Hacker News new | ask | show | jobs
by itake 1548 days ago
> The content you see is determined by people, not a magic algorithm.

Isn't IG's algorithm "determined by the people"? The more engagement (views/likes/comments, etc.) the more the content is shown to other users. Plus its even smarter by customizing the recommendations to your personal preferences (e.g. if you heart a fitness pic, the algorithm will recommend more fitness pics).

Its frustrating to me seeing people being angry at {social media}'s content recommendation engine. The reason why all these companies have this algorithms is cuz they work... At worst, you're re-inventing a less efficient wheel and at best, you do what Twitter, TikTok, Facebook, Youtube, etc. all realized is a better user experience (the one you're trying to replace).

3 comments

The problem with complicated recommendation algorithms is: 1) They aren't transparent—the service could be doing anything behind the scenes, like down-ranking content from competitors and political opponents and up-ranking content from advertisers. Without publicly available rules, there is no way to know 2) The algorithm is designed to get the user to spend as long as possible on the site, so they can see as many ads as possible. But getting the user to spend as many hours as possible on the service isn't necessarily in their best interest; the best thing for the user is to give them the content they are most interested in quickly, in a way that isn't addictive.
1. I don't see why this is a problem. I don't need to know how my car works. I don't need to know how Google works. I just need it to work. Lets assume they do all of bad things you mentioned. I still don't care as long as Google returns the search results that answer my question and isn't filled with spam that games the transparency.

2. Your statements are conflicting. Currently, {social media / search} gives user the content they are most interested in quickly and as much of it as they want. The alternative is what? Showing them less of the content they want? Showing them content they don't want (so they leave)?

In the ideal world, {social media}'s algorithm perfectly aligns with human interests, delivering exactly the content the human wants and as much of this content as the human wants.

I find taking a break from consuming personalized content refreshing. It sometimes feels like I'm missing out on a lot of good stuff simply because the algorithm tries to show me what it thinks I'll like.
I enjoy architecture and interior design photos on IG.

In the future I imagine that computer generated renderings will be impossible to distinguish from photographs.

Sometimes I wonder if next generation algorithms will not only show you what it thinks you like but also create the visual content it thinks you like as well.

The point is that they work...for the company. In most of these cases the business model of the company is ads, which means they want to get as much time on the app as possible, regardless of what that means for the user's mental health or their overall enjoyment. This isn't necessarily what's best for the user.
> regardless of what that means for the user's mental health or their overall enjoyment

So they built an app that people want to use?

Diving down mental health rabbit hole, what is bad for your mental health? Seeing people that are prettier or more successful than you? If so, how does this app discourage recommending content of pretty/successful people? Would you want an app that only has ugly people?