It's the good old problem of system centered design.
All viral social media products the past decades have all been built around human centered design: Looking at how to better solve the problem of helping people find the content they're most likely to engage with, all whilst making the process fun, enjoyable and addictive based on human habits and reward mechanisms (the dopamine you get from getting a like, or from feeling entertained from landing on a fun TikTok video that results in you keeping on clicking what's essentially a slot machine.)
Meanwhile, Hive (and Steem before it) was created by techies with an idea of a Reddit like platform and a complex reward system behind it. However, that came with a lot of assumptions about the platform being used just like reddit but with a reward mechanism working in the background to trend good content, reward early curators as well as the best creators. But this comes with the same failure that always happens when you assume all else equal, because the mere introduction of rewards based on votes changes how people vote (as they care more about getting high rewards at minimum effort than elevating good content). Thus making the trending, hot, and community pages terrible for content discovery, and engagement goes down the drain with it. Now, some might say that the platform offers a solution in the form of downvotes, allowing the community to self-correct by voting down over-rated content or that which is simply auto-voted. However, this then makes the platform dependent on people putting in work and effort that is not enjoyable, opposite to the structure of successful social platforms.
A proper social platform would be built the other way around: Understanding first what humans want to do, and then building a system to enable, support and maximize that. Imo that can only be achieved by completely disregarding the reward pool idea that currently exist, and instead focus on Hive being a platform where users really own and control their own accounts, where the content is stored decentralized, and where there is a fast and fee-less token designed to work well on the plaltform. Sites like Peakd.com could still choose to reward users in this currency, and we could create demand for it that way.
But most importantly: We need to build a social platform centred around user experience where our product adapts to the user, rather than expecting users to adapt to the product (which never works.)
What changes do you think we could make?
I think first, just make it really easy for people to join. That's it. Worry about the other things later. Most people don't speak Blockchain and DPOS. The complex reward system shouldn't be the selling point. KISS (Keep It Simple Stupid). Reddit doesn't reward anyone anything and yet everyone uses the platform.
Why is that? They've gotta be doing something right. Why not reverse engineer their success and somehow get it to work here. It's not about how great a platform Hive is, we know it is but why aren't regular people using it?
Have you tried to sign up a friend using https://hiveonboard.com/create-account ?
First a shift in mentality and approach:
To begin understanding and analysing what users want, and how to build a compatible and sustainable token economics and blockchain around that. Rather than being stuck working on a fixed system and trying (helplessly) to make it function.
One start would be for sites like peakd.com to ignore stake in how it filters content. And instead use user data on how they engage to present content they're most likely to want to consume and engage with.
Yes the frontends have to develop their own content algorithms which are unrelated to stake. That and lite accounts. And the focus being on owning direct relationships with your own followers (rather than via a corporate intermediary who can block access to your followers at any time) so you essentially own your own reach.
@intrepidphotos @fredrikaa That would be great. we would also consider using API from other 3rd parties that offer really good solutions for this.
How cool would it be to have several open source independent algorithms that the you could select on the frontends just like you select your node.
Sounds like pretty good ideas that can be implemented by frontends to help content discovery (and I guess community discovery - if I find a cool piece of content, I might enjoy following the author but also joining the community where it was posted).