How hard is it to do things? Often it turns out to be pretty hard.
There’s an infamous Hacker News thread about Dropbox that crops up whenever there’s a conversation about how difficult it would be to build a replacement for a well-known service. The forum note, which you can read here, comes from back in 2007, when Dropbox was a “YC app” that had a very simple pitch: “Throw away your USB drive.”
The poster argues that what Dropbox built at the time was something potential users could replicate “quite trivially by getting an FTP account, mounting it locally with curlftpfs, and then using SVN or CVS on the mounted filesystem,” provided they are on Linux. The commenter helpfully included instructions for Windows and Mac as well.
This is my last Exchange for some time. Starting Monday, I am on leave for a few months.
Thankfully, my amazing colleague Anna Heim — who already co-writes with me weekly — will be at the helm.
Back soon! — Alex
Perhaps the comment was right at the time; Dropbox’s core user group when it was just a little YC company was likely tech heavy. Some of them must have had the chops — and interest — to build a Dropbox clone on their local machine.
It was a small group in the end. 
Dropbox grew to become a massive company, went public and in its most recent quarter (Q3 2022) posted $591.0 million worth of revenue. That’s nearly $200 million per month.
The real argument that we might pick with the Hacker News scribbler is that building a single piece of software for one’s own use may be, at times, doable as an alternative to using a paid, hosted service. Hats off to folks who tinker, build and write their own code; may you make really cool things that are beloved.
But for us who don’t get paid to noodle with code, we’re going to use services that abstract away the technical challenges, as well as all the complicated company-building work that comes with highly usable and portable digital systems.
Some argue that Slack is another app that could be replaced by homegrown coding. It’s a chat app, after all, and those have been around for ages. How hard could it be, right? Hell, there are online classes that claim to provide the “step-by-step process to building a Slack clone without writing a single line of code.”
The Salesforce-owned enterprise chat service, however, is a bit more complicated than it might seem at first blush. This famous image crosses my Twitter feed every few weeks, attempting to show how Slack decides whether to alert users with a notification.
If coding that for yourself is simpler than using Slack, which even has a free tier, one more round of hats off to you. You must type like the wind.
Behind the Dropbox and Slack examples is a tendency by folks outside a particular system to underestimate how hard it is to build and run. I got a bit of this lesson firsthand by watching the data team at Crunchbase work through thorny questions. With smart folks, good intent and a platform with plenty of history, it was still a tough job. Good luck replicating Crunchbase overnight, no matter how good of a scraper you write. It may look easy, but it isn’t.
Andreessen Horowitz’s Future outlet also comes to mind here. How hard can it be to build a new publication that is less tech-critical than mainstream publications? All that’s needed is a blog-style website (easy), some folks to write (already got ’em in the portfolio) and distribution. Given how much money venture capital shops have to spend on marketing, and the generally extensive reach of venture investors on Twitter, it’s a slam dunk. Right?
We all know how that’s going.
This brings us to Twitter and running a moderated social media service, which is another way of saying a social media service that works. When electric car baron Elon Musk bought Twitter, he came in with a bushel of ideas. These included “free speech” that “matches the law.” That didn’t last. Now Musk is neck deep in an argument about where to draw the lines concerning what is and isn’t allowed on his platform. Put another way, Musk is working in real time to find a set of moderation policies that match his own politics and personal desires.
That’s fine; it’s his company and he can do with it what he wants even if I may find myself critical of some of his choices. But what the mess shows is that moderating Twitter is not easy.
The complexity of building and running digital services is massive. This is why Dropbox has cybersecurity teams, why Slack’s chat app is way more than just that, why an online publication is more than just words on a website. And it’s why Twitter’s internal teams and negotiations about how to handle content moderation were there not to placate some sort of political group hellbent on teaching the capital class a thing or two, but to chew on hard questions and tight deadlines to try and get the rules right.
I don’t know how the Musk-Twitter epoch will end. And frankly, I don’t want to get into a back-and-forth with you about what constitutes doxxing. Which is the point. But since Musk seemed to think that running Twitter’s moderation efforts himself was going to be easy, perhaps next he can code us a Dropbox clone for Linux machines that will take down Big Storage.
Comment