DoGE "cut muscle, not fat"; 26K experts rehired after brutal cuts
arstechnica.com124 points by jnord 2 hours ago
124 points by jnord 2 hours ago
This is coming out the same day two DOJ cases led by a US attorney with no previous prosecutorial experience were unceremoniously tossed out. DOGE sent in a bunch of 20 somethings to "fix" the technology while cutting entire groups of experienced technologists like 18F. To say nothing of the CDC, whose communications are starting to look like a bad, late-night infomercials.
I understand having a problem with a authority that manifests as a distrust of experts, but the combination of ignorance and arrogance is breathtaking.
Hopefully 2026 can be a year of restoring some adults to positions of responsibility.
Nothing wrong with being 20 somethings in itself regardless of the rest. Average age for the Manhattan project was 25-27. We can focus on the merits or mistakes no need to focus on age.
The average age for scientific contributors was 29. https://blog.nuclearsecrecy.com/wp-content/uploads/2014/09/L...
Did the DOGE 20-somethings also have the benefits of supervision from PhDs in various specialties? It's not the age alone, but the age in combination with other factors that make it concerning.
I’m ok with age being used as a partial proxy for experience when we’re talking about highly specialized roles with massive implications like the ones that DOGE staffers were dropped into.
You are right — I shouldn't have been dismissive about the age, but rather the complete lack of experience around governance, why some of these rules exist, and why some of the technology is the way it is.
I think this misses a point.. hiring adolescent hacks with after-midnight chops and thrill-trophies on the walls IS part of the selling point of the DOGE raids. It was a raiding party. That is as old as pre-history, in itself. But the playing fields are terminals and web browsers. Age is a "partial maybe just a little bit proxy for experience" no it is judgement and some healthy understanding of the weight of historic events, and the financial weight of some of the systems.
All that said there is another side of the coin. That is that there were under-the-radar payment systems and not quite audited channels of money in those systems. Built with care, you bet. Essentially diagramming the tech stacks, documenting admin systems, getting and using root and root equivalent at all times possible.. those were the scalps taken, and the targets were actually rotten in some ways in some places. /rant
> DOGE sent in a bunch of 20 somethings to "fix" the technology ...
Elon Musk claims that the vast majority of decisions were made by AI modeling.
Synthetic 20-somethings that cost a few bucks per megatoken may technically be different from flesh 20-somethings working for free for the exposure or whatever it was, but it's not an important distinction.
Much as I find LLMs useful, even today I'd only rate their competence in any given domain like a 21 or 22 year old in that domain. The Penguin Island* tariffs comes to mind as an example of probably-AI; I can think of a few mistakes of this level before the days of AI, the only one I'd like to mention is having had to explain to a real human that someone saying they're in "London" doesn't automatically mean they're in the UK.
And that's if I'm being generous and assuming Musk's statements on this topic were based in reality — given Musk also asserted that savings of 1-2 trillion dollars were possible when this was not only beyond the powers of the executive, but obviously so with minimal research, I don't trust his word.
* Heard and McDonald Islands, IIRC
The transformer architecture was introduced in 2017, so they send in a bunch of 8 year olds to "fix" the technology ;)
This is one of many frequent reminders: In some environments, how you brand and market your work (Mush with a chainsaw cutting spending comes to mind) is often more important than the work you do. Most wont bother to look at the actual results of your work.
I guess I'm what they derisively call a normie. Both the sink and the chainsaw seemed like red flags.
I'm a massive weirdo, what I like and dislike is almost anti-correlated with public success, or at least success in American markets.
You should therefore be unsurprised to learn that I laughed about the sink and had a completely neutral reaction to the chainsaw (other than who it was in support of). I should have noticed in advance that my reactions were warning signs. Unfortunately, I can also add the visual design of Cybertruck to that list — not that I'd want it on the road, obvious safety hazard from the shape, but I liked the look of it.
I don’t understand. Are you claiming the actual results of Musk’s work here were good?
I think rather he's saying that Musk loudly declared he was slashing government waste and firing do-nothing bureaucrats, and the people who supported him never really bothered to see if that was the case or not.
Its one of those things that's a hard lesson to learn; ideology isn't greater than experience.
One of the biggest lessons I learnt when I was a younger dev is a living allegory that my manager told me:
"one day the new boss came in to a budget meeting. The boss was out to make a good impression, and come out winning. The boss looked for any 'useless spend'.
Looking at the budget, the Boss saw how much was being spent on cleaner.
Looking around, the Boss boomed 'The place is spotless, why the fuck am I paying for cleaners. There's nothing to clean'
The underlings laughed and clapped. Oh how clever the Boss was, saving such a big amount at the first budget.
Needless to say the Boss was most put out when the invoice for pest removal, food standard violation and toilet cleaning landed in the next budget. "
There's a reason why things are done that way. It might not be a good reason, but its still a reason. You need to find and evaluate the reason for something existing, before you fuck it up. Yes, before you ask, I did fuck up, more than once.
I have a big legacy code base as part of my responsibility and Chesterton's Fence comes up at least once a month.
At the risk of stretching the metaphor - fat is useful, too. In animals, a layer of fat will help you withstand 'lean times' of less nutrition or higher work. Run the body too hard without fat, and you burn muscle for short-term gain, or worse, die.
Similarly, in organizations, 'fat' helps out when the workload increases or productivity decreases. Run an organization too lean, and when you need to respond to a new situation, you burn out your muscle (workers) and/or go broke. This is similar to the concept of "slack."
The metaphor extension is valid. If they had succeeded in just cutting fat, it might have been merely a mistake -- failing to understand how redundancy works in an organization.
Instead, they cut without regard to fat content. Many of the organizations were already operating on a shoestring. We didn't have an abundance of park rangers. It wasn't "merely" a mistake. It was the application of ideology, without regard to either the principles of good governance or the law.
Who’s going to prosecute them? It won’t be the Trump DOJ. They’re safe, sadly.
Its one of those things that's a hard lesson to learn; the bell is run, the canary in the coal mine is yowling yet people do not listen. Then a country, or large organization, or a business is at the end of its ropes and hard decisions have to be made.
The bumbling idiots who lead us into the situation won't take the blame, the "mean guy" who makes the cuts to save the country does.
Was the USA failing before Trump took power? Most would say no. World's largest economy, and while it did have problems the nation was still an aspiration for many outside it.
Now it looks like the aspirations go elsewhere. And they do so because the cuts increased the problems, they did not save anything.
Core point aside, a nit to pick:
> the canary in the coal mine is yowling yet people do not listen
If the canary is making a noise, the coal mine is safe. They get sick and die quickly in bad air.
[dupe]
'Suddenly exposed' DOGE employees fear prosecution after Musk abandoned them
https://news.ycombinator.com/item?id=46024983
Doge 'doesn't exist' with eight months left on its charter
[flagged]
What I saw was that DOGE spent all their time chasing non-existent (or at most minor) problems imagined from the conspiracy theories they heard about in the media (e.g. woke contracts). They drank their own kool-aid and when they were actually given the keys they came up with nada. A total waste of everyone's time.
I would be interested in knowing what specific 'major waste' DOGE found in your agency. I would also be curious, given how much latitude they were given, how your management made sure they could now go any further. What I saw was senior managers escorted from the building by security and put on administrative leave if they offered anything other than complete cooperation.
[flagged]
> this stat in particular is not proof they were dumb or didn't implement their objectives.
No argument on that point: their objectives were always graft, mixed with a sprinkling of revenge and self-promotion.
Isn’t this part of Elons „process“: Delete until you deleted too much, then restore enough to make it work again, hopefully in a leaner state
That's what the supporters said. The problem was, and is, a complete lack of deliberation, which his process doesn't provide room for.
As mentioned in one of the linked discussions by ChrisArchitect, they didn't go out and actually talk to the groups they were cutting (or not cutting). The people in the field know where a lot of waste is, and having an organization, theoretically, at the level of DOGE take interest in it would have gotten things moving that just don't happen when you're 10-20 levels from those with actual authority to change policy.
The fallacy here is assuming Musk actually knows what he's doing (or even where he is) most of the time.
He is objectively, measurably spending the majority of his waking hours tweeting, not learning or performing work of any value. There was a whole project to install a bigass gaming rig in his government office dormitory[1], because the remaining time when he wasn't tweeting he needed to play video games.
[1] https://www.polygon.com/opinion/532455/elon-musk-gaming-pc-d...
Unfortunately, 600K people and counting are no longer in a condition to be restored...
> As of November 5th, it estimated that U.S.A.I.D.’s dismantling has already caused the deaths of six hundred thousand people, two-thirds of them children.
https://www.newyorker.com/culture/the-new-yorker-documentary...
no one gives a shit if a tweet fails to post. can't bring the same energy into running a country, can you
