Quantcast
Membership Signup
Singularity University

Exclusive Interview With Ray Kurzweil On Future AI Project At Google

In an exclusive with Singularity Hub, Ray Kurzweil gave one of his first interviews since the December announcement that he joined Google full time as Director of Engineering. Speaking with Singularity Hub Founder Keith Kleiner, Ray discusses his new role, how his research interests connect with his latest book How To Create A Mind (which Keith recently interviewed Ray about here), and how technology will advance to produce a “cybernetic friend”

“The project we plan to do is focused on natural language understanding,” said Kurzweil. “We want to give computers the ability to understand the language that they’re reading.”

Check out the interview to hear more:

Regarding the specific kind of artificial intelligence that a Kurzweil-led project will aim to do, he said, “It will know at a semantically deep level what you’re interested in, not just the topic…[but] the specific questions and concerns you have.” He added, “I envision some years from now that the majority of search queries will be answered without you actually asking. It’ll just know this is something that you’re going to want to see.” While it may be take some years to develop this technology, Kurzweil added that he personally thinks it will be embedded into what Google offers currently, rather than as a stand-alone product necessarily.

Now if you’ve been following Singularity Hub’s coverage of personal assistants like Siri, Evi, and the latest, Maluuba, as well as Google Voice Search, then you know that natural language recognition is one of the highest priorities for tech companies today. That’s exciting because it means that holding sophisticated conversations with computers — in much the same way that Dave Bowman does with HAL 9000 in the movie 2001 – is going to become a reality very soon.

As Kurzweil points out, the hurdle currently is that language is hierarchical, and the human brain processes language in a hierarchical way, depending on what stimuli it receives during key stages of development. Computers like IBM’s Watson are just now being programmed to process human information in a related way. Inevitably, the sophistication of this software will grow — slowly, at first, but in all likelihood become exponential, as with many other technological trends that Kurzweil himself has identified.

Though the video is only 10 minutes, it’s great to hear Ray download some more tidbits about what he’ll be doing once he enters Google’s doors. Odds are that when he reemerges, the ability of our computers to understand us is going to take a quantum leap.

Reply

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>

17 comments

  • vmagna says:

    I want to know if there’s a ROS PR2 bot for programming sakes at google.
    Also want to know if googles effort in A.I. could make Nuance irrelevant.
    Would google make their own robot if they succeed? ( besides the cars)

    Would very useful and personal AI service require advances in chip tech or is it possible now with some programming chops?

    What’s going on with asimo? What’s going on with nao? What’s IBM’s Watson up to?

  • Harley Jacobson says:

    “i had a long talk with ray kurzweil last week. we had coffee together. and he was telling me that he would like to live to see the day when computers replace clothing.

    so, we were sitting there and talking about this amazing technological advancement – when ray just pulled his pants off right there in the coffee shop.

    i said, ‘ray! you have no pants on!’

    ray replied that i was only imagining that he was unclothed. and that billions of tiny neuro-clothing activators were actually clothing him with invisible pants.

    transhumanists are just like Emperor Ray with no clothes on.

    you are naked before God. God sees your Evil.

    i am not against technology. i am not against you. i am against your ideas.”

    • jdudley says:

      Interesting that an omniscient, omnipresent, omnipotent being would need you as its personal spokesperson. That’s a serious design flaw.

      I pity any soul who claims to speak for God when (if) they actually meet their creator.

      • SA23 says:

        1) There is no good reason to believe in “God” as presently described.

        2) Some transhumanist ideas are problematic from utilitarian and libertarian perspectives, but silly jokes don’t contribute to the conversation; they lower the level of discourse.

        3) If the present course continues, clothing may be composed on nanobots. This is the least of our concerns — the greatest concern is that the state gets ahold of the more militarily and informationally potent technologies and uses them against its subjects.

        • DanHegelund says:

          I am perfectly comfortable with the Singularity as a believer.

          

The Son in the Trinity is also called the Logos, Wisdom, and the Word. These are just other words for Intelligence (Proverbs 8:23-35, John 1:1).

          

“Until now I have spoken to you figuratively in parables, but a time is coming when I will no longer use this kind of language, but will speak plainly about my Father” (John 16:25)

          The evolution of man through Intelligence had only reached a stage of maturity where narrative/parables was what man was best capable of understanding. “Oh how I wish I could give you meat, but I give you milk” (1 Cor 3:2). So until now Intelligence (the Logos/Wisdom/Word) had been speaking to us through parables. Now the “Kingdom of God” (Singularity?) has been born, which is the age of Intelligence (Logos, Word, Wisdom, Christ).

          God did use parables and narrative before. Now he is using the thing itself. Intelligence was called the Son or Logos or Wisdom or the Word. This was in part to show us what Intelligence was capable of. Jesus Christ was capable of curing deceases, raising the dead, overcoming death, changing water into wine. etc.

          I could elaborate, but this post is already too long.

  • dobermanmacleod says:

    BTW, GREAT GIT Singularity Hub! What Ray hasn’t said is that this “cyber friend,” will also serve as companion and therapist, and that characters like movie theater gunman James Holmes will be tracked by such a computer mechanism, as he would have disclosed his feelings and thoughts to his “computer buddy” for months and years before acting on his darkest impulses. Another example of our lack of privacy in the digital age corresponding to increase in security. It ought to be noted that passive surveillance need not be intrusive, and programs like Mind’s Eye will monitor all cameras for “suspicious behavior” which will then be flagged for further scrutiny. I love it, but some people using the old paradigm of fear will hate it.

  • Alex Murphy says:

    Thanks for posting the interview! Very excited to witness his work at Google :-) I do wish there was better audio in the interview…I’m sure even a basic shotgun microphone would have improved the quality tremendously.

  • Vector says:

    I love Ray. His presented analysis on technological trends is astounding, to say the least. Although I know he was not and is not alone in “predicting” them, he sure is, in my opinion, the right man to be doing it. He does it so well and with great devotion.

    I admire his tenacity in attempting to understand the nature of consciousness (human consciousness in particular) to possibly replicate it through a non-organic medium.

    I hope all bodes well for Ray on his personal journey into Google innovation.

  • Sergio123Cabral says:

    Well, This is really a doubt to me. Instead programming computer to understand humans , it will much better programming humans to understand computer. Human Natural Language are not good to describe precise logical process. All of our Natural Language are embeebed with a high dose of subjetive components and also very level of sintatic conflict. THis is the reason we ( Humans ) have invented some other languages like mathematics and physics. So. I think taht is much better go in the oposite , or at least , in a hybrid direction.
    — Sergio

    • fireofenergy says:

      The transparent cybernetic friends will help us to search for what we want to learn, a million times faster. Say, “I want to learn something”. It weeds out all the unrelated stuff that, today, merely wastes our time as we may spend countless hours searching page after page of results, to no avail, and much frustration.

      Eventually, “it” will put 2 and 2 together and tell us how to save us (and the rest of the biosphere) from ourselves!

      I would think God would WANT us to do that!

      • fireofenergy says:

        Sorry, this post was meant for general response.

        After reading yours, agree that training people to understand how AI works would help elevate our language abilities…
        And our ability to transition into (what I believe should be) the machine economy based upon the efficient extraction and distribution of energy and resources in an enviro friendly manner limited ONLY by physics and our yet unknown abilities to do so.

        I must add that planet Earth has WAY more resources than environmental hopefuls can even want to imagine (once we learn the proper extraction and distribution techniques, and when fossil fuels are a thing of the past)!

  • fireofenergy says:

    Cybernetic friends will be the best way to (figure out) RE < coal.

  • emoshape says:

    Emoshape Announces First Generation Emotions Processing Unit – EPU
    LONDON–(BUSINESS WIRE)–January 28, 2013– Emoshape (www.emoshape.com) announces to have started the prototyping of the first Emotions Processing Unit — EPU on DSP in the prototyping of a first generation of processor representing a significant advancement for the purpose of Artificial Intelligence Devices and Technologies.
    http://online.wsj.com/article/PR-CO-20130128-906426.html

  • buzz says:

    This has all (Kurzweil’s claims) been said before, nothing has worked (involving ‘deep’ AI). Kurzweil’s current claims are simply old wine in a new bottle, updated for the 2.0 era. BFD.

Singularity Hub Newsletter

Close