4 ms·

> Uhm a cigarette you cannot change the ingredients of, it's tobacco. You can soak the tobacco in solution which contains additives, such as more nicotine. Wh

by RajT88 5y ago

> Uhm a cigarette you cannot change the ingredients of, it's tobacco. You can soak the tobacco in solution which contains additives, such as more nicotine. Which is exactly what cigarette companies have done in the past (and not just the tobacco, the filters, and the paper as well). The parallel here is filling people's feeds with divisive political news and posts, even when they have tried to opt out.

hunterb123 5y ago
The point is tobacco itself is a carcinogen, you cannot make a cig not cause cancer because it needs to burn tobacco at least. A social media website does not need doom scrolling or private algorithms for the feed, you can change how it works instead of adding a useless banner.
- prancer_or_vix 5y ago
  2 questions: 1) What would you expect be implemented to reduce/eradicate doom scrolling? 2) What would making the algorithm public do for us? I'm not an ML engineer, but presumably their algorithm isn't just an algebraic equation where x is how toxic the post is and y is how inflammatory it is and y is the number of kids who will think harder about suicide because of the post. Maybe I'm just super naive and that _is_ how Facebook made their algorithm, but my understanding is that the algorithm is a little more of a black-box and is a little abstract. How is a lay-person supposed to evaluate something like that?
  - hunterb123 5y ago
    I don't want them to do anything regarding doom scrolling, it was just an example and came from another user. I do want them to publish their algorithms and moderation logs so we have insight on how they are serving and moderating content. I don't care about organic user content, I do care if FB is pulling the strings to make it either more salacious or being biased in one way or another. I also care if they are banning certain users or content but not others.
  - iamstupidsimple 5y ago
    The input to these algorithms are usually human understandable and quantifiable signals like likes, text sentiment, maybe engagement history -- and the output is probably a score than can be ranked. Ultimately though even if the algorithm is a black box (entirely possible it's not ML based!) we can still evaluate it in a lab environment. Some of the signals might be generated by ML also, like photo labels, but ultimately these things are very understandable if you have the model and data.

by RajT88 5y ago

hunterb123 5y ago

The point is tobacco itself is a carcinogen, you cannot make a cig not cause cancer because it needs to burn tobacco at least. A social media website does not need doom scrolling or private algorithms for the feed, you can change how it works instead of adding a useless banner.

prancer_or_vix 5y ago
2 questions: 1) What would you expect be implemented to reduce/eradicate doom scrolling? 2) What would making the algorithm public do for us? I'm not an ML engineer, but presumably their algorithm isn't just an algebraic equation where x is how toxic the post is and y is how inflammatory it is and y is the number of kids who will think harder about suicide because of the post. Maybe I'm just super naive and that _is_ how Facebook made their algorithm, but my understanding is that the algorithm is a little more of a black-box and is a little abstract. How is a lay-person supposed to evaluate something like that?
- hunterb123 5y ago
  I don't want them to do anything regarding doom scrolling, it was just an example and came from another user. I do want them to publish their algorithms and moderation logs so we have insight on how they are serving and moderating content. I don't care about organic user content, I do care if FB is pulling the strings to make it either more salacious or being biased in one way or another. I also care if they are banning certain users or content but not others.
- iamstupidsimple 5y ago
  The input to these algorithms are usually human understandable and quantifiable signals like likes, text sentiment, maybe engagement history -- and the output is probably a score than can be ranked. Ultimately though even if the algorithm is a black box (entirely possible it's not ML based!) we can still evaluate it in a lab environment. Some of the signals might be generated by ML also, like photo labels, but ultimately these things are very understandable if you have the model and data.