[aserafini]

Why does selling information mean 'large amounts of data must be shared'?

The explanation is in your point (4). Let's expand on it:

I click an ad targeted at African-American, Christian homosexuals, over 40 years old, living in Boston, earning >$100,0000 with a custom audience set consisting of 100 email addresses the advertiser scraped from a forum.

By the way, I never told Facebook my sexual orientation, salary or religion, it was inferred based on other sites I visit.

When I click the ad, Facebook reveals to the advertiser that I match this description: and this is exactly what the advertiser pays them for.

[wsetchell]

Thanks for the explanation!

Short of turning off ads or ad targeting, do you have any ideas about how to make this better?

[aserafini]

I'd like to see the 'Wikipedia' of social media emerge: a centrally managed, decentrally moderated, non-profit, 'nagware' funded social media.

Most users don't care about federation or decentralisation. They want low-cost and convenience but they're starting to realise they don't want it at the expense of their mental health or privacy.

Nothing can be simpler for the end user than a centrally managed service: I just go to wikipedia.org and start reading/writing. Censorship issues are avoided because moderation is decentralised: the site has a governance framework.

So I believe there's a space for a similar concept in social media. There's no question social media fulfils a genuine human need. But the profit motive inevitably forces commercial social medias to make decisions that detriment the user. For example, 'engagement': a santised term for addiction. Is it good for the user that social media should be continually engineered to increase 'engagement'? A non-profit has no such conflict of interest. It may even take steps to reduce engagement if it believes users are at risk of depression, addiction, anxiety etc.

I would love to talk to anyone who has ideas about bootstrapping something like that (email address in profile).