I feel like WeTransfer has got this right. They have wonderful full page ads and are monetizing by selling premium accounts to corporates, which have the ability to change the ads to fit their branding. https://wetransfer.com

WeTransfers user experience is great. I'm not sure the economics are ultimately going to work for them; the funding [1] gives them some runway, but they are already transitioning to more of a cloud storage provider. Simple utilities usually add bloat to drive value for premium plans. [2] In my experience, as they approach Google Drive and Dropbox in features, the product will bloat and become less functional and useful.

[1]: http://www.techtimes.com/articles/33710/20150220/wetransfer-...

[2]: https://wetransfer.com/plus/

Agreed. One of the few websites where I actually enjoy the advertisements and even click on them regularly.

I also ran some file/image hosting sites in early 2000s and another reason not to do them is that people post the most vile content and moderating can cause ptsd.

How did you deal with this content?

Basically by not running another file sharing knockoff.

This is the second, possibly third, time I've seen you link to your post about public file hosting - and this is the second, possibly third, time I'll agree entirely.

There are tons of pomf.se clones - many that have been going on for years as well. The important part for many of these hosts is that (a) enough people donate to keep it online and/or (b) the community remains small enough to remain affordable for the dev. Personally I use a combination of safe.moe and mixtape.moe.

Often times I've noticed these are "labor of love" projects where a dev can work on learning how to (a) keep a server running and (b) upgrade services without any/significant downtime, etc. Standard DevOps things. The monthly cost to these sort of people is seen as a "learning cost" in a real environment where they have hundreds of, if not thousands or more, users depending on them to not mess things up.

>I hate that if I look at a forum post written >5 years ago, odds are all of the images will be broken.

Inactivity pruning is largely to blame for this, not just new websites going under. IIRC, even Imgur does pruning after 6 months of inactivity (no views) which is totally possible on niche technical forum posts.

I think that in the end, the way to really address these issues is with WebTorrent or a similar service. When you're online, you should be:

    1. contributing your own bandwidth to the images you upload

    2. contribute your own bandwidth to the images you look at.

If you look at an image, maybe you should be required to seed it twice over. If you upload your image, maybe you have to seed it at least 10 times for it to stay up past a certain deadline.

As a user of image websites and, well, websites in general, I'd gladly contribute my bandwidth to help the services run. Or even act as a mirror rather than a shared peer.

What do you think about that? It could lessen the load on the image host and help scale things. I just can't seen this happening if you embed an image directly unless you embed iframes or require users to go to the site itself.

Although the idea is great but I'm not comfortable with the legal implications of "seeding" the stuff I view. It might work in some legislations but I'd be putting myself in too much risk under my local law.

What if everyone "seeded" random content that others are browsing? Sort of like a Tor Relay for P2P traffic. Everyone assumes some small responsibility and yet could argue plausible deniability.

I believe Freenet and IPFS are both built on this idea.

Upload metering is happening again in Australia. Especially on mobile devices, this could get super expensive.

I liked everything you said.

I'd question the assumption that all having web pages be immortal is a good thing. There is great value in letting natural processes erode items or things that don't get maintained.

Throughout history, we've learned to write down and preserve important stuff. Sometimes we've gotten it wrong or we create huge single points of failure (see, e.g., Library of Alexandria), but we should be careful in assuming that preserving everything is inherently better than the historical approach.

This was an interesting read and we will do our best to avoid this fate.

Right now we're focusing on patching bugs and providing a stable service, but in the future a premium service will be created which targets heavy users.

PS You can change the status of the last active host in your list (Minus.com) to shut down as well.

I've been running a public file hosting site for 11 years, I've always treated it as a side project to use for playing with technology.

The primary reason it's still alive today is out of obligation to the URLs (2 million or so) that call it home. If you don't already find this fun and don't feel the same obligation, then you'll join the thousands of others littering the web with 404s.

Also, be aware that people will upload bad shit, and you can look forward to phone calls from the FBI and others.

The site is open source, maybe you'll find it useful: https://github.com/kudos/hostr

The problem is one of capabilities; unrestricted image hosts delegate too much power to uploaders. An uploader should not be able to create a valid URL with any data under the hoster's domain.

In addition to what's been mentioned, another possibility is to force users to bring their own domains to the table when doing hosting; the user is free to host things from the service at their own domain, and the hoster's domains remain distinct and untainted.

"Currently I run a file hosting service with controlled growth."

Why do you do this ... that is, run a file hosting service for 130 people ? Is it a friends and family IT enabler sort of thing that you do to be nice ?

Or is there some kind of business here that I can't envision ?

Yep, it's a friends and family thing. Often friends of friends. Anyone who gets to know me a little bit can get an account. Since growth is controlled, costs are low. Since users are expected to donate, that completely stops being a concern.

The world is a smaller place than it seems. I've noticed this site having a measurable impact on the web around me, and lots of files that might have disappeared are sticking around now (23,946 of them, in fact).

All users are expected to donate. Ok, but how can this work better than just selling storage space like S3, Dropbox or whatever?

>I hate that if I look at a forum post written >5 years ago, odds are all of the images will be broken.

Could be interesting to sell these placements to advertisers: if you don't donate, we'll change all your linked images to image ads after 12 months.

That's stretching 'donate' a bit...

Sounds like you could use P2P caching to offset traffic costs.