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

Awesome example of a company that doesn't try to be everything to everyone and just focuses on the one thing they do very well.

Just please don't try and emulate their success in your startup, bc it likely won't work. These guys won the side-project lottery so an amazing kudos to them.




> These guys won the side-project lottery

This is far from the truth. There was a clear demand for free image hosting that is easy to work with, and imgur fit the bill.


I remember the post that launched imgur.

It really was a scratch your own itch kind of thing. So many redditors complained about the crappy image sharing sites...lots of them were verboten to use on reddit because they were so spammy, or slow, or whatever. There was no clear leader in the image sharing space. Facebook was impossible (because it wants to own every user and won't let you show someone something without logging in), even if it were an acceptable place to upload weird random pictures (I know many people who do, but they just don't know about reddit and imgur yet). So...imgur's founder saw a very clear need, and humbly pointed redditors to it.

It took off because it was amazing in its simplicity and sincerity. imgur's founder clearly wanted you to share funny pictures in imgur, even if he couldn't figure out how to make it pay him money for every image. He just wanted to solve a very real problem that caused millions of people tiny amounts of pain. (When picking a problem, pick one that causes either a few people a lot of pain, or one that causes a lot of people a little bit of pain. You're unlikely to find a problem that causes a lot of people a lot of pain, since those are already usually very competitive.)


I also remember the post that originated imgur... because I was addicted to reddit back then. The founder called imgur a "gift to reddit"... http://www.reddit.com/r/reddit.com/comments/7zlyd/my_gift_to...

Like most redditors I was shocked when I couldn't use my karma to buy stake in imgur. After becoming depressed that imgur was not really a "gift", I quit reddit, and decided to found a startup. And now I'm here. So thank you for that imgur.


> Like most redditors I was shocked when I couldn't use my karma to buy stake in imgur. After becoming depressed that imgur was not really a "gift", I quit reddit, and decided to found a startup. And now I'm here. So thank you for that imgur.

I'd be surprised if many readers know what the hell you're talking about.


> Like most redditors I was shocked when I couldn't use my karma to buy stake in imgur. After becoming depressed that imgur was not really a "gift", I quit reddit, and decided to found a startup. And now I'm here. So thank you for that imgur.

wat


People want things for free. News at eleven.


And yet, 'MrGrim' is still using reddit, namely this nice post last week: "I'm finally going to build the computer of my dreams" http://www.reddit.com/r/buildapc/comments/1mjbvl/build_ready...

Nice.


It's almost as if reddit karma is totally meaningless!


> It really was a scratch your own itch kind of thing.

I agree. They make it look lucky because they executed so well. Good entrepreneurs have a habit of making it look easy to outsiders (Zuck was lucky because he was at Harvard and had the right connections, etc.)

There is a new problem of this type at reddit and other sites if somebody wants a project idea to work on. The problem is animated GIF's. At the moment most animated GIF's are too large, not optimized and the sites are slow.

This has been a constant long-term complain on some of the sports subs that I frequent that post highlights from games. You press the button to display the animated GIF inline and it loads a couple of frames a second and takes 10-30x the GIF length to load. There are two factors here:

1. A lot of people don't know how to optimize animated GIF's. You don't need 24 or 30 frames per second where 8 will suffice, and you don't need full image quality on mobiles or tablets, or even on desktops. You can cut down the frame rate automatically and scale up the compression for some clients to load faster.

2. A lot of the regular hosts are overloaded and slow.

A service idea for animated GIF's would be to specify a video to cut from then have the service automatically create and load desktop, mobile and tablet versions with low frame rates and hosted from a decent CDN. The regular posters are now using Google+ since it loads, but they haven't solved the optimization part.

You don't have to spend a lot of time online to see and identify these pain points, you just have to be in the right mindset where you notice them and can identify the opportunity.

edit: If anybody wants to do this, email me (in profile) and i'll send you my notes. It is something that I considered putting together but I don't have the time. I did do some research though (i'm a mod at some sports subs). If you serve optimized versions for different devices then that would be a good reason to get submissions linking to an HTML page rather than hot-linking to an image.

Pipeline would be:

    upload video -> convert to 3 optimized version -> visitor lands on your HTML page -> serve version based on UA
running these services isn't as expensive as what it used to be, CND's are now cheap. imgur is profitable with 12 staff and running on AWS, with no ads.


No idea if it's technically possible, although it really feels like it should be, but for me the biggest single improvement a site could do to animated gifs would be to allow at least basic video controls - particularly pause or step backwards.

I hate watching a 30 second animated gif where the fancy skill or whatever that it's trying to show is about 2 seconds long and I have to keep looping round and round again to try to catch the critical moment.

Maybe it's already out there somewhere, but I've not seen it.



404 Not Found


A lot of people got all excited when sites like YouTube and Vimeo added support for HTML5 video, believing the <video> element was The Future Of Special Sites You Visit To Watch Videos. Given that YouTube still uses Flash exclusively for a whole lot of content, I'm not sure that's going to work out.

On the other hand, I believe the <video> element could be The Animated GIF Killer. GIF compression is terrible especially for video, and I'm sure with some tweaking of compression options you could make a .mp4 video with higher quality and a fraction of the size of the equivalent .gif. If there were some site like Imgur that let you upload some small video fragment, let you do the appropriate timeline snips, and automatically created a nice small .mp4 and a fallback .gif, that would be truly excellent.

Give it a maximum video length of 10 or 20 seconds and pre-load the site with a bunch of the more popular reaction GIFs like the Picard Facepalm and Orson Wells clapping, and you should be good to go.

EDIT: Looks like https://mediacru.sh/ does at least the GIF->MP4 encoding, although it misses imgur's handy "here's how to use this on your site" examples.


> and I'm sure with some tweaking of compression options you could make a .mp4 video with higher quality and a fraction of the size of the equivalent .gif.

No tweaking needed. Test it yourself: Download any of those huge gifs and encode it to h.264 with ffmpeg defaults. With no perceptive quality loss it's immediately 60-90% smaller. And then h.265 comes along soon.


Except part of the reason GIFs are popular is that everything at least sort-of supports them, even Internet Explorer 4.


I expect that will be a common belief long after it ceases to be a true belief. There's still people that avoid PNGs because "they don't work properly in Internet Explorer" but most vaguely web-savvy people wouldn't bat an eyelid at a PNG today. I want the same thing to happen for HTML5 video: the latest version of every major browser supports it, including Internet Explorer, and including mobile browsers; the biggest limitation I can think of is forum software that doesn't yet have <video> on its whitelist, but time (and a bit of targeted complaining) should solve that too.


Vine and its competitors offer this kind of HTML5 video fragments.


And you constantly see posts on reddit where the Vine video was turned into a gif, uploaded to ingur, and shared.


I never understood that. You'd think that reddit users are savvy enough to adopt html5 video. RES could just as readily inline video as they do animated GIFs, and we would all save a hell of a lot of bandwidth.

The advantages of animated GIFs made sense pre-video tag.


It's probably more of a perception thing than anything else. An animated GIF feels like a lot less overhead than a full on video, even if the file itself is smaller as a video.

Also user forums like reddit are a lot more likely to permit images in posts than a video tag, so you get the added portability of being able to post it almost everywhere.


Vine videos only reliably load for me on Chrome. They're almost totally broken on every other browser.


In case anyone is wondering; this guide does wonders with GIF size reduction: http://blog.room208.org/post/48793543478

It helped me to go from 6 MB to 100 KB for my specific type of video to animated GIF conversion.


So it's funny you outlined exactly what I created 2 years ago. I have all the code for uploading videos or gifs, running them through a combination of Imagemagick and ffmpeg to generate mp4/webm videos and image spritesheets for mobile devices. The only hitch now is probably the big cost of running these libraries on elastic computing services. ffmpeg is a beast, and it'll probably cost quite a bit of money to handle all the transcoding.


It's interesting that you say that. Not to long ago I built something kinda like that for gifs because eof exactly what you said: it wasn't easy enough to make them. So, I built http://gifmachine.xwl.me to scratch that itch. It works using youtube videos as input, so you only have to have a url to get started.


A friend of mine is working on http://www.giflike.com/

It's "like animated gifs", but implemented with JPEGs + javascript. JPEG encoding is much more efficient for photos, and doing the animation in javascript enables fun features such as reversing and scrolling through frames in the browser.


So, he is reinventing MJPEG, in Javascript...

I can't find the proper words to comment this.


How about "Those who cannot remember the past are condemned to repeat it"? :)


There was a codec in the mid-90s that porn sites and some illegal warez groups used that was essentially this. At the time your options were either AVI or a very expensive commercial solution, so someone came up with splitting video files into JPEG's and streaming them over the web and using javascript to slice them back together.

I remember watching star wars on one of the earliest web streaming sites - the image was around 150x150px and would refresh once per second. It was still the most awesome thing I had ever seen at the time, though.


It was probably multipart/x-mixed-replace, rather than javascript.


Imgur has ads unless you're a pro member.


Just want to point out that when hotlinking to an image you can still serve seperate versions depending on the user agent.


Pretty sure imgur doesn't run on aws and does have ads.


> Facebook was impossible (because it wants to own every user and won't let you show someone something without logging in)

Well that's not true. Just get the image url and you're good to go.


Facebook denies images based on their privacy controls even if you are viewing a .jpg URL. People try to share those links often on a forum on which I'm a member, and when you view the link not logged in to FB or not connected to the proper proper 'friend', you receive a 'this content is not currently available' message.


I made the very first imgur upload plugin: https://addons.mozilla.org/En-us/firefox/addon/imgur-uploade... I like to think it helped imgur growing bigger. Too bad this was the time reddit gold was not a thing yet.


> I remember the post that launched imgur.

So do I.

It's kind of funny to see some of the things that have been born out of reddit or other social communities, having read the posts that resulted in them.

On that note, I wonder how the history of project origins is going to change, in a sentimental way. We hear folklore of person X talking to person Y at dinner, which birthed projected Z. Now we can link to where MrGrim first promoted one his project.

For the interested, he did an AMA on reddit[1].

[1] http://www.reddit.com/r/IAmA/comments/y81ju/i_created_imgur_...


Well, they are either very competetive or extremely hard to solve. Think cancer.


Well, they are either very competetive or extremely hard to solve. Think cancer.

Have you seen the price of cancer drugs?


Others tried and largely failed. Min.us, photobucket, ... were all OK at 10 or 20 viewers, but when you hit the front page, the top-rated post was an Imgur mirror that [usually] responded to direct .png/.jpg links without fault.


> There was a clear demand for free image hosting that is easy to work with, and imgur fit the bill.

There are plenty of these around, so I doubt it was being free and simple that made them "win" (we have an ad-free one with even simpler UI, for example). There were 2 factors that contributed most to their success in my opinion: a) being used much on Reddit, b) being awesome at implementing many useful, well-designed features. Without a), they would probably still have won this race, but at a much slower pace.


On the contrary:

- low latency (images served instantly)

- no ads (initially)

- upload, get link

Initially there were zero features. The pace was entirely based on serving images fast. Everything else you think of imgur came later.


That doesn't mean it wasn't a side project. I'm not sure, but I'll bet Alan Schaaf got pretty deep into Imgur before he quit his day job for it.


I've met Alan a few times - he told me he was in college when he started it, and the original imgur was run off of a server in his dorm room, and he later moved to a CDN run by somebody else he met on reddit. He's a cool guy, really humble on the rumble.


I think nonchalance is arguing more to the point that this was not a "lottery winner." It was a good idea from the start, because it filled a need people actually knew about, instead of filling a need that people didn't know they had yet.


It was a need that people had, but there was no chance of making a profit without the good will of Reddit users trying to help them get ad views and bury submissions from other image dumps.

The founder also made a thread claiming it was a "gift" to Reddit while also saying the same thing to Digg.


Ahh poor reading on my behalf. Valid point, although many people build side projects that fulfill a valid need that never get adapted. You usually don't have the luxury of a marketing budget or the time to promote your idea effectively when it's just your side project. Then again that's the beauty of Reddit - 1 front page post = 10's of thousands of hits.


What part of their success shouldn't you emulate? Are you suggesting people shouldn't do side projects because they might not be wildly successful? That doesn't sound like good advice. Sure, if you never try, you can never fail -- but you can also never succeed.


Solid questions and the point that is a risky piece to be emulated is the only have one thing.




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

Search: