Ezra Klein discusses the collapse of American political discourse, the attention economy’s corrosive effects on both parties, and his vision for a liberalism that builds—arguing that the path forward requires political virtue, institutional reform, and a public agenda for AI that goes beyond fear.
The attention economy as a tragedy of the commons
Attention is a finite collective resource being exhausted by algorithmic competition, much like a shared grazing field in the classic tragedy of the commons.
Political communication has degraded because the loudest, most extreme voices get noticed—exemplified by the official Democratic National Committee account replying “Shut up, you ugly fuck” to Stephen Miller, a post that reached roughly 50 million people.
Every medium changes its user: Marshall McLuhan’s idea that “the content of a medium is the juicy steak thrown to distract the watchdog of the mind” explains how platforms reshape how people think, not just what they think about.
Ezra’s own strategy for protecting his judgment: maintaining a strict “backstage” life, avoiding Twitter, limiting his weekly output to two podcast episodes and one column, and being intentional about when and how he absorbs criticism.
How algorithmic media deranges political discourse
Whoever dominates Twitter pays for it three to four years later: progressives dominated in 2020, convinced themselves of ideas that became liabilities in 2024, and now the right is repeating the same pattern with conspiratorial figures like Nick Fuentes and Tucker Carlson.
Politicians are now selected for their ability to earn attention rather than for policy competence—figures like Zora Neale Hurston, Spencer Pratt, Graham Platner, and Donald Trump all succeeded attentionally, while “boring talkers who are great at policy” cannot compete.
The left gave up on virtue politics; the right rejected it in favor of vice-maxing—creating a vacuum where young men seeking self-mastery and patriarchal guidance were left without constructive voices, opening the door to figures like Andrew Tate and Nick Fuentes.
The Abundance agenda and the fake civil war in the Democratic Party
Ezra and Derek Thompson’s book Abundance argues that liberalism has made it too hard to build things—housing, clean energy, infrastructure—in the places it governs, and that supply-side reform is essential.
The book became a flashpoint between the “populist” and “liberal” wings of the Democratic Party, but Ezra argues the fight is largely fake: figures across the spectrum (Mamdani, Nithya Raman, the new Democratic congressional caucus) are all embracing abundance-style ideas about cutting red tape to build more.
The real division is not between left and right but between those who want to make government functional and those captured by factional online discourse.
Ezra sees himself as a liberal in the tradition of Obama—believing in universal healthcare and economic egalitarianism, but also in the hard work of making a fractious multi-ethnic democracy function through political virtue and compromise.
Deregulation: not all deregulation is the same
Deregulation is a tool, not an ideology—what matters is the goal. Elon Musk’s deregulation is indiscriminate cost-cutting that guts state capacity; Ezra’s version targets specific rules that make affordable housing cost $1.2 million per unit or that prevent clean energy deployment.
Musk is a tragedy in Ezra’s view: a genius industrialist who built Tesla and SpaceX on government subsidies, then radicalized online and joined the Trump administration to cut without purpose.
The right regulates all the time; the left deregulates—the emotional reaction to the word “deregulation” shuts off thought rather than enabling it.
AI: from speculative fear to present-tense governance
Ezra takes AI risks seriously (he has a “p-doom” sitting in the back of his head) but argues the safety debate has been stuck in speculative future scenarios for too long.
The political system needs to regulate the AI that exists now, not endlessly debate hypothetical superintelligence—you cannot solve a problem whose shape you do not know.
Capability to wield power is more than intelligence; the superintelligence scenario assumes a recursively self-improving system that never makes mistakes in a world full of friction, which Ezra finds implausible.
Concrete regulatory priorities: much stronger public evaluation of AI systems, restrictions on AI for surveillance and kill-chain decisions, keeping a legible chain-of-reasoning notepad in English so AI reasoning is never fully opaque, and significant protections for children.
Ezra would not send his children to a cutting-edge AI school—he would choose a paper-and-pens school like St. John’s University, because what children need is the ability to be human beings, and AI will be available to them regardless.
The need for a public goods agenda for AI
Right now the private market decides what AI optimizes for; the public sector only thinks about what to prevent. Ezra wants a theory of AI goods.
Examples of what public investment could unlock: an IRS AI that does your taxes with you (the IRS already knows your income and the tax code); AI concierges for navigating government services; drug discovery for rare “orphan” diseases where there is no profit motive; solving mathematical problems through synthesis plus tireless computation, as demonstrated by the recent Erdős theorem result.
The model is Operation Warp Speed: the government said “if you invent this, we will buy it and distribute it at low cost”—the same advanced market commitment approach could be applied across many diseases and public problems.
Making government data legible to AI systems is a prerequisite—using the example of AlphaFold, which succeeded because the protein database was one of the cleanest scientific databases in existence.
Men, virtue, and the crisis of self-cultivation
The left became hostile to individualistic explanations for structural problems, which was a betrayal of the liberal tradition of self-cultivation running from Kant through Frederick Douglass and MLK.
This hostility was especially directed at male-coded self-help, creating a vacuum on the right that was filled by figures with no real concept of virtue.
Ezra argues that any healthy vision of masculinity must start with the reality that men are physically stronger and more aggressive due to testosterone, making self-mastery and constructive channeling of those impulses foundational.
He believes the cultural water has shifted—Gavin Newsom, Richard Reeves, and others are now engaging seriously with the problems of men and boys—but that talking about solutions still triggers hackles in online discourse.
A unifying frame may be emerging: rather than men versus women, the competition is increasingly humans versus machines, which opens space to talk about what is innate to human beings of both sexes.
Ezra’s personal practices for maintaining independence of mind
He creates a “portfolio of criticisms”—collecting negative articles and videos, then reviewing them in the morning when his resilience is highest, rather than absorbing them reactively at dinner or on the subway.
He invites critics on his show to deliberate rather than just exposing himself to algorithmically boosted anger.
He has identified what he thinks is productive but isn’t (Slack, unnecessary calls, sitting at his desk with the laptop open) and what isn’t considered productive but is (driving without podcasts, dinner with friends, lying in a hammock, travel).
He reads paper books for extended periods in beautiful spaces as the single most important thing he does for his work—books are a technology of thinking, not just information download.
He has become much more embodied in his decision-making, trusting physical signals (his “skin prickling” during conversations) over question documents, and sees cultivating intuition and taste as one of the most important and undeveloped capacities in modern education.
