Hacker News new | past | comments | ask | show | jobs | submit login

It's really unfortunate that Imgur took such a turn for the worse.

Originally it seemed like Imgur was a benevolent gift to Reddit users... a clean, simple, easy to use image hosting service that everyone could use for free. I guess it was too good to be true, because now it's tricky to link to a direct image (though not impossible), and the standard landing page is filled with unnecessary clutter.

As an aside: if you're looking for an amusing read, the "IgnorantImgur" subreddit[0] is filled with examples of people from the Imgur community who seem to be oblivious to Imgur's origin as an image hosting site for Reddit users.

[0]: http://www.reddit.com/r/IgnorantImgur/top

What really killed them was the GIF explosion. Average image size remains largely static and bandwidth and storage costs steadily decline so if you get your timing exactly right, you can hit that sweet spot where your CPM declines slower than your hosting costs and provide an increasingly profitable business.

However, GIFs completely upended that calculation by increasing the average file size by an order of magnitude and pushed the viability of a cruft free image host further down the road by another few years.

Frankly, I'm surprised Google never just bought imgur and absorbed it under youtube or something. Being the default place on the web to host images seems like something way more strategically valuable to Google than any standalone image host.

Why would Google buy when they can just create one overnight if they wanted? Probably much more stable than Imgur too.

Imgur is just a commodity. Youtube is not.

Imgur has a community. Tons of users that goes directly there to see funny images and GIFs.

Yes, Google can create a infinite scalable image hosting overnight, but creating a community is a lot harder.

Imageshack was in the same position 5 years ago. And Photobucket 3 years before that. Image hosts go through a cycle of upending the tyrant and then becoming the villain.

Ugh. Don’t remind me. I used yfrog (Imageshack’s Twitter-friendly brand) for years, and now I have probably hundreds of old tweets with missing images.


does anyone remember pbase.com for free photo gallery hosting, in around 2002? apparently it still exists with its same basic but functional UI.

From their FAQ:

Is PBase free?


it used to be, a long time ago

Free ones at least.

Creating communities is hard, but Google wants a unified 'Google' community, not multiple disparate ones -- just look at how they handled YouTube.

Even if you throw in some machine learning to figure out if the user's looking at memes or gifs or porn, I can't imagine the marginal value of that data being of use to Google. It's essentially unmonetizable, which is why most image hosts disappear after a few years.

Imgur was successful in avoiding that fate by pivoting to a community, but that community would be of no use to someone like Google.

Especially as Imgur is already giving Google that information for free (from the Imgur source):

    <script type="text/javascript">
        __ga('send', 'pageview');

What makes you think this is free?

And why would Google purchase a forum site for sharing funny images and GIFs?

Why would they purchase a forum site for sharing funny cat videos?

A guess: (and note they kept youtube a wholly owned subsidiary I think) they desperately had to make sure the Viacom lawsuit defense was well funded and aggressively defended or it would severely damage Google's business.

Also, google has now had success building photo sites / photo products, but at the time of the youtube purchase, hadn't had much success building a video product.


Because it's a monopoly. Come back when Imgur becomes the only (or one of very few) place on the web to share funny images and Gifs. Farewell.

Are you trying to imply that Youtube is a monopoly? Aren't there others like vimeo, etc.? Besides, there is nothing from technology perspective to make it a monopoly. All you need is a server rack and a bunch of web developers who can build such a site, right?

Well, there's also massive scaling, which the average web developer isn't experienced with. Then there's creating a consistent experience across various devices and browsers.

So, sure, anyway halfway competent developer could bang out a YouTube (or Facebook, etc) prototype in a few weeks, but going from there to a reliable, usable service at scale is something else altogether.

You're going to need some project managers, SREs, sysadmins, front and back end web developers, programmers who know about video encoding, content experts, lawyers, etc...

Youtube is a monopoly in a way Google is a monopoly. Just like how there's Bing and DuckDuckGo but most people use Google, most people use Youtube although some do use sites like vimeo. If you think it's so easy to build a site that competes with Google or Youtube you should go ahead and build one and overtake them. After all, nowadays everything is open sourced, you can even build your own search engine from scratch overnight.

>Youtube is a monopoly in a way Google is a monopoly.

So not at all unless you change what having a monopoly is.


They did make their own version of youtube at one point, apparently it wasn't enough and they wanted the userbase too.

Does anyone know if it was specifically the community/userbase that enticed Google, or was there some technical aspects to YouTube that Google couldn't or didn't want to build themselves?

Contemporary media coverage suggests they wanted to buy out a competitor which became the dominant video site in a year. This New York Times article [1] from 2006 wrote:

"The acquisition of the privately held YouTube will enable Google to thrive in one area of the Internet where it has so far failed to gain footing. According to Hitwise, which monitors Web traffic, [YouTube] has the lion's share of online video traffic (...) a 46 percent share, MySpace has 23 percent and Google Video has 10 percent."

Meanwhile, this 2006 article from the Economist [2] stated:

"the deal is 'an aggressive, mature move for Google, one that shows that senior management is not too proud or stubborn to see that they can't build everything themselves,' says Henry Blodget, an analyst at Cherry Hill Research. Indeed, Google's Mr Schmidt freely conceded that YouTube is the 'clear winner', especially in creating social networks around its site."

[1] http://www.nytimes.com/2006/10/09/business/09cnd-deal.html

[2] http://www.economist.com/node/8031159

> was there some technical aspects to YouTube that Google couldn't or didn't want to build themselves?

Google had very much the power of building a YouTube-like community themselves, but buying out a competitor is a short-cut since it has the positive side-effect of removing that competition in our business. After buying YouTube, Google didn't have to waste time in building a community around a new product or "converting" the YouTube masses to theirs.

This short-cut way of acquisition is nowadays preferred by large companies rather than risk everything by doing innovation and "real" competition.

If I remember google video was superior in video quality and looser limitations, it was a place where you could commonly find full documentaries while youtube still had a 11 minute video limit. It was actually google acquiring youtube that caused youtube to loosen their limitations and go for HD. I doubt they bought it for any technical reasons, google's infrastructure was already superior.

Google knew that Youtube will turn into the largest video search engine, so they bought it strategically. Otherwise someone else could have swooped in and acquired the largest video search site, which would have left Google vulnerable in many ways.

Summary: userbase/community was what drove the network effect, which made it impossible for Google to catch up, and that's why Google made the decision to purchase Youtube.

Google did build similar functionality. Google Video was a thing for a while. They closed it down a while after acquiring YouTube.

Google had a YouTube competitor at the time. It wasn't very successful.

> Imgur is just a commodity. Youtube is not.

Video hosting + community is a commodity.

Image hosting + community is a commodity.

But most of imgur's audience is not even seeing the site or is just looking at a quick picture and then going back to reddit or some other referring site.

I think you're mistaken here. As a reddit user you might think that, but there's actually a fully functional imgur community that votes on images independently from reddit and comments on them on imgur independent from reddit.

Just scroll down on this page: http://imgur.com/gallery/tWWxi4D

Voting options, and lots and lots of comments.

Parent said "most of", and I think that's pretty much correct regardless of how vibrant of a community Imgur has. Most traffic to imgur is via embeds or links from another discussion forums. The former case obviously provides no exposure to Imgur community, and even in the latter case most people are there only to look at the picture, not to engage as an Imgur community member.

The whole point of this discussion thread is that Imgur is f*ed because they aren't independent enough. Otherwise there's nothing to worry about, right?

wow, that's an awful UI. Reminds me of the gloomy days of 9gag, circa 2010.

A substantial portion of Youtube's audience does that too.

Monopoly is not a commodity.

Non-monopoly is a commodity.

Imgur is exactly as much of a commodity as YouTube. The exception is that they host images instead, and when you consider that the images are silent webm videos a lot of the time, the lines are even more blurry. The value to both communities is just that, their communities.

I can name hundreds of places to find funny images and memes. Heck even buzzfeed satisfies that demand. As does google images

I wanna see the entire game 7 of the nba finals online. Show me where besides YouTube.

I wanna see the Red Wedding scene again. Show me where besides YouTube.

I wanna find a quick entertaining video to show my kid. Show me where besides YouTube.

I wanna find a compilation of the longest homeruns in baseball history. Show me where besides YouTube.

Heck most people try to ignore YouTube comments. It's not really a community at all. The actual content is the value prop.

Agreed; whilst there are some subcommunities that appear unscathed, YouTube comments are typically a cesspool of hurtful spew from the bottom feeders of human society.

> Imgur is exactly as much of a commodity as YouTube

Youtube isn't a commodity. It's the place that even your aunt, mother, grandmother, uncle will visit and knows to visit for entertainment or music. People who would never have a reason to visit imgur.

Commodity is when your product has no differentiation on the market. Imgur does not have enough differentiation on the market and that's why they're a commodity. For image centric community you can go to Reddit, Tumblr, or any kind of online forums really. Youtube's differentiation is that it has created significant network effect around public video hosting. Can you think of any meaningful competition other than Vimeo? (Or maybe a couple of others but they're nowhere as close as all the online forums online)

Buying is cheaper than building in many, many cases, particularly when it requires a userbase.

> I'm surprised Google never just bought imgur

Sex. Corporations tend to keep faaar away from everything related to porn, and a fair share of imgur images is nsfw content. Cleaning it up would be more trouble than it's worth.

I take it you're not familiar with Bing's #1 use case and specialty.

I'm not. Do tell!

Bing's video search has a lot of porn sites indexed and provides in-line preview of said videos.

It is by many considered better than purpose built porn search sites, such as pornmd.

Pron. Bing is better at pron.

> Corporations tend to keep faaar away from everything related to porn

Yahoo's purchase of Tumblr and many bids for SnapChat seem to be evidence to the contrary. Cable companies and premium channels also make a lot off of porn.

Porn is fine for large companies as long as they have some cover (the product isn't just porn).

Some nudie pics or vanilla porn would probably be the least of their worries. I can imagine that a service like imgur spends the bulk of their time cleaning up the nastier bits that inevitably creep their way into such a large site. How many companies really want to inherit a high maintenance mess for such little return? At least with Tumblr and Snapchat you have a strong brand. How many average Internet users outside of Reddit (or the 18-34 male demographic) have even heard of imgur?

Some huge name corporations are very involved in the porn industry. Clearly, that fact is not well known.

What makes you think imgur is dead? They have a thriving community independent of Reddit. Some might even argue it's a better community.

Some might argue that, but it doesn't make imgur's business model any more sustainable unfortunately.

I was thinking it'd make more sense for Dropbox to buy imgur, or offer a similar service: https://news.ycombinator.com/item?id=11906573

>> What really killed them was the GIF explosion.

No kidding. The whole front page is all animated GIFs (or .gifv/mp4/whatever

Eh, they're serving GIFs up as video and low-res, short compressed videos aren't actually all that big these days.

It probably is still larger than normal non-animated images

I've come across animated GIFs that when downloaded in GIF form are 100+MB. That's slightly different than a 100KB jpeg.

Yes but you are supposedly not serving the GIF but the mp4 video.

A premium model for gif hosting might have been an alternative.

As it is:

1. I hate gifs (with very few exceptions).

2. I lack a decent place to host images associated with Reddit.

3. Subscription-based services are a risk on several fronts.

But imgur can control the maximum sized gifs allowed. And they even adopted webms to reduce file size massively.

Imgur started off as a benevolent gift to Reddit users, but then as more and more people started to use it, the expenses became too great to keep it the way it was. There is no money in serving images directly with no advertising, but there is certainly costs. What were they supposed to do?

The only clean route would've been for them to be acquired by Reddit.



Absoutely. See also: Waffleimages after it.

This is cyclical, an image site gains critical mass, builds in size and then buckles under the weight of providing a free, clean, high-bandwidth service.

Imgur was originally offered as a "gift" to Digg users, and when it didn't gain traction there MrGrimm re-branded it as a "gift" to Redditors.

It was posted to both Digg and Reddit on the same day. The Digg post never gained any traction and when the server went down under the load MrGrimm deleted the reddit post (but not the digg post, since it gained no traction).

He made a second Reddit post the next day.

Source: https://www.reddit.com/r/IAmA/comments/9tlwi/im_the_imgur_gu...

Citation? My recollection is that imgur didn't show up until after the Digg exodus and happened on Reddit first, but maybe I just didn't see the Digg side of it.


> I posted it on Reddit, Digg, and a couple forums that I frequent at the same time.

Imgur was never a reddit exclusive thing. It was just mentioned to advertise and gain loyalty from reddit users.

Yeah, unless he was just embellishing when he posted this: https://www.reddit.com/r/reddit.com/comments/7zlyd/my_gift_t...

Which, honestly isn't a bad thing if he did.

In 2014, I wrote a blog post (http://minimaxir.com/2014/02/moved-temporarily/) calling out Imgur's shady redirect practices on Facebook/Twitter, with a warning that things will get chaotic if they try that tactic on Reddit. Guess I was right?

Notably, when that post was submitted to Reddit, it got eviscerated, nothing that "hey, they provide a service for free, they deserve ad revenue!"

This was the reason why I created http://img42.com. They bombard you with ads, and other "features". I wanted something where I could just click, drag and drop, or pate and it would be uploaded.

> I got fed up with all the other image hosts out there so I made my own. It doesn't force you to compress your images, and it has neat things like crop, resize, rotate, and compression from 10-100. It's my gift to you. Let's not see anymore imageshack/photobucket around here ;) I'll be listening if anyone has some suggestions.

EDIT: The server was moved off of shared hosting after about 4 hours of release. It's now on a dedicated server with a 100mb port.

EDIT2: This is an old post and it's no longer on just one 1 dedicated server. It's on many, and utilizes a CDN provided by Voxel.

This was the original post [0] announcing the creation of imgur 7 years ago. Maybe in 7 years from now you have your own img42 image hosting emporium! :)


[0] https://www.reddit.com/r/reddit.com/comments/7zlyd/my_gift_t...

Haha, these posts do look very similar.

I have been hosting img42 off a $5 a mo/ droplet :P. It has served XXm+ impressions, and 1m uploads since its creation. I also have been sustaining a ~3m a day DDOS from Russia for about a year now.

"easy temporary image hosting" - How long do images last before being removed?

It was actually that drawback that lead me to making/using my own image host! With my own host, files that were uploaded back in 2012 when it first started working are still there - with the same URL (despite URL scheme changing since then)

You can change the image lifespan to forever. The site has been around since 2012, and images are still up.

Awesome, also interesting we both bailed on imgur at the same time. It must have been around then it got unbearable!

Yeah they transitioned into becoming a social network... I cant believe they were able to maintain traffic after adding so much friction -_-

Neat service. How do you plan to offset the hosting, storage, and bandwidth costs though?

Costs haven't been that bad, I have floated it since 2012 (only around 1M images have been uploaded).

But I have some scalable plans if it were to take off without breaking the bank.

What an awesome UI, I love it. Thanks for making something so cool.

I have long overdue v2v2 (killed the original v2) update for the UI. I'll try to post it on HN when I complete it. The UI you see now was done in 2012.

Same thing with ImageShack. They use to host images for SA and then got really shady with linking images.

Hosting images is a high cost service that most people don't want to pay for, so there's very little ways to make money.

What is SA? Something Awful?

Ah oops. Yes.

Well they took a bunch of VC money so they could have a nice office in the middle of San Francisco and presumably so the founder could cash in a little bit. You can't blame him for his choice to take VC money, but this was inevitable.

EDIT: It's actually worse if you read their wiki [1]. They were profitable enough early on to have 10 employees AND move to San Francisco, going as far to win a "best bootstrapped start-up award". THEN they took 40 million in funding.


> a clean, simple, easy to use image hosting service that everyone could use for free

How do you monetize such a site without bloating into the current Imgur?

It seems like the only way to have a "clean, simple, easy" image hosting service is to be a billionaire and to do it out of the goodness of your heart.

Imgur has been trying to do more "native" ads. Example:


That's what the lore tells us, but maybe it's just plain wrong and the old incumbents were just greedy and terrible. I think imgur was bootstrapped for a long time and didn't have much in the way of monetization, certainly hotlinking still worked then.

The changes came when imgur took VC money.

imgur worked because a CDN named voxel could readily use the bandwidth bump provided by hosting images to decrease their overall network costs.

Voxel founder here.

That doesn't really make sense, as others have pointed out.

Imgur was indeed a valued customer, and we grew together and had a great relationship with them.

Customers like them helped us achieve the necessary scale to build out our network.

But pushing more bits costs more, not less ;)

Hi! I'll admit it's poorly worded. My intention was to say there was an economic value in providing less-than-cost (or even free -- only you and your accountants knew) bandwidth in order to scale out your network and increase value to customers paying much more.

though not as often today as in times past, there are points where 10Gbps is more expensive overall than 20 or 30Gbps.

Hi. You're absolutely right.

Getting high traffic customers like Imgur allowed us to negotiate better commits, establish new peering and interconnection agreements, and otherwise lower the "unit costs" of the network we built.

It was generally never worth it to do this for free, but you're spot on that in certain cases it was worth it to do it for less than (current) per unit costs.

Oh wow you guys re-branded. You could do with a landing page knocking about somewhere mentioning your old name! I was trying to dig through the web fairly recently trying to find you (Knowing you'd survived hosting imgur for at least a time!)

Ended up with KeyCDN for now

We got acquired by Internap. They killed the brand. I would have loved for it to live on but it was their decision ;)

please explain, doesn't make sense

I guess traffic gets a lot cheaper beyond a certain volume. I suspect it's dollars/TB to cents/TB? So they can cut prices for their offerings and have more revenue and growth due to paying less for the traffic. Having these volumes of data also allows them to approach potential clients with some solid data (e.g. we serve 1/5 of image clicks on the net every day and below 20ms response time... I have no data, just speculation)

I could imagine that imgur traffic paid for itself.

So I should start a burger joint and give away food for free to prove I can handle the throughout and my cost will be essentially zero?

If you have robots in the entire supply chain down to the production of the fertiliser for the plants without any humans then why the hell not?

Burgers are not bits.

how does that work?

I send 5Gbps to an eyeball network (let's say Comcast, for the sake of discussion) it might cost me $5000/month to do that through my transit provider (intentionally inflated cost). Comcast will interconnect, settlement free, to networks who exchange 7 or more Gbps. Three optics and three cross connects are $3000 + $600/month.

How much is that additional 2Gbps of traffic (to Comcast) worth?

Maybe the only people willing to nurse along a free image host are the ones hoping to somehow, in some way, get rich off of it.

Yeah, you either have to convince people to pay to upload or use it to feed into some other money-making enterprise (like serving ads).

We really need a better way to pay the people who create and serve the content we all love to consume.

Limit gif uploads and don't insist on viral growth. Pretty simple.

There was also a muddy middle when Reddit's management was going through various messy leadership dramas.

Reddit gave Imgur an almost impossibly wide berth compared to other "made for Reddit" sites and allowed the owner to get away with all kinds of site policy violations. Imgur also set up various restrictions on content that were conveniently waived for Reddit images.

It turns out it was because Reddit took an investment stake in the service and didn't make it known for a long while.

Then Imgur decided to just become Reddit because, hey after all, they're responsible for some huge percentage of all traffic for the site -- this being entirely contrary to the communication the site put out about what they were and were going to be doing.

Now Reddit is just wholesale replacing it with an internal service -- which is what should have happened in the first place. It's been obvious for years that this should have been a core service, but the terrible previous management of the place had pretty much frozen all improvements to the site.

I think this move speaks more to the sane new management team doing the correct things than anything else. I'm really happy with the job that Steve Huffman et al are doing with the site, the decisions seem reasonable, well intentioned and generally in a good direction.

I was wondering about what you said about linking to a direct image being tricky.

As far as I can see, you can put the link to imgur (like http://imgur.com/gKt0Owa) or directly to the image (http://i.imgur.com/gKt0Owa.jpg) and both (as well as the html, bbcode and markdown versions) appear directly in the image page.

It's tricky because your direct image link redirects me to the full page with ads, and it's not obvious to you. Try hard refreshing your image link.

(I'm pretty sure the way it works is: it shows the image for certain referrers, and when you load the "full" page it requests the image with a referrer of imgur. That way you get the image, and then subsequent requests to the image are served from your browser cache so the direct links work)

Also variably checks http accept headers (<img src> produces accept of image/* iirc)

Weird... I actually tried with a couple of different browsers (FF, Chrome and Safari) to check that it wasn't me logged in or something like that. Could it be ip-based, as I did all my tests with the same computer? Hmm...

Anyway, thanks for expanding on it; I've shared a few images through imgur and never had this issue (maybe it was before this change, or maybe the people receiving them never complained), so I was a bit puzzled by your comment.

I've been able to reproduce it with different browsers on the same computer. I'm not sure of the exact strategy, but it's just a big UX problem that I don't know what I'm sending another user.

... aaaaand now I want to make a malicious image host that always shows you your uploaded images but serves up goatse to anyone you link it to. Or maybe just flips the images, upsidedownternet style.

Imgur was the epitome of setting user expectations way too high. The service they offered was fantastic but costly, but users came to expect they should get it for free without clutter and monetisation wrapped around it.

It's a reinforcement of how important setting expectations with customers early is very important to avoid stepping on their toes later on.

Every single popular free image host goes through that same cycle.

The writing was always on the wall. They needed to start their own community because they were essentially a commodity product that could easily be replaced.

Completely expected as soon as they took money. They are exactly what organic "success" stories online look like. Now they want some of reddit's traffic and so reddit is fighting back at first by embedding images directly on the page and now hosting the image themselves. Reddit is also looking to maximize profit.

Yes. I also think that part of reddit's unspoken strategy with this is to get more people using their mobile apps instead of the high-quality third-party alternatives (which allows them to control the reddit experience more tightly, which they want to try to rehabilitate their image away from 4chanLite, and also allows them to get the ad views instead of giving them to third-party devs).

reddit is trying to squeeze back some of the juice from the third-party devs who use their API without being too overt about it, because they know the community would revolt. Twitter did the same thing much more loudly when they realized that they were giving away the farm by letting most of their userbase access the content via third-party apps.

This works because imgur and other image hosts are already well-integrated into third-party reddit apps, but reddit's image hosting isn't yet, and they can change things around/break it repeatedly until users get annoyed and just move to the official app to try to get a more seamless experience.

If this is their plan they aren't doing a very good job at executing it. Their mobile app is honestly outclassed by others. And their mobile web browsing experience is unpleasant to say the least. Each page takes a very long load time to load, despite it being a single page web app, and the load times cannot really be explained by what it is downloading, as other apps, and the desktop site, can load the same amount of data very quickly.

If the plan of trying to monetize mobile fails, I wonder who, or if, someone will just try to buy them out and try to do it better themselves. I am convenience that the reddit brand is very popular. The other day I started compiling something, and reflectively I alt tabbed and went to reddit, after realizing I was in the middle of doing something I closed it and got back to work, but I went their out of mussel memory - and I think that just goes to show how valuable their brand is.

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact