The Estonian tech-visionaries, led by Ahti Heinla – one of the Founders of Skype, and Rainer Nõlvak are developing what they call “predictive content personalization technology”.
They plan to debut this technology in a project called DailyPerfect. That is a news site that predicts your interests through an automated semantic analysis of information publicly available on you on the web.
DailyPerfect’s site is meant for everyone reading news online and wishing to find interesting stuff easily. They don’t tell too much about this, since the project is at the moment in closed beta-stage. But will be launched publicly at the beginning of next year.
DailyPerfect was initiated by Ambient Sound Investments and Curonia Research. The team is led by Asko Seeba – the former Engineering Manager at Skype, and Ahti Heinla – a partner at Ambient Sound Investments and the former Lead Architect at Skype.
Read more about DailyPerfect here.
Very interesting stuff.
I wonder what content they are grabbing off the web to make their predictions? In my experience the challenge is that while there is a lot of information out there most of it can’t legally be consumed and stored in way that is necessary to do the types of calculations needed for this type of predictive behavior to occur. Without this background data users are forced to input lots of data which usually deters them from using the site.
I’d love to know what “publicly available” data they are using. If anyone knows please let me know.
I somewhat disagree, at this point their target audience probably has a personal blog, facebook and twitter accounts, etc. so it should be easy to do semantic analysis my shared content (checking the text I publish, follow shared links and analyze it’s contents) to figure out who I am and what I like.
Good point – I guess I was thinking it was doing more than just that. This topic of how data that is “out there” can and can’t be used is very intriguing to me.
Let’s say I post reviews on local.yahoo.com would a service like DailyPerfect be able to use that information to help in creating it’s profile?
This sounds a bit like Thoof.com just taking information from more sources than just the site.