The harshest thing about the Silicon Valley “bro” trope, I think, is that the people who create technology products don’t realise, or aren’t compassionate about, the problems that arise when those platforms are unleashed on the public.
For the most part, they do care. But in the same way not all of us are good at math, or writing, or singing, or the long jump… different humans possess different skillsets, some of which necessarily contradict each other. By and large, the hierarchies in the tech industry reward engineers first and foremost, people who — through no fault of their own — are often characteristically not as well-equipped to deal with challenges that fall outside the binary decisions laid before them in code. When extraordinary brains that build products and solve problems based on IF/THEN/OR come face-to-face with a however or maybe–the results can be less than optimal.
I’m minded of this as Elon Musk uses “simple” tweets to lay out how he plans to fix Twitter. Here’s one:
He’s pinned that to his profile, perhaps signaling it as a mini-manifesto of sorts. This is his view on free speech and its parameters. Ok. But those familiar with the content moderation space consider it naive to the point of parody, as though a moderation algorithm is as straightforward as: IF very against law THEN delete tweet.
It’s a classic case of “engineer’s brain” in action. But you can hardly blame Musk for that. After all, he has engineered his way to being the world’s richest man by solving almost exclusively engineering problems: how to securely send money over the internet (PayPal); how to create a high-performance EV and then make it affordable (Tesla); and how to reuse a rocket (SpaceX). None of these are problems in which the great complicator, the human condition with its whims and inconsistencies, are a factor. Nothing on Musk’s CV comes remotely close to the credentials required to know what’s best for Twitter.
In reporting out this story, I spent some time on the anonymous workplace messaging app Blind, where employees — verified using work email accounts — can sound off. It’s been lively on the Twitter board since the Musk news broke, and it’s by no means an entirely anti-Musk environment. But this comment details what Musk is up against — it is an angry response to another Blind user saying Musk’s idea of sharing “the algorithm” is a good one:
Let me clue you in on a little secret, “the algorithm” (which is not a giant if else statement but a neural net) is retrained regularly. And tweets come in all the time. To figure out how something will perform you need almost the entire backend of Twitter. And then YOU are going to try to determine if it is fair? I’ll tell you right now that you can pick a measure and a slice for which it won’t be, not because the algo is evil (ML is morals neutral) but because it is an extremely complex problem. I am sure we don’t have political leaning as one of the thousands of features we train on. But it doesn’t mean that it will be unfair in 3 days or even 1 day after retraining. What does having the algorithm help you with? Moderation policies should be open, I agree 100%, but there really cannot be a world without moderation. I agree that sometimes Twitter probably overenforces and sometimes we underenforce. We probably do both on both sides of political spectrum. Not because we hate one side but because it is a hard problem and we don’t always do it right. One side is more vocal in complaining. And I know you will pull all the incidents they cite because you are completely ignorant of what happened on the other side. Remember when we were going to get level 4 self driving Tesla every year for the past 8 years? Expect a Tweet from Elon in 5 years or so admitting that the social media that allows healthy open conversation is much harder problem than he thought. Jack was a free speech absolutist too. Anyway, I appreciate small minds debating problems they know nothing about, but you have no idea what you are talking about, neither does Elon.
