Thursday, June 29, 2006

The Sociology of the Sociology of Science

Before I dove into the startup life in '96, I was a PhD student at UC Berkeley in the history of science program. I overlapped there for a year or two with a smart guy named Alex Pang, who now works at the Institute for the Future and writes a blog called Relevant History.

As opposed to his regular interesting posts, Alex recently made a [thankfully rare] comment on academic sociology of science, which caused him to quote this passage:

[C]onstructivist sociology of science offers case-based analysis celebrating contingency and locality, favors archival and ethnographic methods, emphasizes agency over structure, and often focuses on issues related to epistemology and knowledge. Neo-institutionalism, on the other hand, searches for patterns over time and space, is more enthusiastic about using statistical and quantitative methods, emphasizes how structure can constrain actors, and returns in part to a sociology of scientists and organizations that was more characteristic of the pre-constructivist, Mertonian era.


Ye Gods. That's just painful to read. I don't know who wrote that, but they're never, ever going to be able to communicate with normal human beings unless they start speaking English. I thought it would be interesting to do a quantitative comparison between that ...stuff... and actual prose. Alex is a fully trained SoS practitioner who's broken free from his chains to rejoin the human race, so I compared the academic passage to one of his recent posts that I consider a favorite:


As an historian interested in Silicon Valley, I'm fascinated by Castilleja. More than any other place, it's given me a sense of just how tightly-knit the area's elites are: for all its global reach and influence, the Valley is still a bunch of small towns, knit together by schools, churches, volunteer organizations, and all the things that turn groups of people into communities. I come to events here, and the parents include people who were recently on the cover of Business Week or Wired. The real power scene isn't Stanford or Sand Hill Road; it's the line of parents waiting to pick up their kids from Menlo or St. Joseph's or Casti.

The students are in their dress whites this morning. Since they're normally in dress blue, it actually isn't that much of a step up in sartorial splendor, but it's a nice gesture. It's hard to raise a bar that's already high. Friedman is pretty high-profile, but the school has a constant parade of Macarthur genius prize-winners, Nobel laureates, people who used to have Secret Service details, and other generally fascinating people who come and speak to the students. (It's hard to raise the bar....) If, as I've sometimes heard, middle school here is like high school in most other schools, the speaker series rivals that of many colleges.

Introduction.... actually being done by one of the students. That's cool. She started an NGO that does relief work in Africa. Okay, that beats my vice presidency of my high school chess club (two years in a row). Would my high school record even get me into community college today?

Friedman's up now. Some anecdotes... why the world is flat... some more anecdotes... references to Lexus and the Olive Tree... I've posted the rest of the talk on Future Now.

Q&A is restricted to students. Love it: I recognize half a dozen CEOs and VCs, people who normally are the center of attention, and they're sitting in the back, listening. And not grumbling. Here, they're just parents.


I don't have to tell you which is better by any rational measure. But here's a numeric measure:

* Passage 1: 509 characters, 77 words, average word length 6.6 letters.
* Passage 2: 1625 characters, 337 words, average word length 4.8 letters.

Winner: Alex.

Social science grad school: Add 37% to your word length and remove 99% from your prospective audience. And it only takes away four, er, six years of your life.

Wednesday, June 28, 2006

Zlango: Didn't we solve this problem 3000 years ago?

Techcrunch is reporting on the launch of Zlango, an SMS icon service from Israel. According to Mike:
"Zlango...has created a very interesting new language...that could change the SMS landscape. The language is based on icons, or pictures. Each icon has a specific meaning - a person pointing to himself for 'me' or a heart for 'love'. There are over 200 icons included in the Zlango language today."



Didn't the Phoenicians solve this problem 3000 years ago by inventing the incredibly flexible and efficient symbolic alphabet? Because it sure looks like hieroglyphics to me.

Zlango's millenium-busting innovation:


State of the Art:

Thursday, June 22, 2006

Zvents 2.0 is launching today

By the time you read this, Zvents 2.0 should be up and running.

Right now, hitting our site gives you this message, a portent of good things to come:



What's new and improved about this version of Zvents?

* Improved search relevance -- when you've got thousands of events in a metro area, relevance becomes a key to happy users. Our results are materially better than before, and we're committed to continuous improvement for relevance.

* Category navigation - refine or browse to events by category, reducing time and effort to find what you're looking for.

* Embeddable calendars are now CSSable version 2, vastly improved - a live example of a 1.5 version is Visit Marin.

* Much faster, lighter code -- less than 1/2 the size -- means faster response times.

* Improved navigation and look and feel

You'll also notice that we've got less features, not more -- we're running against the "features are king" silliness of Web 2.0 here, but we think that great design means focusing on what matters most.

Coming in the next 30 days - major customer announcements, plus bonus fun stuff!

Thursday, June 01, 2006

Lunch 2.0 at Zvents -- the aftermath!

We got a huge turnout for Lunch 2.0 at Zvents on Wednesday - thanks to Mark Jen and the guys at Plaxo for creating this cool event. I would guess that about 40 people turned out - we bought burgers and sausages for 44, and the last two people to eat had to split the final one! sorry, guys :-( But aside from a slight demand-supply mismatch on the grilling front, it was a great place to meet folks, chat about startups, and enjoy the awesome weather and great back yard of our friendly incubators, NetService Ventures.



All the pictures are here.

Zvents powers Palo Alto Daily News

My startup, Zvents, has some cool media partnership news. Check it out!

Tuesday, May 23, 2006

Dear Google: Arabic Gmail to estock is spam

I like Gmail, and I use it as a personal account for this blog and other purposes. One of the things I like best about Gmail is the one-click ease of reporting spam and having it vanish from my inbox, with the sense of accomplishment that due to my small contribution, the Google anti-spam terrier is getting sharper teeth and a nastier disposition.

Except the terrier seems to be sleeping. Unless I am blind to some subtle i18n character set issues, it should be really, really easy for the Bayesian filtering algorithms over at Google spam central to figure out that hey, *every single* Arabic email that estock has received, he's classified as spam! And look, *every single* cyrillic and Mandarin character set email he's received, he's classified as spam as well! Hmmm. What could we do with this information? I think that just possibly, entropy could be reduced a little by incorporating it.

Ten or so of these a day, every day, get old. No, I do not want to hire any stock brokers from Dubai, even when they write to me in English -- and especially when they write in Arabic. Really. Hey, spam terrier -- bite 'em!

Lunch 2.0 at Zvents -- Wednesday, May 31st noon

Come join us for BBQ in the back yard of Zvents / NetService Ventures -- part of the Lunch 2.0 series started by Mark Jen and the guys at Plaxo. Read about the whole Lunch 2.0 series here.

RSVP by commenting on the Lunch 2.0 web site, or by adding a comment to the zLunch event page: Zbutton

Parking will be tight, so please carpool, or park at the Sharon Heights Safeway, which is about a 5 minute walk west on Sand Hill Road (turn at the light at Sharon Park Dr.)

We'll provide all the food and drink, you provider the conversation!

Sunday, May 07, 2006

Daniel Yergin's "The Prize" - History of Oil


I was chatting with one of my self-identified libertarian friends tonight (he started out the conversation with "were you referring to me in that last blog post?") and I ended the evening by recommending Daniel Yergin's fabulous book The Prize to him. It's a history of oil from its discovery in 1850 through to the first Gulf War in 1990, and it won a well-deserved Pulitzer. It's a huge tome (800 pages, I think) but an outstanding history of the single most important factor in world history during the past century and a half. If you read it, you'll be amazed how the thread of oil runs through the waning days of imperialism, both World Wars, and the Cold War... it really has been the fulcrum of power in modern times. Highly, highly recommended.

Airlines vs. Fuel Costs = Dirigibles?

I just read a fascinating article in the New York Times about Boeing, Airbus, and the future of commercial aviation. Boeing, as airplane geeks will know, has bet heavily on the 787 'Dreamliner', a highly fuel-efficient midsize plane designed to serve point-to-point international routes -- San Francisco to Shanghai, Athens to Boston. Airbus, on the other hand, is building the massive A380, a plane designed to carry 500+ passengers on core 'hub and spoke' routes such as New York to Tokyo, with smaller planes serving as 'feeders' to these routes.

This multi-billion-dollar faceoff between two different philosophies and infrastructures is reminiscent of so many other distributed-vs.-centralized battles. FedEx created a business by going with centralized hubs. "Mesh" networking is supposed to be more efficient and resilient by getting rid of hubs. However, it's clear in this case that one simple calculus will decide the fate of Boeing vs. Airbus:

Fuel cost vs. customer happiness.

As recently as 12 months ago, I would argue with friends about 'peak oil' and what an economy looks like as it slides down the back side of a bell curve of production; and I would get signficant pushback that we were anywhere close to such a circumstance. The more libertarian-minded were fond of citing the famous Simon-Ehrlich bet on commodities as a supporting point.

And here we are. Massively rising demand for oil from China and India; potential disruption of significant sources of production ranging from Venezuela and Bolivia, Nigeria and Sudan, Iraq and Iran, to the Khazak basin in the former Soviet Union. No massive new sources of oil discovered, and few likely to come on line soon. And wild cards such as the Canadian tar sands are predicated on... high oil prices to make them relatively cost-efficient to produce.

houston dirigible postcard

Which got me to thinking: If the tradeoff for airlines is fuel efficiency vs. customer satisfaction, why not dirigibles or blimps? Right now it takes about 10 hours to fly the 6000 miles from SF to London, at about 600 miles per hour. An appropriately designed dirigible could do it in 24 hours at 250 miles per hour, at a vastly (90%?) reduced fuel cost -- since a dirigible would benefit both from the cubic reduction in power-required vs. speed flown, and the absence of the need to expend power to keep the aircraft up in the air, which accounts for a large percentage of airplane fuel cost. Imagine that, instead of spending 10 hours on a cramped, noisy, EXPENSIVE airplane, you spent a full day and a full night on a quiet, spacious, dirigible? Broadband internet access would be essential -- not only could you make crystal-clear phone calls, but you could transfer any volume of data. You'd get nice meals from a large kitchen. You could walk around and exercise. You could sleep in a real bed. And in a world of $70 - $140 a barrel oil costs, all of this might be CHEAPER to provide than a miserable 10-hour flight.

The key, of course, is making the experience as business-like for business travelers, and as vacation-like for vacation travelers, as possible. Without doing some considerable aeronautical math, I can't estimate at what fuel price point dirigibles would start to make sene. But it's certainly fascinating to consider.

Wednesday, April 26, 2006

Portfolio Management, the Yale Way


Jason Zien over at Spotsearch recently wrote a post about Yale's asset allocation strategy, which has garnered amazing results - read Jason's post for the details. In about 2001, I read a book by David F. Swensen, the Chief Investing Officer of Yale University, called Pioneering Portfolio Management. It is one of the better investment books that I've ever read, and I highly recommend it. Swensen apparently has a new (2005) book out, "Unconventional Success," which is also well recommended on Amazon.

Tuesday, April 25, 2006

Ponytail Sweepstakes: Ted Waitt, Gateway

I've gotten a few emails based on yesterday's post about Jonathan Schwartz. It has been brought to my attention that Gateway founder and ex-CEO Ted Waitt was ponytailed for much (or all) of his tenure at Gateway:



There's a great quote on Waitt in a recent Fortune article, that foreshadows Sun's challenge:
"But while Waitt, the ponytailed visionary, was looking toward the future, boring old Michael Dell was obsessing over efficiency, thinking about how quickly a PC could be assembled rather than how fast it ran. As it turned out, execution was everything."
The ability of the Internet to quickly and easily answer idle historical questions such as "what was the market cap of Gateway during Waitt's tenure" is still limited, but a quick scan of the stock price suggests that it may have been as high as $28 billion circa 2000. This would significantly eclipse Sun's current $17B number under Schwartz.

The race is on, Jonathan!

Monday, April 24, 2006

"What's a Blog?" - Six Apart at Maker Faire

This was my absolute favorite sign at Maker Faire. I promised Elizabeth (the girl in this picture) that I'd blog about it, so here it is.



There was lots of other cool stuff at Maker Faire, including awesome mathematical castings from Bathsheba, which embarassingly I failed to photograph. Oh, and I-Wei Huangs's steam robots!

Scott out, Ponytail in, shares up!

Sun send Scott McNealy off into the sunset today, after a run of 22 years at the helm. I can't think of too many other CEOs who have run a company that long -- though I bet Michael Dell will get close.



In related news, Jonathan Schwartz is CEO -- does anyone out there know of a CEO of a company larger than Sun run by a ponytail-sporting guy? He may have the market-cap record for that particular feature.

I know a lot of guys rooting for him to hold the "return on shareholder equity" record for ponytails, too.

Search Trends: The Search For User Intent

There are several interesting trends afoot in the world of search. As I have previously written, concepts as disparate as editorial news and personalization are simply different ways to get to a better search answer. I'm now ready to extend this thesis a bit further, and claim that the next big leap in search is going to be driven by a superior ability to acquire and understand user intent.

In any query:response system, the quality of the output is fundamentally limited by the information contained in the query. In a mathematical sense, it's impossible to have more significant digits in your answer than you had in your input data. No matter how well you tune your response algorithm, not only will "Garbage In, Garbage Out" always hold true, but the more general case of "lack of precision in, lack of precision out" will limit how good your results can be.

On a general search engine (Google, Yahoo, etc.) an inbound user query can, quite literally, be about anything. Presented with a blank box that can encompass anything, the user then types in an average of 2.4 words, and the search engine must give back an ordered list of 10 relevant results (having worked very hard to remove spam, porn, etc.) that may approximate what the user is looking for. Suppose the user types in "Peru". Is the user a 4th grader writing a report for school? A college student planning a backpacking trip? A consultant searching for economic data related to a project? A Peruvian looking for things related to their home country in their local area? All these, and many more, are possible.

It's a very tough problem, and a key reason that search quality has ground to a halt, or at least stopped improving dramatically, is that in this completely general context, it's very hard to understand more about the user's intent than they grudgingly give you in their 2.4 words.

So how can search engines solve this problem of paucity of intent?

Answer #1 is to verticalize. Zvents, my startup, is an example of this trend -- as are Simply Hired, Kosmix, and many others. When someone types Peru into Zvents, we know -- simply because we're a search engine for local events, and nothing else -- that the user is looking for something to do in their local area having to do with Peru. In order for Google to get that kind of intent, the user would have had to type, "something to do in my local area having to do with Peru" or a similarly complex query. Zvents gets a great deal of intent information simply by being a specialist. When someone shows up in our search interface -- either directly, or via a media partner -- we can be confident they're looking for stuff to do. Simply Hired, similarly, would know that the user was looking for jobs either in Peru, or related to it; and Kosmix, under its healthcare filter, could infer that the user cared about health issues related to Peru.

Answer #2 is to personalize. Google is the most interesting case emerging at the moment. Go to the Google homepage and in the upper right you'll find a link to the new personalized home. It's a HUGE departure for Google -- who has built a multi-billion dollar business on a purely "visitor" experience -- to be moving to a registered user experience. It's likely that Google is doing this is to get a much clearer sense of search intent, based on personal past search history. At least one very senior technical person involved in the Google personalized home page has a background in mining extremely large log files to extract business intelligence -- exactly the kind of resume you'd want if you were attempting to derive intent from search history. I've also had some discussions with folks at Yahoo (who are also very happy to let you log in while searching) about the extremely sophisticated state information that Yahoo maintains as you move through the Yahoo portal; all aimed at improving the targeting both of ads and products.

Answer #3 is to get more intent directly from users. Many of you might recall the original concept of Ask Jeeves, which was to "answer your questions in plain English". In a recent conversation with some folks at Ask, they mentioned that one of the reasons they'd retired Jeeves was that delivering on this "brand promise" was nearly impossible - but that because of that historical promise, the length of their typical search query was vastly greater than comparables at the other big search engines. That begs the question -- what if you could successfully parse natural language English like "something to do in my local area having to do with Peru"? It's an incredibly hard problem, but the ability to accurately extract enough user intent to deliver a truly better search result would yield enormous market benefits. I know of at least one current serious search startup that is taking exactly this approach -- building a general-purpose search engine that can better parse complex intent, and accurately respond to it. If they succeed in their quest, and establish a "brand promise" similar to the original Jeeves, they'll have a significant advantage in actually giving users what they want.

There are other answers, but they look less and less like search. Aggregate Knowledge, for instance, is a matching engine for supplementary navigation based on heterogenous item:item matching in a "people who viewed this, also viewed that..." sense. At scale, AgKnow can derive intent from the cumulative behavior of multiple users who were exposed to similar information -- and as their scale increases further, they could slice even more finely to begin to distinguish path dependencies as well. I think we'll see an explosion of search alternatives in addition to advances in all three categories mentioned above, as dozens of companies large and small strive to crack the problem of too much information, not enough user intent.

Tuesday, March 07, 2006

What does Web 2.0 mean to me?

I was asked this question for a podcast at the recent Under the Radar conference where Zvents presented. You can test my memory by hunting up the actual podcast, but I am pretty sure this is the same answer I gave:

It means two things -- one technological, and one social.

Socially, it's the return of hope and enthusiasm to Silicon Valley. People may mock "bubble 2.0" and naysayers can always point to some particular point of excess, but the reason we entrepreneurs are here is to change the world, and a lot of dreams lay dormant from 2001 to 2005. Those dreams are being knit into reality today, and many of us can see that we'll be able to make the world a very different, and hopefully better place, soon.

Technologically, it's about data as a platform (just like Tim O'Reilly said) and more importantly, it's about fast and simple integration of that data. Mashups may occasionally be dorky, but they demonstrate a level of interoperability that is simply astonishing when compared to 1999. We've all seen network effects accelerate change and growth to warp speed -- what we're seeing now is kind of Metcalfe's Law of web functionality, where the value of every interoperable web service is the square of the number of other services to which it can be connected.

Hmmm. I know I didn't say that one in the podcast. It may actually be worth its own post.
But, back to work. Big customer meeting tomorrow...

...Web 2.0 means the return of customers. Yee Hah!

Aggregate Knowlege launches at ETech

I may be busting the press embargo by 90 minutes or so, but I'm really excited about this one. e, Tomorrow at ETech, Paul Martino and Chris Law are taking the wraps off of Aggregate Knowledge, their fascinating new play on a recommendations web service.

Paul and Chris were formerly founders of Tribe Networks, and Paul (several lives ago) was a graph theory PhD student at Princeton. They've cracked the problem of providing user-specific recommendations across multiple data types in real time, thanks to some seriously fancy math under the hood. We've been using AgKnow at Zvents to power some of the "related events" content on our site (Disclosure: I am also an investor), and look forward to more customers joining their service so we can begin to benefit from further network-effect driven recommended content.

Most importantly in this mash-up world, it takes from minutes to a few hours to fully implement their service on your site. EAI is soooo dead.

I expect them to do very well in both the content recommendation and non-search navigation business. Rock the house, guys!

P.S. I previously mentioned these guys in November.

Technorati Tag: ETech

Monday, March 06, 2006

Wednesday, March 01, 2006

Ether *is* Keen 2.0: Old Wine, New Bottles

Over at TechCrunch, Mike just posted on Ether, a new company driven out of the pay-per-call innovators Ingenio. My immediate thought was, "is Mayfield funding this company?" because Mayfield is (in)famously trolling through old bubble business plans for ideas, and this sounds an awful lot like Keen.com.

Keen was funded by Benchmark around 1999. I first saw gigantic billboards on the Underground advertising Keen in 2000 when I was living in London, and I remember thinking, "Damn, that's a great idea," as I stared at them from the platform across the tracks. It may still be a great idea, but Keen was relegated to a fairly painful fate, morphing into a psychic hotline and not much else. And certainly nothing that could make back the bubble-zillions that got poured into it.

In poking around to check out the carcass of Keen, I discovered a most curious fact:

Ingenio (Ether) *is* Keen! Changed their name sometime in the last seven years. Look right down at the bottom: "Keen is a trademark of Ingenio, Inc."

What was old is new again...!

Monday, February 27, 2006

Feed reader creators: There is so much left to do

I have a confession to make. Here I am smack in the middle of Web 2.0, a founder of a site that pumps out more RSS than you can shake a stick at, and I don't like feed readers. In fact, until about two weeks ago, I had never really used one, save a single desultory experiment back in 2005.

(Gasp!)

Yes, I have been quite embarassed about this affliction. I must admit, I have hidden it from my peers.

"Hi, I'm Ethan, and I type in URLs by hand."

I felt so bad that I finally caved in a few weeks ago and submitted to Bloglines. At least I could get an OPML file out of it, I thought... since Dave Winer assures me that will be useful soon, and Kevin Burton swears it will help me find all sorts of cool stuff on the web. I heard rumors at the last TechCrunch BBQ that it even attracts girls, but I've already found one of them.

But I couldn't stand it. Sure, it's all my reading in one place, and the atom feeds look sort of like their source blogs, but it's all kind of... attenuated. Like sucking Coca-Cola through a long straw, so the fizz is gone. All the formatting is replaced with grey-on-white arial. The fascinating sidebar tidbits and blogrolls and, yes, ads all disappear. And the urgent boldface of posts as yet unread, marching down the left sidebar like dutiful ants...

I thought I was alone.

And then... I found another! There in the endless grey arial sat a wizened little scrap of text, from AVC, blog home of none other than Fred Wilson, who, if he isn't the patron saint of Web 2.0, is, at a minimum, merry Mercury in its Pantheon. And what sayeth Fred?

"I Prefer Browsing To Reading Feeds"

Hallelujah!

Me too, Fred! In fact, this may be the most important post you've ever written. It's supportive of an inclusive and exploratory, rather than exclusive and reductionist, view of the web. While it calls out feed readers explicitly, it also criticizes by implication all the other exclusionary, reductivist drivers in this fascinating new media landscape -- our tendency to read the same old blogs; our algorithmic focus on the same old hubs and authorities; our analytical adherence to the same old scales and dimensions, and most particularly, our (and MY) fear that somehow, by being human, I was wrong.

A quote from Stephen Pinker's book (yes, an actual paper book!) The Blank Slate:
"The belief that human tastes are reversible cultural preferences has led social planners to write off people's enjoyment of ornament, natural light, and human scale, and forced millions of people to live in drab cement boxes."
I love blogs. I love the conversation, and the chase, and the delight of discovery. I love the humanity of user-created content, and I don't want to aggregate and attenuate that away in some misguided attempt at rational efficiency.

Is BuzzMachine the same without seeing that giant press picture at the top of the site before reading it every day? It reminds you, the reader, of the massive machine that is MSM. Like the three tones of NBC News at dinnertime; like Pavlov's dog; I salivate before dining daily with Jeff Jarvis.

Give me that ornament, and human scale, that connection with the writer and their carefully tweaked site, and at the same time make my life easier - and you'll have a dedicated customer. But until then...

...this new/old human-based web needs to remember us humans.

Feed reader creators: There is so much left to do.

Great new maps from Ask.com

Ask announced that they're going to fight for search, and launched some cool new maps functionality. Much of what you see here is a quality re-implementation of Google Maps, but there are some cool new features - for instance, Tyler pointed out that after you enter a route, you can click a "play" button and Ask will walk you through the steps onscreen.

Yawn, you say. Google did most of the useful bits a year ago, and Yahoo more or less matched this last fall.

True.

What I find fascinating is that the leading edge of search is all happening at startups, and that the offerings of the big boys are increasingly undifferentiated.

Why isn't Google building Plazes on top of Maps? Why isn't Google building Zillow on top of Maps? Heck, why isn't Google building Zvents on top of Maps?

It's a very interesting time to be a search startup. It feels like the calm before the storm -- with an absolute explosion of functionality coming very soon.