Comment

Comment by personjerry

I worked at Facebook for some time, and did a bunch of data work. We had this culture of building something, looking at the results between the experiment groups, and then choosing the statistically more successful one -- i.e. a newsfeed algorithm that had better engagement.

This sounds great at first, and certainly is straightforward if you want a promotion. But behind the scenes some of us had this thought that our observations only amounted to short-term gains. Although we had small long-term experiment holdout groups, the truth is they were rarely reviewed because it was unsexy.

My current thinking is that features like the echo chamber effects from Facebook's algorithms, Snapchat's snap streaks, and clickbait like this, all serve to optimize short-term engagement. Yeah, I want to watch that sexy new show or keep my streak going or have my opinions validated. But there's a diminishing return on clickbait, hollow articles isn't there? I can only fill up so much time with garbage like that before I'm bored. I can only like so many posts before I feel like they're all the same. And once my snap streak is broken I hate snapping.

The data/engineering/product loops at tech companies favor boosting short-term metrics; The employees are incentivized to do so and this is what they measure, so this is what they build. That's why we end up with features like this. That's why Snapchat fell off. That's why Facebook fell off. And that's why Netflix feels increasingly stale (despite there being a lot of quality content if you dig).

parent personjerry 5 years ago

Discussion

Live from the official Hacker News API

A modern reader for Hacker News · SvelteKit + Skeleton