Social media platforms: what’s wrong, and what’s next \ stacker news ~tech

pull down to refresh

Social media platforms: what’s wrong, and what’s next www.scottgoci.com/social-media-platforms-whats-wrong-and-whats-next/

643 sats \ 6 comments \ @k00b 22 May tech

I haven't bothered to read the whole thing yet, but I've enjoyed the critiques so far - they're more detailed than what you'd usually find.

view all related items

20 sats \ 0 replies \ @LibertasBR 18h

This article and its source, which mentions the tools social networks use to retain audience attention, reminded me of a simple post I saw on Nostr (which I unfortunately couldn’t find). What it basically said is that Nostr is much better than other social networks because it lets you directly reward content you liked—instead of useless content being rewarded just for attention. Even if something doesn’t deserve a single sat, the algorithm will still deem it useful, mirror it, and push engagement. Giving users the power to directly reward what’s good is the best approach.

21 sats \ 4 replies \ @Undisciplined 22 May

Anything you're taking away from it to help SN?

126 sats \ 3 replies \ @k00b OP 23 May

It mostly just confirms my biases.

Part 1 - ad dominated businesses cause platforms to abuse customers in oblique ways Part 2 - imo content and content rules are poorly organized because there's no hierarchy/composability/scarcity Part 3 - bad content doesn't cost anything and it gets elevated because it costs nothing to upvote a car crash Part 4 - people are especially bad when it costs them nothing to be bad Part 5 - power to curate/moderate should be expensive to gain, exercise, and very limited in scope/effect

He hasn't written about his own solutions yet, but his hints are things that we're already familiar with:

You get one account – that’s it. (In practice, this probably means he's going to KYC you)
There is no “one” algorithm. (Great in theory ... in practice, I suspect most differences in algorithmic preferences, when the algorithm isn't optimized for the wrong things, are marginal ... i.e. we might want different condiments on a hotdog, but we're still at this hotdog stand for a hotdog)
There are no community moderators, only sherpas (It's unclear what this means, which is why I suspect they think it's a solution to anything.)
It’s all attribute tags (I've seen this tried many times, and while AI might fix the problems with it, it's usually chaos ... because at root, it comes from a rather naive wish for no information hierarchy to exist and a belief that somehow no opinion and preventing opinion is not itself an opinion ... and I think it's better to support primitives for hierarchy formation and just make it expensive and therefore self-organizing.)
AI acts as a helper, NEVER as a creator (I suspect this mostly relates to preventing people asking dumb questions (which I don't think are that bad and are probably extant for good reason) and with the tagging part above.)
A centralized platform ("My house has a toilet." Congrats bro.)

252 sats \ 1 reply \ @Undisciplined 23 May

You're out on the frontier. At some point, you need to be the one writing these think pieces, because very very few people have given this as much thought as you.

10 sats \ 0 replies \ @plebpoet 23 May

retweet

31 sats \ 0 replies \ @Scoresby 43m

Your summary of his 5 parts is better (I don't think he even gets these conclusions). I read his thing and, while interesting, I wish I had read this comment first.

I second @Undisciplined comment: you should write more of your thoughts about social media platforms - how they work and how to build them. Write them here and let us boost them in other places around the internet.