I can't support IPFS anymore.

If money was good, I wouldn't be saying this, but that's not my reality. My money situation is dire, and I need to take action. I've taken down the IPFS cluster and am reclaiming its maintenance cost for improving Futureporn user experience.

This is a last ditch effort to save Futureporn. My debt has become too high and I'm having trouble making interest payments. Something has to change. I have to drastically reduce operating costs.

I don't make this decision lightly. I've been working with IPFS for 5 years and I like parts of it, but overall I have developed little faith in it's durability.

What's wrong with IPFS

The same usability issues came up again and again. "This CID isn't loading", "nobody's providing that CID", "it's spinning forever why won't it load?" The reason? It's mostly a network of leechers, and it's not even their fault-- the design of IPFS does not reward seeders.

My IPFS nodes running on budget VPSs were overloaded for most of their life. Even if they had a CID, they weren't able to provide it fast enough for other people to pin it. This resulted in a terrible user experience.

IPFS is for tech nerds, not for users. It's hard to setup. It's hard to run. It uses a ton of CPU and bandwidth. It doesn't play nicely on home networks and has no built-in throttling or scheduling.

Because it's not designed to run in homes, IPFS has become very centralized, and this made it very censorable. We had multiple issues with public gateways blocking CIDs we created. We had issues with IPFS pinning service providers outright banning us.

It's impossible to run a 5TB+ IPFS node in the cloud on my budget. To be successful here, we would need big CPU, big storage, and big bandwidth-- $1K per month minimum, with an operating cost that grows as the content size increases. What I was working with was bottom of the barrel budget VPS that worked with all sorts of caveats like the biggest disk they offered was 4TB and they would throttle my bandwidth.

Self-hosting was never an option for me since I live in in a rural area with limited internet options. I share my internet with family so I have to restrict my usage.

IPFS is only getting worse. Cloudflare dropped their IPFS public gateways, and so many other companies followed suit. Lightrun, Celestia, Fleek, and Brave browser all dropped IPFS support. Brave reported that less than 0.1% of their customers actually used it.

IPFS's founding visionary Juan Benet moved on to work on other things. His company Protocol Labs then sold IPFS to another company. This was a huge signal for me.

IPFS marketing promised to end 404s. They did that in the worst way possible-- by replacing 404 with 503.

So that's it, Futureporn will no longer use IPFS. It's promising tech, I recognize that. In practice, the incentive structures for people to run their own nodes isn't there. I believe that we are witnessing a format war first-hand. Just like Betamax vs. VHS, LaserDisc vs. DVD, HD-DVD vs. Blu-Ray, we're seeing the battle of IPFS vs. Torrents play out.

Technically there are a lot of IPFS features that are superior technology to Torrents, but again, IPFS is failing to make a network effect like Torrents have, and I don't see any signs that the team behind IPFS is going to change that.

Honestly I hope IPFS proves me wrong. I hope they fix all their issues and make a comeback. But for now, I have to think of Futureporn first, and focus on running it in a way that doesn't send me into bankruptsy. I've been spending a lot of time trying to get IPFS to run nominally and I think that time would better be spent on improving Futureporn the website.

Futureporn's user experience must come first, and then we'll sprinkle on whatever good tech we need to make it robust and long lasting. For now that's multiple CDNs and torrents, all of which run on a fraction of the cost of an IPFS cluster.

What's new with Futureporn

As part of the extreme cost saving measures I'm undergoing, I rebuilt the website from scratch. I am going for minimalism and a site that purrs. I shut down the neglected Strapi server, the NextJS server, Postgres, the experimental networks, clustered VPS, distributed filesystem, and fancy deployment tools.

This time, I'm placing constraints on myself to keep my creativity in check. I've got to focus on the right things-- delivering a performant website that people find useful.

We're now running on a Pocketbase server with Pocketpages for server rendered HTML. There's a backend bullMQ with worker instances that do background task processing. No Kubernetes, no docker, no complicated build system-- I deploy new changes using rsync, and restart services with systemd. We've gone lean and scrappy, we're rebuilding for the gratification of the patron, and the longevity of the site.

Also we have RSS and JSON feeds which I think is pretty cool.

https://futureporn.net/feed/

What's next

Keeping updates small and focused. For the next 30 days I'll be working on the following. My goal is to finish the year strong with the long awaited vibrator support I've been writing about.

Primary objective

1. Vibrator support

Secondary objectives

• Redirecting 404s to updated pages

And then?

The universal constant of change. Futureporn is changing now, and it will no doubt change in the years ahead. There have been a few instances lately where I realized how important it is to record on more than just Chaturbate. I'm envisioning creating an easy to use DVR that people can run on their home computer; a program that records their favorite lewdtuber streams and syncs them with Futureporn. Just an idea, would love to know what you think.

Thank you

Thank you for everyone who has supported this project thus far, and everyone who is continuing to support on this new leg of the journey. I'd love to hear what you think.