The one thing that really surprised me was how polite the disagreement in opinion was voiced by the 6 to 7 year old.
That is, as evidenced daily in abundance at work, the internet, etc ..., something that seems completely lost with a significant proportion of the grown up population.
Makes one think why that is so?
Your friend Erik
In Germany it is rather rare that a kid would go to a private school. State schools used to have a formidable reputation, but things seem to be changing. Parallel to a public discussion about rising inequality that has been heatening up in the last 5 years, the level of kids going to private schools is rising, nearing the 10 percent level.
Ok I had some sleep. I agree with all of your points except the measurement issue. Yes you can compare Search Engines the way you described, but just because of the size of the underlying corpus, the web, there is no way to measure the goodness of a system over the whole corpus. And yeah my post was a bit brash, subjective opinion typed in by a guy who really lacked some sleep. :-)
Yes I agree you do some cool stuff, and I applaude you for what you have achieved, but you really start to get on my nerves. The reason I am posting this here is: I had such high hopes for you to be a different kind of corporation. - to be able to make the transition from cool startup to cool company. - not to be evil ;-). That's not how I see you now.
First of all, how and why does a company that was started by two guys without a PhD so strongly favorise PhDs when it comes to the recruitement for all research and other "cool" positions. I am halfway through on my own PhD in CS, and honestly I just don't see your point. I mean a PhD is not a magical transformation into a creativity boasting super- hack-thing. Au contraire, I wonder how you are able to keep your engine running with so many PhDs around. There is some truth to the stereotype of the slightly mad and sometimes overly complicated CS academic person. I experience that daily. And then statements like: "We want to hire people who are smarter than those already working for us". Just hire people that have amazing skills, are cool, and enthusiastic. That's what you were built on. Please don't just look at GPA and titles.
Second and that applies not only to you but to all search engine companies, please stop giving us the impression that you know what's going on with your algorithm. Come on there are billions of documents on the Web - and you have an algorithm that ranks those according to millions of different queries. So please tell me - how do you figure out if your algorithm is doing it's job? How? It's all educated guessing. Nothing more.
Third, please G get over it with the "Do not be evil thing". AOL might have had an employee unfortunately release a dataset, but at least they allow you to delete all your personal data. While you do not, and boast about anonymizing our data after several years. And yes you censor in China, and I also remember the story about your employee who you fired for blogging that you are actually not as cool as he imagined you. Yahoo and MS do stuff like this too, but at least they don't pretend to be saints.
And fourth. Please your marketing hype about being different, being still like a startup - you overplay it. You going green, and the sudden popularity of that - a mere coincidence? And yes - sue microsoft for being monopolistic - but at the same time acquire doubleclick and argue it's ok?
Ah, dear G. it might just be that I worked too much and yes I only slept 6 hours in the last three days. Dear G. above all - just stop pretending to be something else.
I don't think that you can generalize it the way he did. Sure Yahoo, Google, MS and Amazon are building huge data-centers. That's what makes it into the news. And yes there is a lot more processing to be done since we started to see wide spread use of applications (social networking, search, storage, video, etc...) in the Web opposed to the former mere HTML plus some e-commerce routines here and there. But on the other hand, a lot of software is not physically shipped anymore, and I think this trend will continue. More important, you have things going on like Akamai, and Amazon's S3. I would therefore argue for the exact opposite: Unless your Google or MS, it will probably not make any sense running your own data-centers. My guess is that we will see the creation of more S3 and Akamai-like products. Software companies will develop software but it will not be physically shipped, nor in most cases be self hosted. Specialized companies will due to the economies of scale be much more efficient at that. Software companies will be more lightweight. They create the software and nothing more. Just my personal analysis.
Personally I find trying to get kids to drop out of a college a little bit unethical. Well yes "last time those kids could buy houses in cash", but Slide or Slider is not Paypal. Levchin sure is confident about it but no one knows where it will end. If a college student drops out on his own account - fair enough. To lure kids to drop out, and then if the whole thing sinks - have to kick them out? Personally for no amount of money, success or fame would I want to place myself in that situation. I don't know about the U.S but in Europe, applying as a dropout does not really improve your chances to get a job.
"I find the concept of the long detour fascinating.It says that the company believed strongly in the original vision, but recognized that the time was not ripe to realize the plan." Recognized that the time was not ripe? Personally I don't see the point of this article. What concept is he talking about? They have the idea - they raise 80m $ - they try - does not take off. Fine. They "scale down"- "re-evaluate" their "product line" - find something that sells - survive on that. Ok. Then they sell that something that sells to concentrate on the original idea (which they had not given up upon but had "maintained" all these years). Reads like a marketing effort. I don't see the point. So I suggest: Unless you consider "a long detour" a fascinating concept - Timing is still everything.
Personally I don't think that P2P video from a browser is probably the best utilization of this technology. Someone mentioned that even the top videos on Youtube only have 4-5 concurrent users. That alone generates a lot of difficulty for that approach. I supppose that the distribution of traffic on Youtube will have a rather fat long tail. Also from a user's point of view for this approach to work it would require storing not only one video but many. I don't see the point of using my diskspace and my bandwidth for Youtube. That's different for Joost. They get users to download, install, and share the bandwidth as they have more valuable less fat tail content to offer. What's more if you are watching fullscreen you probably won't mind that Joost is making use of your upstream. However if you interpret P2P in the browser as a potential Web cache, and don't restrict yourself to the video approach there are lots of interesting things one could do. You just have to look out for Power-Law shaped traffic around big chunks of data. Maybe internet radio with high quality streams where you could time-shift streams to make optimal usage of the P2P approach. Or the distribution of software. Once you have several millon users using your applet or whatever, you could become the next akamai. To get there one probably would have to go to those guys where bandwidth really hurts. Thinking of sites like the underdogs.com, shareware distributing sites, web radio, etc ... . If you can make them happy by potentially saving bandwidth, and the users are happier by potentially increasing download speed, you might get those guys to ask user's on their web sites to download your little thingy. I think there is room for a P2P based web cache, probably not in distributing Youtube like videos. Perfecting the P2P part, and being able to aquire a sufficient install base, a lot of opportunities might arise. Go speed up the net ;-).
Yes I totally agree with you. If they manage to take 50 percent, all of the informational queries from Google. Then they've kinda yeah done it. I just think that just achieving this will be super difficult. To get people to move their bottoms away from Google, PS will have to blow them away. And even if they manage to get 50 percent, it does not equal 50 percent revenue. So much less money to be made with "Who won the academy award in 2001"-style queries than with a dull "mortgage refinance" query.
No need to apologize. I think your question was clear enough. Probably my lack of sleep plus me being not a native speaker. No rudeness intended from my side either. I'm not in a YComb startup, but was part of one (non YC :-)) that went down 2 years ago. So it's probably off-topic here but I think it might be useful so I just post (also writing about might help me get over it ;-)). Our project was a rather complicated Java thing. We started off with XP and everything was sort of fine. Running short of time we decided that we had to code at twice the speed(!). No more XP, no more Java doc(!), 6 months all went fine, then 2 new people in the team, and two old people out. Actually it was 2 old out (rather unexpected) and 2 new in, which made it worse. A total mess. A generic advice: Always keep an eye on the codebase and try to spread the knowledge of it. It's not only that the wrong person might quit for whatever reason, people get sick, stranded while travelling, crazy etc ... . Best luck with your venture
I wish they would ..., but here's my humble opionion why Powerset is not going to kill Google. In 2002 a guy called Andrei Broder, who now happens to be Vice President of Emerging Search Technology for Yahoo!, published a paper titled "The Taxonomy of Web Search". Broder analysed clickthrough logs (think of AOL) and came up with a taxonomy to classify what people search for in the Web. He identified three types of queries: Informational (people search for a specific piece of information: Who is Snoopy?), Navigational (people know where they want to go, a specific homepage or web site: Snoopy creator homepage), and Transactional (people search for a site to perform some kind of transaction: Buy Snoopy Case Mod). Long story. However what is remarkable: He and several others after him found that less than 50 percent of all Web queries are informational. Powersets search engine clearly seems to target informational queries. In order to lure users away from Google, Yahoo, and maybe Answers.com they have to offer a significantly better experience. That's hard. But they might do it, and kill Answers.com by doing so. However in order to be good at Navigational and Transactional queries (and that's where the money is) you don't need sophisticated Natural Language Modelling. Those queries are short for a reason. Users have a target in mind and want to go there as quick as possible. And in order to not dissapoint users there, you have to crawl as much of the web as you can. And then build a system that is still able to deliever queries at least as fast as google. And that's really really really hard, and probably the reason we haven't seen a Google killer so far. Hail to Powerset - the Answers.com killer.
Is it also what you wanted to hear ? Personally I have had good and bad experiences with XP. XP can tremendously improve the quality of the software and moreover helps to spread knowledge of the codebase if the right guys are working together. If you have the wrong guys sitting together they will stall each other on discussions of coding style and for versus while loop efficiency. You should probably just sit together and see why the guys want to do it. If they have the right arguments and are not merely argumentative about it; you might wanna let them do it their way.