Social media algorithms also by and large absolutely fail to implement the one characteristic that makes, say, Netflix or Amazon's recommendation algorithms work so well: user feedback. I can't tell Twitter "no, actually, you shouldn't have recommended this tweet to me" or Instagram "eh, I'm only a little interested in this photo". Feedback on the success of the algorithm is crucial for its correct development.
@noelle Internet companies especially social networks are forced to care about 'you'. But only in the plural sense of that pronoun--on the aggregate, that is to say averaging all of us out. To them saying "I desperately need x" is totally immaterial. It's not about ability to demonstrate need. They "listen" in that they accept your data point and if they hear it a threshold number of times, they check the last 5 years of numbers to see if ‘you’ do or not in fact need x. Bad way.
@noelle
Also, if the algorithm doesn't have a feedback loop, it usually comes off as some sort of creepy big data project. Which turns people who know what that's about off, because they correctly conclude that their activity is being data mined, likely for other purposes as well.
https://arstechnica.com/business/2017/05/facebook-helped-advertisers-target-teens-who-feel-worthless/