this post was submitted on 05 Oct 2023
499 points (92.4% liked)

Lemmy

12542 readers
52 users here now

Everything about Lemmy; bugs, gripes, praises, and advocacy.

For discussion about the lemmy.ml instance, go to !meta@lemmy.ml.

founded 4 years ago
MODERATORS
 
you are viewing a single comment's thread
view the rest of the comments
[–] sugar_in_your_tea@sh.itjust.works 6 points 1 year ago (1 children)

I largely agree, yet I'm more like 70% lemmy, 5% Reddit, 25% working on my own Reddit alternative. Why? I refuse to give Reddit more of my data when they've demonstrated that they're more interested in monetization than making the best community (huge shift from when I first joined Reddit). I also think lemmy is doing interesting things to try to foster a great community, and I want to see what works.

But at the end of the day, I think lemmy is architected wrong. It relies on people spending a lot on hosting, which I really don't think it's sustainable, and it is also confusing for users, which is going to reduce adoption. My project attempts to solve both:

  • decentralized, so very little hosting costs, just a few relays
  • single namespace - less confusing for users, so /c/community means the same thing regardless of the relay you choose

There are some potential downsides, so I'm interested to see how bad those are.

[–] ALostInquirer@lemm.ee 3 points 1 year ago (1 children)

But at the end of the day, I think lemmy is architected wrong. It relies on people spending a lot on hosting, which I really don’t think it’s sustainable, and it is also confusing for users, which is going to reduce adoption.

Have you considered that while those may be genuine technical issues, addressing those alone won't in turn help much in building good communities? Imo one of the common problems across all social media is that a lot of smart, capable folks build their backend systems but neglect to bring on community relations teams (or in the context of entire platforms, community governance teams, maybe?) that coordinate with the people that use those systems.

Probably the big reason for this is that thus far large social media platforms have been built with a corporate mindset, and so the people aren't viewed as people, but an audience for adverts, subscriptions, products, etc. Lemmy has a different yet similar issue insofar as technically capable folks building backend systems, but they don't (nor others deploying their tech) have the resources to bring on any additional community-facing help to then coordinate and collaborate with people in governing their spaces.

[–] sugar_in_your_tea@sh.itjust.works 3 points 1 year ago* (last edited 1 year ago)

Thanks for the feedback! And yeah, it's absolutely something I've been thinking about. I'm not sure I'll even publish it once I have it working because I'm worried about a bunch of nonsense like CSAM or bigotry, much less the more mundane issues of not spam.

And that's the rub, building a good community is hard, especially on a digital platform, and requires a very different skillet from building good software. I'm not sure I'm cut out for that part, but I can learn from the issues lemmy runs into and try to solve them with technical solutions, namely quality moderation tools.

That's especially challenging in a decentralized system, and it seems to have caused a lot of people to leave, so I'm trying to have a good solution out of the gate.

Since it's decentralized, I can't force anyone to recognize any given moderation without breaking the whole point of decentralization (i.e. nobody has control, even me), so my plan is to rely on a web of trust type system. For example:

  1. you flag users you trust
  2. users report content for violating certain rules
  3. if enough people in your web of trust flag content, you won't see it

I'll probably include a default list, but users would be free to choose their own moderators if they think mine suck. But I have no idea how well that'll work, but once I get a prototype working, I'll post it somewhere (probably here on lemmy) to solicit feedback.

I think this idea is different enough to get people interested, and hopefully robust enough to keep people on the platform. To get content, I'll probably bridge it with lemmy or something so it'll look like another instance (again, not sure if that's feasible or even wanted). It's early days, and the more frustrated I get with lemmy, the more I'll work on it.