hiroprot

Of course, after they have your personal details, they'll now do an Internet search and find this post. You just gave them your game plan ;)

I'd love it if I could preview what everything looks like right on that page after I change the individual colors at the bottom.

Nice job...you just made me work out for 7 minutes :) That made my day.

Okay, then my current config is probably as good as it gets for now.

Thanks for the help :)

Thanks, this is helpful. Which CBC suites work in TLS 1.1, but not 1.0?

These are my test results: https://www.ssllabs.com/ssltest/analyze.html?d=forkly.com

"So, the way to go with BEAST is to force RC4 with TLS 1.0 and earlier, some CBC suite with TLS 1.1, and GCM suites with TLS 1.2."

Is it possible to configure nginx like this? From what I can tell, there is no way to select a cipher based on the protocol.

On a related note, running this against google.com shows that they support Forward Secrecy as well.

It doesn't seem to be a bug, it has to do with their post-signup invite friends screen. By default, all contacts are selected, and if you just hit next, it will invite them all.

You have to explicitly hit "unselect all" first :(

Obviously, this is really bad UX design (for the user), and it really surprises me that Path would do this just to get a few more users, especially considering that they used to market themselves as a social network for a limited number of close friends.

So, they essentially handed each inmate a key to their freedom (in the form of the handbook) upon arrival?

Am I the only one who thinks that's somewhat philosophical? :)

I simply replaced the eval line at the end of the decode script with console.log.

I hadn't seen this obfuscation technique yet, learned something new :)

As others have said, I think the project isn't super useful. The most interesting thing I saw on the page was an attempt at hiding the source for the script by putting it into an image, and then decoding it.

Pretty easy to decode and just console.log though. Here's a Gist of the source for anybody who's interested:

https://gist.github.com/anonymous/5169037

What do you use to do the recommendations? Apache Mahout?

I don't want to speculate too much around this, but I think it shows that the space is starting to heat up, so this is actually good for us.

Obviously, we would have liked to see a higher price tag for that acquisition ;)

Full disclosure: I work for Forkly (http://forkly.com), a FoodSpotting competitor.

I think that it was probably an okay exit for the founders (but definitely no homerun), and for most of the early angels. As for BlueRun Ventures (who put in the bulk of the last round), it probably wasn't that good of a return.

Of course, these exits often don't get paid out based on equity, but on some negotiated deal that will make investors at least somewhat happy. As for the founders, I'm pretty sure they got signing bonuses, probably tied to staying at OT for a year or two.

It will be interesting to see what happens to FoodSpotting (apparently it will stay standalone), and how OpenTable will make use of it.

Slightly OT, I used to obsess about the accuracy and actual data that these devices provide (had a BodyMedia FIT and now a FitBit), until I realized that the data itself isn't what I'm most interested in.

IMHO the main draw for these devices is their ability to motivate me to be more active, sleep better, etc. Historical data is interesting, but not crucial to that function (although I'd like to see FitBit make the data available). Far more important are real-time feedback mechanisms, such as notices that I achieved a goal, motivating messages ("you're almost there"), etc.

Where I went to school, they taught Turbo Pascal as the first language, and I thought it was a much better beginner language than Basic.

I think this will be really cool when we can index and query on parts of the JSON structure, similar to hstore.

Missing from the list: Knoll Life Chair

http://www.knoll.com/products/product.jsp?prod_id=188

I tested most of the other ones in this post, and liked this one better.

Mmm...so this was a test that compared a production server that is presumedly under load against a test server that is seeing no load other than the testing?

Wanted to follow up and say that while including the proprietary Instagram shaders is wrong, the author did a nice job with this.

We're currently working on some filtering stuff using GPUImage as well, and it's nice to see this stuff progress.

One filter that is missing here BTW is Lux, which is Instagram's equivalent of the Camera+ Clarity filter. The reason (I suspect) is that the Lux "filter" is actually not implemented using shaders, but rather CoreGraphics.

Would love to hear your take on it if you know more.

Yes, they had the plain text shader files embedded in their app, easily extractable. It's been like that since they first released 2.0. Path has (or had) them embedded, too.

It's still their IP though, and publishing them as "open source" is unethical and probably illegal. Just because something is feasible doesn't mean you should do it.

The GPUImage-based execution environment is fine, but you should remove these shaders ASAP.

You have a valid point, given such a constrained space makes it easier to recover that stuff. For what it's worth, we used salted SHA-1 hashes, which would make it harder if somebody were to get a hold of the data, but of course, it wouldn't prevent us from doing the hash reversal ourselves if we were to turn "evil" some day.

So, given all this, why wouldn't AT&T offer a competing free conference service where they keep the termination fee?

Is there a Ruby SDK?

Following David's lead, we've published our stats for Forkly as well:

http://blog.forkly.com/post/15190891337/lets-get-honest-abou...

This made me think of Ender's Game (finished re-reading it last night).

Good call on the display of errors...why is that not available in today's TVs?

Or is it? Maybe in some special diagnostic mode?

My invite request got approved right after the request, but the service isn't working for me :(

All I get is "Oops! We have run into a small problem. Please try again." over and over again.

Overloaded?

Word of warning: I ran this on my MBP with external display, and switched to integrated mode. After that, my display wouldn't update anymore, and I had to VNC into my computer to change the setting back to dynamic.

A "test mode" which reverts to the old setting after X seconds might be useful.

Great API docs...some of the big boys should take note!