Dominic Cummings - COVID, Brexit, & Fixing Western Governance
What this covers
Dominic Cummings sits with Dwarkesh Patel for a wide-ranging conversation on structural failures in Western governance, spanning his roles in Brexit and Number 10, the COVID response, defense policy, and what reform might look like. The central argument is stark: Western governments are paralyzed by architecture in which power and responsibility are decoupled, the permanent civil service has become a closed caste that weeds out its most capable members, and both politicians and media have optimized for theatrical appearance rather than results. Cummings contends that this is not incidental incompetence but systemic—rooted in 19th-century reforms that replaced aristocratic patronage with what was meant to be meritocratic permanence but evolved into self-perpetuating mediocrity. Only crisis-driven leaders willing to break these structures, he argues, can force meaningful change; absent that intervention, the baseline expectation is slow institutional rot punctuated by collapse and bloodshed.
The conversation moves across several distinct terrains. On civil service dysfunction, Cummings details how the most entrepreneurial and idealistic young talent self-selects out of Whitehall, leaving HR-compliant types to lead, and how a Secretary of State lacks authority to remove even one permanent official. On electoral dynamics, he argues the first past the post system makes replacing failing parties harder than capturing them. On talent and governance culture, he draws contrasts between Lee Kuan Yew's Singapore and the West's inability to recruit high-ability people into government, invoking Von Neumann thriving in Washington while Gödel was "completely useless practically." He turns contentious on nuclear strategy—calling Taiwan deterrence non-credible and dangerous—and on defense spending, dismissing aircraft carriers as vulnerable relics. The discussion also covers the Treasury's hidden power over Number 10, how media incentives trap Prime Ministers in reaction mode, and the paradox that systems resistant to autocracy are equally resistant to competent reform.
Cummings argues that Western governments are structurally broken because power and responsibility are divorced, the permanent civil service self-selects for incompetence, and politicians optimize for media theater rather than solving problems—so only a crisis-driven elite that prioritizes results over insider social status can reform them, and the baseline expectation is slow rot, collapse, and bloodshed.
- Only the PM can break bureaucratic bottlenecks; responsibility and authority are not delegated together, paralyzing the state even in crisis.
- The civil service weeds out its best people through HR-friendly self-selection, producing a closed, anti-meritocratic caste.
- Politicians and media operate a 'potemkin' theater that rewards appearance over results, leaving no one optimizing to actually solve problems.
This asset isn't compiled yet
You're seeing its claims, ranked. Compile it to build the argument threads, weight them, and check each claim against your library — the full view.
The civil service contains many very able people but mostly young; the most entrepreneurial and idealistic are weeded out by self-selection because they look at the promotion process and their bosses and refuse to make those compromises, leaving the most HR-compliant types to be promoted to lead supposedly fast-moving agencies, creating a self-reinforcing leadership culture that cannot be escaped.
“The most entrepreneurial, the kind of people who actually want to get on and do stuff now, leave and the most HR compliant, disastrous people to be in charge of supposedly fast moving agencies are the ones who are promoted to take over.”
Contrary to the rationalist/EA assumption that elites are rational and voters are dumb and easily manipulated by emotion, the truth is the opposite: educated elites are by far the easiest to manipulate with emotional propaganda, while voters are much tougher to fool.
“educated elites are by far the easiest to manipulate with emotional propaganda, and the voters are much tougher to fool. The rationalists think it's the other way around.”
The entire British constitutional and management system is built on the opposite of the principle that responsibility and authority should be delegated together; almost nobody ever has the authority to build or do something, which is why the state is paralyzed and bottlenecks cascade up to the PM as the only person who can give certain orders.
“responsibility and authority are always delegated together. The entire British constitutional system and management structure is based on the fundamentally opposite principle.”
The 1850s reform replaced an aristocratic patronage system (which was corrupt but faster and more efficient in crises) with a supposedly meritocratic permanent civil service that over time became a closed caste promoting almost 100% internally, thereby excluding nearly 100% of the world's most talented people and producing paralysis when crises hit.
“this supposedly meritocratic system ends up actually just being a closed caste. It's not actually meritocratic. It's a system that promotes practically 100% internally”
There is a fundamental paradox that the very characteristics making the British and American systems resistant to a Stalin-type figure taking over also make them chronically incompetent and equally resistant to a good statesman taking charge, which is why a Bismarck or Lee Kuan Yew type is far more likely to emerge in China or Russia than in the West.
“the very characteristics that make it very hard for a Stalin to take over also make it chronically incompetent as well.”
Close to 100% of the reality of conversations, motivations, and calculations behind events like the Brexit referendum and his time in Number 10 is lost and never recorded anywhere, which should make one far more skeptical of historical accounts; paradoxically, the digital age may preserve less than the letter-writing era because WhatsApp archives are rarely saved.
“close to 100% of the reality of conversations and motivations and calculations is basically lost. It's history. None of it is really recorded anywhere.”
The ability to synthesize across fields and function in a political environment depends on temperament and personality, not just intellectual ability—illustrated by von Neumann thriving in Washington while Gödel, of comparable genius, was completely useless practically—so reform needs multiple experiments: new broad courses for the young and bringing established elite talent into government.
“It totally depends on temperament and personality.”
Following Colonel Boyd's maxim 'people, ideas, machines, in that order,' Vote Leave built superior election tools not because technology came first but because they genuinely wanted to win and obsessed over voters; a culture with Bezos-like customer obsession will build and exploit tools far better than the old parties, so culture precedes technology.
“I would stress Colonel Boyd: People, ideas, machines, in that order.”
A British Secretary of State can legally fire only the three personal aides they bring in; they cannot remove a single one of the hundreds of thousands of permanent officials, and only the Prime Minister has the constitutional power to remove a senior official from a post.
“He legally cannot fire anybody else in the building.”
The first-past-the-post system makes it structurally very hard to replace an old party even with a superior new entity, because voters who hate both old parties still fear wasting their vote and letting the worse side win—making it easier to capture and transform an existing party (as Brexit nearly enabled) than to build and win with something new.
“even if you imagine this wonderful new entity run by the best and brightest in America with an agenda you actually agree with, even then you might not vote for it because you fear that it might let you.”
The Treasury (Number 11) has in many ways more power over Whitehall than Number 10 because it legally signs off the cheques, and it hides huge amounts of financial information from the Prime Minister, making it structurally anti-growth and preventing coherent economic decision-making.
“Number 11, the treasury, basically hides huge amounts of financial information from Number 10, which is obviously completely insane.”
Aircraft carriers are a massive waste of money because in war games against serious peer opponents they immediately flee to the edge to avoid being sunk; cheap asymmetric systems like drones can now destroy multi-billion-dollar platforms, a realization growing behind the scenes even as senior officials cannot admit it publicly.
“in war games, if they're dealing with any kind of serious peer opponent the British aircraft carriers immediately flee to the edge of the war game so that they don't get sunk immediately.”
Winning elections is actually the easy part because the more divorced from reality the old system becomes, the easier it is to win simply by focusing on the voters—as Vote Leave did in 2016 and 2019—while the old parties cannot do this and are not even optimizing to win, doing things daily that make no sense if winning were their goal.
“the more divorced from reality that the old system gets, the simpler it is to win an election just by actually focusing on the voters.”
Since Thatcher, Number 10 has been built as a press entertainment service that reacts to whatever the media says, rather than maintaining focus on what is actually important, which is the fundamental structural reason Prime Ministers fail to address long-term problems.
“fundamentally reorient Number 10, away from what it's been since Thatcher, which is a kind of press entertainment service.”
Extending a nuclear umbrella or strategic ambiguity over Taiwan is non-credible and dangerous because, unlike NATO's mutual defense which was credible for America, Taiwan is not existential for the US, so threatening nuclear war over a Chinese-populated island China sees as existential is either bluffing or insane—and 1914 shows how strategic ambiguity ends, now with weapons that kill a thousand times more, a thousand times faster.
“you can't go around threatening nuclear war over something where it's not actually a credible threat.”
In 100 years, what will matter most about today is what contributes to war, revolution, collapse, or regime change, and ideas matter most of all—but almost none of those ideas come from government; they come, almost by definition, from people who are fringe at the time.
“And ideas matter most of all. But almost none of that really comes from the government, right? It comes, almost by definition, from people who are fringe at the time.”
The single biggest cause of Britain ceasing to be the global superpower is World War I, which ended naval dominance, poured huge financial reserves into the trenches, killed a generation of leaders, and triggered the biggest transfer of property in Britain in 500 years—a decline from which Europe as a whole never recovered.
“the most important thing, obviously, is World War I. At the end of that the naval dominance was gone. Huge financial reserves poured into the mud of the trenches.”
Every regime is constantly changing and everyone in their own time wrongly believes their system will persist; the most any system persists is a few generations before chaos and change, so we cannot yet say democracy has proved to work—Western democracy is a tiny blip in the sweep of history, comparable to judging the Roman Republic at an early stage.
“regimes are constantly changing and everyone thinks in their own time that we're not going to change... But it's always wrong.”
There is no system that reliably both empowers a leader to control government effectively and reliably hands off power to capable successors; everything has its time of growth and decay depending on the culture, ideas, and public-spiritedness of elites, which normally implodes slowly, hits a crisis, is blown up, and is replaced amid bloodshed.
“There is no such system. That's why history is what it is.”
Lee Kuan Yew's most important achievement was creating a genuinely meritocratic leadership culture in Singapore where people actually try to solve problems and there is moral reinforcement for that culture, without which everything else collapses; its survival depends generation by generation on whether that moral leadership is maintained, since entropy defaults toward corruption.
“the most important thing he's created is a certain kind of culture amongst the leadership in Singapore, a genuinely meritocratic culture, a culture where people are actually trying to solve the problems.”
In many ways the British Prime Minister has more power and is less legally constrained than the US President because of ancient assumptions that the PM operates with the authority of the Crown and royal prerogative, allowing the PM to act in a crisis without necessarily getting parliamentary approval.
“the Prime Minister has more powers than the president does and is less legally constrained than the president is”
The West has simultaneously pushed Ukraine into a war of attrition with Russia, ratcheted up diplomacy against China over Taiwan by abandoning One China policy, thereby driving the world's biggest manufacturer into alliance with Russia, all while leaving a rotten defense procurement system unreformed—a scenario so incoherent it would seem too ludicrous to invent.
“We've pushed the world's biggest manufacturer into alliance with Russia. We have escalated a war of attrition. And we have left a completely rotten dysfunctional procurement and manufacturing system for defense continue.”
The incentive structure judges ministers by media performance rather than departmental outcomes, which incentivizes them to make friends with the media, leak everything, and treat plausible-sounding talk as the goal rather than truth or implementation—a self-reinforcing negative culture.
“By approaching government like that, you're incentivizing them to think that their goal is making friends with the media. So then they get good interviews. That also incentivizes them to leak everything.”
The reasonable conclusion from the public record on nuclear safety is that we survived the Cold War by extraordinary luck—such as a hydrogen bomb falling from a plane with all but one safety device failing—and there is no reason to expect that luck to continue, especially given how willing actors in Ukraine are to dance on the edge of the abyss.
“the only reasonable conclusion is that we've got extraordinarily lucky through the Cold War.”
A fundamental problem with the British state is prioritization of the Prime Minister's time, which is the single most important asset because, given a bureaucracy that cannot move quickly, the PM's decisions are the only thing that can break bureaucratic resistance—yet that time is constantly pulled in all directions by crises and media.
“the Prime Minister's time and the Prime Minister's prioritization is the most important asset.”
In Whitehall, writing policy and spinning it to the media is high status while implementation details are low status and formally separated, so flawed policies are written, handed off, and only later found unworkable by implementation people; co-locating policy and management people, as tried in 2020, surfaces fatal flaws on day one instead of 18 months later—fixing a broken debug loop.
“the policy people should be brought together and actually physically sit with the management people, which is completely revolutionary idea”
Because all bureaucracies end up operating for themselves, simply replacing old bureaucracies with new ones will reproduce the same dysfunction; the only way to sustain change over a century is to build institutional mechanisms and elite/public acceptance of constant reinvention—sunsetting, closing, and refounding institutions—as Singapore does by honestly reporting its own failures.
“you have to build into the system institutions and public acceptance and elite acceptance of a kind of constant reinvention and rejuvenation and closing and refounding of things.”
The madder the old political and media system gets, the stronger the cultural, financial, and personal disincentives for competent people (e.g. Silicon Valley builders) to get involved, so they retreat to walled gardens; this is a Catch-22 because reform requires precisely the able people the system is driving away, as illustrated by the backlash against Elon Musk.
“the more and more mad the old system of politics and the old media get, the stronger are the cultural, financial, personal disincentives to competent people getting involved.”
Classification in intelligence, nuclear, and defense allows the bureaucracy to hide extraordinary public disasters, budget problems, and chronic neglect; the British nuclear weapons infrastructure has tens of billions in unaccounted costs and parts that don't work safely, and China's infiltration of UK critical infrastructure is far worse than almost any MP comprehends.
“you have literally tens of billions of pounds that are going to have to be spent on the nuclear weapons infrastructure that don't appear in the official accounts at all.”
Real change cannot come from one person; it requires a subset of the elite—even as few as ten able people with a common set of goals—who decide their fundamental purpose is to solve a defined set of problems rather than maneuver the social hierarchy of insiders, and who then build a network beyond themselves; this rarely happens except after a disaster.
“It can't be just one person. There needs to be some subset of the elite... that actually decides that its fundamental goal is to solve a set of problems.”
Thatcher's single biggest mistake was never getting to grips with the permanent civil service; she did only sporadic, half-hearted, ineffectual firings and purges, which contributed to her downfall, and she did not actually control the government in the way Churchill did.
“The single biggest mistake Thatcher made was that she never actually got to grips with the permanent state, the permanent civil service.”
Even leaving aside escalation to nuclear conflict, on a conventional basis a war over Taiwan would be a catastrophe for America, as much classified Pentagon analysis and people inside the system already know, documented in works like 'The Kill Chain'.
“if America has a war over Taiwan then it's going to be catastrophic for America.”
The Cabinet is bullshit theater: meetings are literally scripted, ministers read out talking points written by officials, and conclusions are pre-written, while the people actually running government are unknown officials including 28-31 year olds in the PM's private office who hold more real power than the televised ministers.
“The meetings are literally scripted. People are given their talking points to read out by the officials.”
Western intelligence agencies have extraordinary technical capabilities (phone access, secret information acquisition) but in Britain these are badly prioritized and not aimed aggressively enough, while the quality of analysis—predicting how people will behave or whether they will start a war—is much worse, and there has been almost no focus on AI.
“amazing capabilities in various ways but badly focused, badly prioritized. But also the quality of analysis is much, much worse.”
You cannot reform a dysfunctional department like the Ministry of Defence by sitting down inside it to discuss procurement reform; the only thing with a chance of success is to build a parallel new system, then close the existing one and remove on the order of 90% of the people involved, while culling communications teams by 95-99%.
“you have to set up a parallel thing... and then you have to close the existing procurement system and get rid of 95% to 99% of the people involved with it.”
Fighting NIMBYism through a communication campaign to persuade the public to cheer house-building is a fool's errand; the correct strategy is to impose a communications blackout, just do the reforms quietly while talking about other things, and only later point to the visible results—affordable houses, thriving businesses—as the argument.
“I actually basically had a communications blackout on the whole thing... my approach to it then was just do it, don't talk about it, and talk about other things.”
History tells you that what happened to everyone else—disaster, collapse, bloodshed—is likely to happen to us, and Cummings' overall assumption is failure; but that is not a reason to give up: you have to try even though you don't think it will work.
“My overall assumption is failure. That the Western world will go the same way that everyone else goes in history... but still, you got to try.”
Bismarck's career is a living illustration of the AI alignment problem: a far more able agent that defines success as expansion of its own power and freedom of maneuver, treats every attempt to align its goals with broader goals as enemy action to be preemptively destroyed, defeats every safety feature, and cannot be switched off until it dies of old age—after which others felt nostalgic for its rationality, since at least it was negotiable rather than a berserk pirate.
“You see something that's much more able than its competitors and it defines success fundamentally as expansion of its own power, which also means its own freedom of maneuver. And it treats all attempts to align its goals with broader goals as enemy action to be destroyed.”
Britain should not merely copy what works in the US; it needs frontier research and to build things—reforming zoning laws, the startup ecosystem, and heavy regulation—not only for economic growth but for security and national independence, so that in 20 years its choices are not limited to buying American or Chinese.
“How much of Europe in 20 years time is actually going to have meaningfully independent capabilities in lots of areas and how much of its choice will only be — do we buy American or do we buy Chinese?”
Russia is a mafia state where most participants in his airline venture wanted to steal from it and move money offshore rather than make it succeed; the lesson—that people's actual incentives diverge sharply from the ostensible purpose—was a formative failure that later made Cummings more careful about questioning his own assumptions in politics.
“what most Russians wanted to do was steal from the airline and move the money offshore. And I never realized that.”
Because politics is the hardest, most complex domain, the people who do well at the top must look across many fields; an Odyssean education would train synthesizers by combining a broad mix of intellectual subjects (Thucydides, statistics, Tetlock-style forecasting) with constant movement between the theoretical and the intensely practical (working at startups, the military, ER wards), though success depends heavily on individual temperament.
“you also need some people who can look across multiple domains at the same time.”
It is impossible to be a good Prime Minister while accepting the way the job currently works, because you are buffeted by media events and do not actually control the government; even FDR, Bismarck, or General Groves would fail if forced to operate the way the prevailing system requires.
“It's impossible to be a good PM if you accept the way that the job currently works.”
At the COVID peak in March 2020, ordered PPE was not due until summer because it was shipped by sea from China; only the PM's personal authority to override EU and Treasury rules allowed grounded airliners to be commandeered to fly the supplies back immediately, illustrating how only the PM can cut through bureaucracy in a crisis.
“only the Prime Minister could actually cut through all the bureaucracy and say, Ignore these EU rules on Blah.”
After COVID, most of the people who were most wrong were given awards, honors, and promotions, while those who were repeatedly right have almost all left, sending a clear signal that officials are rewarded for following insane processes and punished for trying to do better.
“most of the people who were most wrong were given awards and honors by the system and promoted to new jobs. The people who were most obviously repeatedly right have almost all left.”
The fundamental golden rule of politics is that the people in it are almost never actually trying to solve the problem and don't care about solving it; EA/rationalist plans go wrong because they wrongly assume the political people they talk to genuinely care about solving Problem X.
“the people in politics are almost never actually trying to solve the problem and don't care about solving the problem.”
The PPE degree path, which most PMs and many leaders pass through, instills 'wordcel bluffing'—skimming a few books, producing plausible-sounding sentences with social polish—and trains people to value the plausible conversation-winning phrase over truth, the actual answer, or multi-year implementation, which is a terrible toolkit for a problem like COVID.
“I think it encourages this whole kind of wordcel bluffing. That everything just becomes that you spend a week, you skim through a few books, you come up with some vaguely plausible stuff”
The British government functions like a company where the Prime Minister is a CEO dealing with constant operational fires but there is no CEO with a long-term plan and no background team plodding on with the genuinely important agenda.
“it sounds like it's a company with the Prime Minister as a CEO who's dealing with all these operational things that come up, but there is no CEO who has a long term plan.”
Having actual priorities one genuinely cares about, combined with not worrying what the media says or about being hated, is extremely rare and a powerful advantage in politics, because insiders will seek deals with someone who clearly cares, and opponents who dismiss such a person as crazy repeatedly fool themselves and fail to learn from defeats.
“Having priorities, actual priorities, is incredibly rare in politics. Almost no MPs have them all.”
Very able people are highly interested in government and would take roles like running MI6, but they refuse because they expect to be stuck in endless stupid meetings unable to do anything; the 2020 chaos proved that when the system opens up and real impact is possible, exceptional people flood in, and they leave again when it closes back down.
“The fundamental problem is that they don't want to get involved because rightly they say, "I'll just be in endless stupid meetings. I can't actually do anything."”
There is roughly a financial crisis every decade, and it would not be surprising if within the next two years there were a crisis worse than the 2008 crash, a basis on which many hedge fund people are reportedly planning.
“It wouldn't surprise me if within the next two years there's a worse one than the 2008 crash.”