Sign into an existing account

@Big_Fat_Dangus has the answer: PO Box with money orders. According to a @paul_neri post, @system made a post saying his costs went from $400 to over $1K yearly (haha, I thought it was way higher than this). I promise $20 a month (paid one time yearly as $240 because fuck sending twelve money orders). Others on this site will do it as well, I am certain.

Also, system, don't consider it "donations" or "begging for money." You should considered it hell for being forced to continue as our hobby wagie cagie, haha!

One Voat commendation for your proposal.

All fine and good til a bad (((actor))) comes along and posts something heinous and illegal, then uses that to go after anyone else "hosting" it, and thus break the whole site.

The first question is whether the current server code has room for efficiency/bandwidth improvements. An easy way to check is compare its performance to other board software for similar loads. A P2P architecture might not be necessary to significantly reduce costs.

In your diagram not everyone is connected to a chad, so there are sub-chads. You'd have to make sure they were reasonably reliable. This is actually kind of what go1dfish did with notabug.

My usual solution to reducing db costs is read and write caching. If the Chads really do reduce the cost and it is more efficient then you can just run your own Chad, in a much more controlled and reliable way, and better clearance of caches when it's appropriate.

What most of these sites that have huge db costs are doing is they are accessing the database directly within the request handler. That's how people would write sites in php so it's the typical hello world everyone sees and then builds more feature on. Then they might do some read caching on top of it.

I've been lucky that I wrote my own caching system a long time ago that's allowed me to build everything cached first. The beauty is that in most systems the database is considered king and then the cache reflects it. In my system the cache is king and the database is a persistence store for it. It makes it so when you design cache first you are guaranteed to get the cache coverage on everything you need when you need it. Is it technically a performance imperfection to be caching things that don't need it? Yes. But it means I never need to solve a major caching problem after the fact only to find out I can't. Everything I build automatically has the chads.

But if system doesn't take you up on the offer I will. I only have a few things programmed right and there is a lot more to get right. If you really do want to work on something and you don't end up working with System, my door is always open.

I do like this idea and want it to work.

There have been promises of pure p2p sites since torrents became a thing, but the drawback has always been the central servers that are needed to bootstrap everything.

The user can't find a movie torrent without going to a central website to get the link, and that makes the central servers easily targeted.

On the flip side, if there is an automated download of all the torrents for every startup, then the user gets bogged down with a long wait to find anything.

I definitely don't have the answer to this problem, as many attempts have been made to make the pure p2p system, and the ones I've tried have had significant barriers for casual users.

Already mentioned is the possible content issue, where an uncensorable system will face serious scrutiny for such content.

One idea that was floated around somewhere a while ago was to have all the files encrypted and spread out around multiple hosts, so that any user/host has no idea what is stored locally, and only serves bits and pieces, and no complete file of anything. No idea how to do that bit either.

I am interested in how the tech will work.

Yeah, sure. But isn't kubernetes a little overkill for 56 users and Paul Neri?

What software does upgoat use?