The Decline of Imgur on Reddit and the Rise of Reddit's Native Image Hosting

There is a Mature filter on imgur already that's human-determined. The problem is that people posting pictures will mark things mature unnecessarily, and there have been trolls who put up hardcore pornography without marking the image as mature. The problem is, as always, trusting humans to do the right thing without massive consequence for doing the wrong thing.

I agree that should be the default but getting the data to do it automatically is hard if you're going to ask users to mark things NSFW on their own accord.

I'm really not sure what point you're making. It's not surprising that Reddit rolled their own image hosting. In general, most major websites have done that, because it's easy to do. If your business relies upon image hosting, why would you trust others to do it?

Anyway, Reddit gets paid via ads and gold, so people are paying for it. They just won't pay for it on it's own, because image hosting on it's own is not really worth much. You could make your own perfectly adequate image host on a raspberry pi. Reddit's value is not in it's image hosting. Nor is Wikipedias or 4chans.

Imgur picked a shit business model, the internet will survive their inevitable passing.

That doesn't make sense to me. You could do this now with S3 (although you might need to write the slick interface); scaling shouldn't break anything, though pricing might need to correspond to how much you upload.

I tend to see this as personally beneficial. Once the patois diverges enough, I stop wasting time attempting to decode it, unless I'm bored enough to treat it as anthropology.

I have to use http://knowyourmeme.com/ more than I care to admit.

Off the cuff comment to indicate that my observation isn't a serious proposition, but just me sharing my general feelings, and acknowledging that my feelings should be considered fairly cynical.

it's the original poster displaying some self-awareness about his rationalization of pointless contempt and constant misery as "healthy cynicism"

FWIW /r/no_sob_story/ is dedicated to highlighting content that is only interesting due to post context

When was the last time you heard someone actually debate about text editors?

Eh, people regularly get around this by splitting their comments in much the same way as on Twitter, dividing them into multiple 1/2

Not only that, but they already have cloud environments that allow them to get all of that data anyway. Android phones come pre-configured to automatically upload every photo you've ever taken to the cloud, Amazon offers unlimited personal photo storage, and Google and Amazon offer the two largest enterprise cloud computing environments available, so they can likely just analyze imgur's logs.

I think the parent's point is that they ultimately remove this content. Imgur basically removes things that are illegal in the USA, and pretty much nothing else.

That's a feature, but it's not a feature most conglomerates want to provide.

Simpler times. Public facing IP address, web server running on my machine and serving from /var/httpd/public_html or some similar directory. People go with 3rd parties nowadays for the slick web interface and that they do the "hard" work of setting up bandwidth caps and security for you.

Back in the day, the bulk of Internet users were on university networks, and they gave everyone access to Unix systems where anything you put in public_html goes up on the university's website under your username.

For example, I went to UTD from 2003 to 2007. This was long after the days of "the bulk of Internet users were on university networks" were over, but the infrastructure was still there. All students had ssh access to a handful of machines, the main one being apache.utdallas.edu.

So I could do this:

    $ scp picture.jpg ajb0##000@apache.utdallas.edu:./public_html

And then anyone would be able to access it at:

    http://www.utdallas.edu/~ajb0##000/picture.jpg

(FYI, at UTD, everyone's username was their initials followed by a six-digit number where usually only the second and third were nonzero... I ##'d that part out because I don't feel like giving out the exact username I had back then... especially since my initials were different)

Edit: And even if you weren't at a university, back in the dialup days it used to be that even with home internet, everyone's computer was directly on the Internet with no NAT or any kind of firewall, so you could just install Apache on your box and serve what you wanted. On paper, most ISPs banned that practice in their TOS, but it was never enforced. The real impediments were that a) dialup connections would usually disconnect if you went idle too long and b) most people had dynamic IP addresses that changed every time you dialed in (for the latter, DynDNS was a godsend). And of course dialup bandwidth was shit, so hosting anything substantial (especially an image!) would slow down your connection (seriously, people would actually buy second phone lines from their telcos and maybe even get a second account with their ISP so as not to tie up their main connection).

Broadband made it easier: even with a NAT router, you could just forward port 80 to your desktop, and you won't have to worry about bandwidth, your connection would stay up 24/7, and even if your IP was technically considered dynamic it would almost never change. What really sunk the concept of hosting your own was a) mobile (good luck running Apache on your phone, and getting port forwarding on a mobile network is impossible) and b) the rise of services like Imgur that let you host shit with no hassle (Facebook killed self-hosted personal home pages for the same reason).

Everyone used to have a public IP back before the days of NAT. Your modem would dial in, and your computer would be assigned a publicly accessible IP address.

In the unix world, iirc, many systems were configured to expose a folder on the user's home directory (www). Which would expose that to something like: www.mit.edu/~<username>

A link to your server. I have servers, at home, on the internet with no NAT, through a business cable ISP.

For me it's quite the opposite. I like browsing the web in the evening on a 5 year old tablet. Ever since imgur started redirecting mobile clients to the full site for all direct image links the experience gets worse and worse. They keep adding more scripts and shit over time so currently when opening an imgur link the tablet freezes for about 4 seconds until it starts showing the loading spinner and then it takes another couple seconds depending on image size until I can finally see a plain old jpeg. Every fucking time. I mean I don't understand the decision to redirect mobile but not desktop at all. Desktops are faster (cpu and connection) and got more screen space to show ads. Why not the other way round?

> Reddit's motivations

It's effectively a tracking "pixel" for what you click, even if you block their javascript/redirect.

It also keeps that data in house which they can sell (if they need to) rather than give all that data away to imgur for free. Though I wonder if the bandwidth costs will break even.

They can also use it in order to craft personalised suggestions.

That link does not go to the image. You have to manually edit it.

Or, for now, use imgur.

No. Dropbox used to have a folder that let you share direct links.

https://www.dropbox.com/help/files-folders/public-folder

It's been retroactively removed from accounts.

Thanks for the great article.

> Except the end-user ISPs are quite content with requesting traffic from the providers; they like the imbalance as it gives them something to negotiate with (that's why inbound is free on cloud - they're happy to fight the imbalance). Why would they start caching your images for free?

Normally that would benefit ISPs (they do have a network of routers so it would naturally help them, since they would not need to send the same thing over and over again, they do pay for traffic to their uplinks).

But, on the other hand ISPs artificially reduce their bandwidth and expect companies like Netflix to pay and do peering with them. Well, this won't help against it, but while it can replace TCP/IP it can also work as an overlay network and you can skip the ISP.

> Also doesn't that mean everyone in between know you're requesting degeneration.jpg? Vs now where they'd only know the domain and maybe could do transfer size analysis? Sort of the opposite direction of TLS everywhere (CloudFlare exempted)?

The protocol itself doesn't leak that information. Remember, there are no addresses to individual machines to people, only addresses for the data. When you request something the routing protocol will route the request to the source of data. Each router remembers the previous hop the request came from. On each hop a router checks if it already caches the data and responds, otherwise forwards request further. So even the source website doesn't know who requested it. I suppose the issue might be that router immediately would know what you requested.

You can mitigate it in two ways:

- encrypt the data and only give key to decrypt it to subset of people who supposed to be able to access it

- encrypt data individually per user, in that case you no longer can rely on caching properties of the network and the source server needs to be aware who requested what (since it is encrypting the data). i.e. you falling back to the way how TCP/IP now operates.

> Also how do you bootstrap it? If I upload some image and it's instantly popular on reddit and generating many Mbps of traffic, don't I have to stick around long enough to get it up and cached? Or would there be free seed services that will get the first several GB out?

Well you're expected to host it all the time. The network won't be caching it forever individual packets have TTL and also routers can purge data if they need to make space for other data that is now popular.

I suggested NDN as a response to that S3 bucket doesn't handle highly popular content well. The NDN properties allow highly popular content being hosted even from places with low bandwidth, because the network will handle the load.

In a way, it also was created by people who contributed to developing Internet and is NSF founded. The head behind it is Van Jacobson[1] who was responsible for implementing congestion control in TCP/IP also creator of popular tools such as traceroute.

Unlike FreeNet it was designated that while it can work as an overlay network it is actually capable to run without TCP/IP so it could even replace it.

Also unlike FreeNet it's not trying to obscure who is hosting the content. Basically you get a prefix (which looks like unix file path) which you start advertising to neighboring routers (like IP is advertised in TCP/IP) so then all requests under it (as long as no cached version is available anywhere) will be forwarded to you.

[1] https://en.wikipedia.org/wiki/Van_Jacobson

Not 140 words, but: https://named-data.net/project/execsummary/

Essentially the big change is that instead of using addresses for servers what if you put addresses on the data itself. That approach gives some properties:

- network can cache the data - no CDNs needed, no slash dot effects etc - free multicasting (a hard problem at scale with TCP/IP) - better handling of slow/lossy network - easy multipath routing

etc.

It's very efficient for things that can potentially have many people interested in. Like streaming popular videos, hosting images (like discussed here imgur). The benefit for these sites is that they would have lower costs to operate because they would require less bandwidth, because anything that's popular will be cached by the network.

Edit: Also: https://en.wikipedia.org/wiki/Named_data_networking

Not really. IPFS from what I understand is a P2P application, with filesystem like functionality.

NDN on the other hand is a network protocol that is capable to run without TCP/IP (it's actually designed to be able to replace it). They actually built a testbed: https://named-data.net/ndn-testbed/

Similarly to TCP/IP you can build application that utilize the protocol. In fact I suspect implementing IPFS in NDN would be much easier than in TCP/IP.

In fact looks like someone did just that: https://named-data.net/publications/techreports/ndn-tr-27-nd...

and source code: https://github.com/named-data/NDNFS

There was one that I recall: The Deck, which shut down recently [1] after more than ten years. Could be seen originally on sites like A List Apart where its ads were small, felt relevant and were of good quality (they had a limited monthly pool of decent brands and artists that used them last time I checked).

I actually liked them, though after switching to uBlock Origin years ago I'd forgotten about the network and to whitelist them, sadly.

[1] http://decknetwork.net/

> Do you think it's possible to create an ethical ad network? It's something I wonder about at least once a month, but I'm not sure if it could compete.

I'm highly skeptical, in that I believe the nature of ads makes this inherently unfeasible.

The point of ads is to be seen by as many people as possible. In dramatic terms, you could say the purpose of ads is to invade your mental space against your will on behalf of product makers. How could that ever be ethical?

That aside, ads also present a conflict of interest for content makers. You want your site to be nice and high-quality. But you need money, so you put up ads and comprise on the aesthetics. You want more people to see your ads so you're tempted to make more content, or arbitrarily divide your content into smaller pieces (fighting the end user's ultimate desire which is to peruse your stuff).

And then there's the whole "biting the hand that feeds" problem. I personally don't believe anyone showing ads can be truly objective regarding them. So maybe Google is evil, maybe not, are you going to turn down their money either way? Most smaller ad networks seem shady to me.

I think this seemingly theoretical problem is why ad blockers are still on the rise, going so far as to be baked into browsers (Brave on mobile is awesome). Not only do people dislike ads, I think they don't trust advertisers, and I think most users nowadays are aware of these ethical conflicts at least on a semi-conscious level.

In short, this is why I've personally made a promise to myself never to rely on ads. Maybe that means I can't get rich off of "killer apps", or even earn a living. But part of me believes that anything useful is worth paying for, so I'm focusing on building something useful.

Edit: Didn't mean for this to turn into a rant. Oh well!

There is projectwonderful, which is serving ethical ads and has a uncomplicated bidding method (bid for a day on a website).

They're explicitly on all my adblocker whitelists since their ads only amount to a PNG image with a link.

The advertisers know where they are advertising, the site owners can control who advertises, the payment is transparent and everyone is happy.

No. See http://www.cluetrain.com/ - paid advertising is inherently unethical on a network where genuine mass interaction is free.

"For me a parasite is a startup being a copy-cat,"

The parasite metaphor does not include the idea of being a "copy-cat"; parasites in the real world do not do the same things as their host. Indeed, that is the entire point of a parasite, in some sense; to take advantage of the host setting up various expensive systems to survive and just using them instead of developing their own.

The characterisation of Imgur as a parasite isn't a value judgement. Imgur isn't doing anything wrong (nor reddit for that matter).

Honestly before RES came out the proportion of images on reddit was very low. Going back further before subreddits I remember it as largely just news articles and discussions, much like HN.

That goes to the post, when I'm sharing an image I don't want it to load Reddit.