Month: October 2007

Trigger: I went for another weekend to Boston. I was thinking about Artificial Intelligence and how I can make a simple Artificial Neural Network to learn some simple concept. How could I get in touch with someone that knows how to do that? I am in Boston after all, there are plenty of amazing colleges (MIT, Harvard blah blah) that have professors teaching this stuff. And as a matter of fact as I am walking around the streets of Boston, I am probably seeing them all around…except, I don’t know it.

The Problem: The big problem, I thought, was that no one knows anything about anyone on the streets. These are people that just come and go without giving you one bit of information (the look of a person is exception, which if you think about it is the prime way of meeting people online – pictures). At the same time, plenty of people that go past you in the grocery store are probably very interesting people with whom you can connect very well. The only way to find anything about strangers is to talk to them. But how do you know who to talk to? Should you act like a nuts job and talk to anyone?

The Idea: It would have been nice therefore to have a website where you can put specific information you are characterized with and information you are looking in other people. In my case I am looking for someone that is working with ANN. Once I setup those preferences I walk out to the open world. I carry a phone/watch/device that connects wirelessly to other such devices, checks the preferences of two people walking by each other and beeps if they match.

This way, I can walk around in CompUSA and meet ‘randomly’ a professor in ANN that was looking for a flat screen, he would probably be more than happy to talk about it, since he posted it on his profile for that specific reason. This taps into a totally new way of dating/meeting services, because it has the following very important advantages:
. . . (1) It is much more natural. There is a very big problem with dating sites: it’s the notion that if you are signing up on one, you are probably doing it out of despair b/c you have tried all your real life alternatives (or so you think). This moral underlying guilt prevents many people from using such services. But if you could meet someone in such a natural random way (on the street while shopping) then things look a whole lot different.
. . . (2) Confidentiality of information. There are many cases in which you want to make a piece of information available to the public but at the same time really only to certain people and not to others. Take a weird but true example – HIV positive person. He wants to find a date (most probably HIV positive too) so he has to state that fact and make it available to public. However it would have been nice if only the other HIV positive people new, rather than just about anyone that can exploit that information. Thus your device will only beep on the right occasions.

Difficulties: I discussed it for few days with Alek. We had the idea of making it work with cell phones because it will be easier to start the service. However, it turns out GPRS and bluetooth will not be a very likely option. Wifi could work if the area is fully covered, but again, I think cell phones are at a disadvantage. Also anything else that is more practical (a watch or a device specifically designed for the purpose) will be too expensive and will require investment and experience, neither of which we have. You got better ideas on execution, get in touch with me! You have that ready? I will definitely sign up!

Trigger: I was walking on Times Square and I saw a huge advertisement that looked like a canvas with random drawings on it. I don’t remember what it was exactly, but I thought, “What if I could make a huge gynormous or endless canvas where everybody would be able to draw whatever they want. Somewhat like an endless graffiti? How about if you could purchase of your area? how about if you could add multimedia? How about if you allow user created flash applications? How about if you make it a real estate business?…”

How: I wanted to make it on my own, but needed time and money. Requires extensive communication with server and the mouse movements need to be well sync-ed with users. My friend Alek needed to help me with the backend, I was going to make the fronted with flash.

Mistake: It was nearly impossible to explain it. It takes me hours and hours to explain the complicated idea and the possibilities that arise. I need paper to draw, I need then to always answer many questions and often by these questions I would think that the person facing me totally does not understand what kinda of digital world I am talking about.

The Story: As we thought we had to get some funding. One of the few things we did was apply for Ycombinator.com. Paul Graham did ask me some stuff after we submitted the application (which didn’t surprise me, given the very limited paragraphs in which we had to explain the idea). We didn’t get accepted for one reason or another. Between the feeling of rejection, the ambition, and the excitement we decided that we should try it out even if we didn’t get the funding. And so we started. It took us only 2, 3 days to give up. Over those days, we argued each trying to impose their own vision of how this idea can be done and what it should be like. It turns out it was so complicated in terms of concept that both of us founders had different idea to the point we didn’t even write one line of code before we let go. After all PG had made a good choice.

There was thousand ways of explaining it. “Social networking with a spatial perspective”, “Facebook on 3d”, “Drawing canvas where you could lay your web page at coordinates 300:320”, or “The useful Second Life.”

Every time I tried to explain it with a different term, it would confuse the listener even more. If I had it built maybe people would get the idea. But I can’t do it on my own, which is why I am posting it. It would look like a mix of GE’s drawing site,
http://www.imaginationcubed.com
Eric Deis’s experiment which works better when there are more people:
http://ericdeis.com/content/beautyandchaos/beautyandchaos.php
Google Maps:
http://maps.google.com/
and SwarmSketch:
http://www.swarmsketch.com/

If you could post a music player on Boston and hear it everywhere on the canves within a radius of 1000pixels it would be quit fun. If Nike posted an add at 400:400, then the value around it would appreciate due to traffic. And little by little, according to the content around, cities, villages of virtual information would emerge. A new market for spatial website management emerges. Instead of domains that are virtually attached by links, everything would be like in real world marked by dependency of distance.

This is how we tried to explain it at YC: “A spatial application in the form of an unlimited interactive whiteboard, that is divided in X and Y coordinates. People will not only be able to freely roam and draw, but also upload music, images, videos as well as Flash applications anywhere on the surface. This way, users will be able to upload media, and also tools to modify it. As they create new content they will want to keep it by purchasing physical space on the X:Y plane. In time, popularity of owners, physical positioning, and content of the ???land??™ will create dependencies among the citizens of the now populated ???cities??™. Hopefully, we will be able at a later point to introduce Z dimension for 3D.”

I think its a cool idea, but its difficult to be done and its almost impossible to be explained. If you can get the idea, or if it sparks something similar of yours, and you think you can do it – please feel free – make it. I’d love to see it alive. Tons of money there!

Moral: Sometimes simple things may turn out to be a better choice…