Alec Saunders SquawkBox

Riya is a new photo search tool which premiered at DEMO 2006.  It’s just gone into open beta.  The premise? As founder Munjal Shah said at DEMO "I have 13000 photos on my hard disk, and they’re all named things like P00109.JPG.  How can I find the photo I want?".  It’s a common problem.  Riya’s innovation is its ability to recognize, and automatically tag faces and text in photographs. 

Getting started with Riya is pretty painless.  Simply go to the website, create an account, and download the photo upload appplication (Windows only at this point).   Start the photo application, point it at your photos directory, and let it do it’s work.  It scans for faces and text, and then uploads each photo. That’s what it’s doing in the photo below.  And, as you add new photos to your PC it automatically scans them and uploads them.

Now, there’s a certain amount of irony in that photo.  I’ve actually got it stored on Flickr.  Riya, being beta, doesn’t yet have the ability to upload an individual photo.  It can only upload a collection.  Three days into the process, I’ve got about a little over 5,000 of the 9,000 photos in my collection uploaded.

The folks at Riya warn you upfront that the process of uploading your photos is slow.  They claim to be able to upload 4000 photos in a 24 hour period.  That may be true.  In practice, we found that the Riya uploader was such a resource hog that it had to be run at night only.

Once Riya has your photos online, it then asks you to train the engine to recognize the people in them.  This involves repeatedly looking at sets of photos, and identifying the faces Riya has found in them.  You can either have Riya show you unidentified faces, or browse your photos and identify people in them, or have Riya find faces that it thinks look like a face you’ve already identified, and then correct it.  Riya then takes the training you have given it, and uses it to recognize and tag faces in other photographs that you haven’t told it about.  According to Munjal, Riya will also take your collection of photos and compare it with your friends collections of photos, speeding up the recognition process dramatically.  I couldn’t test this feature.  I have no friends… on Riya, anyway.

So how good is it?  If you have only one or two photos of someone, it’s pretty hit and miss.  At 10 photos it’s getting pretty good.  At 100 or more photos, it’s hard to fool it.  For example:

• I have hundreds of photos of my wife.  Riya always recognizes her. 
• My two oldest sons are very similar in appearance.  Riya sometimes gets them mixed up.
• I have pictures of my Dad from when he was a young man, now a grandfather, clean shaven, heavily bearded, with and without hat.  It’s pretty good at recognizing him, despite all the variations.

Riya can also recognize text, and tag photos based on the text in them. Like Flickr, you can also manually tag photos. I haven’t used either of these features much yet, though. 

Riya also gives hints about features that are in the works.  For instance, it has a blogging feature, and a print feature, and a URL feature coming.  These are the table stakes that other photo sharing systems already have, though.  Although required, they don’t really differentiate Riya from the competition.

All in all, Riya works pretty well.  It’s a great tool for anybody who takes lots of pictures of friends and family. Thumbs up!

Techdirt wrote this morning that the VoIP space is getting more crowded with Internet era bubble models.  This was in response to an AP story about Jajah and Lycos titled New Ways to Call Over the Internet Debut.   Meanwhile, on Friday TechWeb News published Gartner Sees Rocky Road Ahead For Telecoms.  Elroy Jopling, the Gartner analyst quoted says:

Similar to the disruption music download sites, such as Napster, initially brought on the music industry, the telecommunication carriers, especially those in the "voice telephony market, will suffer the biggest shock," said Elroy Jopling, research director at Gartner.

All three of these pieces point to the same theme: the old telecoms are dying.  Businesses may emerge from those ashes, branded the same way, but they won’t be the businesses we know today.  The networks are becoming less and less of a barrier to applications.  Models built on network scarcity are doomed.  It’s the emergence of the Voice 2.0 model. During the transition period, we’re going to see lots of whacky business models. 

I’ll be speaking on some of these themes at VON Canada next week.  Monday at 10:45 I’ll be delivering an industry perspective speech, focused on Voice 2.0.

And in the meantime, as Andy Abramson says:

The key is for the goliaths to start to work like the new upstarts and move faster than battleships, while still covering the same ground.

Michael Robertson was interviewed last Friday for SoCal TechNews.  He talked about SIP, the kinds of businesses he likes to start, and entrepreneurship.  I loved this paragraph where he says that the things he does are obvious:

What I think about, is I have two sons—and what they will think about their father and what he did. I think they are going to say “you had it so easy Dad” and “Your businesses are so obvious.” And they are so obvious. It’s not hard to figure out my business strategy—I am focusing on industries that can be completely digitized. Music is a perfect example—you can move it around, make playlists, play files. There’s no question all music is going online. Software, with Linspire and click and run – all software is going to be delivered online. Phone calls—no doubt they will all be delivered on the internet. These are undeniable truths. You can argue how quickly it will happen, but getting back to my kids—I’m sure they will say it’s so obvious. When they are adults, they will not be able to say that it wasn’t obvious that all voice calls will be on the Internet, that calling would not be free, or that instant messaging and email is free. Though it might have been revolutionary, they will say that it should have been pretty obvious to anyone looking around. That cuts to the core—I like businesses that are purely digital, which can really change the whole economic structure. What I mean by pure digital is things you can digitize and shoot around on the Internet, such as news, music, videos, and phone calls. You can’t shoot a pair of shoes, or a loaf of bread, or physical goods like that. If you can digitize it by moving to the Internet, it’s going to happen. And with that change, it knocks out old incumbents and makes new room at the table for new leaders. You get a seat at the table with adults and make some money during that transition period. The businesses I do are pretty obvious to me because they are all about digitizing an industry that previously was more offline, and moving those industries 100 percent online.

Worth a read.

Andy has just posted the full text of the Streamcast / Skype RICO lawsuit.  It’s kind of a grubby fax.  Who knows where he got it from?  He would only tell me "sources".  Nonetheless, the suit alleges that:

1. Streamcast and some of the predecessor entities to Skype had a right of first refusal agreement allowing Streamcast to purchase the peer to peer technology which is the basis of Kazaa, and which is alleged to be the basis of Skype, in the event that an offer was made by a third party.
2. Streamcast was not offered this right when the technology was sold to Sharman.
3. Through a series of transactions in offshore havens such as Vanuatu, the technology was illegally transferred to Skype, in breach of the original contract.

Streamcast is claiming ownership of the technology which Skype sold to EBay last year.  They’re demanding the usual remedies such as injunctions and the like.  And, they’re demanding a share of the $4.1 billion sale price.

Wowsers!

No doubt there will be a lot learned in discovery.  The allegations, however, are stunning. The whole suit begs the question "How much, and how carefully, did EBay do their due diligence?".

Most of all, I am in awe of Andy Abramson.  He told me last week that he had a big story coming. Never in my wildest dreams did I imagine that it would be this big.

Update:  Andy just dropped me a piece of email at 6:31 AM Monday March 27th to say that he may not have received the final suit in his fax.  So, he’s updated his posting, and taken the fax down for now.  Stay tuned. 

Whoa… I need to read and understand this in more detail, but Andy has a blockbuster post.  Skype’s founders have been named in a racketeering suit.  Go check it out.