Twitter set to accept Musk's $43B offer – sources

Any machine learning experts out there willing to weigh in on what they expect to see if Musk does open-source the algorithms used in Twitter's individualized recommendation and overall top-ranking systems? What kind of datasets are used to train these machine-learning models? What's their technology stack like anyway?

Personally I've always had a low opinion of Twitter and have never used it or even looked at it, except by accident when a tweet gets memed somewhere. More often than not my response to reading some 'tweet' is "I am now stupider for having read this." I also loathe the typical recommendation algorithms (Netflix in particular), which assume that you only want to see more the same.

Has anyone ever considered that at least some people still like to see things outside their little bubbles, some novel content, something new, creative, interesting? Those algorithms are reinforcing siloing - here, get in this little box, here's what you and your fellow box-mates like, here's some more of that box content, here's what to think and believe...

Really, the knobs and dials on these recommendation algorithms need to be directly exposed to the userbase, and the overall ranking algorithm and the datasets used to train it need to be open-sourced.

Nobody who supports Twitter's (imho responsible) use of censorship thinks that Musk's plan to libertarianize Twitter should be illegal, unlike the folks who used to shout "First amendment! Free speech!" about Twitter's moderation. They're just saying that his actions are childish and will make the site substantially worse.

And having seen the arc of many social media sites with "libertarian" moderation, I'm firmly in that camp.

I mean, Hacker News would not be what it is today without firm-handed moderation.

I can't say that I've ever felt the emotion of rage when using twitter. People say things I don't agree with all the time, even here. It just doesn't warrant an emotional reaction.

My experience with twitter is great. Things happen and the people that I follow react to them. I see news, machine learning papers and one liner jokes / pithy observations. I've no problems with freedom of speech or any other aspect of it.

Free speech absolutism isn't really the norm in places where people come together for any shared purpose.

Courts and parliaments have rules of procedure so it's not, whoever's loudest wins.

Free speech is never 100% free, there are laws against libel, fraud, conspiracy, copyright, trademark, which create crimes that consist only of speech, or civil liability. And then there are social norms.

It's always a balance between letting 20% of hateful crazy trolls hijack all rational conversation, on the one hand, and blocking unpopular opinions on the other hand. Even HN moderates a lot.

Same applies to all the rights enumerated in the US Constitution, you have freedom of religion to the extent it doesn't infringe on the other important rights and provisions of the Constitution. Polygamy is banned. If your religion says servitude of women or Black people is God's will, you don't get to practice it. 2nd Amendment however broadly interpreted doesn't let you build a nuclear weapon in your backyard.

There are people who want to block legitimate speech they don't want to hear and these should be resisted. A first step toward fascism is indeed people not caring about free speech and thinking their personal discomfort is the most important thing, starting with the most powerful. There are also liars and extremists who want to take a first step toward fascism by extinguishing the ability to have the sorts of rational debate we need in a democracy.

You need to find a balance. You need to protect free speech by having reasonable rules and norms.

"Arbitrary power is most easily established on the ruins of liberty abused to licentiousness." - George Washington

We could just filter that stuff the same way we filter email. Naive Bayes works at email scale why not comment scale? It should all be done on a per user basis so the filters only filter out language you personally find offensive.

It never made sense why try to force me to use the same filter as the Karen that complains about kids skateboarding.