Ezra Klein22 June 2026

The Collapse of American Politics - Ezra Klein - #1114

2Frameworks

15Insights

Frameworks in this episode

Self-Mastery6 steps

The Criticism Portfolio

Batch, time, and metabolize criticism so it sharpens you instead of steering you

Read →

Productivity5 steps

The Productivity Reclassification Audit

List what feels productive but isn't, and what doesn't feel productive but is

Read →

Insights & moments

The myth-busts, hot takes, explainers, and tools worth keeping.

Myth Buster· 1

Myth Buster53:30

Deregulation isn't owned by the right, and regulation isn't owned by the left

Klein rejects the idea that adding or removing rules is inherently good or bad — it depends entirely on the rule and the goal. His and Musk's uses of 'deregulation' share a word but have opposite aims. He illustrates with affordable housing costing far more than market-rate because public money triggers layered rules, and warns that emotional reactions to the tools themselves shut off thought.

Whether adding or removing rules is good depends on the rule and what you want more of.
Public-money affordable housing can cost 2-4x market rate because of stacked requirements.
Trump deregulating coal/oil isn't an abundance win because Klein doesn't want that output.
Treating deregulation as right-coded or regulation as left-coded is 'a way of shutting off your thought.'

“The idea that deregulation is owned by the right... it's not true, but it's a way of shutting off your thought.”
Ezra Klein · 59:30

#policy#regulation#housing#abundance

Hot Take· 3

Hot Take13:30

Why political speech is collapsing into a race to the bottom

Using the official Democrats account replying 'Shut up, you ugly fuck' to Stephen Miller, Klein frames the exchange as a tragedy of the commons: it's hard for an institution to be noticed, so the extreme bids win attention. Trump was the first mover in shifting what political communication is allowed to sound like, and everyone else is now aping it.

The vulgar reply reached tens of millions precisely because extremity earns attention.
Trump shifted the baseline of what political communication can sound like.
Boring, sober institutional accounts get ignored, so norms erode.
It's a 'way in' to attention, but Klein thinks the winning move will be the 'way out.'

“it is a degradation of the entire system”
Ezra Klein · 16:30

“the winning move in politics in the next couple years is going to be the way out not the way in”
Ezra Klein · 25:00

#politics#social media#norms#trump

Hot Take36:00

Whoever dominates Twitter pays for it three to four years later

Klein's theory: the faction that dominates the platform talks itself into wild ideas that come back to bite it. Progressives dominated in 2020 and paid in 2024 when old posts became attack ads; now the right dominates X and is attaching itself to conspiratorial voices that he predicts will hurt it in a few years.

2020 progressive dominance produced ideas that were used against Harris and Talarico later.
Musk buying Twitter drove the left off and opened the floodgates to the right.
The right is now talking itself into wilder, more conspiratorial positions.
These fast rises and falls can't be separated from algorithmic social media.

“Whoever dominates Twitter pays for it three to four years later”
Ezra Klein · 36:00

#politics#social media#twitter#prediction

Hot Take1:11:00

Regulate the AI that exists now, not the speculative superintelligence

Klein says the AI safety debate has been stuck in speculative future scenarios for too long. Just as labs argue they can't make AI safe without building it, regulators can't learn to regulate it without regulating it. He's skeptical of fast-takeoff scenarios and argues intelligence doesn't translate cleanly into power — the world is full of friction and even geniuses make mistakes.

You can't figure out how to regulate AI without actually regulating it in the present.
Klein doubts the extraordinary fast-takeoff scenario and its frictionless takeover logic.
Wielding power takes far more than intelligence; even Stalin and Trump made many mistakes.
His concrete buckets: fund evaluation capacity, regulate AI around kids, define AI goods, keep a legible English chain-of-reasoning.

“You cannot figure out how to regulate it unless you regulate it.”
Ezra Klein · 1:11:30

“The capability to wield power is more than intelligence, a lot more.”
Ezra Klein · 1:12:30

#ai#regulation#safety#policy

Explainer· 6

Explainer07:30

Every medium changes the user, not just the message

Drawing on McLuhan, Postman and Ong, Klein explains that we think we are consuming content when in fact the medium is always using and changing us. The algorithmic feed reshapes how long ideas should be and how they should feel, and the change runs bidirectionally through both what you absorb and how you output.

The content of a medium is a distraction while the medium rewires your expectations.
Optimizing to win on X can quietly degrade the way you think long-term.
Dostoevsky/Nietzsche's prose changed when they moved from handwriting to typewriter.
Local maximums on a platform can be long-term minimums for your actual work.

“The content of a medium is the juicy steak thrown to distract the watchdog of the mind.”
Ezra Klein · 08:00

“the trade you're making on that influence is that the way you think is degrading”
Ezra Klein · 09:30

#media theory#attention#technology#mcluhan

Explainer10:30

Attention is a commons, and we're all being 'attention fracked'

Klein proposes treating attention as a collective resource subject to a tragedy of the commons. As competition for it intensifies, everyone gets more aggressive to capture it, exhausting a shared public good. He describes political emails escalating to constant siren-emoji emergencies and X as 'gain-of-function research for takes.'

Attention is finite and held collectively, not just individually.
More competition drives everyone toward louder, more extreme bids.
The voices that break through the cacophony are the extreme ones.
Occasionally an engineered take escapes containment and does real damage.

“we are being like attention fracked”
Ezra Klein · 11:00

“X is like gain-of-function research for takes”
Ezra Klein · 11:30

#attention#politics#social media#commons

Explainer26:00

The Democratic 'civil war' over abundance is mostly fake

Klein's book Abundance became the text people used to stage a fight between the party's populist and liberal wings, but he argues the fight is largely invented online. In practice figures from Newsom to Mamdani to the New Democrats have all embraced making it easier to build, so the online factionalism doesn't match the real cross-cutting consensus.

Abundance is about why Democratic-run places find it hard to build homes and clean energy.
The people it criticized (Newsom, Obama) grabbed onto it; insurgents reacted against it.
Mamdani's 'block by block' housing plan is itself about cutting red tape to build.
Online discourse manufactures factional fights that don't reflect real practice.

“I just think most of this fight is fake”
Ezra Klein · 28:30

#politics#abundance#housing#democrats

Explainer30:30

Trump collapsed Republican purity to a single axis: loyalty to him

Klein contrasts the two parties' purity structures. Trump reduced GOP purity to one dimension — personal loyalty — accepting almost any view so long as you praise him. The left, lacking a single leader, has a plurality of ways to get in trouble. But loyalty is now getting more complex as Trump takes real actions like the Iran war, pushing out even MAGA figures like Greene and Massie.

The right has one big test; the left has many programmatic ones.
A single ordinating principle keeps a coalition cohesive but is now fracturing.
Loyalty used to be simple; now it means signing onto specific policy actions.
The left's programmatic tests keep it coherent but make it hard to welcome people in.

“he collapsed purity on the Republican Party down to a single dimensional point, which is loyalty to him”
Ezra Klein · 31:00

#politics#coalitions#trump#republicans

Explainer1:02:30

We have a theory of AI harms but no theory of AI goods

Klein argues the public conversation obsesses over what we don't want from AI and neglects what we want from it. For AI to solve a public problem, the problem needs money behind it and needs to be made legible via clean data — as the Protein Data Bank enabled AlphaFold. He proposes government advance-market commitments, like Operation Warp Speed, to steer AI at orphan diseases and other neglected public goods.

AI can only solve problems that are funded and made legible through structured data.
AlphaFold worked because the Protein Data Bank was an exceptionally clean database.
An IRS LLM could do your taxes; government could be a concierge to its own services.
A prize/advance-market-commitment system could target orphan diseases the market ignores.

“you also need a theory of AI goods. And right now we don't have one.”
Ezra Klein · 1:08:00

#ai#policy#public goods#science

Explainer1:31:00

The left gave up on self-cultivation and the right pushed toward vice

Klein argues liberalism came to see individual self-improvement as an excuse for ignoring structural problems, becoming hostile to self-cultivation especially when male-coded. Meanwhile the right's masculinity discourse moved from Jordan Peterson's focus on virtue toward figures Klein sees as having no real concept of virtue, leaving a vacuum that pulled in young, often fatherless men.

Liberalism has a long self-cultivation tradition (Mill, Kant, Douglass, MLK) it abandoned.
Self-help went down easily when therapeutic and female-coded, met pushback when male-coded.
The right moved from Peterson's virtue focus toward 'vice maxing.'
Opening a market for young men's guidance and then vacating it invites worse actors in.

“the left gave up on virtue. And the right rejected it.”
Ezra Klein · 1:34:30

“self-mastery and the constructive channeling of those impulses is like foundational to any healthy masculinity”
Ezra Klein · 1:46:30

#culture#masculinity#virtue#politics

Takeaway· 5

Takeaway01:30

Guard your backstage or the public version of you eats the real one

Klein argues that the input to good work is independence of mind, and that once the world's idea of you gets into your head it becomes poison. He urges anyone having a moment in the press to be intentional about preserving a private 'backstage,' and worries paternalistically about streamers who put their whole lives on camera.

Seeing yourself in the third person is dangerous for doing good work.
Everything you do creates more of itself and eats your time, so cut ruthlessly.
Klein organizes each week around just three outputs plus family, friends, and self-care.
Williamson's defense is deliberately keeping his private life boring so the internet moves on.

“Once the world's idea of you gets into your head, it is poison”
Ezra Klein · 01:30

“you really have to be intentional... about maintaining as much of a backstage as you can”
Ezra Klein · 02:00

#attention#privacy#creativity#fame

Takeaway38:00

Politics is endless pluralism; posting is just energizing your side to hate the other

Klein draws a sharp line between doing politics and doing posting. Politics is the constant balancing of disagreement and the work of making your ideas legible enough to win support. Posting habituates people to a weak form of politics: rallying your own side and hating the other, while mistaking that for progress.

The danger isn't holding certain ideas; it's not doing the work to make them legible to others.
Getting consumed talking to your own side blinds you to where everyone else actually is.
Posting optimizes for your specific echo chamber, not the median voter.
That optimization surfaces later as an attack ad in El Paso that 'wasn't for them.'

“instead of doing politics, you're doing posting”
Ezra Klein · 38:00

“Politics is an act of endless pluralism in a liberal democracy. And posting is not.”
Ezra Klein · 38:30

#politics#social media#persuasion#echo chamber

Takeaway49:30

Attention is now its own political competency

Klein argues that earning and wielding attention is now a standalone skill required to compete at the highest levels of politics — you can't just be great on policy and boring on delivery. He names Newsom, AOC, Buttigieg and dark-horse Jon Ossoff as Democrats who have figured attention out, and diagnoses a deeper flaw: Republicans are under-formed by institutions, Democrats over-formed.

If you can't break through on attention yourself, you can't compete at the top.
You have to sound like a real, authentic person, not an optimized institution.
Democrats risk becoming a party of 'Tracy Flick' — molded to sound like the institutions.
Keir Starmer is cited as someone who 'speaks like he is the government.'

“attention is its own competency now”
Ezra Klein · 50:00

“Republicans are under formed by institutions and Democrats are over formed by them”
Ezra Klein · 51:00

#politics#attention#authenticity#leadership

Takeaway1:21:00

Read paper books — they're a technology of thinking, not information

Klein's simplest advice to students asking about AI is to read books and paper. Books aren't primarily for downloading information; they're a scaffold that trains sustained attention and lets connections form in your mind. He warns that AI offers a constant simulacrum of productivity, and that people who lean on it heavily are often producing worse work, not better.

The value of a book is what happens in your mind while reading it, not just its content.
Reading trains sustained attention without reaching to resolve every question instantly.
AI's seduction is the feeling of productivity, not actual output.
Klein would school his kids on paper and pens over a cutting-edge AI school.

“the value of a book is not just the information on it. It is what happens in your mind when you read it.”
Ezra Klein · 1:22:00

“AI makes you feel superhuman and it's making you less than human”
Ezra Klein · 1:23:00

#attention#reading#ai#learning

Takeaway2:02:00

Feelings are intelligent — don't outsource your taste to AI

Klein makes the case that the body's signals are a form of intelligence worth cultivating, arguing his skin prickling, not his questions document, is what makes him a good interviewer. But he adds a Buddhist counterweight from Pema Chödrön: knowing what you feel matters not so you always obey it, but so you aren't unconsciously led around by contractions you never noticed.

'Facts don't care about your feelings' is wrong — you should care about your feelings.
Intuition and taste are hard to teach, idiosyncratic, and among the most important capacities.
The value is partly in noticing feelings so you aren't unconsciously steered by them.
AI can't feel how things feel; that embodied capacity is a human edge to develop.

“Feelings are very intelligent. And the dismissal of them... is a mistake”
Ezra Klein · 2:05:00

“Do not outsource your taste to the AI.”
Chris Williamson · 2:03:00

#intuition#emotions#ai#self-mastery