Leaving Twitter

Twitter always used to look a lot like Craigslist. It stumbled into something that a lot of people found very useful, with very strong network effects, and then it squatted on those network effects for a generation, while the tech industry moved on. Twitter, as a technology company, has been irrelevant to everything that’s going on for a decade. It was the place where we talked about what mattered, but Twitter the company didn’t matter at all – indeed it did nothing for so long that people got bored of complaining about it.

Meanwhile, lots of people tried to build a better Craigslist and a better Twitter, but though a better product was pretty easy, the network effects were too strong and none of them really worked. Instead, we unbundled use cases one by one. As Andrew Parker pointed out in 2010, a whole range of people from Airbnb to Zillow to Tinder unbundled separate pieces of Craigslist into billion dollar companies that didn’t look like Craigslist and solved some individual need much better. This is often the real challenge to tech incumbents: once the network effects are locked in, it’s very hard to get people to switch to something that’s roughly the same but 10% better – they switch to something that solves one underlying need in an entirely new way.

Hence, Mastodon has been around since 2016 without getting much traction, but slices of conversation, content or industry have been unbundled to Reddit, LinkedIn, Instagram, Signal, Discord or, more recently, Substack, which someone joked was Twitter’s paywall. 

Meanwhile, Twitter itself drifted aimlessly for a decade, becoming known in Silicon Valley as a place where no-one could get anything done. This is a big part of why Elon Musk was able to buy it – $44bn was a top-of-the-market price, but even Snap was worth $75bn in January 2022, when he started building a stake – how much bigger should Twitter have been? And so, when he made his bid, there was, briefly, a lot of enthusiasm in tech: pent-up frustration with the existing product and a sense of how much better it could be; enthusiasm that there could be innovation and new product ideas (and, from a small but noisy group, frustration with the politics of Twitter’s content policies, of which more in a moment).

It didn’t work out like that. The last year swapped stasis for chaos. Stuff breaks at random and you don’t know if it’s a bug or a decision. The advertisers have fled, and no-one knows what will be broken by accident or on purpose tomorrow. The example that’s closest to home for me was that the in-house newsletter product was shut down – and then links to other newsletters were banned. Pick one! It’s hard to see anyone who depends on having a long-term platform investing in anything that Twitter builds, when it might not be there tomorrow.

There are various diagnoses for this. Tesla has sometimes been run in chaos as well, but the pain of that is on the employees, not the customers: you can’t wake up in the middle of the night and decide the car should have five wheels and ship that the next day, but you can make those kinds of decisions in software, and Elon Musk does, all the time. Perhaps it’s a fundamental failure to understand how you run a community. Or something else. But whatever the explanation, Twitter now feels like the Brewster’s Millions of tech – ‘Watch One Man Turn $40bn Into $4 In 24 Months!’

Meanwhile, beyond the chaos, there has been no sense for the actual users of where we’re going. There was a plan, both ruthless and chaotic, to reset a broken and grotesquely overstaffed company culture and turn it into a place that can execute, but no coherent sense of what it should be executing. What should those newly hard-core engineers be shipping? A ‘super app’? A universal content platform with no external links? Your financial life? Seriously? 

And then, there are the Nazis. 

This is a debate with baggage. Part of the criticism of Old Twitter was a perceived tendency to trigger-happy moderation, and there is in fact a pretty mainstream view in the content moderation world that you shouldn’t (or indeed can’t, practically) try to ban and block anything you don’t like (unless it’s actually illegal), but instead you should have a spectrum of what’s objectionable and control things within that by controlling visibility. Keep things out of the recommendations and suggestions, down-rank them in the feed and replies, and don’t let them monetise or advertise. There will be some bad stuff, but the worse it is the fewer people will see it. Meanwhile, pour your effort into stopping scammers and state manipulation, and think about how your product design might encourage or discourage the rest of us from being mean. Reasonable people can disagree about that. But. 

But it didn’t work out like that. The teams that looked for bots, scammers and state actors were mostly fired, and the scammers, Nazis and propagandists all bought the ‘Blue Ticks’. These little badges used to mean ‘notable person’ (in a chaotic and inconsistent way typical of the old Twitter) and are now supposed only to mean ‘real person’ (but often don’t) – and they give you both amplification in all the algorithms and a share of revenue if you drive a lot of replies. The more you troll, and the more furious replies you generate, the more Twitter promotes you and the more Twitter pays you. We saw this at its logical conclusion in the last week, with deliberate misinformation promoted by what we used to call ‘fake accounts’ that now get promoted by the algorithm because they pay their $8/month. It turns out that social networks are harder than rocket science.

And then, there’s Elon. 

I once called Elon Musk ‘a bullshitter who delivers’ – he says a lot of stuff, and yet, there are the cars and the self-landing rockets. People generally struggle with one or other of these – they will refuse to accept the problem in selling a car that can’t drive itself as ‘full self driving’, or they will say ‘he didn’t found Tesla!’, forgetting that he’s run it for the last 15 years. Most of what you see at Tesla or SpaceX really is his creation – but half of what he says is bullshit. 

Until recently, though, the bullshit was mostly about cars or tunnels. It wasn’t repeating obvious anti-semitic dog-whistles. It wasn’t telling us that George Soros is plotting to destroy western civilisation. It wasn’t engaging with and promoting white supremacists. It wasn’t, as this week, telling us all to read a very obvious misinformation account, with a record of anti-semitism, as the best source on Israel. Of course, it had bought a Blue Tick. 

In talking about this, I am reminded very much of talking about the last leader of the UK’s Labour Party, Jeremy Corbyn, who had somehow spent much of a career devoted to anti-racism, well, supporting and praising anti-semites (‘the world’s most unlucky anti-racist’). The Chief Rabbi declared that British Jews were afraid of a Labour election victory, and yet too many people with a tribal loyalty to the party just refused to read, see or hear any of this. They decided to blind themselves.

If you see a man claim that he’ll have ‘full self-driving’ working ‘next year’ for half a decade and can’t make fun of that just a little, you are probably blinding yourself too, but it does’t matter much. And maybe you don’t care much about this, or have decided not to see it. But I was on Twitter since 2007, and built a meaningful part of my career on it, and I won’t be posting at all, for the foreseeable future, because I think it does matter.