p. Missed Stewart Butterfield
h3. Dick Hardt from SXIP on presentation on identity 2.0
* .net passport rejected expensive, proprietary, sing signon
* that web services need identity protocols separte from username/password
* that identity 2.0 is inevitable
* that users need to own their identity, currently sites own various visages of identity
h3. CIO of City of Philadelphia, Dianah Neff
p. ad for Philly played...ergh
* access to affordable high-speed internet access
* in industrial areas can be $400-$1400/month to get internet access
* expect to create 3,000 jobs as a result of the city-wide wifi
* enable partnership to build on skills of private sector
h3. UI Minute w/Seth Sternberg on Meebo
* With AJAX you can make I/M available in the browser
* completely dynamic
* as fast as regular IM
* making IM available anywhere to anyone
h3. Conversation with John Highland and Vinod Khosla
* investing his own money in his own account, but not starting a fund
* that kids move amongst media more fluidly, more flexibly than we realize (media "mag?dl?")
* skype not web 2.0, more liner along web 1.0
* Q: feels very boomy in the room like 1996, where are we in the cycle
* A: That a lot of companies got funded in pre-2000 but then got stuck and lost their exit possibilities. Illiquid bulge of capital. Some companies have since found there way out, others have gone out of busines. The amount of money in the VC business is still more than is needed. Just because you get funded doesn't mean you're successfull. Just because you have money doesn't mean you can spend it like before.
* A: only when you're pretty certain you have traction and you're worrying about scaling, then you consider funding. That the more moeny you give to a startup the less likely they are to succeed.
* The older a company gets the more traditional it gets.
* That excite turned down buying Google's technology for $1M due to belief in their (excite's) own technology
* That authoring top-down content is no longer relevant. Far more creativity out there, small groups can be much more resonant with what the create
* psychology of this generation has completely changed...that kids think everything is clickable, always on IM, always on friendships
* that the total picture from blogs is more comprehensive, though any single blog may be less _trustable_
* trust rating sellers / vendors, building repuation systems and review systems
* that we should look past ringtons and simple games and ask "what's next?"
* Q: What's coming out of India/China that you think is really interesting?
* A: No one thing but multiple indicators of how more creative the world is getting.
Example: learning English pronunciation via cell phone.
Another example tutoring from India for US, European students.
* Q: What does the education system need to do to change to this new context?
* A: no longer a need to teach kids certain facts, but there is a desperate
need to teach critical thinking, how to assess information and filter it.
Not as urgent in the past because we had less information to deal with.
Why can't we use the same model of Wikipedia to make textbooks open and
* Q: You are fond not just of open source text books but open source everything, can you give some examples?
* A: that researchers in australia developed alternate GM seed methods and turned the resulting patents over to the public domain
* Jeff Jarvis: look at wikibooks. delighted to hear you say that the top down content model is over. But in five-ten years where is the value going to be in media?
* A: I suspect the value is going to be in the company that can create and maintain audiences and not in the content. Aggregating audiences in interesting ways. The most adaptable company.
* Q: do you see the appearance of simpler devices beyond the PC or cell phone to take advantage of broadband or wifi?
* A: we will always want to do more computing than we have processing power or battery power. notion of a network proxy that does the work and interfaces with a personal thin client
* Q: what will happen with search?
* A: more innovation in search, more use of collaboration and filtering, many flavors of algorithms to be applied to different vertical areas.
h2. From the Labs
p. HP, Yahoo...
h3. Gene Becker of HP Labs
* ubiquitous and utility computing
* "break out of the kazoo interface"
* web getting diverse, social , creative, mobile, contextual, physical, experiential
* printable hyperlinks, eg QR Codes, http://activeprint.com/
* resulting in physical world clickthroughs, already big in japan
* mediascapes: digital landscapes overlaid in the physical world. Combine media authored on or tied to physical locations, access via wireless, results in rich, engagin experiences in the real world. Premiering mediascape in san francisco see http://www.dstory.com/dsf_05/
* misto: social, tangilble furniture
* that most computer applications today assume only one person driving at a time, how do you create applications where multiple people interact with the same application interface at the same time?
* virgil: context-aware extensible media browser
* utility computing: commoditization and democratization. created rendering service for use by places like dreamworks. saves companies who don't need all the equipment after the movie is done. Allow more general users to use the HP service to use 200-300 CPUs to do rendering.
h3. Usama Fayyad and Prabhakar Raghavan from Yahoo
* drive innovation out of Yahoo at a much faster pace
* introduce 100s of products a year
* creating the science of the web
* Reference Yahoo Buzz game from etech (http://buzz.research.yahoo.com). Developing leading indicator to leading indicators.
* forecast error decreases over time. Humasn beat randomized bots.
p. Now Prabhakar on Yahoo! Mindset
* http://mindset.research.yahoo.com/ -- real time tuning of search results in browser
* Data deluge 10Tb of data in to Yahoo daily...transactions, crawls, tags
* TagLine ...combining social data mining, microeconomics ; demo of tags from flickr over time
h3. Alan Eustace, VP Engineering of Google
* focused on innovation, small teams, 20% time
* gender identification of photos. create a system that's easy to implement and maintain
* information overload
* Introducing Google Reader ...rss reader, with search engine, tag feeds, ajaxy, in page rich media support can share, gmail, blog items
h2. The Alumni Report
p. Kim Polese, SpikeSource, Joe Kraus, JotSpot
p. Interviewed by John Battelle
* kp: rules have completely changed about how you build a software company
* jb: how big were you at launch vs now
* kp: 13 at launch, 45 today
* jb: what is jotspot?
* jk: it's an enterprise wiki, ad-hoc document management, information publishing
* jk: I think a lot about startups. the biggest mistake was that we didn't focus on revenue early enough
I think startups make a mistake, partly from the success that google had. we were in beta far too long, by not focussing on revenue we were focussing on the wrong things.
* KP: I went in with an assumption that the rules had changed and how to learn. That a lot of enterprise customers already had complex stacks and challenging support issues around their open source infrastructure. A lot of our assumptions about what they were looking for from us ended up being very difference.
* JK: that jotspot when from 5 to 22 people.
* KP: don't feel too big yet
* JK: "no false positives hiring philosophy", optimize for never getting anyone in the company who won't be successful.
Nothing inspires great people to work at a company better than other great people.
Use contractors for areas outside core of company.
* KP: our contractors are largely in India, invested a lot in relationship with Cognizant in India, they are part of the team, we were formed as a global startup from day one, we are sourcing and selling code globally, from day one
* JB: venture capital, being thrown under a bus
* JK: we're venture backed, $5m via redpoint and mayfield. Going good [like he is going to criticize his VCs in front of the conference?]
There's a set of business that need VC and a set that don't.
* KP: we were formed out of a VC firm, great board, a lot of the challenges or horror stories have to do with the partner you get
* JB: who should take, who shouldn't take VC?
* JK: spent some period of time talking to entrepreneurs. don't think how taking VC can reduce your chance of a successful outcome. Clear matter of economics or supply/demand.
A lot of what I see are features or products wrapped in company's clothing, they should definitely not take VC since their exit strategy is to be bought.
Far more buyers at the $1-$5MM level then the $30-$50MM level demanded by VCs.
The VC is not the prize, it's the wrong thing to focus on and can be harmful to your outcome.
* JB: Kim, the arc of a Kleiner backed startup is well understood, are you on that arc?
* KP: It's not on our mind [IPO]. One of the biggest value ads for KP is an understanding of what we're trying to do is solve a big problem, not build a billion dollar company.
We talk about scaling the company, our support model, what we get is a lot of great input. They are also interested in building a new model for how you scale a software company. Emphasis is on getting it right.
* JB: Joe, did you enjoy excite going public?
* JK: The private piece of the market feels very frothy now...it's a bootstrap froth. Not necessarily a bad thing, it can be a problem if you inject a lot of money into these companies.
* JB: Bootstrap fratricide?
* JK: it's so cheap to start a new company these days, it took $100k for jotspot to go from idea to market. Costs are going down or to zero.
Ability to access offshore labor as opposed to in 1990s when only IBM could offshore.
Fourth thing is search engine marketing.
Venture Fratricide was when multiple companies would go into the same market trying to use the same momentum and end up killing all off.
Being cheaper to start a company causes its own problems but nothing to avoid it.
* JB: Is it different this time personally? Or is it more crazier?
* KP: I joined a team of seasoned entrepreneurs. We take things in stride that were curve balls in the early days with Marimba.
People have families, not burning the midnight oil like we did in the 1990s. Don't need to be killing yourself to build a successful company.
You derive energy back.
* JK: I'm equally neurotic and paranoid as I've ever been. The nervousness, the desire, the drive, the paranoia,
all those startup qualities, I think I'm wired that way. The want to make this successful. I'm no more calm than
the early days of Excite. But I think that's the way I am.
h2. Show Me/High Order Bit/UI Minute
h3. Rolf Herken / founder Mental Images
* created Mental Images in 1986, specialize in component technology for visualization
* OEM business model
* see a need for 3D on the web
* many applications do not require realtime interaction but do require interactive access to original 3d data
* challenge in that complex data is often measured in Gb
* data is often valuable assets to be protected
* mobile clients can't handle such large data
* currently no collaborating except by changing files
* introducing server based 3d via RealityServer platform
* allows integration of 3d applicatoins into enterprise software and web service environments
* documents become the UI for 3d content
* becoming easier to deploy because of saturation of broadband
* scalability a problem
h3. Jim Lanzone / Ask Jeeves / High Order Bit
* A look at statistics of search and feeds
* reports available online at bloglines or ask.com
* Only 5% of searches can be classified as popular
* most of the revenue is made in the head, in the tail there are words but no one is competing for them so they cost less
* 30% of searches drive 70% of revenue
p. Update: "Powerpoint version":http://blog.searchenginewatch.com/blog/ppt/web2_aj_bloglines.ppt , "Online version":http://blog.ask.com/2005/10/deep_thoughts_a.html
h3. John Kish / CEO Wyse Technology
* Thin computing for a flat world
* use in rural economies, third world countries
* solid state devices that sit on the edge of the network. no moving parts, cheap to mfg
* challenges apply in not just enterprise but a wide variety of markets
* easier to manage
h3. Satish Dharmaraj / Zimbra
* Demo of zimbra (again) www.zimbra.com
p. break for lunch
Posted in Conference Notes