BikerJared

  • 29 Posts
  • 25 Comments
Joined 1 year ago
cake
Cake day: June 7th, 2023

help-circle





















  • I agree - decentralized is the future. Fwiw, I’m really enjoying lemmy and want it to succeed. I’m just hoping to help somehow. I’d love for centralized social networks to become decentralized so that the perverse incentives that exist currently can evaporate.

    I’m nervous posting this because I don’t want to come off as telling you “what to do”. This is coming more from a desire to support the awesome work you’re doing to help lemmy grow on the fediverse. If this is too “forward” - I’m happy to back down and continue posting everything I’ve got to /c/earthporn, /c/photography, and /c/jeep. ;)

    It seems in the short term, we’re just trying to capture and retain users so that communities can grow and become new homes for reddit refugees.

    • If the signup process is difficult or confusing, the “capture” part of that goal will be diminished.
    • Small instances have kind of a “ghost town” feel. I understand why this is, but new users don’t and this affects the “retention” part.

    Some ideas:

    • A getting started guide – for example, creating an account on a new/small instance is different than a large instance because of how federation works. The information in the fediverse is largely available, but knowing how to find it is nuanced. Having a guide or video or something could help with this and support the “retention” goal.
    • Having a few preferred instances to handle the surge of new users. This will take efforts from server admins (like yourself) to communicate to the community the needs they have so that we can provide support. This addresses both “capture” and “retain” goals really well in the short term. Some sort of strategy would be needed long term to “decentralize”.
    • I’m donating on patreon to lemmy.ml (just $10/month for now), but I wonder if other users realize that money is needed to handle the new load (for lemmy.ml, beehaw, etc). It seems reasonable to support the people that are crucial to making instances like this work without requiring them to take on enormous financial risk. (I’m saying this without understanding anything about your hosting solution or backend infrastructure, but assume that to scale, you need to pay Amazon/Microsoft/Google mo money).
    • If you need help building out infrastructure, there are those of us here that would be willing to take some time to help as well - we just need a way to know what you need.
    • Some kind of “invite” feature - it would let me send “invite” codes to my friends. This eliminates the “what instance should I use” question and potentially the “manual approval” process. This could potentially be used to create nefarious bot accounts, and may just need to exist initially (but not long term).


  • Long term, I agree – the whole point of the fediverse is to distribute the user base, moderation capacity, etc. Initially though, we’re just trying to make it as easy as possible to for folks to discover lemmy and use it.

    Sending them on a wild goose chase to find an instance and sign up complicates that. Getting them to come back the next day is also way harder when that experience sucks.


  • If folks can sign up on your instance and use it as their gateway to the lemmy fediverse, its tremendously helpful for distributing load.

    The challenge is, letting people know your instance exists, and when they finally do and you get 30 signups per hour, scaling your instance to keep up.

    Long term, you also have to deal with all the sysadmin crap (scaling up/down based on load, security and updates, backups, assholes that DDOS your instance because they don’t like your moderation decisions, copyright take downs, legal requests, etc).




  • Yeah - its tricky. What I’m seeing though is instead of communicating that need at all, lemmy.ml and potentially other instance owners are just trying to push new users to smaller instances.

    I am running my own mastodon instance in my basement - I’ve got other personal projects running in AWS, and work professionally in Azure. It sounds like you’ve got some great cloud experience as well. There seem to be lots of other similarly skilled folks here that can assist with deployment and scale automation (if that’s what they need), or others that could assist by just signing up for $5/month on patreon to cover server costs. That call to action needs to happen though or else people wont do anything.


  • The context I’m using to predict the “big change” is the way the twitter/mastodon migration happened. Musk would do something, and literally within the same hour, thousands of new accounts would be created on mastodon and all those new users were flexing the various features in the platform to find accounts to follow and post content. With that history, I feel like “a drip” of users is unlikely - it’ll be more like waves from multiple tsunamis.

    If things play out the way I’m expecting they will, we really will need instance owners to stand up, ask for the help they need and coordinate those efforts. Otherwise, users like me will just post meandering comments like this one, wondering what we can do to help.








  • I agree – the FOMO thing is real.

    However having a clear path to follow would really help capture/retain those users. In other words “Wow, Lemmy is so popular it can’t handle the traffic – here’s an up-to-date list of instances to try anyway”, or some other option.

    Possibly a pinned post where instance owners can comment info on their own instances without having to wait for someone to update a master list?


  • Mastodon during the twitter migration had a problem where folks were standing up instances, but new users were having trouble finding them, or overwhelmed with lots of “interest-based” instances. Most users don’t care about a “furry” specific instance, or an instance focused on San Francisco Donut enthusiasts. They just want an instance.

    Another problem is as time went on, some of those instances shut down abruptly and wiped out accounts for folks. Having a few centralized and supported instances at first will help solve that problem as well.

    I have the technical know how, given a “GettingStarted.md” file to read to stand up an instance, but I’d rather put my time/energy into supporting an existing one. Dealing with the horde of people currently joining and about to join will require some coordination.