Could two smart computer science Ph.D students create a search engine that unseats Google? How vulnerable is Google to this possibility?

Anyone anyday could come up with something that can unset the foundations of existing framework. Too much jargon.......but the idea is its all about relevancy. When google started yahoo has already captured the market. Search engine is about relevancy of the information and speed at which you show the results. Initially speed can be compromised but relevancy can't be. Google still is a centralized mechanism where they collect information and process its meta to obtain a list of results. But the source of information i.e. croud is forced to work according to google's standard protocol to show up in organic listing. If you decentalize the core and process results geocentrically the results would be more relevant. Example. Fashion is a word in english. It could mean something different in visual terms for someone in India and someone in France. If you decentralize the core and break it down to geolevel the meaning of words would translate into different prespective and people would get a better result of the query. You never know, a new peer to peer protocol would turn the world around........

Yes. But I agree with Graham. It will not be a traditional search engine, but something different. Let's face the truth: Google sucks a lot in a lot of ways. So, if those weak points are attacked, a better search engine can be devised. First of all, Google doesn't display much intelligence. That's why you can't use the I'm feeling lucky feature. When Google provides you with a result, they already assume that it will fail. That's why you can request 1000 answers. Google does not understand the content they index. They just eat and eat gigabytes and index them in a very efficient database, but they know almost nothing about the true quality of the information. Yes, some tests are run to avoid fake web pages to scale in the page rank, but that's all. Google may be defeated by a system that displays some kind of intelligence. A system like the Watson computer of IBM will destroy Google because it will do all the work that nowadays you have to do... for you. The problem is that solving big a number of requests in real time might be quite difficult. On the other hand, an intelligent system will not need to solve every question from scratch. It may even create its own content from what it has learned already. What I mean is that Google will be beaten, and the rival will be an expert system. Do not lead me to the content, just provide me with the answer. There's another option: Google is starting to follow the mainstream of thinking exhorted by the lawyers of the great multimedia companies. They are starting to bury torrent links. They follow the robots.txt convention scrupulously. A non compliant search engine could win a terrible amount of users, too.

Google is not a search company, it's a keyterm auction company and the mountains of money in that arena have forced the corporate entity to harden around that core. True search would be language independent and include the result that something does *not* exist in the known searchable domain by accurate identification of locus. It would also require affirmative privacy (think video surveillance that is always encrypted and only opened by court order with cause, versus video surveillance to look for cause) because individuals would also occupy discoverable locii. If individuals are not empowered to own their locus, free and original speech becomes a fiction of the past. Google is likely no more or less evil than any other money-driven automaton (workable definition for corporate "people") but it is dependent on a dangerous direction. Callow cleverness is certainly not intelligence, and perhaps we shouldn't be focused so much on what we now think of as genius. How about letting educated adults have some say? Politics, not genius, will unseat Google.

There are two parts - search is pretty much 99% of what we could possibly want nowadays - bing, yahoo and duckduckgo pretty much deliver the results that are good enough - admittedly it's not as good as google but the gulf is not very large - certainly not when google started - search was not good and google really delivered a clear cut lead in accurate information. A small portion of changes  is googles current desperation to optimize revenue - ad searches go on top, yelp gets downgraded for googles food search - same with hotels and airfare ... Etc, etc but of course, these actions cause some users to drift away. But googles main strength was leveraging & monetizing search results that ms, yahoo and others could not figure out. Basically Google made search ad sales a professional business. So it's not just the results but how google managed to make the most amount from it  ... Going forward, googles problems in search ad sales are: mobile ads are worth less (less clicks) People search within Facebook and Amazon They locked themselves out of China With mobile apps, people do not search Google but just launch the app - think webmd, yelp, etc ... They have not really figured a serious profitable model for YouTube So Google faces cuts by a thousand nibbles ... And after 10 years, they have not come up with a real second business unit ... That is the future google faces ... So, yes- it is possible to come up with a better search and people are willing to switch BUT  the big question is whether you can scale up and build a professional company but Facebook scaled up in a few years - taking a big chunk of ad dollars that would've gone to google. So, it's possible.

Well, a PHD would be nice, but not really required as a law of the universe.  Still, it could happen, but seems unlikely.  Part of the googol business plan is blatant crushing of innovation in order to lock the market for its own profit.  Here it is, ten, fifteen years later, and they still give you millions of useless answers in usually less than a second, with no attempt made at common understanding of the English language. They can sort thru porn and yootoobe videos with near perfect success, so it's clear that their algorithms could be more capable than they are.  They are so intent on attempting to trick you into impulse purchasing that they regularly disrupt your search with distractions.  Their marketing algorithm deliberately conflates popularity with fact by the inclusion of idiot features, like "I'm feeling lucky".  The image engine is worthless for all but the most popular celebrities. Type in "Brown v. Board of Education of Topeka, 347 U.S. 483 (1954)" and you will not get, as the first entry, a link to the actual Supreme Court of the United States of America, for an actual transcript of this actual case.  Which is where the true answer to this search should lead first. This is not to say that the service is useless, because many times it can be spot on.  The term "Simplicity part # 1690039" returns salient info, but other queries do not.  We all have our lists of inane search results, and time does not permit that list of useless search results to be referenced here, but for this one: Using on the term "inane search results", you will find, well, inanity only. Future searches are limited by erroneous assumptions one one's previous searches.  The quote function is ignored most of the time.  Seemingly random responses float to the top of a search by some unknown process of focussing on words in one's search term, taken out of context completely.  This list of problems is enormous, and the googol has simply not made any steps to correct them, being focused, for one thing, on suppressing the number of accidents its driverless cars have; for another, floating baloons to provide internets services where none may be desired.  And on and on. What's needed is for these two CS students to work on a search algorithm which features a growing look up table of the meanings of common phrases, which can be saved, in English, on one's individual computer, such that one can teach one's search engine what to do.  the look up table could also include rules of engagement and inference, based on Cobb's twelve rules for relational databases. The students would have to work, first, with friends and family, to get the system up and running.  Perhaps, with a quiet, nominal paid subscription, they could expand their base and provide actual humans responding to questions raised by their subscribers, with all questions categorized and made available online.  For them to succeed, they would have to create a grass roots movement that would take corporate America by surprise. It would probably take ten years for this network to be grown, and the authors would have to make a living, but also be honor bound not to sell out.  This last thing would be very hard to do.  Personally, I could be bought for only a million dollars, after taxes, assuming that I would maintain moral control over the work that I would contract for.  Hard to imagine any corporate takers who would concede control. The sums they would be offered would be hard to resist, and they'd have to accept that their work would be corrupted. So..... While possible, highly unlikely.  Not that I'm some sort of saint, mind you.

Maybe the Real Question should be : Can two smart computer science Ph.D students create a new way of ranking pages and displaying adds to people so they don`t have to use addblocker -- better than Google does  it now? that is the focus, business and investments to make your ideas reality . Everybody can program a complex search algorithm and feed it to some rented supercomputer  !

Of course they can be unseated. Just like they unseated somebody else, and Apple unseated somebody, and IBM unseated somebody else, and then were unseated themselves. It's just hard to picture because if it were easy, somebody would have already done it. But it is possible, and the odds of it happening are quite good. The answer to your question is: innovation.

Google is vulnerable just like any other incumbent company is vulnerable.  It's not like there weren't search engines in 1999 (Alta Vista, Lycos, Excite, Yahoo! and Infoseek come to mind).  Someone else will come along with a better way of finding information and depending on whether Google is asleep at the switch, could potentially gain a toe hold in the market.  Right now DuckDuckGo would like to be that contender, but they haven't differentiated themselves enough (yet) from Google to be a credible threat. This is shear speculation, however the most likely threat would be a technology like Watson with an interface like Siri, but with the ability to scale to millions of simultaneous requests.  Google's biggest weakness is that it doesn't understand the context of what you're trying to search for, and it needs to make inferences based on very limited information.  As a result, it has biases which make searching for certain types of information really difficult.  The funny thing about Google is that we usually _feel_ like it is giving us great results, but I think that is in part because it has trained us to only search for things in certain ways and to expect certain results.  You can't ask it a question like "What will supplant Google?" and have it spit out a reasonable answer, unless it has indexed a page like this which matches parts of the text (most likely "supplant Google").

Adam's answer recalls the joke about two economics professors out taking a walk. One of them sees a $20 bill on the ground and reaches for it. His companion restrains him, saying "if that $20 were real, someone would have already taken it." Yes, search is now profitable and yes, Google etc. have invested huge sums in developing their approaches. But investment is not a guarantee, and profitability is not a factor that reduces Google's risk  exposure; rather the contrary. Profitability means that any better search system will immediately earn real money for its creators. Brin and Page invented something and then had to convince people it was worth money; their would-be replacement will invent something and have the ability to cash out immediately. Brin and Page were lonely pioneers, with few or no competitors; there are a thousand sets of people trying to do exactly what the question describes. So yes, Google is vulnerable to such a development. Search is not a mature field, there is no reason to believe that the existing players have identified every possible revolutionary development in the near term (let alone over the long term), and in fact institutional inertia and the sunk cost fallacy are likely to impair Google's ability to nimbly move to newer and better ideas even if they develop or acquire those ideas. ("We can't deploy this new algorithm, we just finished spending a billion dollars deploying the last one!" vs. "Cool, I've invented a new algorithm, might as well put this baby to work.") The probability is unknowable, but it is not zero and I doubt that it is low; human genius is unpredictable and someone with a better idea could be five minutes away from hitting the "deploy" button as I write this paragraph.