February 14, 2003

Google and Larry Page

Google: 25000 computers all held together by velcro. How much--one cent? ten cents? twenty-five cents?--for each ad click-through. How many searches a day? And now they cannot be unseated from their king-of-search position unless someone has a truly much better software idea...

Scattered notes taken that have very little to do with the substance of Google founder Larry Page's talk:


"Inkjet printers made of legos" "Disk drive cases made of legos" "Actually, they weren't even legos. The point was to save money by building cheaper disk-drive cases. They were knock-off duplos from CostCo."


Larry Page: "Keynote would be really outstanding if you had a fast machine to edit your presentations on." Smart-Ass: "A machine faster than those at the disposal of the founders of Google?" Larry Page: "You know what I mean: a machine faster than this laptop here."


Larry Page: "Google has been profitable since the first quarter of 2001. Why did we make becoming profitable such a priority? It's good that we did, because we might well be gone if we hadn't. The real reason is that we became profitable in the first quarter of 2001 because Sergey Brin made it a priority. You see, Sergey would try to go out on dates. He would call up women. And to impress them he would say, 'I'm the president of a money-losing dot-com.' But in Palo Alto in 2000, a huge number of people were presidents of money-losing dot-coms. And so they would not call him back. And he thought, 'If only I were president of a money-making dot-com, things would be very different...'"


Larry Page: "It wasn't that we intended to build a search engine. We built a ranking system to deal with annotations. We wanted to annotate the web--build a system so that after you'd viewed a page you could click and see what smart comments other people had about it. But how do you decide who gets to annotate Yahoo? We needed to figure out how to choose which annotations people should look at, which meant that we needed to figure out which other sites contained comments we should classify as authoritative. Hence PageRank.

"Only later did we realize that PageRank was much more useful for search than for annotation..."


Larry Page: "Lucas Pereira: 'You idiots, you spelled [Googol] wrong!' But this was good, because google.com was available and googol.com was not. Now most people spell 'Googol' 'Google', so it worked out OK in the end."


Larry Page: "Our original hardware acquisition strategy was non-standard. We would go out and wait on the loading dock and beg for computers. When new computers arrived at Stanford we would go up to the people taking delivery and say, 'Surely you don't need all ten of these computers. Surely you can give us one. We have a really interesting research project..'"


Larry Page: "We used half the bandwidth of Stanford during our research phase..."

Eric Brewer--another search engine expert--is here.

Larry Page: Google Co-Founder; President, Google Products; Computer Science Ph.D. Student "On Leave"

1995-1998 academic work at Stanford. Terry Winograd advisor. At start interested in telepresence and all kinds of wacky things. Interested in link structure of web. A large graph. Big graphs are fun things. A large graph and the web might together make a doable dissertation.

Sergey Brin. He was interested in data mining. I had this big database of weblinks. We have worked 24 a hours a day together for the past eight years.

"Chance favors the prepared mind" --Louis Pasteur

Interested in reversing the web structure to view annotations--transclusion.

First system: BackRub: "not having an art department": indexed titles and anchors and did pretty well. Showed it to Eric Brewer.

An art museum got upset because we downloaded their content. No robots.txt file. "Occasionally, we would shut down the nameservers at Stanford, and nobody could log into anything."

The point at which it became too much was when we had so much traffic coming to google--10,000 searches per day, one per second. At this point we decided to start a company. Andy Bechtolsteim, "Who do I make out the check to?".

Pizza ovens: roughly that size and that temperature. Heat still a huge problem. Brewer: "I remember you had lots and lots of fans all over the place in your cage... pretty humorous." "The fire marshall got really made at us at one point."

We missed both things. We didn't go public during the boom. We didn't go bankrupt during the bust.

Today: profitable since 2001:Q1, 700+ people, 60+ Ph.D.s, 86 languages, 14 offices, 150M searches/day

AltaVista spent over $120 million in marketing in one year...

Why is google still around:

  • We're lucky.
  • Deep technical understanding of what we are doing.
    • Not true of many companies (not Inktomi, of course).
  • Everybody searches.
  • Everything successful in terms of traffic has stayed successful in some way (except for napster; although kazaa's doing pretty well: kazaa is well-designed to be illegal; but not my idea of fun).
  • How we make money: ads. I would never have guessed that we had ads for steel buildings...

Engineers said we shouldn't run ads from double-click. If the ads were what people were looking for, people would click on them. Ads. Syndication (AOL, et cetera). Providing search for other companies.

The little yellow box... We sell hardware. So we can guarantee performance. And we can guarantee to be up and running in hours...

World Graph of where the searches are coming from... India pretty amazing... Lots more searches coming from India than electric lights at night...

Information wants to be free? Copying doesn't cost anything. Distributing another copy costs basically zero. Google surveys the free part of the web.

Posted by DeLong at February 14, 2003 04:45 PM | TrackBack
Comments

So transclusion is like a pointer that also displays a preview of what it's pointing to, doing so by pointing at a document (in google's cache, for instance: guaranteed to stay the same, too) and saying "go get these paragraphs and put them here".

Then you let the browser handle it.

Hrm, people are saying things about Xanadu and transcopyright to me. More research required.

Posted by: Richard Soderberg on February 14, 2003 08:28 PM

here's a talk by brewster kahle about his plan, not to just make a profit, but to digitize Everything, and make it available to Everybody. neat!

http://www.loc.gov/rr/program/lectures/kahle-pres.ram

copyright be damned, cuz it's cheap :)

Posted by: kenny on February 15, 2003 05:35 PM

here's a talk by brewster kahle about his plan, not to just make a profit, but to digitize Everything, and make it available to Everybody. neat!

http://www.loc.gov/rr/program/lectures/kahle-pres.ram

copyright be damned, cuz it's cheap :)

Posted by: kenny on February 15, 2003 05:36 PM

here's a talk by brewster kahle about his plan, not to just make a profit, but to digitize Everything, and make it available to Everybody. neat!

http://www.loc.gov/rr/program/lectures/kahle-pres.ram

copyright be damned, cuz it's cheap :)

Posted by: kenny on February 15, 2003 05:36 PM

Actually AltaVista spent $120 millon in about six months. All of the ads were stupid too. They had example searches that did not work well, or sometimes at all, on the search engine.

Posted by: MattJ on February 16, 2003 07:40 PM

NIG

Posted by: NIG on February 17, 2003 08:58 AM

You guys have to work on your spelling. You know what I'm talking about...

Posted by: Billy G. on February 17, 2003 09:28 AM

You guys have to work on your spelling. You know what I'm talking about...

Posted by: Billy G. on February 17, 2003 09:28 AM

Work on your spelling dude

-BillG

Posted by: Billy G on February 17, 2003 09:29 AM

Wow, great write up. Thanks for taking notes!

Posted by: Chris Snyder on February 17, 2003 11:26 AM

./

Posted by: ./ on February 17, 2003 01:13 PM

./

Posted by: ./ on February 17, 2003 01:13 PM

./

Posted by: ./ on February 17, 2003 01:14 PM

I pinged your trackback URL and my trackback was registered OK, even though I got the following error back:
"Rebuild failed: Build error in template 'TotW Last Five Template': Error in tag: No such category 'TotW-thoughts'"
You might want to look into that...

Posted by: Rainer Brockerhoff on February 17, 2003 04:00 PM

what can i say? i love google.. its amazing! *speeachless*

Posted by: Majed on February 17, 2003 10:03 PM

I have no idea why Google chose Blogger than any one of a dozen equally, or more effective Blogging businesses, that is a business decision. But it is pretty clear why they would do that.

The breathless story from Dan Gillmor does all the usual, but invalid stuff about the Internet.
Google is known best for its search capabilities, but the Pyra buyout isn't the company's first foray into creating or buying Internet content. Two years ago, Google bought Deja. com, a company that had collected and continued to update Usenet newsgroups, Internet discussion forums. More recently, it created Google News, a site that gauges the collective thoughts of more than 4,000 news sites on the Net.

But now, Google will surge to the forefront of what David Krane, the company's director of corporate communications, called ``a global self-publishing phenomenon that connects Internet users with dynamic, diverse points of view while also enabling comment and participation.''
and yadda yadda yadda.

All that may be true but I don't believe it. This is not about "content" and Google isn't interested in "content", the news service is the result of the reliability that Google brings to other people's resource and content and the Blogger deal is the next stage in the development of Googling the net.

Think about it, Google's king hit is PageRank which generates highly reliable sources of whatever kind of information you ask for, based on search terms generated by idiots like you and me. Bloggers have developed a whole new set of tools that is making the reliability of information much higher and is using massively distributed human resources to find and rank that information very quickly.

A story hits the net and the Blogosphere, using newsreaders, Blogs with Trackback links and the usual power rule processes of the network evaluates that information and weaves it into it proper place in the knowledge universe very quickly.

By making Blogs the preferred system for publishing information in the first place, Google will be able to help improve the weaving and ranking processes even more reliably and in the otherwise untrustworthy world of the Internet, that will keep them on top. While other search engines are still trying to figure out how to turn their spidered information into a business, Google is focusing on what really matters, and that is reliability. The Blogosphere will benefit because Google will fund the development of the tools and they will be open source because the more of them they have out there the more valuable they are, because, can I say it again, Google does not sell search, it sells reliability, and every blogger and surfer and webmaster in the world is contributing to that. We do not get free services from Google, we pay for them with our clicks on their linked information. Pretty soon we will also be paying with our Blogging and we will be paid back with reliable information. That's an information economy; the information is the currency, knowledge is the payoff and reliability is marketable to those whose reliance on it is highest.

I've said for a long time that Google is not a search engine. Yes, it spiders, but that's not what it does. Now we have Larry Page's word for it. Thanks Brad

Here's another bit that makes so much more sense than most people get.
Information wants to be free? Copying doesn't cost anything. Distributing another copy costs basically zero. Google surveys the free part of the web.
Get it? Google surveys the free part of the web. Everyone wants to be on Google, because if you aren't in Google you don't exist. So you had better be free and good and referenced, and linked to the universe, or you don't exist and if you don't exist you sure as hell don't do business. Can we call that part of the debate settled please?

Posted by: Earl Mardle on February 18, 2003 03:24 AM

I have no idea why Google chose Blogger than any one of a dozen equally, or more effective Blogging businesses, that is a business decision. But it is pretty clear why they would do that.

The breathless story from Dan Gillmor does all the usual, but invalid stuff about the Internet.
Google is known best for its search capabilities, but the Pyra buyout isn't the company's first foray into creating or buying Internet content. Two years ago, Google bought Deja. com, a company that had collected and continued to update Usenet newsgroups, Internet discussion forums. More recently, it created Google News, a site that gauges the collective thoughts of more than 4,000 news sites on the Net.

But now, Google will surge to the forefront of what David Krane, the company's director of corporate communications, called ``a global self-publishing phenomenon that connects Internet users with dynamic, diverse points of view while also enabling comment and participation.''
and yadda yadda yadda.

All that may be true but I don't believe it. This is not about "content" and Google isn't interested in "content", the news service is the result of the reliability that Google brings to other people's resource and content and the Blogger deal is the next stage in the development of Googling the net.

Think about it, Google's king hit is PageRank which generates highly reliable sources of whatever kind of information you ask for, based on search terms generated by idiots like you and me. Bloggers have developed a whole new set of tools that is making the reliability of information much higher and is using massively distributed human resources to find and rank that information very quickly.

A story hits the net and the Blogosphere, using newsreaders, Blogs with Trackback links and the usual power rule processes of the network evaluates that information and weaves it into it proper place in the knowledge universe very quickly.

By making Blogs the preferred system for publishing information in the first place, Google will be able to help improve the weaving and ranking processes even more reliably and in the otherwise untrustworthy world of the Internet, that will keep them on top. While other search engines are still trying to figure out how to turn their spidered information into a business, Google is focusing on what really matters, and that is reliability. The Blogosphere will benefit because Google will fund the development of the tools and they will be open source because the more of them they have out there the more valuable they are, because, can I say it again, Google does not sell search, it sells reliability, and every blogger and surfer and webmaster in the world is contributing to that. We do not get free services from Google, we pay for them with our clicks on their linked information. Pretty soon we will also be paying with our Blogging and we will be paid back with reliable information. That's an information economy; the information is the currency, knowledge is the payoff and reliability is marketable to those whose reliance on it is highest.

I've said for a long time that Google is not a search engine. Yes, it spiders, but that's not what it does. Now we have Larry Page's word for it. Thanks Brad

Here's another bit that makes so much more sense than most people get.
Information wants to be free? Copying doesn't cost anything. Distributing another copy costs basically zero. Google surveys the free part of the web.
Get it? Google surveys the free part of the web. Everyone wants to be on Google, because if you aren't in Google you don't exist. So you had better be free and good and referenced, and linked to the universe, or you don't exist and if you don't exist you sure as hell don't do business. Can we call that part of the debate settled please?

Posted by: Earl Mardle on February 18, 2003 03:26 AM

hulloooo, hulooooo! is there an echo? echoo?..:D

Posted by: kenny on February 18, 2003 07:25 AM

Cool comments by Earl. They need to be blogged somewhere.

Blogger will become the equivalent of Seti@home, except it will use your spare brain cycles instead of your cpu.

Posted by: Craniac on February 18, 2003 07:23 PM

I just love Google. I am still disapointed that Yahoo gets more hits tho.

I also like the fact that there are no ads. please keep it that way. please.

Posted by: Henry Singleton on February 19, 2003 12:05 AM

Google + Blogger = Go_Ogle, the Mother of All Online Dating Sites

Here's how I think it will happen:

First, Google will improve the searchability of the "blogosphere" by making it easy for bloggers to append a file containing information about themselves and their blogger friends. This information will be encoded in an RDF dialect called FOAF (Friend of a Friend).

Soon after, it will start to dawn on people that the FOAF file is effectively a static online profile, while the associated blog is akin to a living profile (in the 'living document' sense).

One tipping (i.e. inflection) point later, usage of Google by date seekers will grow to an such extent that our (grand)children will read about it in their history texts. Online dating is at 26M users and growing, after all.

Google will then acquire the best RDF query toolmakers and launch Go_Ogle, the mother of all online dating sites.

Once Go_Ogle is in place, the possibilities are absolutely mind-googling :^)

More on this, including a pointer to foundational code for GPLed Go_Ogle, at www.opportunityservices.com.

Thoughts?

Enjoy,

Frank Ruscica

Founder
The Opportunity Services Group :: Have Fun to Get Ready
www.opportunityservices.com

Posted by: Frank Ruscica on February 19, 2003 10:50 AM
Post a comment