Archive for March, 2007|Monthly archive page

Etech Conference

I’m at the opening session of this usually excellent conference. In the executive briefing today and will post my thoughts and notes on this later, so far there have been several companies up that I’ve seen before and am familiar with, not as stimulating as I thought so far.

One new thing, or at least a new product from an existing concept is Atten.TV which is based on attention trust, Mac only. Not sure I see the difference between it and Slife so far but not had a chance to look too closely. More later….



Etech Executive Summary

On the first day of Etech I attended the Executive Briefing session in order to get an overview of the important themes coming up over the following three days.

The format of exec briefings is one I like, more about discussion than simple presentation hence it was relatively ‘ppt lite’

The On-Demand Manufacturing Revolution

In the first session some widely ranged topics are discussed, at one point Tim wonders is the meme of MAKE is getting ‘old school’ in the same way Rock music has over time, and as in fact anything does over time.

Dale Dougherty responds that it’s not just kids, geeks and retiree`s looking for something to do now

One Issue with the last few years is that people can are decreasingly able to fix things themselves, 30 years ago everyone could fix a tractor or their car, now the systems are simply to complex for most to consider, the rising trend in making your own equipment is putting the power back in to peoples hands to repair things

Kite level mapping is now being done, it’s possible mainly because the cost of the cameras has come down so much that it’s ok if the kite crashes and you lose it (later in the conference Microsoft showed the Ultracams they use for virtual Earth data, you wouldn’t want to loose one of them)

Brian Warshawsky from Potenco runs us through the design process and iterations of the hand crank charger they developed for the OLPC project, the main advantage now is that you can make a design in 3D software, send it to US companies and have the parts back within a week, this allows for a much faster process, it’s also important he felt that they built each iteration themselves so that they could amend the designs. They are also using manufacturing in China now where costs are less but it takes two weeks

There follows a quick overview of the OLPC project and the current model is shown, it’s the same one I saw at Kiwi Foo as far as I can tell. The hand crank they have developed will power it for several hours due to the extremely low power consumption it has, unlike a Macbook which you wouldn’t be able to charge (they also don’t charge on planes as draw too much power), he wonders if current laptop makers are paying attention to this, more on energy later

The next speaker to join this discussion is Andrew ‘Bunny’ Huang from Chumby industries. Tim uses a popular description for the Chumby ‘the next-generation wifi-enabled clock radio that you can “hack with a seam ripper”‘ this is due to the fact that the hardware is encased in materials which can easily be modified to your tastes, the one I got at Foo Camp is only recently working but I’ve always liked the concept, Tim likes it so much he says that he keeps it in his toilet, the newspaper being replaced by technology once again.

And finally John Hagel joins the discussion, he is giving a talk on “The On-Demand Manufacturing Revolution later but the main point he makes is that we are in an age where we print circuit boards including the electronics and one everyone can do this then the playing field changes

From Manufacturing to The Attention Economy

This session starts off with an overview of Threadless, Threadless is a user-generated-design, on-demand manufacturing business. Users submit designs for t-shirts and other items, other users vote with their wallet, and when there’s enough demand, the products are created. Threadless has sold out of every product they’ve ever created.

The key point from this is that they allowed a community to form around a shared interest and built trust with that community from the ground up as opposed to a a Big brand down approach, in fact they have been approached by several major brands that want to be successful on Threadless, however they haven’t done anything since it’s not what Threadless is about.

On top of this is the question of why users work so hard designing and running a community for a company they don’t work for, in a similar vein the next speaker Seth Goldstein introduces us to his latest product

Built on the ideals of AttentionTrust (which he founded) and the collection of users Clickstreams the idea is that one persons Clickstream is another’s entertainment. As more people expose more data about themselves at the same time as giving away more privacy the question becomes where is it all going? Seth thinks that is one part of the answer, it’s another way for people to broadcast themselves (the tagline of YouTube of course), while he as showing us some early screens I felt that this was going a step too far, for me at least, I don’t really want to broadcast every webpage I visit, there are controls built in to his product but still, I send URLs I think are interesting to still which is as far as I want to go for now. Seth goes on to explain that with big corporations can now see what users are doing explicitly and therefore can offer much more targeted advertising, or they can subscribe to a group of people to get a better idea of what they are interested in. Some of these principles are covered in my Personomy paper. It also feels similar to Slife

Jeff Jonas gives us some background on his company and the technology they developed called NORA. He is also delivering a keynote on the same topic. The basis of what Jeff does is a new technique enabling advanced data correlation while only using irreversible cryptographic hashes. This new capability makes it possible for organizations to discover records of common interest (e.g., identities) without the transfer of any privacy invading content. This privacy-enhancing technology known as anonymous entity resolution delivers extraordinary new levels of privacy protection while enabling technology to contribute to critical societal interests like clinical health care research, aviation safety, homeland security, fraud detection and identity theft.

Without going in to the specifics which are very in-depth I could see this sort of technology being very useful if applied across the vast databases of users across FT.

Ain’t Dead Yet: Skyrider and the Next Gen P2P Economy

The references to Skyrider are the hook here (not the worst example of ‘exciting titles’ though) but the basis of XX discussion was around how he belives that P2P wil be the next application, so far it has had a bad reputation by association with Piracy however most people have wrongly labelled the technology/protocol with the applications that people chose to build with it. That statement reminded me of the IMS Debate at Etel and how many people are doing the same thing there, mixing up technology with intention.

The proposition here is that his company are developing the ability to provide search based marketing around P2P searches performed across the major networks, this enables them to create targeted advertising based on popular searches (reminiscent of above) They are able to do this by becoming a supernode in the relevant network and collecting the queries ‘as they go by’

We finish with a fairly open statement around the fact that as Broadband penetration grows and processing power becomes increasingly cheaper that (and using P2P to avoid and client/server infrastructure costs “the kids will be doing amazing stuff”

Energy Innovation

Tim opens the post lunch session with a joke about the Irony of moving on to Energy while suffering post lunch Torpor

However the topic itself is an increasingly important one, both politically and culturally.

New data centres are required to support the ever growing need for data in the cloud, Web2.0 is responsible for this in many ways, Amazon and Google need more data centres, or more precisely more data centre space and they are looking very hard at how to reduce the costs, it’s not so much about cpu power but efficiency now.

Climate issues are well versed currently but what are the suggestions to resolve this? O’Reilly are putting on a new conference dedicated to this theme as it’s ‘on the Radar’

During the course of the discussion which revolves mainly around US power consumption and issues (related to any major industrial company though) some new thinking is proposed (new to me anyway)

One idea I liked is P2P power distribution, if I got the idea right the suggestion is to allow people to plug in their cars when not using them so that the stored energy in the batteries can be distributed back out over the grid, the more you contribute the lower your bill, in fact could there be a time when you wouldn’t have a bill? However I suspect it may never get there dues to the inherent nature of some P2P users, they leech and never return, however if the business rules are well thought out then this shouldn’t be a problem.

One thing that struck me is how current grid distribution is built around the maximum capacity model, for the peak times of usage, and how P2P distribution may flatten oout that those peaks and mean less infrastructure costs, the same model can be applied to Mobile networks although I’m not sure how technically that would work just now.

Investing in Data Center Construction

The discussion moves on to what the data centre of the future may look like with the ever growing needs of Web2.0 and the explosion in User Generated content in mind, one point made is that since the 9/11 disaster and the effect it had on the financial systems that there are also additional regulations that have to be born in mind.

Tim asks whether we will see new towns growing around these new data centres like we did in the past around power stations, the answer is it is happening already in the case of the new Google centre for example but to a lesser scale perhaps

The Q&A session revealed one additional concern for new data centres over and above the power consumption which is weight, not such a concern in the US though where land is in ready supply but more for Asian markets

Metaweb: The Semantic Web Meets Web 2.0

Robert Cook runs us through Metaweb, the mission is to create ‘World Scale’ databases, the first instance of which is available now as Freebase. Freebase aims to consolidate massive datasets and treat them as Topics, the example used was Arnold Schwarzenegger, there are 3 main datasets already, as Body Builder, as Film Star and now as Governor, Freebase brings all this data together as a Topic, the main advantage of which is the creation of one place to search for any data. Once again the Skynet reference is made since one ‘machine’ will ultimately hold all human knowledge if Freebase continues to grow.

Metabase provides an API to write on, MQL API, and all applications are built on this API, in the same way Flickr does with its API.

Their music info comes from MusicBrainz which brings up another key point, they want to release the information from existing data silos (rather than license CDDB they went with MusicBrainz) and hope to do this with other sources.

The issue of data spamming came up, it’s always a risk but Wikipedia is held up as a good example, the session ends with some very cool demos of services built on Freebase, I look forward to its more general release.

Web 2.0 and Wall Street

Peter Bloom and William H. Janeway hold a discussion around the financial markets and how Web2.0 relates.

Todays trading systems are now doing trades at 1 per 30 milliseconds whereas only a few years ago they were in the 9000 range. Give up using a mouse to do trades is the quip.

New Direct market access systems are also at the 1000 requests per second speed now meaning a trade can be done in a single millisecond.

One key point I picked up in this conversation, which was mostly beyond me in that it used lots of financial talk, is that latency is compressing, society in general is less tolerant of latency, this means new tools are required that take advantage of reducing latency, however if latency is almost zero and tools are created to take advantage of this then there needs to be conditions built in so that human intervention is possible, i.e. humans are still required to make sure a decision path holds true and these systems need circuit breakers in them

Backing Up Instinct with the Numbers

Hitwise track internet usage to provide market intelligence, the early part of the session related some tales of prediction that were very US centric (Americon Idol) so I tuned out.

However the interest came back when Bill Trancer started to describe how they track the young Digerati and where they go on the ‘net, effectively the early adopter crowd, those that jumped on MySpace, Flickr and others, this leads to the question of where they are now as many have moved away from those services.

Using their systems they pulled out the following sites:

Imeem – an online community where artists, fans & friends can promote their content, share their tastes, and discover blogs, photos, music and video.
Veoh – Your Internet television network
Wikimedia Commons – a media repository that is created and maintained not by paid-for artists, but by volunteers
Metacafe – serving the worlds best videos

Interesting to see that all of them are content related with a touch of social networks….

Roger Magoulas wrapped up the session with some analysis of the trends in computer book sales, how terms change over time with the popularity of certain coding languages, I sadly missed most of this due to an ongoing conversation re the best sites part of the talk..

The exec breifing gave a good overview of what to expect from the full conference, some more thoughts on this in the wrap post I plan, at this point though I felt that there was more emerged tech on show than emerging, as usual though it is always hard to judge what level to pitch these talks at, however the group for Etech is pretty self selecting and mostly technical or geeky in nature, in which case the level pitched so far is not right, the next few days will tell though