Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

How Ray Kurzweil Plans To Revolutionize Search At Google

forbes.com

Post reply on HN
Print view View original

61–70 of 114 posts

Re: How Ray Kurzweil Plans To Revolutionize Search At Google

#61

Post by OGinparadise »

comex wrote: post #20
rsync wrote: post #11

All grand plans indeed ... but a higher priority should be making search even work at all. Google search returns approximate results, related results, results with your words split (or joined) and other nonsense to drive ad views. Even using the so-called "power tools" like allinsite: and the (barely functional) quotation marks, you still get very shoddy results. Anyone who ever searches for very specific groups of t…

As annoying as the approximation can be, I'd like a citation for "to drive ad views". It seems much more likely that it was added for the obvious reason - to improve search results, on average, for most people, because relevant results don't always contain the exact search term you put in. In my case, as many times as it's burned me, there have also been a few where I remember it presenting exactly the results I want…

>> As annoying as the approximation can be, I'd like a citation for "to drive ad views".

Not 100% proof but ad views and clicks seem to be increasing by 20% to 30% each quarter. Google tests everything to death and they please wall st with those ad click increases.

Re: How Ray Kurzweil Plans To Revolutionize Search At Google

#62

Post by OGinparadise »

DanBC wrote: post #52
threeseed wrote: post #19

Earlier quoted context omitted.

It is just me or is it a recent change ? I find Google Search to be incredibly frustrating to use because it chooses to ignore search terms even when pages exist with those terms. Google better be really careful because crippling the experience in order to sell more advertising reminds me why I switched away from Altavista.

> Google better be really careful because Sadly, you are in a minority. Google's results are excellent for most people. I used to use + often. We know that the + operator was rarely used, and most of those times it wasn't used correctly. (Of all searches, only 1 in 600 were correctly using the + operator.) Being outside Google (and not able to see their data) is frustrating, but they have a lot of numbers and they do…

>>Sadly, you are in a minority. Google's results are excellent for most people.

Sadly, you are in a minority. Bing's results are excellent for most people. Slap a Google logo and I'm willing to bet that most people wouldn't tell the difference. It's the Google brand that is golden, at least so far. But lately people are talking and questioning their honesty.

Re: How Ray Kurzweil Plans To Revolutionize Search At Google

#63

Post by revelation »

georgemcbay wrote: post #30
threeseed wrote: post #19

Earlier quoted context omitted.

It is just me or is it a recent change ? I find Google Search to be incredibly frustrating to use because it chooses to ignore search terms even when pages exist with those terms. Google better be really careful because crippling the experience in order to sell more advertising reminds me why I switched away from Altavista.

Google search for me (and I've been using it a very long time, having been an Internet user since the late 1980s and around when Google first appeared) has generally gotten worse over time, not better. I don't think the changes are to sell more advertising, though, I think they are a combination of: 1) The web is just much bigger and noisier than it ever was. There are so many SEO-bait sites out there now, it is a wo…

The biggest thing for me has been localization. If you happen to use a localized version of Google, it will heavily favor results in that language, and give you mostly crap when searching for stuff in a different language using the localized search machine (so pretty much all the time for programmers).

Thankfully you can still tell it to use Google.com and in english.

Re: How Ray Kurzweil Plans To Revolutionize Search At Google

#64

Post by fchollet »

gavinh wrote: post #5

I'm working on some open source projects that pertain to recognizing entailment in plain text. Entailment is the relation that holds when one text "follows from" another. For example, "Carolina beat Duke" follows from "Carolina defeated the Duke Blue Devils once again". I have a demo here: http://ec2-23-22-22-135.compute-1.amazonaws.com:8001/demo Here's an example app that uses entailment recognition to answer natura…

I like your demo. Do you have references to papers that you used as a basis for your code?

Re: How Ray Kurzweil Plans To Revolutionize Search At Google

#65

Post by gavinh »

fchollet wrote: post #64
gavinh wrote: post #5

I'm working on some open source projects that pertain to recognizing entailment in plain text. Entailment is the relation that holds when one text "follows from" another. For example, "Carolina beat Duke" follows from "Carolina defeated the Duke Blue Devils once again". I have a demo here: http://ec2-23-22-22-135.compute-1.amazonaws.com:8001/demo Here's an example app that uses entailment recognition to answer natura…

I like your demo. Do you have references to papers that you used as a basis for your code?

Thanks. It is based on http://nlp.stanford.edu/~wcmac/papers/nli-diss.pdf.

I use a simpler alignment representation, and use semantic role types in the predictions.

Re: How Ray Kurzweil Plans To Revolutionize Search At Google

#66

Post by joe_the_user »

dude_abides wrote: post #15

Douglas Hofstadter once said of Ray Kurzweil and his grand ideas: "It's an intimate mixture of rubbish and good ideas, and it's very hard to disentangle the two, because these are smart people; they're not stupid." I think Google (and Larry Page) is betting that they will be able to separate out this mixture.

As I recall, Hofstadter said this about "the singularity" in particular. I don't Kurzweil is claiming he will bring this about at Google, at least not before his rough projected arrival date of 2030-40.

Re: How Ray Kurzweil Plans To Revolutionize Search At Google

#67

Post by dave_sullivan »

ot wrote: post #17
dude_abides wrote: post #15

Douglas Hofstadter once said of Ray Kurzweil and his grand ideas: "It's an intimate mixture of rubbish and good ideas, and it's very hard to disentangle the two, because these are smart people; they're not stupid." I think Google (and Larry Page) is betting that they will be able to separate out this mixture.

The irony is that you could say the exact same thing about Douglas Hofstadter himself.

As with so many things, it takes one to know one

Re: How Ray Kurzweil Plans To Revolutionize Search At Google

#68

Post by joe_the_user »

coldtea wrote: post #32
ot wrote: post #17

Earlier quoted context omitted.

The irony is that you could say the exact same thing about Douglas Hofstadter himself.

Really? I don't think so. For one, Kurzweill DOES have "grand" ideas. Extravagant visions of a future with technological immortality, singularity, etc. Hofstadter does not. He merely examines some things, like cognition, and proposes some theories about their workings. Like, you know, every scientist. Kurzweill comes out as grandioze, obsessive and deluded, Hofstadter like a normal writer, no more or less strange tha…

You are correct that Hofstadter doesn't have grand assertions at the level of The Singularity but there is still something like a "Hofstadter" mix - say all the ideas in Godel, Esther and Bach with the implication that these are related and important. One might reasonably say this a somewhat grandiose admixture.

All that said, I think Hofstadter's "mixture of good ideas and rubbish" description isn't a good characterization of the thinking behind The Singularity. Rather, Singularity thinking is extrapolation of growth trends which might easily be false but since it is made on a broad level, it is not easy to formulate why it is false. And Kurzweil, in particular is popularizer of the singularity with a particular version of it.

One could argue there are limitations to the expansions that Singularians have been extrapolating from but characterizing these limitations is itself quite a tricky problem.

I think the best refutation to the Singularity idea is the argument of Paul Allen, that software and the understanding of intelligence simply haven't been amenable to increased computing power.

http://www.technologyreview.com/view/425733/paul-allen-the-s...

It's a good argument but I feel the tone isn't quite fair, the sense that you could more fairly say the singularity looks plausible until you really focus on the software barrier.

Re: How Ray Kurzweil Plans To Revolutionize Search At Google

#69

Post by harshreality »

coldtea wrote: post #45

> How Ray Kurzweil Plans To Revolutionize Search At Google By rigorous hand-waving, as he mostly does things the last 1-2 decades?

I attended a talk he gave on campus when I was in college, and afterward I lurked as some really smart people (grad students, maybe one or two profs) asked him for details. He came across as a con artist, selling ideas that he can't defend. For instance, one that stuck out was that he had nothing to back up his assertion that within x years there would be the ability to non-destructively copy all the information in a brain. Several questions I thought were fairly reasonable were met with the response: "read my book[s]." (in context, as I recall he was promoting _The Age of Spiritual Machines_.)

His basic argument of the inevitable advance of technology to true AI, or ability to upload consciousness or integrate computers with brains, rests on priors that are not clearly relevant.

With his latest title, "How to Create a Mind: The Secret of Human Thought Revealed", how can anyone take him seriously? Scientists (practicing scientists) have been writing books like that for decades, and the lesson should be that unless you've demonstrated that you can build an AI (real AI, not soft AI like machine learning), you have no business using a title like that. If real AI is simply a combination of the right machine learning tricks, then the trick is in the combination, because nobody has figured it out (publicly) yet.

Re: How Ray Kurzweil Plans To Revolutionize Search At Google

#70

Post by ippisl »

I wonder what would be the best strategy for google to monetize such an asset as ibm watson? Surely ads based model is not the best way(since they've already got a large part of the internet ad market and given that only google, ibm and maybe microsoft could build something similar.

If it's a truly competitive advantage, wouldn't offering it for startups/other-businesses in return for equity would offer bigger profits?

Post reply on HN