Werd I/O

While I don’t think the guy who runs racist factories is necessarily the right person to buy a prominent social media company in order to “save democracy”, he’s right about one thing: if Twitter is to truly be a public square, the algorithm needs to be open sourced.

Over the years, there’s been a lot of chat about algorithms: how they’re designed to keep you on social media sites by filtering your feed for the kind of content that you want to see and interact with, rather than just showing you the reverse chronological list of all content you’ve subscribed to. It’s the mechanism that causes Facebook Pages to have to pay to have their posts actually reach their communities; it creates filter bubbles; it exacerbates power laws that help people with large communities reach even more people.

When we’re talking about algorithms on Twitter, though, the subtext is around the work the company has done on harassment and abuse. Accounts that regularly post hate speech are kicked off the platform, keeping vulnerable communities safer and making interactions on the site less toxic for everybody. To some (hint: their demographics are usually not the ones targeted for violence by these kinds of accounts) these are simply “differences of opinion”. That’s the kind of content that would be reinstated in a world with an open source algorithm. Don’t want nationalists on your feed? Use an algorithm that hides them.

And sure, maybe. The web as a whole works a bit like that, after all: if you’re not a white nationalist, don’t visit Stormfront or Truth Social. Those sites exist as niche underbellies where disaffected racists can spew their hatred without being disturbed by the rest of us. When that content crosses the line into illegality - at least, the content that’s observable, which is likely the tip of the iceberg - theoretically the police get involved. (The police themselves have a white supremacy problem, hence the theoretically.)

But speech isn’t simply speech. Speech has the power to organize, to rally, to build movements and cause both great positive change and great harm. Free speech maximalists like to quote Brandeis’s principle that the way to counter harmful speech is with more, positive speech. But past a certain point, once speech has brought together movements and those movements have taken to the streets, the way to counter it has been with armies and force. Long before that point, it’s doxxed activists, invited pipe bomb attacks at abortion clinics, and led to a man firing an AR-15 rifle in a family pizza joint. Racism and violence are not harmless differences of opinion; they are a cancer.

The First Amendment restricts the government’s power to limit speech and assembly. However, tweets are stored on Twitter; they’re entirely in the domain of a private company. Private companies have the right to make rules about what happens on their systems, at least until they become a common carrier. The content that is restricted on Twitter is not restricted in America; other sites exist where it can be posted. That those sites are markedly less popular - and that most hosting providers want to avoid any association with them, as is their right - says a lot about where American hearts generally really lie. When Twitter imposed stronger content moderation, the site began to grow faster.

Musk’s call for open algorithms is not unproblematic, for the reasons I’ve described. But if that’s what he really wants, the solution is a fully-decentralized protocol for social media: one that, like the web, isn’t owned by anybody, so there’s no central organization that can made decisions about allowable content or how the algorithm works. Everyone will be able to choose their own algorithm. It just won’t quite go how he, or other members of the nationalist-aligned, think it will.

As a web user, you probably use a web browser every day. There are tens or even hundreds to choose from, but you probably have never considered using Puffin or Redcore. Most likely, you’ve heard of three or four: Chrome, Firefox, Microsoft Edge (which you might think of as Internet Explorer, even though they’re not the same), and Safari. You might not even know that Edge and Chrome have the same browser engine.

If social media becomes fully decentralized, there will be a handful of “social media browsers” (we’ll probably know them by another name) that people will ultimately used. They’ll compete on providing the friendliest experience and the least toxic environment. If there are other browsers that cater specifically to the nationalist experience, it’s likely that the mainstream browsers will refuse to peer, blocking access to them outright. This has already happened: most Mastodon instances refused to peer with Gab. For its part, Google removes listings to comply with local law, for example to remove Nazi sites in Germany.

That’s not to say that open sourcing is a bad thing: it’s a great thing. That choice between algorithms, the customization of how you receive content, would be a major boon for consumers. Making social media more like the web is a win for everyone.

But there’s no world where nationalists get what they want. If Twitter turns down the dial on its content moderation, the community becomes more toxic and turns more people away. A nationalist-friendly alternative will never become mainstream, as Gab, Truth Social, Minds, etc, have already shown us. If the community turns en masse to a decentralized, open source alternative, any broadly successful entry point to that network will need to incorporate a friendly experience that includes community protections.

Because what they want is for their ideology to be mainstream, and for their words to be heard as loudly as possible. In a world where most Americans support diversity - and where diversity is part of the fundamental DNA of the nation - that message is only going to spread so far. In America, you have the right to free speech, but you do not have the right to be heard. For that, your message actually has to resonate.

Elon is right to want to open source, but he’s wrong about the implications. The world is moving in a more inclusive, more compassionate direction, and there’s no going back. Nationalism and traditionalism are firmly party of the 20th century, and that is becoming an increasingly long time ago.

As part of the ongoing mid-life crisis that brought me electric blue hair and a new car, I’ve been thinking about getting a tattoo. Of course, if I’m going to mark my body, I’d like it to be the most meaningful (and potentially the nerdiest) image possible.

I have parameters. Given my family history, and the history of the twentieth century, I’d rather not use any kind of barcode, QR code (can you imagine?), or identifier. I also don’t really want words or any kind of quote. I thought about a waveform of my mother’s voice, but honestly, I don’t think she would have approved, and it feels a little like the 21st century equivalent of drawing “mom” in a heart. I qualify for a semicolon tattoo, but I don’t want one of those either.

Maybe an and gate? A symbol representing Earthseed’s God is change? A TARDIS? It all feels very stereotypical.

Maybe I’m just too fickle. It seems so permanent, and the me I am now is not the same person I’ll be in two, five, twenty years. On the other hand, I like the idea of marking life like rings in a tree.

Do you have a tattoo? Is it meaningful to you? What did you get?

Startup culture as popularly described is a sham. You can read all the books you want on the subject, but the most successful companies build their culture from the ground up based on the same kind of learning cycle that they use on their customers. To succeed, you’re going to have to attract the best people - people who have a ton of options, many of which probably pay better than you do - and you’re going to have to find ways to keep them there.

Particularly in today’s market, if you’re not treating people well, they’re going to find something better. If you create a hustle-rich, competitive, aggressive environment that makes people feel like they’re under attack, they’re going to go find a place where they don’t. If you create a culture of long hours peppered with inflexible meetings, you’re going to lose the parents and carers who likely also happen to be your most experienced colleagues (as well as one where women, who largely still bear the brunt of parenting, are less likely to feel welcome). Your culture has to be one of deeply-held respect: not just of the expertise of every employee, but of what they bring as a three-dimensional human, and of their lives outside of work. If you think of people as a fungible resource, they’re going to feel it.

There’s no glory in working nights and weekends, and there’s nothing laudable about asking people to do so. Startups are a marathon, not a sprint. All your employees have lives beyond work. None of them are anywhere near as invested - in the literal, company-ownership sense, but also emotionally - as you are. As a founder, you might be burning the candle at both ends, but when the startup exits, you have the most to gain. Generous options help here, but if employees don’t feel like they have a strong say in the direction of the company, they’re little more than a lottery ticket from their perspective; a get-rich-quick scheme. If they lose trust in you, if they don’t have enough options to make a meaningful difference in their lives even in the event of an exit, or if the option price is so high that executing them is out of reach, or if there aren’t meaningful triggers, any kind of motivating factor that options could have brought is lost.

Even for founders, those long days come with diminishing returns: most knowledge workers can muster six hours of focused work at best. After that, anyone’s work is low-quality. In a small team, that means you’ve got to focus on building the smallest, simplest thing you can: a clearly-defined plan you know you can execute well with the time, team, and resources at your disposal. Because all of those things come at a premium, built-in ways to fail fast and learn quickly are incredibly important. A growth mindset and a nimble approach are more important than an “agile” one: paint-by-numbers scrum ceremonies aren’t going to save you, but short work sprints built around learning loops might.

That also means optimizing your workday for flow: removing meetings and interruptions so people can actually get work done. (Talking in meetings isn’t work; at best it’s a tactical huddle, and at worst it’s the performance of doing work.) As Steve Galevski put it in HBR a few years ago:

By cultivating a flow-friendly workplace and introducing a shorter workday, you’re setting the scene not only for higher productivity and better outcomes, but for more motivated and less-stressed employees, improved rates of employee acquisition and retention, and more time for all that fun stuff that goes on outside of office walls, otherwise known as life.

People have to think and reflect on their work to do it at a high quality. To be able to do that, they need time, emotional safety, and rest. If you create an environment of constant interruptions, long hours, and a lack of emotional safety, you’re shooting yourself in the foot and then some. Yet that’s exactly what a lot of startup porn advocates for, and where work has begun to go during the pandemic: a world where you can’t escape work, with numerous interruptions, long hours, and an underlying aggressive culture of hustle.

What modern startup employees are looking for is an inclusive place where they can do great work, live well, be treated with respect, and be compensated accordingly. It’s not hard, as long as you stop to really think and care about them. The catch is that many founders don’t.

Rewriting software from scratch is usually a terrible idea. But I’m thinking about it.

The Known open source codebase is now 9 years old; a PHP kludge that I wrote while my mother was recovering from a double lung transplant still powers my site and many others. It became the foundation of my second startup, and is still an open source project today. But there were a number of years when I didn’t pay attention to the codebase, and there’s a lot to unpick.

Meanwhile, the hosting landscape has completely changed. It used to be that you’d buy some space with a shared host and upload files via (S)FTP; these days virtual hosts are commonplace and getting easier to use. There are one-click installation buttons for Heroku and other hosts.

I’d like to clean PHP Known up, and I’m trying my best in between all the other things that are going on in my life. Probably that should mostly be about getting to another stable release: a lot of the architecture has been changed (by other developers) and a lot of users are having trouble installing it. So bringing that back to accessibility would be nice.

I also want to fix import / export, so that people can take their Known content and use it elsewhere. A lot of folks, rightly, would like to migrate to WordPress or Ghost in particular. They should be able to do that with ease.

But I also like the idea of going back to basics with Idno, the underlying platform, and thinking about it again. The original core idea was that you could create a stream of arbitrary content, set fine-grained permissions on it, and both post to it and consume from it in a bunch of different ways. If you wanted to post via the web, great; via a webhook, API endpoint or common standard like Micropub, also great. Likewise, reading via the web, JSON, RSS, MRSS, ActivityStreams, and so on would all be easily possible. Permissions would limit both reading and writing to a customizable set of people, from everyone on the internet down to one person.

That’s not really where Known ended up going, but I still find that potentially interesting as a project. Instead of PHP, I’d be more inclined to write it as a Node service these days (or use it to learn something I’m less familiar with, like Go).

I wish I had more time to work on these sorts of projects. But it’s something I’d love to figure out how to fit in: I want to clean Known up, and return to Idno as a way to write scalable streams of arbitrary content. In the meantime, it’s fun to think about.