Prof. Anima Anandkumar

Exactly! @Facebook @Twitter @YouTube have no option but to stop #misinformation Self policing has failed miserably. So there has to be regulation. Free speech vs regulation is tricky but we can't ignore it anymorehttps://twitter.com/caitlin__kelly/status/1324730148856410112 …

This is just not so easy. I am writing about this currently. I'm becoming convinced that you'd have to put whole swaths of algorithmic direction off-limits. Yes to handbags. No to political speech. Which ain't gonna fly.

Conspiracy theories like #qanonsense should receive priority. Once something gets classified as #misinformation platforms should not be able to recommend them on feeds. We only care about posts that go viral and have huge following. Focusing on them is important

Someone's got to make that initial classification. And that's not easy. It's not the QAnon of the present that I worry about.

Indeed. We need to set those standards. First amendment gives no right for speech to go viral. They can shout in vacuum but not take advantage of recommendation algorithms

I'm writing a post on this - it turns out that this is a function of Power Law memetics that are generated through social media, much as in the same fashion as "the rich get richer, and the poor get poorer." Previous spread of opinion was more constrained (and Gaussian) locally.

Exactly! Concentrates power in few. #misinformation is more likely to go viral because it is surprising and negative. I have been talking about this with @rmichaelalvarez

Why that happens is what I'm writing about. Need to understand the deep assumptions under how we(and pollsters) couldn't figure this out. Short version - we were working with a stat. independent model, inherently Gaussian. Social media has changed all that -- with status metrics.