Slashdot Mirror
The Man Who (Really) Makes Google Tick
An anonymous reader writes "Like his friends Sergey Brin and Larry Page, Craig Silverstein abandoned his PhD studies at Stanford to become employee No.1 and technology director at Google. While building the search engine in a garage, never in his wildest dreams did he think Google would become what it is today. Not only is it the envy of software giant Microsoft, Google continues to redefine the technology market with its creativity and tenacity. In this in-depth interview, Silverstein discusses a wide range of issues including the backlash against Gmail among privacy advocates, the company's cultural changes and its shifting reliance on PageRank."
Is not having all your personal information in the hands of Google. I don't feel that Google is the threat here. They've proved time after time to be an honest company. I'm more worried about some crazy new law (Patriot Act anyone?) giving the Government / Other corporations instant access to this online archive of some of our most private information.
Like in the WiFi market? Or with Quicken???? Does someone still uses Money?
Lots of companies succeed against MS. Not that it's the easiest thing to do in the world, but it's doable. Google might be another Intuit.
What's even funnier is that Google is not intelligent whatsoever. All it can do is find webpages. It doesn't answer questions, it can't look up facts, it can't verify any informational searches for accuracy. It just finds websites. When you stop to think about it you realize that Google is not all that amazing. Important? Yeah but not that amazing. And nowhere near as smart as a librarian.
Comment removed based on user account deletion
..is the "why is Google so successful?" question. This interview seems pretty focused on talking about "hot" topics (gmail privacy, microsoft, blah blah), and it talks about possible future technologies in Google, but the interview doesn't probe about just why Google got there in the first place. Where's the talk about what Google did differently? PageRank (before its manipulation by spamdemons), clean design, obliteration of banner advertising and "portal" services, clear separation of search results and "related advertising" results... that's the compelling stuff that I'd want to hear the man behind Google talk about. Those were all pretty bold moves from an economic standpoint ("what, you want to remove banner ads?! how do you expect to make money!!" etc etc), and by golly, it panned out and then some. Someone should go back and ask, "how the hell did that succeed, how did you convince people to come on board and work with you on Google when it was so damn different?"
One thing's for certain: The guy does an excellent job of keeping up Google's mysterious aura. When asked if the number of servers was 10k or more like 100k, he said "over 10k". When asked about future technologies and directions for the company, he always answered vaguely ("I can't comment on specifics").
This is pretty cool. The aura that google has that no one knows how it works, and no one knows where it is, and no one knows what it's doing... That's a pretty cool public image to have for something used as much as google is. I just wonder if investors are going to want to know more about what's going on.
~Will
sig?
Much as I love Craig (I'm having dim-sum with him next Sunday) I'd have to disagree with the poster's characterization of the interview as 'in-depth'.
It's got more than a few questions, but few of them are terribly interesting, and (by necessity, I'm sure) many of the answers are vague or "I can't really talk about that".
Of all the companies that Microsoft has ever competed agains, Google is the first one where Microsoft is fighting a battle in enemy territory.
Microsoft's victories come in the software front (Netscape, Quicken, Office, etc.) where they can leverage their operating system dominance.
Google's home turf is massively scalable, reliable web services. Even though much of it is secret, all signs point to an incredible advanced platform that keeps these things running. Its highly redundant and distributed, using some cutting edge research and open source technologies. If Microsoft were to try to utilize Windows to power such a platform, their developers would soon discover how laughable Windows is for such a solution. Not that Microsoft isn't smart, but the culture of Google lends itself much better to success in this field than the culture of Microsoft.
I am, however, looking forward to Microsoft going up against them, as it will allow us to point out yet another failure in them trying to move beyond their core business.
-- Fighting mediocrity one bad post at a time.
They can't practically do a full search across everyone's email for a particular keyword. To do so, the providers need to offer this kind of service, which they haven't been built to do (data persistence, indexing etc.). Alternatively, the FBI/CIA could just install snooping boxes at network hubs, but again this isn't practical for realtime searches given the volume of mail going around the world every day.
On the other hand, Gmail is (the first system yet) specifically designed to make searching across its datastore as easy as searching the web. Now, for the first time, large scale email snooping is practical. The FBI/CIA can just get a special privilege account from Google, with the ability to search everyone's email for keywords just like we do now when searching the web.
Laws don't mean much if enforcing them is impractical. Gmail and similar systems if they catch on make new laws practical.
Google is known for their new and interesting technologies. I stumbled across this search engine right before reading this article actually. A search engine that clusters your results! It makes it even faster and simpler to get right to what you want. It's nice to see new ideas like this coming out and helping to change the direction of search engines as google did several years ago.
http://vivisimo.com/
I don't know about anybody else, but I have noticed that, lately, Google seems to be rather polluted by people who are exploiting the PageRank system to get higher listings. You know the types -- the url is of the form www.domain.com/your-exact-keywords.html, and the page doesn't actually have any content. Google is fantastic for anything that hasn't gone main-stream, simply because advertisers aren't Google-Bombing (heh, I can't believe I actually said that) those particular words yet.
Has any other GMail user noticed that there seems to be a limit on the size of attachments that GMail allows? For instance, it currently won't accept my 60+ MB attachment consisting of a gpg-encrypted tar of my Documents directory. I'm going to tar and encrypt smaller and smaller chunks until I find the limit. People could save me time if they already knew the limit.
And besides, why should Google do this!? They are the ones giving out a Gig. If I want to send a friend the ISO of Debian Disc 1, why should Google stop me? Why be choosy about how I use the Gig? Again, I didn't ask them to give out Gigs. They freely chose to. Let's remember: Don't be evil.
One, Silverstein acknowledges that AI problems are basically hopeless (gonna take "about 200 to 300 years").
Two, when asked if PageRank is dead and what they are doing to fight false popularity, he says they are "tweaking it in new ways".
Three, when asked how ("do you have algorithms?) he answers,
OK, they are looking at the anchor text. Then what? As long as HTML is the language, I'm afraid there aren't that many more things they can do.Time to cash in, perhaps?
This is...
O
U
T
R
A
G
E
O
U
S
!
Agreed! I wrote a blog entry about this the other day and emailed it off to Google as well. Basically I suggest a preference to exclude sites selling stuff and exclude training courses (as well as wishing for improved indexing and ranking for content in Wiki's...)
--Murray Barton
Yeah, and that's just salary. The real money is in research grants. I have an acquantance who works as a lab tech. Her boss, a PhD at UCLA, usually has about a dozen or so research subjects running at once, with each being funded by one or two grants. In fact, the only work that individual does is think of new things to research and file the appropriate papers, grad students and lab techs do the actual work. Pretty good gig, if you can put in the 20 years of school and work it takes to get it.
God invented whiskey so the Irish would not rule the world.
I had a similar problem the other day when searching for a digital camera, and indeed the other day when searching for a slightly exotic piece of HP Fiber Channel hardware. Who would have thought there were so many "Internet Entrepranurs" wanting to sell $5k PCI cards?...
I don't think the spammers can be actually selling the cards; they presumably want to bring you in just to show banners or to sell something else. I suspect the spammers got the product name by gobbling up HP's site or some other reseller.
Anyhow, here is an amusing conspiracy theory: Google are happy for product keywords to get totally spammed out, because it makes it more likely that people will just click the paid links. They might not be the best value, but at least you know they're enough of a real business to pay their advertising bills.
Of course this is a bit tough if you don't actually want to buy the thing, but just to find the manual or drivers or linux support information.
I don't think Google are really doing this, because they seem to be sticking to "don't be evil so far".
The DoT, namely C.S., used to be on the list of Google Executives.
Any comment on *this*, I mean...hello? Mr. Brin? Mr. Page?
Did Mr. Silverstein just dematerialize or what?
"In an interview before Google's IPO filing, Silverstein discussed [...]"
*yawn*
If I only I had applied...
I don't know,
If Microsoft wanted to they could develop clean interfaces too. Microsoft's reputation is not that bad, except that it is known that they don't play nice with others.
But still, if they do develop a worthwhile engine, on par with Google or better (they have very good researchers, they are certainly capable of coming up with something) and put it as the default search engine in the next version of Windows and the next service patches, then Google could find itself in trouble.
Microsoft has tremendous leverage with its users by virtue of most of them not being very educated and not caring about interfaces all that much as long as it sort of works.
My points are still valid, even if my psuedo-code is not 100% correct -- but you miss my point: any of the large webmail providers (Yahoo, Hotmail, etc) will still be able to search all the email in any users mailboxes, almost as easily you can log-in on their respective homepages. It would be a fallacy to think otherwise. Of course, these services already do -- just like Gmail -- have one huge store for all their users' mail, even if it is distributed, as you mention.
It is my belief, gained through knowledge of mail servers -- and too many years real-world experience writing high-end web-services/front-ends of one kind or another -- that SQL is the most scalable solution for the back-end of a web-based email system with this quantity of users, the idea of using any kind of file based mailstore is unpractical for web-based email for a number of technical reasons.
Furthermore, if I remember correctly, in the past I have read articles about the big webmail provider's back-end systems being SQL based (sorry, I can't remember which company the article was about -- I think I've read about more than one..(?)).
Your analogy about searching everyone's email is moot: we are not really talking about searching everyone's email spool, rather, people are arguing over whether Google's webmail -- Gmail -- is any less private than any of the other big webmail solutions (Yahoo, Hotmail) that are already out there -- and it's not. It's no better, and no worse -- but they are being more upfront about things (i.e. explicit about their business/technical processes) in their privacy policy than some of the other providers care to be, which has brought this matter into the eyes of the general user (who probably do not realise that when they click 'Delete' on Hotmail, a copy of their message may indeed still reside on another of Hotmail's systems in an archived backup, unaccessible to the all but the sysadmins -- and the respective law enforcement agents/agencies, if they have the right paperwork).
In this statement (and possibly inferred in some other statements) you make it sound like Gmail/Google will index everyone's mail-server's mailstores like it indexes webpages -- it won't. Gmail only indexes the mail of Gmail users.