I finally got the chance to play around with Riya, the new social facial recognizer system. There were quite a few quirks, most of which can easily be forgiven since it’s in early beta.

The uploader didn’t work on my computer at home. My machine is a bit old (a 1GHz machine from 2000), but I haven’t had much trouble running other modern applications. The uploader would grind and say everything was successfully uploaded, when in fact, no pictures were uploaded.

The core functionality is definitely there. The facial recognition is fairly accurate, and I got the faces in my albums detected without too much trouble. There were about five false positives where a face was detected in an object that wasn’t a face. Given that a fair amount of faces were discovered, the false positives weren’t too annoying.

I really like the point and click interface–it makes it easy to tag single and multiple faces with people’s names. It’s still buggy, but once they get it worked out, Riya will be a valuable tool in organizing pictures. It’s really stunning how organized your albums become when metadata about people’s presence in photos are inserted. It makes it really easy to surf through and find the photos you’re looking for. In fact, I would be big fan of a desktop version of Riya that I could use offline.

Facebook, Flickr, MySpace, and all the rest better perk their ears up if Riya gets big. It would be wise of them to try to get Riya’s technology licensed to them (if Riya will even consider that), because metadata about people in a picture is valuable. For these types of social services involving pictures, it might well be one of those technologies that creates a significant paradigm shift in the way people use photos to connect to others.

I noticed that CNN has redesigned their front page. The familiar left hand menu has been made inline, and is now at the top of the page. This leaves plenty of room for the three columns which are chock full of information. Some new additions are the promotions for their pipeline video service, and AJAX tabs under Latest News.

The relocation of the menu to the top is in direct opposition with Google’s potential redesign, which will relocate their menu to the left hand column.

Overall, I think the top half of CNN’s frontpage is looking better. The bottom half, however, is obviously in need of some rework to make it cleaner.

Riya, the media darling of the Web 2.0 community, is opening up their service to people who signed up during alpha. Rumor has it that they will be opening for the public later this week. I’ll probably be uploading some photos tonight.

The biggest question in my mind is how much time and effort is needed by the user to bring Riya’s image recognition software up to speed? And secondly, how will the difficulty level change in response to wildly different photo collections?

In addition, the sheer processing power needed by most intelligent image processing algorithms can be large. I’m sure that when they open up their service, the whole population of digg and slashdot will be wanting to parse their photos. I hope their servers don’t melt.

In the mean time, has anyone tried out the latest version of Riya, which was opened today? I’d be curious to find out your experiences.

Once again, MSN proves to be a site catering only to internet explorer users. The new Beta MSN breaks both of my favorite browsers: Opera and Firefox. It’s definitely not a pretty sight. Divs are crashing into other divs, smearing content across borders. Oh, and don’t forget the extra wide right margin which forces me to do the dreaded horizontal scroll in Opera.

Here’s what the Opera render looks like:

And here is Firefox’s view:

Lo and behold, the layout looks fine in IE:

You gotta give it to Microsoft for building a crippled browser, and then building its pages to only display correctly on their crippled browser.

hngry, one of the latest additions to the Web 2.0 gamut, aims to solve the age old riddle that beguiles many of us every night–What should I eat? Basically, you sign up and add your favorite restaurants to the database, and, when you need to decide on what to eat, you enter in criteria about your budget for the night and your preferences, and hngry will suggest a restaurant from your list that is the best match.

My friends and I have long talked about such a magical restaurant decision program. Heck, we’ve sometimes been so indecisive that we use the roll of a dice to determine our final gustatory destination. Despite this, hngry does have areas which need improvement.

First, it’s extremely tedious to enter in restaurant information. The hallmark of such a decision system would be to have accurate and complete information about the restaurants. hngry provides many data fields, including opening and closing times, favorite meals, phone numbers, and price ranges. The problem is, none of this is done automatically–you must look up all the information and type it in by hand. A better solution is for hngry to parse through online directories and automatically retrieve the requesite information when a user selects a restaurant. And better yet, the interaction between users’ databases would be able to provide even better suggestions through network effects.

The second problem is deeper, and the fate of hngry will depend on it. It’s a nice thought that an online database could pluck a restaurant suggestion for me, thereby relieving me of the stress. But, will I actually like the suggestions? Probably not. And that’s when I’ll just hit the suggest button again, hoping for a restaurant ephiphany which will seldom happen. The best that I would hope for is to reveal a restaurant which I had totally forgotten was on my list. Given the current input process, my personal utility from such an application is quite low.

The idea behind hngry is good one. But, it’s going to take some hard work to transform it to a more automated system, which will be essential.

It looks like Google is going back to its old tricks again by providing a Google maps interface to images of Mars. In the midst of the clutter of betas, it’s nice to know Google still has time to mess around with science and other interesting quirks. I’ve actually been a little disappointed with Google’s offerings lately–they just seem to lack the tight focus that the company used to have.

What I’m seeing is little annoying things like Gmail being down for 20 minutes at a time and the new Reader interface showing huge bugs. I’m all for releasing prototype products and calling them beta, but, there is a point where releasing something too early can kill the product. It’ll be interesting to see where Google will be focusing its efforts in the coming years. The GDrive certainly looks promising, if not a little scary.

I’ve released the first version of Cloudee, an automatic feed aggregator and filter. It’s basically a place to go for your daily fix of tops memes and news. Right now, it’s tracking mostly technology and “web 2.0″ sorts of sites. You can follow the development on the Cloudee Blog.

For now, this will be a little pet project of mine–hopefully people will find it interesting enough to visit. Drop me a line if you have comments or feedback.

A co-worker of mine noticed that YouTube has removed some copyrighted videos from their website (including Natalie Portman’s Rap SNL short). But, we noticed that several other potentially copyrighted materials of the same sort are not removed, like Natalie Portman on the Inside Actor’s Studio. Since then, the Natalie rap short has been uploaded again by another user. It’ll probably be a game of whack-a-mole as YouTube tries to please the copyright holders while keeping the users at bay.

Is this the beginning of the end for the free exchange of videos over these web 2.0 sites? I still get my daily fix of short clips from the Simpsons and Family Guy from Google Video and other similar video services. Will YouTube start being a copyright police as it gets larger? I surely hope not.

If you’re a potential engineering student at Stanford, I highly suggest taking a look at the SCPD Free Previews. There, you’ll find some sample lectures for the first week of a number of engineering courses. From these, you can better assess the professors and the teaching styles of the engineering school.

Stanford is well known for their services catering to off-campus students, which are usually people who work in the Bay Area, and are sponsored by their companies to take a few courses (or earn a degree). Stanford probably has one of the best infrastructures for online instruction in the US. A large number of engineering lectures are all online, allowing easy perusal for the off-campus (or, lazy on-campus) students.

I was doing a few searches on Google today when I noticed a stark change in the layout. I switched to another computer, and found that no change had occured, so perhaps my IP or computer was being chosen as a “beta” tester for the new layout. In any case, here is what I saw:

Click on the image for a larger view. The big difference is the placement of the search types into the left column. Also, gone is the fat bar at the top that used to say which kind of search you are conducting (whether it’s web, video, etc.)

Finally, the already slim page margins are shrunk even further–I believe that the logo and text are only a few pixels away from the edge of the page. Personally, I would want to see more margins, not less. But I do think the left-hand placement of the search menu is a step in the right direction.

So what does everyone think? Is Google really redesigning? Or is this just a fluke? Anyone else seen this?

digg this