The alternative to all of these institutions is currently social media, which is worse by any metric: accuracy, fairness, curiosity, etc.

I am more optimistic about AI than this post simply because I think it is a better substitute than social media. In some ways, I think AI and institutions are symbiotic

▲

qcnguy 4 hours ago | parent [-]

It's better in all those metrics.

Go on X. Claims are being fact checked and annotated in real time by an algorithm that finds cases where ideologically opposed people still agree on the fact check. People can summon a cutting edge LLM to evaluate claims on demand. There is almost no gatekeeping so discussions show every point of view, which is fair and curious.

Compare to, I dunno, the BBC. The video you see might not even be real. If you're watching a conservative politician maybe 50 minutes were spliced out of the middle of a sentence and the splice was hidden. You hear only what they want you to hear and they gatekeep aggressively. Facts are not checked in real time by a distributed vote, LLMs are not on hand to double check their claims.

AI and social media are working well together. The biggest problem is synthetic video. But TV news has that problem too, it turns out. Just because you hear someone say some words doesn't mean that was what they actually said. So they're doing equally badly in that regard.

▲

mattnewton 4 hours ago | parent | next [-]

Last time I went on X my feed which I curated from ML contributors and a few politicians had multiple white nationalist memes, and engagement slop. Fact checks frequently are added after millions of impressions.

I am sure there are very smart well meaning people working on it but it certainly doesn’t feel better than the BBC to me. At least I know that’s state media of the UK and when something is published I see the same article as other people.

▲

chrisjj 3 hours ago | parent | prev | next [-]

> Facts are not checked in real time by a distributed vote

Nor could they be. We don't even have the tech for trustworthy electronic elections.

▲

biophysboy 4 hours ago | parent | prev | next [-]

As I said, AI is better than social media. AI is trained on and references original sources, which makes it better than reading and believing random posts.

	▲	chrisjj 3 hours ago \| parent [-]
		Original sources include random posts.

▲

throwaw12 3 hours ago | parent | prev [-]

X is not a good representative to free speech.

1. It censors some topics. Just for fun, try to write something about Israel-Gaza, or try to praise Russia and compare the likes/views with your other posts and over the next week observe how these topics is impacting your overall reach even in other topics.

2. X amplifies your interests, which is not objectively true, so if you are interested in conspiracy or Middle East, it pushes you those topics, but others see different things. Although its showing you something you are interested in, in reality its isolating you in your bubble.

	▲	qcnguy 3 hours ago \| parent [-]
		1. Are those topics being censored? You don't seem to know that is true, you're just making assumptions about what reach should be. They open sourced the ranking algorithm and just refreshed it - can you find any code that'd suppress these topics? 2. The media also amplifies people's interests which is why it focuses on bad news and celebrity gossip. How is this unique to social media? Why is it even bad? I wouldn't want to consume any form of media that deliberately showed me boring and irrelevant things.