people who wrote memos about what they were doing and saved them in presidential libraries.

The response to it, if there is one, requires the same clarity about what's actually being built and why — something that neither party's current leadership has demonstrated much appetite for. Awareness of the mechanism is the beginning of that work. It is very much not the end.

Part One: They Built the Swamp — The Machine

Want to support me but not ready for a monthly commitment? Why not buy me a coffee!