The RAGE of the DOGE

Rise of the Real Technocracy

During the wilderness years of the Obama administration, the software engineer turned neo-reactionary thinker and general gadfly Curtis Yarvin proposed, in his plans for a future monarch-type figure, to fix late democracies with a bold initiative he deemed RAGE: Retire All Government Employees. The Trump administration did just this in its first two weeks, offering a buyout to as many federal employees who wished to assure some form of severance before the risk of being summarily dismissed.

Yarvin, a one-time fringe thinker whose top intellectual influences were Robert Filmer, a 17th-century defender of the divine right of kings in his anti-Lockean work Patriarcha, and the anti-modern philosopher Thomas Carlyle, played an outsize influence on the intellectual development of the Young Turks of the Trump regime in exile from 2021 to 2024 and their public avatar, J.D. Vance.

Yarvin himself was skeptical of these efforts at populist politics, believing any change must come from the caste of “elves” — existing highly intelligent elites — rather than the MAGA “hobbits” of the retail Trump movement. The somewhat depressive Yarvin has since apologized for his skepticism, citing his natural pessimism, which he describes as a protective personality trait that causes him to never be wrong and disappointed at the same time.

The interesting question here is what exactly Yarvin got wrong. It is hard to believe that populism per se works, as the idiocy of all retail political movements, MAGA or woke, is self-evident. The key seems to be the unique figure of Trump, who is smart enough to channel populism rather than have it burn out with hare-brained, overly simplistic policy proposals. Bottom-up populism can only diagnose, never cure. Trump is more FDR than William Jennings Bryan in fostering the development of a new elite to replace the old one.

Yarvin’s mistake was placing too much faith in the existing elites. They were farther along the path to degeneracy than he had anticipated and completely captured by their own idiots in the woke movement: largely corrupt, lazy, asleep at the wheel, and convinced that history was over. They were hollow in the face of men of action. Elon Musk, perhaps coincidentally or perhaps sensing weakness, saw an opening for a new elite. It turns out that the elves must know how to code.

And these elves will be naturally drawn to Yarvin’s politics due to the computer science axiom of “garbage in, garbage out.” If the average voter is an idiot, how can the aggregate of idiots be much better? One needs legitimacy in a political movement, which Trump provides as a figurehead. However, as all good engineers know, what the user thinks he wants is different than what the user truly needs; good engineering is inherently anti-democratic and aristocratic.

The Rise of DOGE

Trump, seeing the “proof of work” in Elon’s technical accomplishments at SpaceX, wisely accepted his offer to run the Department of Government Efficiency (DOGE). Even more unbelievably, compared to the ineptness of Trump’s first term, DOGE is being given the tools to do its job, striking terror into the old elite. Personnel is policy, indeed. Day after day, DOGE is discovering new abuses in the federal system, a drip feed of negative publicity that eats away at the old regime’s legitimacy like acid on limestone.

As Wired recently profiled, Musk has recruited “hardcore” software engineers to splay open the books of the government. Musk is mostly guiding this work but is providing a critical boost of confidence to these young engineers as they seek to uncover the old regime’s abuses.

Musk has the lived experience of a talented engineer. It is a character arc, when successful, that leads to disorientation, then cynicism, then confidence, and finally to a semi-justified arrogance.

Disorientation occurs as the young technical mind discovers that most normal humans are, to put it not so delicately, full of crap: incoherent, imprecise, irrational, biased, illogical, with a ChatGPT-like LLM attached to spew words designed to preserve their status more than discover the truth, and that’s extremely adept at the emotional manipulation of others. The young techie, having often experienced an inability to navigate social situations, will sometimes at first wonder if he’s the crazy one and often back off when normal humans attack emotionally when their inconsistencies are made evident by questions or attempts to clarify.

Over time, however, the initial hypothesis is confirmed, and it’s noticed, with sometimes crushing cynicism, that in most human organizations the ability to manipulate other humans is more highly valued than the ability to discern ground-level truth. Engineers and similar personalities in particular value truth above all, and are both bad at and uncomfortable with lying as a way of life, the key skill for any successful bureaucrat.

Software engineers often have it the worst. In developing software, say internally for some private or public bureaucracy, they will have requirements meetings with highly-ranked managers, at which it will become clear that none of the people leading a given organization, with much larger paychecks, have any idea how anything actually works or is supposed to work. The higher up the tree one goes, the less anyone knows, and the more the job is not about the putative purpose of the organization but rather mutual nest-feathering operations with their managerial peers at other firms.

To be fair, many engineer types make little effort to escape the ghetto of geek culture. Finding a niche in math and science, which they find easy but others find impossible, they can be prone to self-satisfaction early in life. Many are in permanent arrested development, holding on to childish obsessions like comic books, children’s movies, toys, and literature, and board games, as lampooned in The Big Bang Theory.

Some struggle with basic grooming, creeping out normal people who lump them in a genus with the mentally ill or homeless. Few seem interested in applying that same rational, engineering approach to interactions with normal people, whose psychology is both known and predictable, i.e., subject to rational inquiry and useful as a raw material for achieving rational objectives. They can be stuck in an irrational loop of both knowing others are irrational and being angry that they will not change, instead of incorporating this fact into a social algorithm that anticipates and corrects for it¹.

Thus, most technical folks lead lives of quiet desperation, watching the clock until a usually comfortable retirement, serving people dumber than them, who are better at lying and manipulating while fooling themselves as to the value they add, who do things that make no direct rational sense.

A few, however, manage to break out of corporate jail with their own businesses. When this happens, watch out, because cynicism, once liberated, becomes extreme confidence. Luckily, this confidence is at least matched with competence, as compared to the modal managerial sociopath. A great example of the liberated engineer-turned-entrepreneur is this video of Thomas Massie (BSEE, MSME, MIT) grilling John Kerry (BA, Political Science, Yale) over carbon emissions:

Massie comes across badly in this clip, as he can barely contain his contempt; his lived experience has been with lightweights like Kerry, with their fake education, and hopelessly out of his depth in handling any technical problem, promoted to elite positions over smarter guys like Massie². While Massie is free, he remembers his friends at MIT who now work for people like Kerry.

The old managerial elite, emerging from FDR and of which Kerry is an exemplar, has at times been described as “technocratic.” That, however, is generous to their actual intellectual capacity. Most had no actual technical knowledge, relying on experts in various fields such as sociology, psychology, and economics, better described as pseudosciences compared to the hard fields of physics and engineering.

Since soft fields can be made to prove nearly anything, the “techno” part of “technocrat” became a mere rationale — though an effective public one based on fake “expertise” — for the same old graft and incompetence that plagues every growing bureaucracy. It hid the ball successfully in the controlled media era but cannot withstand the assault of autistes digging through their dirty laundry on X. Interestingly, DOGE is finding natural allies in many honest line workers of the federal government who have been frustrated by the fraud of management for a long time, a modified demonstration of the midwit meme.

Technocracy may have failed because it has never been tried. With Elon and his crew in charge of DOGE, we, for the first time, have actual technocrats in charge, and I expect they will make quick and easy work of exposing the federal leviathan. A quick demonstration of scale might be helpful.

When I was discussing DOGE with friends after the election, I said that its success would depend ultimately on whether Musk got the root credentials to the system that cuts federal checks, including payroll. There’s a lot that can be done by just not paying people. A few weeks ago, he got it, and the press immediately protested about the deep technical expertise of the federal bureaucrats formerly in charge.

But by tech standards, the scale of the federal government is a joke. Estimates vary, but the total amount of individual transactions on the federal ledger is perhaps one billion per year. By contrast, Instagram’s algorithm generates something like tens of billions of data points per day across all users — scrolls, likes, follows, etc. For software engineers accustomed to this kind of scale, a billion annual transactions can be queried on a decent on-premises server; it’s just not that complicated.

The relationship between Trump and Elon is perhaps a template for a new kind of management. It is necessary, due to the demands of a democratic system, that Trump be the nominal head; Elon nor his crack engineers are charismatic enough to be elected. But in outsourcing the actions of his movement to the most technically capable, he makes his movement maximally effective in achieving its objectives.

Type 1 vs. Type 2 Leaders

I referenced Ben Horowitz’ The Hard Thing About Hard Things in my last post for elucidating several extremely valuable frameworks for managing organizations. One of those is his distinction between Type 1 and Type 2 leaders. A Type 1 leader is someone who knows what to do. A Type 2 leader is someone who can get the team to do what he knows.

Horowitz presents this as a fundamental, often unresolvable tension. Very few leaders have both qualities, and this becomes a problem as an organization scales. A founder is typically Type 1 but will struggle with Type 2 tasks as the staff grows. There are no easy solutions, as Microsoft found when transitioning from Type 1 (Gates) to Type 2 (Ballmer).

A related nugget from the book is Horowitz’s analogy to offensive linemen in his advice to “give ground grudgingly.” An offensive lineman knows he will eventually lose the battle to protect the quarterback, and his goal is to protect him just long enough for the play to be executed. A lineman can not simply stand his ground, as defensive players will go around him. But nor can he move back too quickly, crowding the pocket. The only solution is to give ground grudgingly, allowing the defensive player to make some progress, but forcing him to fight for every inch along the way, delaying the inevitable as long as possible.

Many situations in life and business are losing battles best managed with resistance and delay. Horowitz specifically cites this in making decisions about scaling. That is, there are benefits to keeping a Type 1 leader as long as possible and delaying the type of scale that takes a founder out of decision loops. Scaling must be both accommodated and resisted at the same time. Band-aids and half-measures are acceptable solutions if they give you a few more months of avoiding bureaucracy, and do not indicate some fundamental lack of vision or courage. The applications of this analogy are enormous, from fighting aging — a battle we all eventually lose – or raising kids in a hostile culture.

The Chief of Staff Model

For organizational leadership, however, I wonder if Horowitz inadvertently stumbled across a solution in another part of the book when he recommended the “Chief of Staff” (COS) model for founders. Type 2 leaders are often the stereotypical, less technical management types like John Kerry. They critically have people skills to get the team to do things, but unfortunately, they lack the technical chops to know what should be done yet have more confidence in their ignorance than a Type 1 might have in an informed evidence-based hypothesis. They are the type of people who run organizations into the ground by doing lots of the wrong things.

Most critically, why are Type 2 leaders usually in charge in most large organizations, rather than Type 1 leaders, who are typically technical staff? Why is it normal that Elon Musk and Thomas Massie types work for John Kerry types instead of the other way around? I suppose it has something to do with the superior politicking ability of Type 2 leaders and the dispersed nature of authority in most organizations, requiring coordinating with many stakeholders to support one’s ascent to the position. That means telling a lot of people what they want to hear.

By contrast, Type 1 leaders with technical mastery are often grumpy and blunt, as a better understanding of reality, the thorns and thistles that attend man’s work, are the opposite of happy talk. Yet most founders of new businesses are Type 1, and presumably, organizations would benefit with decision-making authority lodged in their most technically informed rather than the most gregarious.

For founders with operational control of their businesses, Horowitz advises against ceding authority to a Type 2 leader but rather using these folks' strengths as chiefs of staff across the organization. That is, the founder and his highly technical managers maintain all decision-making authority, with the more socially adept Type 2 managers serving as assistants to multiply the most competent leaders’ directives while smoothing over human interaction problems that technical types find most difficult to manage.

One benefit of this approach is that Type 2 leaders tend to be more personally jealous of their ranking in the pecking order than the organization’s mission and extremely protective of “muh authoriteh.” Even when they’re wrong, they’ll exact an emotional tax, gaslighting you as a control freak, if you “micromanage” them out of bad practices and decisions. By stripping them of actual authority to a mere delegated authority from the Type 1 leader, the Type 1 can reserve the right, as he often must, of directly engaging with line workers when necessary. Hence, we see Elon sleeping on the factory floor to address key bottlenecks instead of a slow bureaucratic approach of always respecting the chain of command. “Founder mode” means keeping Type 2’s within their proper limits and talents.

That said, often the need for a Chief of Staff is more due to poverty of time than any real personal deficits of the Type 1 leader. While some minimum of people skills is necessary, it might be better for one’s COS to be a fellow Type 1 person who can essentially clone the founder’s mindset, allowing for some limited delegated decision-making. I can’t see Elon, for example, hiring a glad-handing MBA as a chief-of-staff; I think he’d rather a highly trusted engineer with good judgment like himself.

Trump As Type I/II

Interestingly, this concept is borrowed from politics, notably the Presidency, where day-to-day operations are run by the White House Chief of Staff. Theoretically, the President is essentially a Type 1 leader informed by his election as to what the public would like done, though natural charisma undoubtedly influences who wins an election. The federal government, however, is so cumbersome that the actual work is delegated to staff. That staff in the past have undermined the sitting presidents, especially Republicans, is a secular issue that is solvable for the right POTUS with sufficient force of will to hire the right lieutenants. Trump, in his second term so far, seems to be fixing this problem.

He also seems to have a rare combination of Type 1 and Type 2 skills. He is an independent thinker and his instincts on what to do are often correct compared to previous Republican empty suits like George W. Bush or Mike Pence. His extreme Type 2 skills in using simple language and narratives help him sell it to the uninformed median voter.

Trump, in restoring the presidency to its proper authority, can then delegate directly to another Type 1 leader, like Musk, instead of having his interactions heavily filtered by career Type 2’s obsessed with process and “norms.” In the extremely compressed timeframe between elections, the administration will have a fighting chance to “move fast and break things” that need to be broken. The “flood the zone” approach of the second administration is much more like a startup than the first.

For the first time since the 18th century, founders are back in charge of the United States, and politics will never be the same.

1. It doesn’t help that engineering schools train their students as technicians, not as a discipline for the rightful leaders of society. An enhanced curriculum would include aspects of the classical liberal arts and social engineering, teaching how to modify one’s communications to lead the majority of irrational people rationally. This is why it is rare for innovators to get paid and reap the fruits of innovation; rather, they are usually intermediated by a non-technical social butterfly who captures most of the value (e.g., Jobs vs. Wozniak). Such a change of focus might result in more socially talented, smart students pursuing useful fields over parasitic careers in finance, which, to use Thiel’s terminology, is at best 1 to N instead of 0 to 1.

2. Massie may be particularly bitter because he is still, to all public indications, a naive libertarian who has not accepted the second-order effects of normie irrationality and thus moved past quixotic hopes of democratic persuasion. He is still at the Ron Paul level of the libertarian to dissident right intellectual funnel, and the cognitive dissonance entailed by straight libertarianism can be supremely frustrating. “People be crazy,” obviously, which doesn’t work for a philosophy that demands they not be for the non-aggression principle to work in practice.