LeapTag

Big Brother Technologies, Inc.

Google announced that they will anonymize user information in 18-24 months. I think this is a great step in the right direction (though some think that this may be a PR maneuver). Why 2 years? John Battelle thinks this is “because Google is all over personalization and doesn't want to hobble itself.” I tend to agree. I also think that two years is a very long time to lose one's privacy.

I think what needs to change is the basic model of how the user and the environment interact. Let me illustrate this by telling you a story that sounds a bit like the movie “Minority Report”:

“It is the year 2012. It has been a good year for me and business is doing well. I am rushing to the airport, late for a flight to Texas to meet with some customers and to close a $3 million deal! Thanks to the BBT (Big Brother Technologies, Inc.) equipment the airport has installed this year, I am instantly recognized by the airport the minute I enter. My wrist-screen immediately displays my gate information. I pass through the security screeners, and as I walk to my gate, I am targeted with personalized advertisements on the BBT screens along the corridor - including a number of screens that notify me that my plane has been delayed. What a relief! This is followed by other on-screen suggestions to stop at the Starbucks for my favorite coffee and to catch my breath. Good idea! These things know me too well. I arrive at the Starbucks, grab a latte and plop myself down for a minute. A couple of BBT officers approach me. Very gently, they ask me if I will go with them.

…

I come out two days later. I was in custody, trying to prove that I was not connected to a bombing in Texas the week before. It turns out that a couple of terrorists hacked into the BBT computers and impersonated my digital signature. They entered the airport, pretending to be me, to plant a bomb. I was digitally framed!”

Of course, this is just a story, a fabrication to illustrate a point - that when your environment knows your presence it opens up a whole new set of vulnerabilities. As suggested in my previous post "Google the Spy?," this model, however well intentioned it may be, is prone to hackers, terrorists, and yes, just like in Minority Report, frame-ups.

We all would love to enjoy the conveniences that come from a world that adapts itself to us. However, I prefer a vision of the future which involves a fundamentally different model with a similar outcome:

My “wearable computer” knows everything about me but never discloses my identity unless it is absolutely necessary and with proper authorization. The airport does not know that "I" have arrived. Instead, the airport first gets a query from my wearable computer for the gate number of my flight. It may also get a query for the location of my favorite coffee place, and my wearable may communicate with the screen I am passing by to tell it which of the ads I am likely to be interested in. Later, the message of my flight delay is communicated in response to a query from my wearable, letting me know that I'm free to relax at Starbucks for another hour.

While this is not a guarantee against identity theft, I am more likely to discover if someone has stolen my wearable computer, enabling me to take action quickly and making it more difficult for someone to impersonate me. Technology can also guarantee that my computer can only be used by me (such as through fingerprinting). On the other hand, when someone steals my identity online I may not find out about it until it's too late.

In short, my computer discovers my environment and makes my environment conform to my needs without revealing my identity.

Same outcome, different model. I feel much safer with the second version.

- Cuneyt

ETech

As you probably know, ETech will be held in San Diego next week. I will be at the conference Monday through Thursday. While we will not officially be demonstrating LeapTag there I will be available to chat about LeapTag and to show it to anyone who is interested. If you have been a beta tester for LeapTag or if you are just interested in learning more about it, please let me know (I will be checking the conference message board). I would love to talk to you.

- Cuneyt

Google the Spy?

The issue of privacy has long been a big interest of mine and lately I have seen it become an important topic in the world of free internet services. You can see some of my views on privacy in a previous post.

 

John Battelle asked a crucial question about ownership of Google search profiles on his blog, which was followed by a very interesting comment discussion.  That question was: who owns my personal search history? 

    

Something happened to me recently that turned this discussion upside down.

 

A couple of months ago, l was pitching our product to a VC in a coffee house in Palo Alto.  He wanted to compare LeapTag to Google Personalized Search, so he logged in to his Google account from my laptop. We had a nice discussion. 

    

A week ago, I received an email from him asking me if the following searches were ones I made recently, and he included a list of searches. 

    

My jaw dropped! Those were all searches I made a mere half an hour earlier - two months after our meeting.

 

I realized that he had never logged out of my laptop, and that all the searches, personal and professional, that I had done for the last two months using my laptop were now part of his search history.  He had access to all of it – as if it were his.  Now who owns what?

    

I cannot begin to express how violated I felt.  He had access to all my searches which could potentially reveal a lot about me. I had no idea what he now knew.  I cannot even access those searches myself to find out.  So, right now, I have no idea what my exposure is. 

    

I think with Personalized Search, Google has created a tool that not only allows Google to spy on people but also allows people to spy on each other.

    

Since then I have performed a test to verify this behavior.  First I created a Google account called "GoogleSpy" :-). I logged into that account from a different PC at work and then I logged in to the same account from my laptop.  Google was happy to let me log into both machines and did not log me out from the other (we know from experience that Google does not log you out even after two months!).  On my laptop I performed a couple of searches.  Here is a screen shot for one:

    

         

Now, the only indication that something may not be right is the login name in the upper right hand corner.  Busy people that we are, who notices anything in the upper right hand corner?

    

Then I went to the other PC.  Both of my searches appeared immediately in my search history (pretty fast service I have to admit :-)).  Here is a screen shot:

    

    

The first search is one I made from the PC and the last two are searches I made from my laptop. 

    

Yes, I did give my permission for someone to log in to their Google account from my laptop.  However, I reasonably expected Google to log him out after a while even if he did not log himself out.  Then I realized that this is probably not a bug, but rather an architectural limitation.  Google cannot tell when a person has finished using a particular computer or if in fact if that person actively uses multiple computers.  For personalized search to work well, Google needs to capture all of a user’s search activity.  While doing that aggressively, Google became a tool for compromising my privacy. 

    

As a result, my search results are not only "owned" by someone else, I don't even have access to them. 

    

I don't want to accuse Google of building spyware because I don't really think that is their intention. But in effect, Google Personalized Search can be used as spyware.

    

So who owns my search results? Apparently not me.

    

- Cuneyt

The World According to del.icio.us

The tag-line for our product is "The world according to you" since we strive to create a world full of information relevant to you and your interest. When we were thinking about a way to convey this idea to people who have not used our product yet, we decided to experiment with a concept we called: The world according to del.icio.us.

Our concept was to create a world full of information relevant to the del.icio.us community based on some of the tags in del.icio.us. So we took the top 20 most popular tags in del.icio.us, we used the URLs the del.icio.us community provided for each tag as positive examples for the corresponding tag, and we selected a few negative examples for each tag in order to create a more realistic scenario for our application (since user's interests are usually defined by both positives and negatives).

You can check out the results at: delicious.leaptag.com

The site will provide you with a very realistic scenario of how your tags would look if your interests were aligned with the del.icio.us community. On the left you will see the top 20 tags and if you click on any of them our suggestions will appear in three different panes (news & blogs, websites and books) to the right. You can see the associated del.icio.us URLs in a list underneath each tag and if you click on a tag's Interest Preferences you can explore the negative examples we selected as well.

Unfortunately you can't vote or add new tags. For that, you need your own copy of LeapTag so that you can personalize your own tags based on your own interests by voting.

We would be delighted to hear what you think about this experiment. Later I'll share with you what we found out. In the meantime, enjoy!

-Cuneyt

Some thoughts about privacy

We are spoiled by getting most Internet services for free. Of course, we had to give up something for it. Then we realized that what we gave up was our privacy, and we did not like it. Who wants to be monitored by Big Brother, whether or not we have something to hide?

The current business model for most Internet companies today is to collect information about people, centralize it and then monetize it. This model naturally results in a loss of privacy and it has started getting more attention lately and bothering people. For example, a San Jose Mercury news article points out that Google is now tracking people. John Battelle, in a post yesterday, remarked that he is deeply concerned about the erosion of press freedoms in this country. There are some pending legislative measures that can make things worse.

What are the alternatives? We can pay, but that may be worse than losing our privacy, at least to some of us :-). Somehow we tend to keep our wallets deeper. Another approach is to decentralize. Instead of a pure web service, we go to a hybrid client-server model. The client knows everything about us and that is where the monetization happens. The server only supplies supporting information for the client to perform its functions properly.

A good example of this is Skype. A server, or a distributed set of servers, is involved in setting up a call. After the call is set up, the conversation is peer-to-peer, and encrypted (even if some conversations go through intermediaries, they are still encrypted). Wait a minute, this is the same approach used by the telephone companies, so we have not invented anything new here. We have gone back to the distributed model and made it more secure. First of all, the call traverses the Internet (as opposed to the phone company's network) so it is much harder to tap. Second, it is encrypted, so it is hard to break. However we do have to realize that there are levels of privacy.

So if we build applications in this distributed fashion, and keep the personal information stored on our computers, then we are much more secure. Someone can always break into our computers, but as we all know, there are a number of tools designed to prevent that.

With the personal information kept on our computers encrypted, we may be able to regain our privacy while enjoying the wonderful world of a free Internet.

- Cuneyt

Multiple personalities, multiple domains

We all have multiple personalities. For example, many of us have a work personality and an "outside of work" personality. There is the side of ourselves people see at the business conference, and the one they see at the rock concert. Each facet of our personality can make us seem like different people, depending on what we are doing and thinking about.

Depending on our professions, we have a variety of different topics we are interested in at work. For example, I am interested in the technical aspects of what we do, like classification, the programming languages we use etc., as well as marketing aspects, such as competition, or the concept of viral growth. I am also on the boards of several companies, and since each company is quite different, each represents a different aspect of my personality.

If you are a developer, perhaps the different programming languages you use each day constitute an area of interest. If you are a VC, each of the companies you have invested in or are thinking about investing in may be a different area of interest, or a facet of your personality.

We also have different personalities when we are socializing, exercising, or relaxing. We all like different types of sports, some as spectators, some as participants. We may like mental exercises or games. Some of us may like movies, and others may like the theater.

These are all distinct "personalities," or interests, of ours. None of the various interests are necessarily related to each other. I may like both wrestling and ballet. Furthermore, since we are individuals, each of these interests is unique to us. For example, you and I may both like soccer, but you may like Brazilian soccer while I like European soccer.

Each of our unique (and perhaps potentially conflicting) personalities is in constant thirst for information. We always want to be on top of our work-related news while simultaneously wanting know how our favorite teams are doing, or what the latest ski equipment is. The information we look for may come from multiple domains: news, blogs, websites, a new book, etc.

There is also so much information available these days that it is very easy to get completely lost in it. For example, just for the news and blogs domain, my RSS reader produces hundreds of new items every day based on 10-20 feeds. I know there are at least a hundred feeds I need to be monitoring, but I don’t even have time to monitor the 10 or 20 I set up.

I need a personal assistant! One who knows exactly what I want, in fact what each of my different personalities want, and who can monitor all the different types of domains and report to me what I need to know exactly when I want to know it. Wait a minute though, this assistant needs to monitor hundreds or thousands, or maybe more sources for all the different personalities that I have.

I need an army of assistants!

-Cuneyt

The world according to you

We are at the dawn of the age of collaborative filtering.  But then, what happens to our individuality?

The old adage "no two people are alike" comes to mind.  While collaborative filtering is fun, we also need to go back to our individualism, at least sometimes.  So, we would like to take you on a ride to individualism, but don't worry we will get back to communities in more ways than one, just bear with us.  This ride will be about content discovery, but since it needs to cater to individualism, our goal is to discover algorithmically. 

The first thing to notice (right after "no two people are alike") is that people have multiple personalities.  Any algorithmic approach needs to capture that.  Well, in the world of tagging this is easy.  Let’s say we associate a tag with each personality.  These tags will be a little more involved than the current notion of tags.  We will need to select them a little carefully and we will need to nurture them.  Eventually, this collection of tags will represent me, the individual. 

So let’s say one my personalities likes cooking, and I create a tag named cooking.  Now I need to define what cooking means for me.  It is important to realize that keywords may be inadequate to describe my "cooking" personality.  Creating a long description may be too laborious and besides my personality may evolve over time.  For example, my taste in cooking may change.  I need to be able to define my "cooking personality" dynamically. 

So, we will define this "by association", by selecting some examples of what cooking means to me.  The easiest way of doing this is to tag web pages that I like.  Well, actually this is not quite enough.  A good "definition by association" also needs some negative examples.  So I would also need to "negatively tag" some web pages that I don't like (as far as my kind of cooking goes).  Now, I feed these "positively and negatively" tagged items to an algorithm along with a whole bunch of "to be discovered" web sites, and I get back a set of web sites that "fit" my personality.  I look at these results, positively tag the ones I like, and negatively tag the ones I don't like, feed them into the algorithm again, and get back a whole bunch of new web sites, hopefully with a better fit.  If I do this for a whole set of dynamic tags, after a while I end up with a machine that knows exactly what I like and what I don't like, and it can discover content continuously – even while I sleep. 

- cuneyt

What's in a name?

Hello World!

Welcome to the Yoriwa blog. We are the team* behind LeapTag, a personalized content discovery tool we think will make a huge difference in how people discover, organize and share information on the web.

What's the story behind the Yoriwa name? Well, naming is not an easy thing to do these days, not only do you need a good name, but you need a URL as well. And let us tell you, all the simple 4-7 letter names are g-o-n-e. But that's ok, those kinds of constraints just beg for creative solutions and that's what we have here.

The name Yoriwa comes from the Japanese word “yori wakeru” which means “to classify.” Our patent-pending algorithmic classification technology is core to Yoriwa and used in LeapTag, so it made sense that we would use this word in our name. 

This is our team blog, the place where you will find posts from us about technology, philosophy and LeapTag.

*Our team: Cuneyt, Ulas, Karen, Eugene and Rao

-Cuneyt