Regime Change #2: A plea to Silicon Valley - start a project NOW to write the plan for the next GOP candidate

The goal is not 'reform' but *a government that actually controls the government*

- If the rule you followed brought you to this, of what use was the rule?

- I dont know what you’re talking about.

- I’m talking about your life. In which now everything can be seen at once… How did you let yourself get in this situation?

Cormac McCarthy, p174 No Country for Old Men.

‘Party associations are not based upon any established law nor do they seek the public good, they are formed in defiance of the laws and from self-interest…’ Thucydides Book III.

‘[Trump] may not be good for America but he’s damn good for CBS’. CEO of CBS.

[1/10 Some thoughts added after feedback at end]


  1. In Washington as in London, the golden rule of Government is — the government does not control the government and anybody who tries to change this is seen as the enemy by the bureaucracies that actually control ~99% of the government. Politicians talk as if the government controls the government and fundraise as if it does. The media reports as if it does. It does not.

  2. Iraq, the 2008 financial crisis, covid, the Afghanistan debacle — all are impossible to understand unless you understand this rule.

  3. Almost no ‘conservatives’ in the UK understand this rule, why ‘reform’ is so extremely hard, and what to do about it. There is almost no public discussion in conventional UK media about how power actually works in the British state and this seems to be roughly the same in the US. For example, even after >100k avoidable deaths the entire UK political media provides no analysis of the crucial bureuacratic engine, ‘the Covid Taskforce’, which actually controls covid policy (who controls it, how does it work). Instead covid policy is portrayed as the product of ‘Cabinet discussions/rows’ — but Cabinet is actually almost totally irrelevant, and usually not even consulted. Even the greatest researchers and extremely able people who have built multi-billion dollar companies often talk as if the President/PM controls the bureaucracies and can instruct them to do X or Y, and they do X or Y, just as they routinely discuss politics as if a) there must be a hidden plan that renders the visible chaos more rational and b) politicians are ‘at least’ rational about electoral/communication strategy (they usually aren’t).

  4. America needs a government that controls the government, as it did under FDR and Lincoln, and shatters the party structure which is a plague. You don’t control the government unless you can shut down parts of the ‘permanent’ bureaucracy and you can’t legally do this in the current regime without grabbing control of a party. It’s hard to imagine sane politics over the next 50 years without somehow closing (or at very least ‘changing beyond recognition’) the GOP, Democrats, Tories, Labour.

  5. Biden is rubbish and will deteriorate. 4 years of him will be bad, 8 will be worse. Many bad trends will deteriorate the longer he is there. If he dies the VP will be even worse. The Democratic Party is firmly in the grip of a generation of activists deranged by Ivy League insanity, BLM etc. It will not be able to focus usefully on serious issues until that mad energy has burned itself out. Even if it did the White House could not make the DC supertanker turn more than a fraction.

  6. Trump has good reasons to run again. Structural factors mean Trump, despite being Trump, will start close. His chances of winning in 2024 are, now, higher than most realise, partly because the media exaggerated the competence of the Biden administration in its zeal to trash Trump.

  7. Whether Trump wins or loses his candidacy will be terrible for everybody. He demonstrated no interest in actually controlling the government. He didn’t drain even a corner of the swamp, he just annoyed it.

  8. A tiny and cheap (<$2-3 million) project — independent-for-now of any candidate — should start now to figure out how a GOP candidate could win, how they could actually control the government after they win, and who the candidate should be.

  9. The sooner it starts the greater the chance that someone serious can be found and helped early enough to stop Trump. There is an obvious danger that various people think — leave 2024 alone, I don’t need Trump attacking me and whether he or Biden wins they’ll both be gone in 2028.

  10. Even if the project fails to propel a candidate to the Presidency in 2024, the predictive model of the electorate it should build will be valuable so the project can hardly fail to repay the original investment. If it fails short-term — no viable candidate/plan — at the very least you’ll be able to sell the model to hedge funds (and/or use yourselves for trading) and make a ?10X return.

  11. If the project hits the jackpot and propels a candidate to President, it could also make the difference between another failure or a rare historical change. I built such a discrete project 2017-19. I’d watched how Insiders blundered in 2016 and lost a referendum they should never have lost. I thought the government could collapse and another opportunity present itself. It did in summer 2019. The preparation proved very valuable when we captured No10 and it arguably made the difference between us escaping a once-a-century constitutional crisis, winning the election and having a plan — or not. The rate of return if you do something really cheap (my project was <500k) that affects hundreds of millions of people and trillions of dollars is greater than any VC bet.

  12. A project like this must be done away from DC, the conventional GOP power centres and the consultant class. While there is a need for new institutions, this project should not be one. It should disappear in a puff of smoke, not try to justify permanent existence to donors. The Valley is the natural place to build the best model of the electorate and some weird subculture there is more likely than DC to look at the problem with the fresh eyes it needs. It’s also the natural place to think about how to have startups replace parts of the US government and the sort of policies to pursue if you can control the government.

This is a sketch to get people thinking about what a detailed plan for a serious project would look like, it’s not a ‘plan’.

I want some feedback and I’ll post a public version of this shortly…

Biden is useless, the next Presidential cycle starts in 15 months, Trump could win

After the 2020 election, as usual with the political media they swung from one emotional wave (Trump is the end of the world) to another (Biden is brilliant) with little sense of realism.

Biden is Biden. He’s rubbish relative to the level of Bill Clinton or Obama. The Obama campaign and White House staff told people over the years that he was not a serious figure. He was marginalised by Obama’s staff for good reasons. And he’s old, mentally frail (putting it kindly), clearly even less up to the job than in the 1980s, and getting older/frailer. Biden will be an even worse candidate in 2024 than he was in 2020.

Mid-terms almost always lead to losses for incumbents. Many structural factors make it hard for the Democrats to keep control. And many structural factors have pulled and will pull Biden towards positions that will alienate millions. For example, those who work in Democrat campaigns are much younger, much more educated and much more left than the median voter. They routinely push Democrat candidates to do things that are objectively irrational from the perspective of winning, such as running ads against Trump that actually help Trump (cf. @davidshor for details, e.g HERE which will tell you much more about campaigns than every NYT/Guardian oped combined for the next four years). As Andrew Sullivan has pointed out, if Obama gave some of his 2007-8 speeches now he’d be cancelled by his own activists.

The next US Presidential cycle starts by Christmas 2022 at the latest — just ~15 months away.

Trump has many reasons to run again. If he wins he can squash lawsuits and make a lot of money by exploiting data and generating a media business.

Trump could have taken another branching history in February 2020 and stomped on covid — closed the borders (with the Democrats screaming ‘racist’), forced strict lockdowns, controlled the virus, and fired/appointed people to do vaccines at war-speed (with the Democrats screaming ‘fascist’). If he had he would have won. He wasn’t far off winning in our branching history even after achieving little, making an epic farce of 2020 and running a dire campaign.

He has a lot of data, a lot of activists and a lot of people scared/wary of him. The GOP lost the plot and shows no sign of regaining it. While he’s always capable of blowing himself up, he could also, as of now, easily win the nomination again.

Trump as GOP candidate is a disaster whether he wins or loses

If Trump is the GOP candidate in 2024, it’s a disaster for America and the world.

If he loses, it’s four more years of Biden/Kamala.

If he wins it’s four more years of Trump.

And Trump showed:

1. He does not understand power in Washington.

2. He doesn’t have a CEO mindset or skillset in the Bezos/Gates/Jobs/Musk sense of being able to execute at scale and speed.

3. Like Boris Johnson, his insecurities mean he can’t face his lack of skills and trust/empower anyone to build the team to run the administration for him.

4. He has some showman skills, a good nickname game and a sporadically good twitter game. But like Johnson, he prefers to spend his time babbling about and at the media rather than the (often mind-numbing) problems of institutions and incentives you need to focus on to change big things.

This combination meant Trump made a lot of noise but got very little done.

He could not control the government. He was sometimes right, sometimes wrong, sometimes idiotic, often right in his complaints that the media were lying, but very little he said mattered because his words did not connect to power. He annoyed the swamp but he couldn’t drain the swamp — not the tiniest corner. From the intelligence services to the public health bureaucracy, he ranted and threatened but he could not persuade them to change, he could not appoint people to make them change, he couldn’t close them, he couldn’t replace them. He could not even force the deep state to vet his appointments some of whom were given the run around for over 18 months by the FBI before giving up in despair. If you can’t force the FBI to vet your appointments, you can’t do much of anything. And though he hated the media, was seriously weakened by them, and wanted to weaken them he actually flooded his enemies with a wave of energy and money! As the boss of CBS said, Trump ‘may not be good for America but he’s damn good for CBS’.

He does not understand what happened to him. So even if you understandably hate the NYT and CNN and have some lingering sense of ‘at least Trump was against the forces I hate’, backing him again is guaranteed failure and disappointment.

America needs ‘a government that controls the government’: close & replace with startups, don’t ‘reform’

America, like the rest of the West, is mostly run by permanent bureaucracies. Elections are to a large extent bad showbiz. The noise is high but the stakes often amazingly low. The parties scream about each other but generally whether X or Y wins changes an amazingly small fraction of policy, money, or real power — and has little effect on the permanent bureaucracies. (One of the reasons the Brexit referendum was different is it led to much actual weeping across Whitehall on 24/6/16 as the permanent bureaucracies faced something new — real change for them. Trump’s victory was sold as the same but clearly was not.)

The governments don’t control the governments. Conservative parties don’t want to control the governments and don’t know how to even if/when they do want to. Anti-conservative parties largely support the permanent bureaucracies and want more of them more than they want to escape the effects of being bogged down by them. The permanent bureaucracies certainly don’t want anybody elected controlling the governments, and they don’t even run themselves themselves! — nobody ‘runs’ them, everybody can veto everything but nobody has the authority to run them in the way effective organisations are run. The media portrays a ‘conservative’ government actually controlling the government as proto-fascist. And in the US/UK the courts increasingly use administrative law and judicial review to make it impossible for the government to control the government (‘the rule of law’ is now often used as a slogan to justify judges deciding political issues, which is a novel idea and an excellent device for Harvard/Oxbridge/media/officials to control/slow any executive acting outside their Overton Window of acceptable behaviour).

When nobody is in charge, you have chaos. Nobody is in charge of western governments. We see chaos everywhere. We see a chronic inability to think about hard problems under extreme uncertainty, decide and act at speed and scale. We see governments unable to escape the delusion that government largely involves chasing the media all day then cocktails with them by night. The political media is dominated by a subset of graduates who, like Oblonsky in Anna Karenina, largely cannot think for themselves and simply absorb and emit leftist political ‘views’ like clothes fashions.

Oblonsky never chose his tendencies and opinions any more than he chose the style of his hat or coat. He always wore those which happened to be in fashion. Moving in a certain circle where a desire for some form of mental activity was part of maturity, he was obliged to hold views in the same way he was obliged to wear a hat. If he had a reason for preferring Liberalism to the Conservatism of many in his set, it was not that he considered the liberal outlook more rational but because it corresponded better with his mode of life… Liberalism had become a habit with Oblonsky and he enjoyed his newspaper, as he did his after-dinner cigar, for the slight haze it produced in his brain.

In the UK the hacks at the Telegraph or FT are much more similar to hacks at the Guardian/NYT than they are to entrepreneurs or people on average incomes, just as ‘conservative’ MPs unconsciously absorb Oblonsky-like the prevailing morality and Overton Window of Oxbridge. En masse the big political media organisations that largely set official focus and ‘ideas’ cannot understand and report seriously on complex systems, business, regulation, science, technlogy, war, the Taliban or anything else. Their incentives and culture program idiocracy. They spew a rapidly changing kaleidoscope of nonsense memes and ephemeral emotions as one story and scandal succeeds the last. Orienting towards this closed loop of ‘the media watching the student-politicians and can’t-execute-officials watching the media watching…’ is orienting towards insanity.

Every winning campaign knows in its bones that political journalism is Epistemological Mos Eisley — ‘You will never find a more wretched hive of scum and villainy’ (Obi-Wan Kenobi) — and that orienting to it is the fast train to disaster and defeat. But every winning candidate immediately jettisons this lesson in government and instead wastes a huge amount of their and their team’s energy trying to bribe, cheat, trick, flatter and fuck over the media. Those close to power in Washington or London never think seriously about cutting off the media’s information, creating alternative communication channels and a completely different communication model, and focusing their attention differently. (When I made clear I intended to do this it was seen as a declaration of war.) The media is treated by those with power as if they are a semi-official part of the extended constitutional order, hence, for example, why they are allowed to steal and publish classified secrets while citizens cannot.

As we’re all seeing on TV yet again in Afghanistan, this closed loop cannot cope with serious issues. And we’ll see the same on the next big story and the one after that.

So, the most important thing for the next GOP candidate and their team to appreciate is this: if you want to have a serious effect, if you want to have a chance of dealing with serious crises, if you want to bend the arc of culture rather than just have a cool-sounding job, then you should generally not be trying to ‘take over the institutions and run them’, you should not develop a traditional ‘reform’ agenda, instead, like Lincoln and FDR, you need to be the once-every-~70 years sort of a President who actually controls the government.

This means appointing people to many parts of DC not to ‘reform’ the department/agency but to close it while someone else runs the startup to replace it (if it needs replacing) operating on completely different legal and management principles and staffed by completely different sorts of people. This does not of course mean closing everything — there are parts of the regime that can carry on pretty much as before — it means closing and reopening those parts necessary to control the important parts of government, such as the Pentagon.

Why is closing/startups a better model than ‘reform’?

People in politics seldom have serious experience in being truly responsible for managing a complex organisation nor have they usually watched someone who knows what they’re doing do this.

If you do, you learn three crucial connected lessons.

First, high performance execution in complex operations requires a constant process of simplification, of stripping out unnecessary/foolish requirements and processes, of stopping things to simplify and focus efforts on priorities. Look at Steve Jobs’ relentless efforts, look at how Elon Musk runs SpaceX/Starbase and Tesla.

Second, turning around failing organisations is incredibly difficult even for very talented people and it’s often 100 times easier to close X and start something new than try to ‘reform X’.

Don’t believe me? Listen to the world’s most successful investors:

I’d say the history that Charlie [Munger] and I have had of persuading decent, intelligent people who we thought were doing unintelligent things to change their course of action has been poor… We really don’t believe in buying into organisations to change them. (Warren Buffett)

Third, high performance execution in complex operations requires combining responsibility and authority but by design government bureaucracies separate them, which guarantees delays, chaos and waste.

Don’t believe me? Listen to one of the most effective people in the entire western world in the 20th Century describe the crucial principles behind his success:

There was positive, clear-cut, unquestioned direction of the project at all levels. Authority was invariably delegated with responsibility, and this delegation was absolute and without reservation. (General Groves, director of the Manhattan Project)

Everything about modern government makes these three fundamental principles of high performance somewhere between extremely hard and legally and/or politically impossible to apply.

Just as bureaucracies resist changes to their current direction as if governed by Newton’s First Law, so they are governed by Newton’s Third Law: the more force you exert to simplify and remove ludicrous proecess, the more demented the bureaucratic resistance becomes.

How demented?

In March 2020 I had meetings in the Cabinet room to strip insane processes out of the way. What did I discover one day? That PPE equipment would not arrive in the NHS for months because we had ordered it to be SHIPPED. Why? Because ‘the rules’ said that FLYING it was ‘not value for money’ and we always follow ‘the rules’ even when they kill people. So I told them to call the airlines and commandeer their planes (all grounded) and fly them to Asia and collect the PPE and fly back. And they asked for a letter from the PM’s office indemnifying them in the event of legal challenges and/or disciplinary action. And now much of Westminster is fighting NOT to remove the insane rules but to ensure that the insanity revealed by covid is NOT used as an excuse to remove them. And they have had much success, and got the High Court to agree that avoiding killing people is not a good enough reason to move quickly. And the lesson has been duly learned — even in a crisis killing tens of thousands, make sure you prioritise the insane rules if you want to keep your job.

Throughout my time in government, the only thing that sparked as much frenzied resistance as ‘political interference’ in personnel decisions was saying ‘let’s just close X’. And if you try to ‘reform’ a bureaucracy so that responsibility and authority are combined, you are told it is … UNCONSTITUTIONAL AND ILLEGAL. Why? UK government departments legally separate responsibility and authority, making ministers ‘responsible’ but largely without authority. They are unable to hire or fire anybody other than a handful at most of ‘political’ advisers who are not supposed to have any input to personnel or management.

Imagine going into the Pentagon in January 2025 and starting discussions on ‘reforming the Pentagon’. How far do you think you will get?


There will be a process. Powerpoints. Lots of smiles, flattery and civilised chats. Lots of trivial ways in which you seem to be getting your way. And lots and lots of leaks. And no progress on anything fundamental.

Don’t try it.

Instead say: ‘US defence will be run from our new startup agency ABC, here is Alice who is closing down this building and making many of you redundant and here is Bob who will rehire a small fraction of you for ABC’.

And do it on Day 1 when you can say: I said in the last week of the election that I would do this and that’s what people voted for and now, the first day of my Presidency, I’m doing what I promised.

(If you want to understand how hard it would be to ‘reform’ the Pentagon, and why anybody who really wants change would close it instead, read The Pentagon Wars by one of Colonel Boyd’s allies, one of the few books I’ve read that give a realistic description of how big government bureaucracies really work and how evil actions are programmed by career incentives.)

Can it be done?

To those who say it can’t be done — look at Lincoln. If you think ‘well that was a war’ then look at FDR in the 1930s. Under FDR the government actually controlled the government. When his staff, like Hopkins, called the head of a department/agency and said ‘do this’, everyone jumped. Why? Partly because the FDR project was propelled by the Zeitgeist — everywhere 19th Century liberalism was in retreat. Partly because everyone on the end of the phone knew FDR’s White House wasn’t screwing around — they were merrily closing things they didn’t like, purging officials who opposed them, and creating startups to execute their priorities. The FDR government remade the republic. This should be your goal.

Little is more in tune with the Zeitgeist than attacking the big government bureaucracies — everybody hates them except those in power and those making money out of them, and you can use the energy of any resistance to power your campaign! Since FDR winning campaigns are run against Washington. And everybody is already programmed to think of the Pentagon and ‘evil arms companies’ as the enemy by decades of Hollywood! (And the best companies will cheer the closing of the Pentagon.)

Further, look at the record of elites recently. Look how they failed in Iraq and in 2008. Look how they blew the 2016 referendum (which a competent oligarchy would never have let us win) and blew the campaign against Trump (whom a competent oligarchy would never have let win).

Academics like Krugman and lawyers are the centre of gravity for the Democratic establishment and they are not formidable political opponents. Look at what Krugman was tweeting the night of the 2020 result:

Or something! A work experience 16 year-old who worked with me for a few weeks would have a better grasp of political reality than the NYT’s star pundit. These people are not good at politics! And as we proved in 2016 and 2019 they crumble when faced with people who know what they’re doing and do not play by the normal rules. They lost their minds over Trump who couldn’t grip anything — imagine how fast they’ll collapse if they face a team that can execute. Especially when the campaign … comes after them!

The hardest battle will not be the general election against Biden (the easiest bit) — the hardest part will be the next 18 months, creating a startup that prepares in 2022, can build a core team with agreed goals, then scale in 2023 and blast through the GOP primaries with a weird blend of democratic energy, technical skills and political strategy. In the 2016 referendum, by far the hardest problem was not ‘beating the government/Remain’, it was dealing with our ‘own’ side, just as in 2019 the hardest problem was not beating Labour/Remain, it was controlling the PM and the bureaucracy.

What if you don’t do it?

Even if a GOP candidate wins in 2024, without a plan and the people to control the government, Washington, Harvard, the New York Timeset al will simply delay everything, start investigations into you, and generally kick you from pillar to post, like almost all Republicans (and conservatives) for decades. After struggling through your first few months, delayed on every front by conscious enmity and normal bureaucratic inertia, with many of your appointments still unconfirmed but the FBI promising ‘the delay will be dealt with shortly Mr President’, you’ll only be a year from the start of the mid-terms (as Biden is now) and your own side will already be losing heart.

The bureaucracy does not need flashy victories. It wins by inertia — by just having dumb meeting after dumb meeting, by having its lawyers declare your idea ‘unlawful’ which takes you months to overcome because you’ve got to hire other lawyers and argue with officials over which lawyers to listen to, during all of which the clock is ticking. Before you say ‘honeymoon over’ your own side is panicking about the next election, prioritises that — and ‘it’s not the time for structural changes the public doesn’t care about, let’s focus on re-election…’

And the time for controlling the government, which would allow x1000 more progress than anything else, never seems to come...

As I’ve repeated ad nauseam since my 2013 essay, endemic government chaos combined with the closed loop of student-politicians/officials/media babbling nonsense at each other — i.e broken OODA loops — is colliding with a world in which growing destructive power is increasingly available to tiny groups and crises can move ~1,000 times faster than in July 1914. This guarantees catastrophe — the only question is what scale when.

Covid was not even a grey swan. It was top of the list of predicted crises and governments had rated themselves well prepared — the UK was ‘the best prepared in the world’, according to our permanent bureaucracy! But across the EU, UK and America the parties and bureaucracies are collectively steering debate away from the fundamental questions of why covid was such a disaster — the fact that the government does not control the government, and our parties and bureaucracies promote people mostly incapable of serious thought and actionand towards fairy tales and surface phenomena (ditto Afghanistan).

Without such a project, you have to assume the next 6 years will be not just policy and execution failure as usual, not just media fairy tales as usual, but disaster if we hit a serious crisis.

What does the project do?

A/ A political plan for the primaries and general — message, money, machine, strategy to maximise chances of >270 electoral college votes. How to define target voters? What do they really care about? How do they think about the things they really care about, what do they want, what do they hate/fear, why? Where are the opportunities for huge leverage?

E.g in the 2016 referendum we identified a) connecting our campaign to the moral force of the NHS and b) using Turkey’s slow-motion accession to the EU to sow chaos for Remain and focus attention on the prize of democratic control of immigration policy. People really cared about both — more than they did about all the other media babble. What are equivalents?

B/ A communication plan. There should be deep research in crucial states, the focus group guys/girls should be living in their pickups driving from swing state to swing state getting the language, rhythm and psychology of target voters embedded in their minds, with the music the voters listen to playing in the pickup. ‘Live in the village don’t attack the village’, as Colonel Boyd advised. When the candidate speaks, it’s inevitable that lots of what they say is incomprehensible to the median voter but some core things have to be absolutely simple and comprehensible by people who (like me) watch The Undertaker for fun. (Avoid all abstract slogans like ‘opportunity’, think ‘peace bread land’.)

C/ A hiring and network plan. The fate of every effort like this depends on the skills, personalities and relationships of maybe 10 people. Who are the equivalents to Plouffe (running the campaign) and James Baker (part chief of staff, part envoy/interpreter to the old power structures) — but versions prepared to dismantle the system rather than be friends with it?! Who is the intellectual motor? Who best understands how to map the DC power structure in order to dismantle it? Who is the ideal diplomat-smoothie? Who will build the network, from policy and money to activists in key states? This project should scope technical requirements and build some tools, but who will be CTO? Who are the very best lawyers that are sympatico (even if only privately), as (D) will need serious legal thinking?

D/ The real plan for government. In particular, there should be an audit of: what to start closing on DAY ONE of the new regime, what startups are needed to replace core functions, what are the likely legal attacks (judicial review), who are the key people to do which part?

NB. FDR’s 1933 Inaugural Speech telling Congress — get in line or else:

It is to be hoped that the normal balance of executive and legislative authority may be wholly adequate to meet the unprecedented task before us. But it may be that an unprecedented demand and need for undelayed action may call for temporary departure from that normal balance of public procedure… But in the event that the Congress shall fail to take one of these two courses, and in the event that the national emergency is still critical, I shall not evade the clear course of duty that will then confront me. I shall ask the Congress for the one remaining instrument to meet the crisis — broad Executive power to wage a war against the emergency, as great as the power that would be given to me if we were in fact invaded by a foreign foe. [Bold added]

If this project morphs partly into a formal campaign structure, then much of ‘the real plan’ should be physically separated from the campaign (with few links between the two). When it comes to a campaign and a candidate, much of the detailed planning for changing DC in the first year should be secret though in this phase (i.e the next year) much can and should be discussed openly.

E/ Which candidate has the best chance of putting A-D together? Who has the combination of political skills and CEO skills? Or if they don’t have CEO skills, they realise they don’t and are prepared to hire someone who does. NB. (D) is impossible without the candidate either having CEO skills or having the personality to hire people with such skills.

F/ Tool building. This project requires a lot of market research and building some tools to see and analyse all the information including predictive voter models. This project must itself embody a shift it seeks to bring to government — a shift from emailing spreadsheets, PDFs and Powerpoint etc to code, dashboards, interactive tools, dynamic models. (I did this, it made a huge difference, and those of you reading this in a serious company would be shocked by how archaic government systems are.)

G/ A system view. As Bismarck said about his Schleswig-Holstein diplomatic tour de force, ‘The individual actions were trifles in themselves, to see that they connected was the difficulty.’ A project like this needs to look across different ‘game levels’ simultaneously and see what connects to what in other levels:

  1. What are we really trying to do;

  2. The ‘Bubble’ game: what are Insiders doing (Harvard, NYT, the parties etc), the whole chaos of speeches, interviews, votes in Congress, daily spin — this absorbs ~100% of Insiders’ attention but is ~100% ephemeral;

  3. Who are relevant voters and what do they make of it — this is generally misunderstood inside the Bubble where untrue pundit memes about opinion spread and take on a weird ‘truth’ of their own;

  4. Who and how to execute — this is high status in serious institutions so it’s low status inside the Bubble and gets almost no attention;

  5. How do priorities connect across these game levels and timescales at different levels of abstraction, and how sensitive are plans to inevitable surprises?

To pull off something like this needs a mix of A) ultra-simple things that are popular to help you win and B) an ultra-sophisticated plan for replacing the existing regime. Most conservative parties and governments struggle to do a bit of A and don’t think seriously about B. This project could use the time before Christmas 2022 to figure out not just a lot of both but how they connect. Many capable people will assume that there ‘must’ be some team at the top of governments where people systematically examine the system view… NO! This does not happen. (When I left No10, one of the teams I was assembling was the team to do this — look across politics, policy, communication, project management of priorities and develop a ‘crude picture of the whole’. Such an essential thing does not exist in London and I bet you does not exist in Biden’s White House.)

Political discussion tends to zigzag across the piste — one minute people are talking about ‘what’s the actual right answer on X’ (actually far fewer conversations are about this at the apex of power than you think/hope), then ‘what the media thinks’ (this is far more of the conversation), then ‘what people think’, then ‘what gimmick can we come up with to help us advance tactically’ and so on. This project can use the time it has over the next year to create a schema for such discussions so they are disciplined.


This project must focus on practical solutions to the most important questions. It is not an academic exercise. It is not a think tank exercise. The output is not a report, seminars, media coverage or fundraising dinners. The output is a mix of actionable intelligence on the state of the existing regime and public opinion, tools for deployment in the campaign, ideas for a winning campaign, a set of plans for replacing the regime’s institutions, a map for thinking through the extreme complexity and hardest questions, and so on.

Almost everything will turn out to be near-irrelevant and a very small number of decisions will turn out to be critical — either critical errors from which more errors flow or critical successes from which many more successes naturally flow. The most important things to figure out are these foundational ideas/decisions on which everything else naturally builds. Identifying them, subjecting them to extreme scrutiny, and refining them is the critical task.

(Example. In the referendum we could have chosen to go with conventional wisdom and the MPs who wanted Brexit and focus on ‘trade’. This would have been a total disaster and guaranteed defeat. Instead we focused on ‘cost’. This was extremely controversial and caused ructions with our own side. It was also the foundation and necessary condition of victory. Getting big decisions like ths right is more important than >99% of discussions.)

While a serious political project must plan to control the government, it must not confuse this with a winning campaign. Most people do not care about and will not listen to people talking about bureaucracies and management. This is not as disparaging as it sounds, because if you look at Insider debate in Washington and London, you will see that the Insiders themselves do not care about and will not listen to people talking about bureaucracies and management!

So there must be a clear mandate for controlling and closing parts of the government when the candidate wins, but this agenda must not be confused with a winning campaign message as it will interest only a few, and that’s fine.

How much would it cost?

I haven’t thought about costs for more than 15 minutes but some rough costs.

  • 250-500k buys you a lot of focus groups. Don’t hire a company to do this then read reports. Hire 2 people to conduct the groups and live among target voters, driving from place to place, for the next year through the midterms and, if it works, through to October 2024. After being on the road in key states for a year culminating in the midterms, they will have a great feel for target voters’ true priorities, how they think, why and what works. They’ll be a great asset to a campaign.

  • If you hire the right people, ~500-750k (depending on data costs) will buy you the best predictive model of the electorate. You’ll be able to make predictions and tune it in the midterms. We built a model that in December 2019 predicted we would win 364 seats — the result was 365 (the model did better than the exit poll). We were lucky to be so close but the seat-by-seat predictions/results show not very lucky. And we didn’t have nearly as much data as you could throw at this project. You could easily throw x100 more at it than we did, and there are interesting ideas about how to move beyond the current state-of-the-art in the industry (the future of polling is not in the old polling companies, as we proved in 2016/19).

  • <10 salaries until ~Christmas 2022, by which time the project is morphing into a campaign or folding

  • Some data costs — I don’t know how to estimate these for the US but say 500k.

  • Some cash for specific people to do specific work.

  • Throw in ~200k for some expenses and logistics.

So ~$2-3 million for a tight focused project. Ten people could chuck in 200k each after dinner and fund most of it. And in practice if someone serious decides to do this and gets a few other serious people involved, money will not be a limiting factor. The limiting factor will be — can a core group of people who know what they’re doing agree on goals, can they deal with the pressure etc. High stakes politics is much harder than a normal startup because the former inevitably lacks what the latter has — relatively clear goals. In politics it’s normal for people to go years without ever really considering clear goals because thinking about goals in a disiplined way is a ticket to an argument.

The most important discussions will be mostly free. A whole set of smart knowledgeable people will talk to you for free especially if you share some of the insights from the market research.

Keep it tiny. Keep it temporary. This is not a new think tank that immediately looks to donors, offices and a permanent place in the system. It is a project to reboot the system. Keeping it tiny ensures quality control.

Also the model will be valuable so if the project goes nowhere politically, you’ll be able to make more than the cost of the project by selling the results to a few hedge funds or your rich friends so you literally can’t lose!

Why am I addressing this plea to the Valley?

This requires a project outside the normal penumbra of conservative politics in the US, just as it did in 2016 and 2019 here.

It’s almost inconceivable that the world of Washington think tanks and political consultants would be interested in, or could execute, a plan for a GOP President to control the government — it’s an oxymoron in DC, that’s the core problem. It’s more likely that people in the Valley, or mentally adjacent, will think about politics sufficiently outside conventional wisdom and be able to execute a project like this.

Should the funding be mostly hidden?

Those who undertake a project like this will, if it’s taken seriously, be seen as dangerous by some live players. It can be dangerous to be seen as dangerous by live players!

While some of those conducting the project will have to interact with others and be known, arguably it’s best if the money is raised and spent in a way that shields most of those involved from publicity.

While it’s a fascinating project, many will not want to be associated with it because of the blowback. This will make it trickier to execute but only a bit. People competent enough to do the project will be competent enough to manage the sensitivities.

Further, done right this project will create a platform so that many interested in politics but a) busy and b) unwilling to be publicly associated will be able to contribute privately, as with an OS software project.

Is this a pitch?


I would be interested to talk to people about something serious, but for personal and professional reasons I’m not the right person to run this.

I haven’t paid much attention to US politics for a long time and have just started looking at it again. A few people I’ve found interesting are:

Scott Alexander

Richard Hanania

Marginal Revolution

David Shor

Andrew Sullivan

Curtis Yarvin

Please leave comments.

I’ll review next week and post a public version.

I imagine over the next few months I’ll write more about what this project could do, questions about a winning strategy, the policy priorities for a serious government…

If you think what I’m doing is useful please consider gifting.

Give a gift subscription


Added 1/10

The above was written 1/9.

A few points given some feedback…

  1. The goal is not ‘GOP dominance’ — it’s a government that controls the government. Such a government would incentivise some of the most able people in the world to rewire the basics of the state — as the Founders, Lincoln and FDR did — including to close disastrous entities such as the Pentagon, change incentives in public services and welfare, focus on technology and productivity, and deep freeze the Ivy League/NYT et al focus on identity/race/sex. It isn’t to ‘win the culture war’, it’s to ‘make most of it disappear by withdrawing oxygen’. It isn’t ‘GOP dominance’, it’s ‘break the current party structure’ with a message, plan and execution that leaves existing party dynamics behind. (Yes the same could be done in the UK.)

  2. ‘Close and do startups’ is the safe option, carrying on as we are is the ‘risky’ option. In the US and UK we have done the ‘close and build startups’ model many times in our history. After Pearl Harbor George Marshall fired a load of generals promoted in peacetime and created a lot of startups, including the uber-startup known as General Groves’s ‘Manhattan Project’. Our leaders and the parties have no memory of watching this happen and they don’t know the entrepreneurial world where it’s normal. All they know is normal failure. Remember how astonished people in politics were when the Obama health website crashed? Washington had forgotten basic principles of systems management that its own projects had created in the 1950s. So they had to get Silicon Valley to fix it. And SV came in and did everything differently — not 20% or 100% better, but 100x or 1000x faster-better-cheaper. Did DC learn? Of course not! Like with the UK and the Vaccine Taskforce, the permanent bureaucracy is programmed to ignore, close and rewrite the story of enterprises that expose normal failure, even though many honest Insiders would like to behave differently. Systems can be smarter, dumber, or more evil than people. Amazing results are an existential danger to legacy systems and the officials promoted by the legacy systems, and the legacy political parties don’t want to tangle with these legacy systems. People will say ‘oh startups are so risky’. No. Continuing with the Pentagon and the UK Home Office is ‘risky’. Continuing with the leadership of the Met police and its management is risky. Replacing them is safer. And yes — we should sunset the startups so that in 15 years maximum, they are closed and replaced too. (15 years seems about the maximum time you see peak performance? Even PARC’s golden age was only a decade. Discuss…) The ‘costs’ involved are tiny, a fraction of 1% of the vast costs imposed by normal failure in a place like the Pentagon/MoD, and the startups will immediately save the costs of the admin by stopping stupid stuff, so the returns are massive.

  3. History teaches us that ‘close and do startups’ is normal when a system implodes like with the Great Depression and war. Why not try to learn and do this before we have further debacles like Iraq, 2008, covid? If you think the existing systems can be reformed, just watch the CDC over the past few months on vaccines, treatments and fast tests. These entities CANNOT learn. If they can’t learn even in a global disaster killing millions when everybody’s focus is on them, why would you think they will ever learn?! In summer 2019 I wrote about how the British state would fail in the next crisis. Nine months later I was in the Cabinet room watching what I had said would happen actually happen. Do you see any serious attempt to learn now? No, we have a PM and ruling party desperate to reinvent history and forget. Those officials who wanted to use the disaster as a catalyst for a reboot have been left in no doubt that such ideas are ‘career limiting’. Instead of learning those with power have blundered straight into a predictable and predicted petrol shortage because they cannot build the most basic early warning systems and execute. (And over the past 48 hours in the UK the opposition party has supported the ruling party in defending the permanent bureaucracy after a horrific kidnap-rape-murder by a police officer, after which nobody in the senior management has been removed. This episode shows what I’m talking about above: a) the permanent bureaucracy will prioritise defending itself in virtually all circumstances, and b) the legacy parties reflexively defend it even when the public is united in clutching their heads — the politicians are so programmed not to govern they struggle to grab the power to govern even in dramatic circumstances.)

  4. Since I wrote this a month ago it has become clearer that a) Trump plans to run again, b) it will be even more deranged, c) the GOP remains a disaster zone and has no plan to pre-empt this and many of its key people are locked in ‘he’ll get the nomination so it’s carer death to oppose him’ mode. If this psychology is not broken well before the mid-terms, nothing except health or some legal issue will stop Trump’s nomination — and the closer it gets to him announcing his campaign, the less legal issues will disrupt him. And this will trigger a cascade of further bad things because a) the Democrats will adopt the mindset of ‘anything is justified to stop Trump’, b) this will look to many Republicans like ‘Washington is fixing the next election too’, c) many will therefore rally to Trump, d) all sorts of terrible feedback effects will intensify fed by the media which has a commercial interest in amplifying Trump because he is so good for their profits.

  5. A lot of people have said ‘you couldn’t campaign on closing the Pentagon as a Republican’. This is false. What you do is start a ‘Veterans for Replacing the Pentagon’ campaign and you get it focused on things like corruption and have veterans themselves explain how ‘the Pentagon’s corruption and institutionalised incompetence killed my colleagues’. The more that parts of the Pentagon and big businesses campaign against you, the more energy your campaign will gain. They’ll be working for you.

  6. The winning message for a project like this will seem weird to activists of the GOP and Democrats but it will make sense to regular voters. In the 2016 referendum, our message was attacked by some for being to the ‘right’ of the Conservative Party and by others as being too ‘left’. Our approach made no sense from a conventional party perspective and was attacked by almost all high status pundits. We won more votes than anybody has ever won in a British election and a higher percentage of the votes than any party since the crisis of 1931, despite almost every force with power and money being on the other side, and our campaign being effectively insolvent until a month before the end. Message matters more than money. Orienting to the public will inevitably seem weird to the pundit class in DC (‘incoherent’ they’ll write) but it’s the only way to break the impasse.

  7. ‘Just’ controlling the executive branch of government would be very radical, in the sense of ‘hasn’t happened since ?Eisenhower?’ but it’s also not enough. It needs to reach into many other institutions including the media and universities, both gripped by crazed memes intensified by social media. This makes me very pessimistic and I have almost never watched people on ‘the right’ have a discussion about this that makes me hopeful! But … pessimism of the intellect, optimism of the will!

  8. Balaji and others have proposed many interesting ideas summarised as ‘don’t try to fix the current system, build alternatives’. I strongly support the second half of this. I think it’s true that most people should focus efforts on building alternatives rather than trying to influence existing parties and bureaucracies. It’s clear that Balaji himself is an extremely effective person, who was one of the earliest and rightest on covid/CDC/FDA etc. I am not saying he’s wrong. But it’s not either/or. Given the stakes I think the project I sketch is worthwhile as a complement to, not replacement for, such Balaji-efforts. I will write re the Balaji-efforts soon too. Also I suspect the project I sketch would naturally throw off Balaji-adjacent projects.

Politics is path dependent and highly nonlinear. A relatively small effort now could nudge trajectories in a way that will be 10x or 100x in 18 months and almost impossible in 24 months.

Doesn’t it seem odd to forgo the possibility of avoiding a whole cycle of Trump-Biden horror for the sake of $2-3 million and the temporary diversion of 10-20 people?