Hacker News new | past | comments | ask | show | jobs | submit login

> Search, on the other hand, still has massive room for improvement.

This is true, but it's also a really hard problem and most likely any improvements will require enormous scale to implement (because they'll be based on statistical analysis of prior search queries), which only Google and a handful of others have. So no network effects, but huge economies of scale.

Plus, having used some other search engines (DuckDuckGo, Yandex, Bing, etc.), I feel that Google still has a comfortable lead than the competition.




I suspect you could go a rediculusly long way with crated searches. Basically page rank with ~10-100x the hand tuning to more or less kill off SEO.


http://searchengineland.com/google-25-of-queries-are-new-add...

25% of google searches are unique. I bet another 25% would be so rare and numerous you couldn't possibly have time to tune them. In fact, when you think about how much revenue a single search brings, a search is going to have to happen on the order of 1,000 times (at least?) before you can have a human hand tuning them.


That's from 2007. I wonder how much Google's autocomplete has changed that statistic. Even still, 25% of searches may be unique strings, but I doubt 25% of searches are meaningfully/semantically unique.




Consider applying for YC's Spring batch! Applications are open till Feb 11.

Guidelines | FAQ | Lists | API | Security | Legal | Apply to YC | Contact

Search: