3 ms·

"Building a Better Search Engine." How many times have you heard that technology 30 years in the works is about to change everything?

2 points by ivankirigin 19y ago

davidw 19y ago
Enough that I have a suspicion that trawling for stuff that didn't quite work out and was more or less abandoned 20/30 years ago might yield some interesting nuggets.
aswanson 19y ago
This rant pretty much sums it up: http://blog.searchenginewatch.com/blog/061005-095006 http://blog.searchenginewatch.com/blog/061005-095006
sabat 19y ago
I like the rant (on searchenginewatch) and I think he summed up a lot of valid points. I can still imagine a useful natural-language search engine, though. This morning, for example, I was trying to remember the name of a podcast. One of the hosts' first name is Jeanette, and I know she has a Spanish last name. Now, if I could've typed in something like: podcast with host Jeanette who has a spanish last name and it actually were able to figure it out, then I think the search engine could cause a paradigm shift. What I'm really talking about is closer to pure AI, not just natural language concepts. Apple == a company and a fruit, depending on context -- that's natural language. What a Spanish last name is -- that's AI in my book. So maybe we need to be pursuing an AI search engine rather than a natural-language search engine.
- sabat 19y ago
  I'm not even sure we're in desperate need of AI. I wanted to read an article about how Robert Scoble reads 1000 RSS feeds a day, and how he does it. (Something like that, anyway.) I couldn't remember the blog I noticed it on, and didn't remember anything specific about what it said. Google: +how scoble reads rss feeds and it returned exactly the article I wanted (www.fourhourworkweek.com/blog/2007/05/16/how-scoble-reads-622-rss-feeds-each-morning/) as the #1 choice. I have my complaints about Google's search and the others as well. But it's still pretty darn good, and any advancement to this is going to have to be a big leap to get any traction.
nickb 19y ago
To say that I'm highly skeptical is an understatement. Problem with all these new search engines is that they are trying to change the way people input queries into that search box. For 10 years, we've been taught to type in a term or two and get what we want. Now Powerset is trying to change that by forcing us to type in sentences and long queries. I don't think their success will happen overnight.

2 points by ivankirigin 19y ago

davidw 19y ago

Enough that I have a suspicion that trawling for stuff that didn't quite work out and was more or less abandoned 20/30 years ago might yield some interesting nuggets.

aswanson 19y ago

This rant pretty much sums it up: http://blog.searchenginewatch.com/blog/061005-095006 http://blog.searchenginewatch.com/blog/061005-095006

sabat 19y ago

I like the rant (on searchenginewatch) and I think he summed up a lot of valid points. I can still imagine a useful natural-language search engine, though. This morning, for example, I was trying to remember the name of a podcast. One of the hosts' first name is Jeanette, and I know she has a Spanish last name. Now, if I could've typed in something like: podcast with host Jeanette who has a spanish last name and it actually were able to figure it out, then I think the search engine could cause a paradigm shift. What I'm really talking about is closer to pure AI, not just natural language concepts. Apple == a company and a fruit, depending on context -- that's natural language. What a Spanish last name is -- that's AI in my book. So maybe we need to be pursuing an AI search engine rather than a natural-language search engine.

sabat 19y ago
I'm not even sure we're in desperate need of AI. I wanted to read an article about how Robert Scoble reads 1000 RSS feeds a day, and how he does it. (Something like that, anyway.) I couldn't remember the blog I noticed it on, and didn't remember anything specific about what it said. Google: +how scoble reads rss feeds and it returned exactly the article I wanted (www.fourhourworkweek.com/blog/2007/05/16/how-scoble-reads-622-rss-feeds-each-morning/) as the #1 choice. I have my complaints about Google's search and the others as well. But it's still pretty darn good, and any advancement to this is going to have to be a big leap to get any traction.

nickb 19y ago

To say that I'm highly skeptical is an understatement. Problem with all these new search engines is that they are trying to change the way people input queries into that search box. For 10 years, we've been taught to type in a term or two and get what we want. Now Powerset is trying to change that by forcing us to type in sentences and long queries. I don't think their success will happen overnight.