Daily Archive for September 5th, 2006

Captcha Brain Power

I just came across a very interesting Google Video from Google. Luis von Ahn, a Carnegie Mellon professor, discusses the Captcha concept used to stop automatic systems from filling in forms and how currently Spam companies are trying to get around it.
Captcha is used to describe a system that allows a computer to automatically generate random numbers and or letters and then obfuscate them in order to render reading them by a computer highly unlikely or impossible.

This was interesting for me since I had worked on a similar concept for stopping people from requesting lost passwords on Significant-Media.
The video below then goes on to explain how human computation i.e. people solving problems rather than computers can be useful. With the Captcha system for example, computers are not capable of providing correct answers.

Then the ESP Game is explained and the process for it is pretty amazing from a thought process point of view. Anyone who has to come up with new ideas in their company should find this part truly fascinating :

The part I found really impressive, in view of a successive set of procedures, was when von Ahn explains the different combinations of possible player modes, as well as anti-cheating mechanisms…

The guy seems like a great teacher as well, if the seminar is anything to go by !

Placed in context this image tagging system is what early users of flickr were used to when they add tags to their own photos and is now pretty common procedure. Flickr was bought by Yahoo and Google well may have found a novel way of catching up here, if you read between the lines…

Update : interesting article over at the washingtonpost.com concerning a way to ‘turn the tables’ on the phishing techniques using images from the banks

Google SEO Lessons

Google, via it’s evangelist Adam Lasnik is starting to provide SEO classes. As you can see here the seminars are cheap and open to all.
It begs the question about Google asking people to create web sites for people and not for their robots. However, it is true that the best idea is to build your web sites for your target population while respecting W3C standards, and then optimise your code for search engines.
Once you have read a fair amount about Search Engine Optimization, you realise that most people are just guessing what is working and then try to make you think they have the real deal when it’s likely that they only have half the story. Optimising pages is partly logical on the basics and then guess work on what may be used by Google to define the order of their results. I personally think they should spend more time dealing with the weeds like domain name parking with pages full of adverts/links…