Hacker News Reader

Comment by sylware | Hacker News Reader

This has nothing to do with search. See https://news.ycombinator.com/item?id=31490603. And that's not actually how we make our results either -- our results are actually made anonymously from a variety of sources. In particular, when people search, we believe they’re really looking for answers, as opposed to just links. For many categories of searches (restaurants, lyrics, weather, etc.), there is usually a specialized search engine (e.g., Tripadvisor), content site (e.g., Musixmatch), or dedicated source (e.g., DarkSky) that does a better job of actually answering searches than a general search engine can with just links. Our long-term goal has been for over a decade to get you Instant Answers from these best sources.

Most of our search result pages now feature one or more Instant Answers. To deliver Instant Answers on specific topics, DuckDuckGo leverages many sources, including specialized sources like Sportradar and crowd-sourced sites like Wikipedia. We also maintain our own crawler (DuckDuckBot) and many indexes to support our results. Of course, we have more traditional links in our search results too, which we do largely source from Bing, but that's just part of the page. Our focus is synthesizing all these sources to create a superior search experience, and there is a lot of technology behind it. For example, local searches appear on most mobile searches, and none of that is coming from Bing.

Come on Gabriel! Yes people search for weather, lyrics, sport scores and local results, but you do not have a web search engine without being able to search the web. Saying "Of course, we have more traditional links in our search results too, which we do largely source from Bing, but that's just part of the page" is disingenuous.

No, it's really not. What people don't realize about search is things get clicked on in an exponential fashion, with each piece down the page being engaged with about half as much, so nearer to the bottom of the visible page, 100x less. Since instant answers are often on top, the % of engagement on non-traditional links is much lower than one would otherwise think. And as mobile searches are now the majority, local results (including maps, places listings, etc.) occur on a large % of searches. Same for Wikipedia content. And neither of those are sourced by Bing, along with dozens of other popular Instant Answers driven by many different indexes. Put another way, we have a very large search codebase and overall engineering team, and all of this technology is doing something, and we believe something good!

Could you quantify this with some numbers? What proportion of the requests you get every day are answered completely by your own index, without using Bing?

I am pretty sure you won't get an answer to this question, it was asked many times before but nothing was revealed, who knows maybe their agreement with bing binds them, or it is just terrible for their business / illusion.

> In particular, when people search, we believe they’re really looking for answers, as opposed to just links.

I'm really looking for links. In fact I'd rather not have "answers" that the search engine just claims are true with little or no context. I trust search engines to search, and nothing more than that. They should not be a source of truth.

Reply on news.ycombinator.com