Notes from Kevin Rose’s talk about digg.com (at Future of Web Apps summit)

What is digg?

Stories live in the upcoming queues for 24 hours
– Upcoming stories that were dugg by friends are highlighted
– User can identify a store as their #1 story at a particular point in time

My friends
–> filter for upcoming stories

If multiple friends agree on something, then it’s chances of being interesting to everyone goes up a lot so there’s a “shared diggs”

Initial prototype: 20 hours to complete, $10 / hour

$99 a month hosting plan

200+ blog readers

People love to say they hate stuff (ha) –> threshold for comments that have been buried is 4, yet some comments have like 50 or 500 buries!

Give your developers time to experiment with your data (a la Google 20% time)

Today, there are a little more than 500k registered digg.com users
10M+ pages per day
1M+ daily unique visitors
90+ linux boxes

scaling has been a big issue for us
– PDF ‘Inside LiveJournal’s BackEnd’ (bible for us — how to scale and cache a site like this)
– We hired experienced DBAs, the same people who helped friendster, they taught us how to write queries properly.

Be aware of scaling from day 1 or you could have a month (like we did) where it was extremely painful

Future plans
– Learning from users interests story suggestions, friend suggestions
– Digging other types of content online (videos has been the most interesting… third most popular part of the site today)

We have a fully funded business plan
Path to profitability with current ads –> intention is not to do more any more advertising.

Ads provided by Federated Media with extra inventory filled by Google Adsense

People timeline:

  • just Owen and Kevin –> first four months
  • $50,000 investment from Text America friends –> hired Owen full-time, move him out here from Toronto
  • Jay Allison –> convinced him to leave his current company and join Digg full-time as CEO and pursue funding
  • today, there are 15 people
  • majority of people are in operations and development
  • 3 full-time PHP people
  • 4 people that help scale the site, DBAs, UNIX admins, experts at netscalers
  • 1 finance guy
  • 1 business dev
  • 1 graphic designer
  • 1 Kevin Rose (product manager)
  • 1 person dedicated to going through all incoming emails

User’s ability to remove their own data –> we’ll be rolling this out in the future
All of your actions belong to you as the user, your attention is something that belongs to you. You’ll even be able to export your data.

Yes, our search sucks –> we have lots of issues with this, we were using mysql’s text search, moving to Lucene. New technology will add the ability to search by number of diggs (yay!)

Paying top users? The most popular content doesn’t necessarily come from top users — very important that there isn’t outside motivation for submitting an article.

“The Emerging Age of Who” talk by Dick Hardt of SXIP

My notes from Dick HardtDick Hardt’s talk kicked off the Future of Web Apps thing yesterday about identity… it was short talk, only 15 minutes, but it was dense.  I loved Dick’s presentation style of speaking without interruption while a stream of synchronized images flowed on the presentation in the background.  It’s something I’d love to try myself, I expect that doing it well takes a lot of practice.

Anyways, the notes:

Overview
————————–
What is identity?
Who are you?
Identity 2.0

parable: throw a frog into a pot of boiling water, it jumps out right away…
but if you slowly turn up the heat, it just sits there and slowly its dead

If we don’t have identity, it’s like groundhog day –> you start all over every day
credit history, etc –> past behaviour, predictor of future behaviour

digital identity
— lots of online identities, none of them are machine readability

Yahoo
— lots of experiences that are silo’ed, one identity system
— some aberrations like flickr which has an awkward dual login system

Microsoft
— .Net Passport –> name changed to MSN Live Login
— coming up is “Infocard” — this is “Identity 2.0”, it’s good

Google
— bunch of sites most of which are accessible through Google Accounts
— still the same silo thing that Yahoo and Microsoft have going on

eBay
— sepearate accounts — paypal, ebay,
— reputation — past behaviour –> future behaviour (ebay’s “viagra” around their application)
— anti-phishing — starting to do “multi-factor identification”, biometric ID

evolution of identity systems:

domain centric authentication (identity 1.0)
federation of identity systems — circle of trust — silos (identiy 1.5)
circle of trust that scales to the whole globe… need for identity 2.0

opportunities — single account (single point of failure?  no, that’s already there… really strong lock in one place)

Wikipedia–> using openID with reputation
slashdot–> “karma”– what if you could take it over to another site?
ebay–> users could present their ebay reputation to another site

infocard, openID are solutions for single account identity