Ask HN: Will there be a better Google?
I was reading this book called 'In The Plex' which talks about the beginnings of Google, how no one thought of using back-links in ranking web pages and how Larry Page used it to make Page rank it's a fascinating book to read, and surprised to see how everyone missed an idea like that.
Fast forward to today, has search reached its climax? Can't there be any other search engine that will deliver better results than Google? There was a time when people talked about social search for a while but that too died down.
So is there a new frontier for search? Will there be a better search engine in the next few decades? Of course. If history has taught us anything, it's that no company, product or individual has ever remained 'on top' for eternity. Everything gets superseded and replaced by something better, it's just a matter of when that happens rather than if. As for search in particular, yes again. There are plenty of ways a search engine could work better than Google. It could read your mind and figure out what you intended rather than having you try and write out your search query. It could actually answer questions in purely conversation English (or any other language), like what Ask Jeeves tried and failed to do years ago. It might have results somehow tailored to your interests in a much better way, and basing what comes up on both people/sources you trust, relationships between them and new pages and your exact intentions. Heck, it could just be better at finding generic pages than Google is at the moment. I'm sure you've come across tons of situations where you searched for a certain specific technical term, right? And then found all the results were ones that only had the most vague similarity to what you wanted, with the term you cared most about it in grey and strikethrough form underneath the result? Even a search engine that realises you want a certain term in a certain context (say, a discussion, or a long form article) would do better than Google at the moment. So yeah, there will be a better search engine at one point. Quite possibly one that completely changes how searching works in general and finds results that are far more accurate to what you actually want than anything available at the moment. The answer, I think, is in the book you're reading. Search was stuck long before Google came. Then came Google and they made search better. Right now, search again is stuck. Until someone comes up with a way that everybody else is missing, to take search to a new level. There's the key question. What could you do differently than google for a niche or even mass? I think privacy is massive but that leads to how will search engines that focus on privacy make enough money? Google offers a good search engine and a heck of lot of cool free software, all payed for by ads. Frankly, I don't mind targeted ads as they're far less annoying than random ads. I'm into the outdoors and tech and even went so far as to go into my Google settings and adjust it. No, I don't like Tango. What was that about in my settings? So I'm not even 100% convinced that the trade offs in going to a completely private search experience are worth the costs. I don't know. What are some other problems a search engine company could focus on? I'm not even sure what the top concerns are of users that would inspire them to migrate from something that is pretty damn good, frankly. I am yet to come to that point of the book. Where is the search stuck at the moment? I can't seem to understand that. some points: Result quality has not really improved, all the while: more and more content is in silos that are not easily searchable. Spam/SEO/trust in content remain a problem, thus many people rely on social recommendations for content (which has its own problems). More and more content is in non-textual forms. Google (sometimes vastly) prioritizes content on their own platforms. It's not. Search works great. I think that's the wrong way to frame it as being "stuck." Better to say where's the opening? The main one that I see any action around is privacy. There must be other opportunities that nobody's seen and acted on yet but will. It'll be interesting. I suppose you are too young to have used Altavista? Or Yahoo "curated" link lists? The pre-Google search engine I preferred during it's brief period of utility was 'WebCrawler' because it returned dumb results. Google doesn't do that any more, it's back to curated links with advertising being a large factor in curation. Oh god, I remember how exciting it was when Google rolled out. Goodby Altavista and all the rest. Though I do have a place in my heart and gratitude for Altavista. That search engine found me whatever I wanted to find. It worked well. Google just came in and was better. My memory of Altavista was that it was great for discovery of obscure but interesting sites. Sure, you might have to go through 10 pages of results to find one, but the experience was one of serendipitous discovery. It's lack of precision (in favor of recall, making it seem less winner-take-all, SEO-optimized) seemed to correlate strongly with the notion of "surfing" the web. (And the use of the phrase "web surfing" likely declined as Google, and later, the app economy, rose.) Yes I have not used altavista but I have read about it, and I've read the book and I know search was stuck back then, my question was why the above comment says search is stuck at the moment, because I don't see search stuck at the moment. I am not the OP but I interpret "research is stuck" as follows: Remember Altavista (and similar search engines from the era?) it worked, but everyone was using the same approach so you got the same results when you searched for a given word... Then Google arrived and the quality of their results looked like magic. In terms of "user experience" today does resemble then: Bing, DuckDuckGo... They return a subset of what Google finds, not something qualitatively better. It's probably worth pointing out that if anyone has an idea for how to improve on search, they're not going to share it in this thread, they're going to complete it quietly and publish their product. Though there are many better ideas like semantic search, entity search, Q & A, deep search, etc. Startups could crack some market share from big players. Building better search engine is not the big problem for startups, but from business perspective, It's a marathon race - in terms of scale, investment, making the user switching from google etc. also Google have billions of $$, thousands of researchers working to improve the results, competing them is not easy, hence many stay away from search engine development. Imagine a search engine that not only gave you search results, it used advanced AI to tell you in brief how each result applied to the problem you are currently trying to solve. Currently, you have to waste time thinking about each result to classify the contents, as for example, entertainment, product marketing, opinions and rants, purely factual, legal, political, social networking, media, API/interface, et cetera. It would also be useful if the search engine could verify statements made on pages and give an accuracy/honesty score for each result. This is particularly important when the author of a page has a serious conflict of interest. Perhaps the search engine could tell you what that conflict is -- to help protect you from deceptive or incomplete information. Maybe I'm at risk of being a fan, but the Google of tomorrow will likely be better than the Google of today. And while they have expanded and evolved I mean this relative to search specifically. "Google’s mission is to organize the world’s information and make it universally accessible and useful." The universally part has a lot of room to grow, and the useful part changes everyday. Not having read the book, but based on what you posted, the example is just one of the giant leaps forward and a way they differentiated themselves as a technology and a company. New leaps will present themselves through continuous improvement and reassessing what is useful. Probably not. Google will likely remain at the top, although they'll continue to improve. It might end up being effectively a new engine decades from now, but it will most likely get there by evolving gradually over time. Absolutely. Search is still in the stone age. A new paradigm will come up, and Google, like most large companies, will be too invested to the old way of doing things, they will struggle to adapt. That leaves plenty of opportunities for startups. For instance: everyone seems to assume that search is about finding the one page on the Internet that best answers the query. Wrong assumption. People want answers, not pages. What about a search engine that combines facts from two pages or more to answer my question. If you are working on a startup in that space, ping me on AngelList. >What about a search engine that combines facts from two pages or more to answer my question. Interesting concept, but why would I (as a site owner) want my content mixed with my competitor's? I can see value to th en user, but what would the value add be for indexed sites? Search engines have never been about what websites want, so it doesn't matter. The idea of a search engine is long since gone. Google is a destination...mail, news, interesting links, etc. So is Facebook [just replace 'mail' with 'messages']. It's really the Yahoo/AOL/Compuserve model that works over the long haul. The future of actual search is something that gets out of the way like Siri or Cortana. But there's not really money in search. The money is in siloing and data collection. Google is largely ads these days on anything commercial, followed by organic. With the introduction of a 4th top placement ad its amazing how little organic search is showing on a screen. Type in a competitively commercial term like 'credit card' and think how it was 10 years back. Its amazing how they have trained us to accept this. Or maybe its just the lock of better options... I'm not saying they're better but there are alternatives to Google that address specific concerns. For example, DuckDuckgo seems to be getting some traction with their promise to give you privacy. That pitch resonates with a lot of people. I hope they don't do something that would degrade our trust in that. I think they have a future. Whether their seach measures up to Google's I don't know. I'm guessing they're not there yet as Google is good but they'll get there and if they stay true to their vision, they'll win over a lot of people. I worry because there are a lot of temptations that could throw a company off it's core promise, especially a solemn promise to protect your privacy, while others offer money to go different directions. I consider search an essentially solved problem and I think most people agree with me. When I want to find something I can usually use Google or Bing to get to it within 10 seconds or convince myself it doesn't exist. Will search get better? Probably, but going from 99% to 99.5% isn't very exciting to me. If anything, I see Google and others moving in a direction of using context to reduce the amount people search (e.g. google now). I think this pattern is going to get much bigger in the coming years with better AIs and more information about what the user is doing. I agree with what you've said but I wanted to add that there is room for improvement on Google's model. For example, selling page 1 results is unfair, but that is something google does. I think DuckDuckGo is an example of how Google's methodology can be refined and improved upon, specifically by not spying on users. It's a subtle improvement, but I find the results I receive are actually superior and more relevant than Google's, plus I don't have to worry about who is watching my traffic. I don't think it's sustainable to not spy on users or to not serve ads. Keeping user data adds to relevancy and if there is no revenue, there are no relevancy engineers. That might be true, but web-search is an absolutely vital service for cultural health in the information age, like the telephone system used to be. Every search engine does sell ads, which seems to be working for them all so far. If that ever stops working, we might see web-search become a public utility funded by tax dollars. > plus I don't have to worry about who is watching my traffic. Millions of people (myself included) use google and don't worry about this already. Online privacy isn't an issue (or a deal breaking feature) for the overwhelming majority of people. DuckDuckGo may get niche business because their privacy stance, but it's not going to help them unseat Google. No you are right. It will be Google's continued inching toward a revised "do evil whenever it's profitable" business model that is going to unseat google. I mentioned DuckDuckGo because it seems it would be most likely to take the top search spot if Google were to die today. I think there is an area of social search which remains largely untapped. This is because major social networks doesn't allow search engines to search their networks. I think Facebook tired to do this but they used Bing but have they used Google mmm maybe they would have been at social search but I don't think social search will be a thing if Facebook and other networks open up their network Try to find a funny gif you saw a while ago and tell me again if you think search is a solved problem. I don't think there will be a search engine better than Google in the near future (may not forever). Best search engines today will have to throw results based on user preferences to stay relevant and compete with Google's results. Cracking the perfect semantic search (chat bots) may change this but Google may have advantage here just because of the amount of data it has.
Any great algorithm, unfortunately will be limited by the data on which it will be trained. Google has this unique advantage. I don't think search needs to be made better. There are more important problems. That doesn't mean there aren't innovations to be had in search engines but it does not have to be better search algos. I am not sure about general search but I think an option based on ranking of author + knowledge domain would do well.