Ask HN: What would a social media *for humans* look like?
8 hours ago
The problems I (and presumably others) have with social media:
- too much spam/bots
- -> it feels like the dark forest[1] is already there
- current social media platforms create echo chambers
- furthermore, disinformation is rampant
- quantity > quality, when it comes to following people
- same with likes: people are addicted to them; some base their self worth on them
- it feels like a big city; there's people everywhere but you still feel lonely
Some ideas I have:
- login based on WorldID[2] to prove human uniqueness
- twitter like system (x max characters; posts; hashtags)
- up/downvotes instead of likes
- num upvotes/follows etc. are not shown anywhere to discourage farming them
- ...
Gimmicks:
- natualistic UI; paper look; example[3][4]
- dependency graph of followers (like obsidian note graph, but for followers)[5]
- customizeable profile page (think of myspace pages)
- live cursors / other kinds of activity to indicate how other people are interacting with posts in real time
- ...
Links
- [^1]: https://de-de.world.org/world-id
- [^2]: https://maggieappleton.com/ai-dark-forest
- [^3]: https://x.com/samdape/status/1857801963317932435
- [^4]: https://x.com/_jzhao/status/1856103807903773015
- [^5]: https://help.obsidian.md/Plugins/Graph+view
Mostly the same as in early 2010s, but without growth hacking.
But, the problem is every incremental internet user after that was worse, regardless of country.
And adding to that, excluding propaganda, junk content or repost profiles mainly originate from global south as human labor value is low and people are desperate for income. (Yeah, I know, controversial thought, but on IG you can check country of all these reposting profiles and much of them aren't from EU/US). And these locations got connected to the internet specifically during 2010s.
my hopes would be that the proof of human uniqueness (via World ID), combined with moderation, would be enough to combat this problem.
current social media platforms can't cope with it, because the user supply is infinite, as only an email is required to create an account
You're being very naive. That's a classic "simple solution of a hard problem" coming from somebody with tech background.
Today people are not ashamed of posting horrible things under their real name! Let that sink in.
1 reply →
The ideas are good; you do require some sort of KYC; then these proven humans can still be bots but at least the supply is not infinite.
What I would like to see is a 'interest' timeline which does not just show me posts from people I follow, but only the posts that interest me and nothing else from those users (unless i click on User timeline); for instance; i only care about english and dutch posts, i only care about tech posts (a tech poster might have gone to taylor swift; I don't want to see that post as i don't give a rats ass). With AI's language skills, this should be fairly trivial.
The problem is, once you need to make money, you will show me stuff I don't want to see and then I will leave. Rather you just ask me for $/mo and never do that, but we see how that goes with existing networks....
so, in essence, a "for you" feed, based on your "follower feed", which filters posts based on your interactions (e.g. upvotes) with content on the platform? sounds cool
about the monetization: i would strongly be against sponsored posts / the advertising BS we see in current social media. yes, it's what made them so profitable, but IMO it's also one of the main reasons why they suck. IMO, a subscription based model would be the only viable way: it's an optional subscription, that gives you mainly social benefits: icons, decorations for your page etc. furthermore, early access features would be accessible to subscribers only. (yes, X has a similar subscription model, but it barely focuses on social status benefits; having played MMOs, I must say that humans are ready to spend big sums of money for social status items, even if it's just 'pixels on a screen'; however, a simple blue badge is OFC not enough to make someone spend big bucks.
> so, in essence, a "for you" feed, based on your "follower feed", which filters posts based on your interactions (e.g. upvotes) with content on the platform? sounds cool
yes or actually working 'less like this' 'more like this' which does not care about following or not, but only about whatever this is about.
perhaps leaving hints to stuff i Really Do Not want (always a stronger emotion than do want), for instance politics and parenting; no matter how interesting the OPs are, I care so absolutely nothing about these things, I will unfollow. And that is a shame.
Have you tried Mastodon / Fediverse with no promotion algorithm based on engagement? It feels very different to Twitter even pre-Musk.
Also: I run an informational bot, which is why I was on Twitter until Musk turned all that access off. Not all bots are bad any more than all humans. Intent is important.
tying bot tokens to a user account (i.e. the bot owner) would be something i don't see a problem with; I have a problem with "user bots": bots that simulate users with bad intent
> same with likes: people are addicted to them; some base their self worth on them
The number of followers even more so, and sometimes also who the followers are.
WorldID is not the only way to KYC. I joined Proof of Humanity ages ago, and it was a very sound project. It lost its way a bit because of trying to do too many things, but the DAO was essentially sound. https://proofofhumanity.id/.
thanks for the link; I didn't know about the project yet
I'm thinking probably a better chance of an independent KYC system if Mr Altman is not involved. :)
It's not that hard to stop most bad actors on social networks. The key is wanting to. When volume of engagement matters more than quality, then you get what we have. Strong moderation and terms of service enforcement are a prerequisite for productive conversation. If you want to be a member of the Rotary Club in good standing, there are rules to follow, and you will be judged by leadership and your peers. But the worst anti-social behavior is, not just tolerated, but encouraged on most mainstream sites. People act as though all speech is equally valid and are unwilling to make basic moral or intellectual judgements. Until someone has the backbone to define what is acceptable, we'll remain in the woods.
For example, today on HN there was another long thread about homelessness. We had people claiming that Gavin Newsom gave out cash in the 90's. We had people claiming that most homeless people have given up looking for work due to mental illness. We had people claiming that people come from all over to San Francisco because of how much better it is to be homeless there. None remotely true.
Should these opinions be permitted in public forums? This is not a hard question. It's never acceptable to pollute the infosphere with disinformation, either intentional or inadvert. It's never acceptable to pontificate about things one has never bothered to understand. If you censor such hot takes, you open space for people who know something about the topic to weigh in.
What disturbs me is how obvious this conclusion is. We all control our friend circles to avoid problematic people, yet new social media initiatives are still free-for-alls in which nefarious actors and hidden hierarchies call the shots. We build sewers, then complain about the smell.