Live data from Hacker News

Ask HN: Is 'search' a solved problem?

news.ycombinator.com

11–20 of 125 posts

Re: Ask HN: Is 'search' a solved problem?

#11
I've learned that search is very domain-specific and it's only "solved" if you can pin down what "relevant" means for your corpus of documents.

In terms of fundamentals, I'd suggest reading about tf-idf, which is the basis of Lucene (which powers Solr and Elasticsearch).

Re: Ask HN: Is 'search' a solved problem?

#12

I think Google has pretty much monopolized search. You could say it's solved. I doubt if there are any complaints that sound like 'i couldn't find something using Google'

Google ultimately serves the advertisers, not the end users. This incentivizes manipulation of search results in order to maximize ad growth. On that token alone, they have not solved search. Any newcomer developing a search service would do well to make it radically different from Google rather than a clone with added privacy or a Microsoft logo.

Regarding always being able to find what you are looking for on Google: I often struggle to find niche information using Google, and most content on the Web is not indexed by Google.

I'm more of the opinion that YouTube is less possible to compete with in any serious way.

Re: Ask HN: Is 'search' a solved problem?

#16
If you just want a primer on how to think about adding search to a product, this piece by Max Grigorev is a great starting point: https://medium.com/startup-grind/what-every-software-enginee... It's ostensibly for the engineer, but actually feels more like it's written from the POV of a product manager.

Re: Ask HN: Is 'search' a solved problem?

#17

It depends what you are looking for. I have the hardest time searching for laptops that meet all of the specs I want, for instance. Takes days. Minimum screen brightness. 1440p touch. Needs USB-A. 8th gen quad core. 20+ watt TDP. etc. Not too thin, not too thick. decent graphics.

Can't that be solved with a bunch of checkbox filters, generating the proper WHERE statements? Several websites offer this, be it for laptops or other products.

Re: Ask HN: Is 'search' a solved problem?

#18
Google's good for 1 - 3 key words / phrases but anything longer can still be difficult.

Regardless of query length it's strange that I often find the second result better than the first one.

Finding something in Gmail can be a real challenge, I wonder if would be possible to unleash Algolia on your inbox.

Re: Ask HN: Is 'search' a solved problem?

#19
It's not that people have moved on. It's that the entire culture of the ecosystem is built upon a narrative that Google is an all-powerful machine that cannot be stopped or contested. So people don't try to compete, and if they do, they will be ridiculed for it.

And for what good reason? Certainly not because of past attempts. In fact, Google has bought some startups that were involved with search.

I'd be interested in knowing what you find. There are bound to be many relevant texts not labeled as search, if you know what you are looking for. Perhaps I would start with web crawlers and go from there.

Post reply on HN