DOGE is beginning to hit, and this piece, by Mike Masnick, provides the best overview I’ve seen of what’s happening.*
The parallel to Twitter is striking and terrifying. At Twitter, Musk’s “reform” strategy transformed a platform used by hundreds of millions for vital communication into his personal megaphone, hemorrhaging somewhere between 60-85% of its revenue in the process. But Twitter was just a private company. Now he’s applying the same destructive playbook to the federal government, where the stakes involve not just user experience or advertising dollars, but the basic functioning of American democracy.
there are multiple reports of Musk basically taking over various parts of the government. He apparently showed up at the General Services Administration on Thursday, just after his right-hand man in the Twitter shakeup, Steve Davis, told them they were ending a bunch of leases on government buildings (another thing that Twitter also did). Even worse, Wired reports that Musk’s friends are using the GSA to try to get access to a variety of systems, including remote access to laptops, and even reading emails of government employees. … there were even scarier reports of him fighting with the longest tenured non-political employee at the Treasury Department, David Lebryk, leading Lebryk to resign after Musk demanded access to the US Treasury’s payment system.
… Millions of people actually depend on the US government functioning. You can’t just have some random jackass show up and rip out fences and assume shit won’t go south. … it’s being ripped apart by someone with no concern or care for the actual damage he’s doing. … the federal government needs massive reform … But there are smart ways to do it and then there’s this: which is just utter destruction while looking over his shoulder to see if the nihilistic kids who worship his every move are finding it entertaining.
The gloss that I’d add to what Mike says is this. Just as happened with Twitter, we are seeing a Silicon Valley power fantasy playing out in real life. And if we want something better - if in some improbable seeming future universe, we have the chance to make government better as Mike says - we will need to build “the basic functioning of American democracy” into the process.
Over the last few years, there has been a radical shift in the politics of many (not all) Silicon Valley funders and CEOs. They have identified two enemies: their own employees, whose politics often differ greatly from their own, and East Coast politicians and regulators. Paul Graham, the founder of Y Combinator, laid out the standard shtick a couple of days ago:
The problem with employees was that they often tended contrary and woke. When you, as a funder or founder, considered yourself to be an avatar of destiny, with a substantial chunk of the fate of human civilization in your hands, you found it really, really fucking aggravating to have to deal with HR all the time, placating employees who carped, sniped, protested, slow-walked and opposed, and were visibly not as impressed by your divine effulgence as you knew they ought to be.
The problem with the East Coast “paper belt,” as Balaji Srinivasan notoriously dubbed it, was twofold. First, it was visibly archaic, technologically incompetent and incapable of doing its job, and second it was getting in the way of Silicon Valley’s instinctive drive to move fast and break things. Silicon Valley unhappiness with the government got much worse after 2017, after Washington DC became increasingly unwilling to defer to the visionaries of tech, proposing regulations, anti-trust actions and other measures that threatened to get in the way of the awesome post-Singularity future.**
Those enmities help explain the visible giddiness among many Silicon Valley elites when Elon Musk took over Twitter, and fired most of the employees. Finally, finally someone was showing those entitled scolds who was boss. Now that there was a real CEO, willing to strip the organization down to the bones and rebuild it faster and leaner, with employees who were committed to the mission, we would surely see a return to the good old days of unbridled creative destruction.
Of course, this fantasy turned out to be just that: a fantasy. Musk has overseen a lot of destruction at Twitter, but not very much creation. The people who used to defend Musk to me in private have stopped, and are now for the most part maintaining a discreet silence over his unique approach to public communication and business decision-making. It is perfectly possible that Twitter will start making more money again, possibly as a class of an unintended political protection racket, but that is not the kind of business model that one wants to publicly associate oneself with.
Now we are seeing a much bigger catastrophe unfurling. For years, many people in Silicon Valley have been convinced that if only they were in charge, they could eliminate the waste and ideological excesses of government, getting rid of all the obstructionists, and creating a leaner, tighter machine instead.
From this unique perspective, the fact that none of the DOGE people actually understand how government functions is a feature, not a bug. If you understand the workings of the federal bureaucracy you are almost certainly part of the problem, not the solution.
That’s how we end up in a world where a 21 year old and a recent high school graduate seem to have taken effective charge of a crucial engine of government. Unfortunately, “Let’s see what this button does,” worse, “we’ll find out what we need by ripping everything out and seeing what fails,” and worst of all, “government rulemaking is _so_ mid: we’ll have AGI by 2027!” are not viable approaches to reform.
I don’t need to dwell on how badly this is going to go wrong: it will. Nor do I have much that is useful to say about how to stop it. If you want expertise on the gutting of the administrative state, and are on Bluesky, here’s a starter pack for you.
What I do want to say is that small-d democrats should take a lesson, when they start to articulate their own alternatives, as they one day will have to. Power fantasies are a bad idea.
It isn’t only Silicon Valley types who want to want to reform the federal government. As Mike says, it desperately needs reform. There are many people on the center left too who are frustrated by how clunky, unresponsive and unfit for purpose the federal government is. There is a lot that government needs to do, or let others do, and it is easy to lapse into fantasies about shattering the whole sorry scheme of things to bits, and molding it closer to the heart’s desire so that all the good stuff can begin to get done. As matters stand now, there are way too many veto points, and too few gates for opportunity to get the changes made that need to be made. If only the technocrats who really understand what the future might hold for us, could be unfettered from the regulations that hold them down …
There is a lot that is right about these aspirations. But - and this is the unfolding lesson of DOGE in all its likely horrors, there is a lot that is fantastical too. You may not like the forces that prevent your grand vision of change from being realized. And it is perfectly possible in the specific that your opponents have too much power to hold it back - all systems of decision making provide ample opportunities for short-sighted idiocy, though they differ as to which specific forms of idiocy are empowered. But there is usually wisdom as well as stupidity in the forces of disagreement. You want the power to get things done, but you also want democratic feedback and accountability, to restrain your human tendencies toward hubris, and to forcefully correct you when you are wrong. Otherwise, you risk turning into DOGE.
That holds for all the alternatives, including the cybernetic ones that I’ve been talking up in this newsletter. A couple of weeks ago, the Niskanen Center’s Hypertext newsletter ran a short seminar with Dan Davies, Margaret Levi, and David Dagan, on Dan’s book, The Unaccountability Machine (now forthcoming in the US; more here).*** As David remarks, the book provides a very different understanding of how to remake government than DOGE.
Claude Shannon could help them in the here and now, as they swagger into their campaign to reshape the federal bureaucracy through their self-styled “Department of Governmental Efficiency.” Instead of grandiosely promising to cut trillions or targeting individual public servants, the DOGE team could consider how lessons from the fundamental architecture of computing could help us fix government. Cybernetics sounds like an oddball concept when you encounter it for the first time, as I did upon reading Dan Davies’ brilliant new book The Unaccountability Machine … But in the second decade of the “polycrisis,” in which one institution after another seems to be failing, the cybernetic approach is worth a second look.
Equally, as Margaret points out, there are important blind spots in the cybernetic vision:
what I find missing are politics and power. They are mentioned in the forthcoming book, yet get short shrift there and in his essay for Hypertext. More attention is given to the public at large, for Davies makes clear his commitment to recognizing the interests, concerns, and knowledge of “users and consumers.” However, this is largely accomplished by an improved feedback loop that focuses on increasing management capacity. And that takes us back to the relative neglect of the political and power arrangements that can distort or even ignore that feedback loop. Davies’ analysis is focused on improved communication. That is certainly necessary, but it is hardly sufficient to ensure that the perspectives of the affected populace are incorporated into governance decisions and goals. … it was power and politics, not communication issues or an “accountability sink,” that led the authorities to ignore the interests of Mexico, weaker state governments, and small farmers — let alone the expertise and rights of indigenous peoples in the region.
I know, from conversation with Dan that he thinks Margaret’s criticisms are a fair cop. More generally, the cybernetic vision will plausibly only work well in a world of less unequal power relations. Cybernetics is about feedback loops - providing information back to the system about how things are working or not working, whether the environment has changed and such. Power relations are about many things, including basic human dignity and autonomy. But they are also about information, in the sense that systems with horribly unequal power relations are extremely unlikely to have good feedback.
If you want a better alternative to DOGE, it has to involve some form of democratic steering. In the words of Danielle Allen:
When the knowledge and understanding that flows into a political decision is closely controlled by a limited few, their control of knowledge resources pulls decision- making power to them as well. Broadening the engagement of the citizenry in the discovery, analysis, and deliberation processes that feed into policy-making decentralizes power, supporting political equality.
Danielle makes this sound great - but she certainly knows that such broadening of engagement is often, in practice, a massive pain in the arse. Systems that thoroughly incorporate democratic feedback can and will have their own frustrations and squalid compromises. But the grotesque delirium of unfettered technological agency that Elon Musk and his ilk are trying to implement are visibly much worse. As we try to figure out an alternative, we need to avoid our own power fantasies.
* It’s startling that there is nothing e.g. at the New York Times right now trying to pull this together in the way that Mike does. I suspect that they lack the expertise: the only person I know of who they have right now who could do this is Jamelle Bouie, who is on the opinion side rather than a reporter.
** As discussed in a previous post there were an awful lot of other reasons for why these people turned right. Furthermore, as Graham himself noted, in a revealing early essay, “As much as they respect brains in Silicon Valley, the message the Valley sends is: you should be more powerful.” In retrospect, it’s not surprising that Silicon Valley was as willing to bow to power as it has turned out to be.
*** These are my people - both Dan and Margaret are co-authors and friends - so I am not unbiased.
Government is inefficient, because efficiency is not really the point of government. Representation is the point, and it is really hard to represent everyone’s interests and be efficient. We also need government to be resilient, and efficiency is the enemy of resiliency.
I’ve worked in higher education and for-profit companies, served on non-profit boards, and been involved in local government. There are differences in the importance of efficiency, representation, and resiliency in all of them, and differences in how power works in each.
My experience with tech types, particularly, in getting involved with either non-profits or government, is that they don’t understand this, are often resistant to learning this, and as a result, leave in frustration after being ineffective. Unfortunately, nobody is stopping Musk from just kicking everything over, so far.
I see a mix of the Peter Principle and Peter Pan (puer aeternus) Syndrome playing out, people moving into jobs they're constitutionally incapable of doing largely because they're arrested as shit, stuck in puberty or adolescence and saddled with fantasies of omnipotence. And they've never had any motivation to step outside their attachment to the abstract world they love and experience the complexities and limitations involved, for example, in having a conscience and giving a fuck about other people. People are not actors in video games.