Pessoal, estive em NY em setembro para o Web 2.0 Expo.
Vou, aos poucos, comentar algumas coisa que vi por lá. Gostei, por exemplo, da palestra do CEO do Digg, Jay Adelson, (9:55am Friday, 09/19/2008 – Organizing Chaos: The Growth of Collaborative Filters). Presentation: Organizing Chaos_ The Growth of Collaborative Filters Presentation [PPT]
Fiz aqui uma transcrição da palestra com algumas palavras que me escaparam para ajudar ao pessoal. Comento mais sobre a palestra depois. Quem puder esclarecer os buracos da transcrição coloque nos comentários que eu vou aprimorando o texto.
Segue transcrição (em inglês).
Woman – Please, help me welcome Jay Adelson…..Thank you.
Jay Adelson – Hi everybody! I´m Jay Adelson. Thanks for having me here today. I´ve been ask to, in about eight minutes, or so, explain why collaborative filters matter.
Anyone who knows me, knows that for me to get to that the point in one hour is a difficult tAG past?. So I would do to my best to cover some of this.
You know, why does this matter? You know, first of all, how many of you would say would use the collaborative filters the last day, in, say, twenty four hours?
How many of you used Google in the last 24 hours? A lot more of you.
How many of you used Digg? Let´s see. Even more. Well done.
Anytime you take the Internet, the preferences, or as Joshua Schachter (founder of Delicious) said, the foibles of the group and use that information to filter, and ) create relevance for an individual or a larger group that is the collaborative filtering.
Really, search engines is a form of a collaborative filtering. When we searched engines for first create the data set about you, the individual, was much more shallow.
We didn’t know much about you. And so how would you build rank. Well, one way was page rank and bankrupt this idea, looking back, I see, IT managers and web developers would point to these links.
The other way, would be clicks that´s your point of contact, a very explicit action that you type into a box and get IT resolved.
This is change the lot, the news forms of collaborative filters, I mean get all that because as this shift has happen as we´ve become more envolve with the Web.
The data set that you are providing into various collaborative filters in a work all you done in creating Web 2.0 really matters in terms of where they can take of relevance data on the Web
So, what’s change? Ok…so I describe…
( I take my pick up)
…I describe… the world are very explicited. Well, now, we are on the Web 24 hours a day. I think, in 2003, Berkley said that was about 7.3 millions new web sites, web pages, adds to Internet every day, I think now we have something like terabyte of data add to the Internet every day.
Its one reason so much because it is dinamyc, it’s real time changes. We tweater everything. Including paintin xxxxx toino, what we are eating, where are travelling, we… post to a friend networks, in social networks.
It’s data, It’s extremely valioumess. It’s dynamic. And tells something about you. That’s perfect for me, builder collaborative filtering, because I can use that information that some how sort through this and make a better and the more effectless experiences.
The other thing that is changing is privacy, the sense of privacy. And I took me a while the figure this one now.
The one the thing is open flag gates here was that the younger generation doesn’t have the same issues associated with privacy that we´ve had or my parents had.
And so the willing to share more information. Classical example be how I use my e-message AIM. For me its I’m a lunch. I’m waiting. I’m here.;)
My teenager baby-sitter well say, you know, I’m, I’m, you know, I’m feeling down right now. I’m feeling better. I’m fall. You know. I’m eating enough. Whatever, this idea lot of personal in private data sharing in the world has open up in this open flag gates.
And so this is gonna change and we move from, even more from in xxxx culture, to more of we mode of type the word box in the computer an get result back to more we cancel to connected.
So, Why Web 2.0 why in a Digg all these companies?
One many things that breaking for the three parts.
There is the generic collaborative filtering like a search engine or more even Digg front page this way.
Or looking of Zaigais large group of people input something that kind very usefful also that a large people also community consuming the same data.
Right to me in the social network … I create a subset community, a people who are my friends and maybe a watch their activities streams.
It’s filtering content in different way.
Now, I don’t know by you, but a lot of my friends are anything like me. So, I can use them as a judging factor of what kind to my be interested to me.
But, I’m still interesting, I´m still logging to Tweater, Facebook, a watch this stuffs, my friends activites and Digg and so for .
But, really the exciting thing (…) if this one point that I can live today is this hyper-personalization opportunity. Is the next step of collaborative filtering.
It’s idea the instead of looking at social network the you’ve create yourselves the event in the names. I´m … looking all of you, everyone, and I can compare you all together I goona find people that like you and I’m gonna use that, collective wisdom, the actually find things are more specifically interesting to you.
Like a recommendation engine today. You have be this that Netflex, you know, I’m look at everyone whos in the horror movies across the world xx rent horror movies wich one are the best and maybe figure out for you specifically, what youre you interested.
Particularly sent personal data is going to move from one website to the next. What about a like nine millions Facebook users registered use on Digg that’s a lot of personal data.
What can I do of it? How can I take that information that’s your front page experiences is specific to you, versus, you know, the rest of the public.
News values to both this differents perspectives, you know, the zy guys to point of view versus the personalization.
I think that huge important shift and how this collaborative filters will affect the develop websites, because you can monetarizng that, people talk like monetarization of social networks.
Well, you know, if I knew exactly what you interested in, your’re located, you GPS location,
what you did in last hour, what you read of your last three years I’ve be XXX target bether, so you better, better.
So this collaborative filtering really are the key to the monetarization of this web 2.0 applications.
I’m exciting because when I thinking about the future of Digg, in particular, I…I, you know, you’ve 60 thousands submissions a day to the Digg platform.
How we going to take all that data and make interesting and relevant without this idea of collaborative filter?
Initially, all we have diggs and embarress we have a of lot more data.
When you use the system? Where you coming from? What categories you’ve been interested?… And what period of time?
This is the data thats gonna help me change not just up coming section, but the front page we’re extremely exciting about been able to over there all future.
It’s not just Digg. Im thinking that moving for the world by Amazon, you know, going buy a pair of shoes you xxx buy CD, etc.
When you said: you like a new pair of shoes, to world were now because poibolls information you can go back an action I know that you be walking a lot this week.
And I know you be a lot this week because your Nike connect to IPOD, tell us so, you know, assuming of you, a lot of information by with your privacy rules and my been interested pair of shoes, at that point
Anyway.so, I know, eight its not a lot time to talk about collaborative filters. That’s my goal added. Thank you very much.
