Just a teaser, as I really should save the best for the book, but I had a very interesting talk with Jeff Bezos yesterday. A significant insight of his, which came up as I was pushing to understand Amazon’s long-term interest in A9, was his use of the term “discovery” as an umbrella term which incorporates search. I think in the end when I use the word “search” I really mean “discovery” as Jeff uses it. What’s discovery? Well, much more in the book, but in the end, it’s search plus what happens when the network finds things for *you* – based on what it knows of you, your actions, and your inferred intent. Inferred intent? How might the network be smart enough to do that? Ay, there’s the rub….
A Call With Bezos
Just a teaser, as I really should save the best for the book, but I had a very interesting talk with Jeff Bezos yesterday. A significant insight of his, which came up as I was pushing to understand Amazon's long-term interest in A9, was his use of the term "discovery"…
Hi John,
Steve Lawrence formally with NEC research (along with Gary Flake (Yahho R&D)), now with Google I believe (coud be wrong) wrote a good paper called “Context in Web Search” – it i’ll be on citeseer.
“Inferred intent” was the key theme running through his work at the time, and also referenced “Haystack” with I believe was down to Manber over at A9 (wouldn’t it be great to get them all in a room and thrash it out for the better of mankind instead of $$)
Come on John lets have a few more pieces of the jigsaw, whe are going to buy the bloody book anyway 🙂
What I’ve not seen you comment on yet (coud be wrong) was the recent paper out of the googleplex on “query free news search.” The intersection of the broadcast (i.e. streaming media) world & the web via open/closed captioning & efficient algo’s. Real time data mining + real time ads for the bucks – when your watching TV via your DSL connection (ala MSFT media centre etc) oh boy we are in a whole new ball park.
sorry for ranting, any chance you might cover this?
Here are a few maps of Amazon data that can be used for…
* discovery
* inferred intent
* context exploration
http://www.orgnet.com/buzz.html
http://www.orgnet.com/divided2.html
Perhaps it won’t be a network finding things for you. perhaps it would be the operating system.
Makes since for what Amazon’s business is, especially considering using customer searches to fine tune delivery of the correct product for your search.
Discover and deliver.