Elon is Twitter's Chaos Monkey
Elon might not be good for entitled Twitter employees, but he's terrific for 'Twitter the product' and its users
There's been enough articles written about Elon Musk's Twitter takeover that would take years to read them all. Why another one? Because almost all of them missed a critical thing: what Elon is doing with Twitter, the product, is pure genius.
Let's leave aside the layoffs and Elon’s managerial style for a minute and let me focus on what he's doing at the product level.
Anyone who works in tech knows that the product changes he is making would take 10x longer in any other company. Why? “Because he works his employees harder than anyone else!” I hear hoards of marxist-leaning engineers clamoring. Sure, that's part of it, but focus on this and you will miss that Elon is implementing the most faithful and pure version of Lean Software Development & Agile Methodologies that I have seen to date in a big tech company. It's the biggest real time experiment on the virtues of Agile ever conducted in a company this size and influence.
Take, for example, the rollout of the 'Twitter Blue' feature a few days ago. First thing to note is that it was built and shipped to all customers in just one week. That within itself is an astonishing accomplishment given Twitter's size & scale: I work in a tech company which I deem to be quite similar to Twitter in many respects (including codebase size and general product "popularity") and, let me tell you, we wouldn't be able to ship shit to our users in less than a couple of months, being optimistic. More often than not 60-70% of the "development" time is actually composed of non-technical tasks and overhead: first, product and design need to get their shit together and figure out what the hell we should be building in the first place, we then build it, show it to them -- which will magically elicit more requirements --, then we need to figure out what population of our users to ship it to so we can do a controlled test for a few weeks, talk to the respective stakeholders, security... you get the point.
Elon somehow is able to cut through all that BS and ship a new feature in just a week to all his users. How? What does he know that the rest of the industry don't? I think he understands something that might sound obvious on the surface: Twitter is just a fucking website. Literally no one will die if they fuck up.
Let me explain: Elon is used to building rockets and automobiles. That is, stuff that will kill people if built hastily or in the wrong way. A single bug in the Dragon Capsule firmware and a couple astronauts die. Not to mention the huge reputational, financial and political consequences of fucking up at that level. And even in those industries, Elon has famously pushed the boundaries, implementing a much more "agile" methodology than those industries were used to.
Keeping that in mind, you can see why he does not give a flying fuck about breaking twitter's 2FA service, or the "official" / verified accounts feature for a few hours. A bunch of opinionated SW engineers will be triggered because they think Elon has no idea of what he is doing and, well, that's about it. The world will keep turning. Twitter will lose some ad revenue and take a reputational hit, but it's a cheap price to pay for quickly trimming the product fat and iterating the product at light speed compared to its competitors.
And, to be fair, I think it's true that Elon has no idea of what he's doing. But he knows. And he knows how to learn: being like an annoying 5 year old poking at every microservice and layer of the tech stack while asking "what does this do? Do we need this? Can we turn this off and see what happens?". The guy is Twitter's much needed chaos fucking monkey pulling all the cables in the server room and lighting a fire under everyone's ass to get shippin' — and do it pronto. He's making Twitter "devolve" into the startup it used to be: move fucking fast, break shit, and keep what works. After all what matters the most is delivering value to customers as fast as possible — precisely the core tenet of Agile.
But he's not only making Twitter ship like it's a scrubby 2-month-old, 3-people startup in a garage, he's also making them iterate the fuck out of it. When the whole 'Twitter Blue' mess derived in a bunch of impersonation of official accounts after Twitter removed the new "Official" badge, they backtracked the changes in less than a day in order to do another iteration. Talk about a tight feedback loop. The whole thing took like 10 days.
I can hear Product Managers around the world wagging their fingers at me right now and saying "but anyone could have foreseen that outcome, you didn't need to ship the verified feature to learn that it didn't work!". First, fuck off. Second, sure, but the learnings you gather from actually shipping is orders of magnitude greater than whatever you geniuses can come up with. That's why the most passionate founders and CEO bypass Product altogether: they are the product, they have a vision, they want to try it out, and they don't have time for y'all to circle jerk in your PM meetings discussing where to put the fucking call to action for a month at a time without shipping shit.
This is why Elon wants to turn Twitter in an engineering-driven company. There’s less need for product visionaries when you humbly accept that no one can really predict how a feature will do before shipping it. Not one, not a hundred PMs combined -- reality is just too complex. Shipping, iterating and keeping what works is the only actual way to learn what users want, and I love that Elon is brutally remembering everyone about it.