Washington has been blown away by the speed and recklessness with which Elon Musk and his team of engineers have swept across the executive branch. But what, really, does Musk’s Department of Government Efficiency want? Washington insiders largely see it through the lens of policy, or ideology — it’s about “major reform”, or dismantling the administrative state, or concentrating power in the White House while protecting Elon Musk’s businesses. To tech-world observers, that’s the wrong way to see what’s happening. “It’s not a technical victory, but a cultural victory,” says Rohit Krishnan, a Bay Area-based engineer, economist and venture capitalist who has spent his life immersed in the business culture of Silicon Valley. Krishna offered a long download on how DOGE’s code-first approach differs from efforts by previous administrations to cut waste, and how the “disruptive” business model of startups like Uber might work (or not work) when applied to the hulking federal bureaucracy. An edited and condensed version of the conversation follows: What does Washington need to understand about DOGE’s way of doing business? If you go take over a company, you have to go fast, you have to go early and you have to go with some level of certainty that you are right, because you can't wait. You don't dilly-dally trying to follow processes and procedures, because if you do that, nothing gets done. This is a similar playbook to the way Elon took over Twitter. Many things will break, but at the other end there's a leaner, meaner, smarter, more efficient government that stands the test of time. At the same time, the government is not a business, and that has its own implications. Many people have wished for decades that someone would run the government more like a business. Is this what they meant? Or are Elon Musk and his group of 25-year-old coders doing something different? Probably the most salient difference is that they're going to be tech, data and speed first. The focus on these guys being young misses the point. It’s like a McKinsey team. The fundamental concept is just that you throw them at the problem, they gather as much data as possible about where the money is going or who's doing what, very simple, straightforward questions. Arguably difficult answers, but straightforward questions — and then you use that information to create some sort of ranking, saying this is good, this is bad, and then you try to solve the bad. So the playbook, in some sense, is extremely simple. The difference here is that they're applying it to areas like the government, where maybe it has never been applied, or it's really hard to apply because the downstream implications are extremely complex. It could cut off cancer research, or genomics departments are getting axed, or AIDS assistance elsewhere in the world. This reminds me of Uber. It was a problem for governance in cities with entrenched taxi rules. Ultimately everyone realized it was great, and the country retrofitted a governance framework onto it. Is that what Musk and Trump are hoping happens here? Uber helped popularize the idea that you can bulldoze your way through bad regulations and come out at the other end with something resembling a really good outcome. I would argue that Uber is a very good example of that working in the real world. Broadly speaking, it's good for almost everyone except those who are selling [taxi] medallions. If you believe that most regulations are like that, which is arguably a very strong belief, then you think, okay, let's go after a ton more stuff. What I’m seeing from the outside in with DOGE is that they have leaned into this far more than anyone expected. They’ve cut funds that aren’t going to make a dent in the federal budget. But on the other hand, they have also entirely disrupted the way most people think government should work. It’s not a technical victory, but a cultural victory, which is a very weird thing to get out of a bunch of technocrats who are sitting in Washington, and it’s only been three weeks. What are the costs of the business-style approach? Part of the problem is the fact that the government's apparatus needs to serve everybody. And if it needs to serve everybody, then almost by definition you can't afford to slice your customer base into those who are easiest to serve, and therefore reduce the cost of service. That idea that people in the tech world think they know how to run the government — what would it look like in practice? You look at different facets of government, and you can say things like … why should it work like this? You analyze what is actually going on behind the scenes, with things like aid being disbursed. How is it possible that there is so much fraud, or how is it possible that there's so much waste? You can investigate and say “that’s why this works this way, that’s why that exists.” In some cases those rationales might be right. Like why some large banks still run on COBOL, because it’s difficult to strip out an existing office and modernize it by putting it on cloud infrastructure or something. I also don't know whether that is worth it, but that's a separate question. At the DMV you have an appointment booking system, and you have a payment system. When you look around and you see companies like Amazon and Google doing impossibly complicated logistical things at a moment's notice, then you look at the DMV and you go, OK, maybe there is a better way to actually do these things.
| 
