TED Conversations

Jordan Reeves

TED Conferences

TEDCRED 50+
TED StaffAssociateTED Attendee

This conversation is closed. Start a new conversation
or join one »

In ten words or less, what is a question no one (yet) knows the answer to?

Is there life outside of our planet?

What's a bigger factor in our development: environment or genetics?

Is there any real truth, or is everything relative?

What will the earth be like in 100 years?

Topics: answers questions
+19
Share:

Showing single comment thread. View the full conversation.

  • thumb

    Logan Crouse

    • 0
    Dec 9 2011: Will we discover an algorithm for generating true human intelligence?
    • thumb

      Philip Sinclair

      • +1
      Dec 9 2011: speak for yourself ;)
    • thumb

      James Wallis Martin

      • 0
      Dec 9 2011: The answer is no. Human intelligence is comprised of non rational components that comprise our human intelligence. Algorithms depend on all components being rational.
      • thumb

        Logan Crouse

        • 0
        Dec 9 2011: I'm not sure I follow; even the non-rational components of human intelligence lead to definite solutions to a problem. The heuristic approaches the human mind employ, while ambiguous at best, are still effective, which would suggest an organizational heirarchy that can cope with this ambiguity and leads to satisfactory solutions.

        As a result, shouldn't it be possible to quantify the rules that this structure employs and translate them into a machine-code compatible format? Perhaps through the machinations of some rule-based computing language, like LISP? All that would remain would be to identify these structures and reduce them to their simplest forms, wouldn't it?

        Fuzzy logic deals a lot with this idea; values within a fuzzy logic system are not exact, but are approximate. In many ways, the variability of a value between 0 and 1 simulates a neuron and the energy with which it fires. And they are beginning to be implemented in consumer electronics.

        Fuzzy electronics:
        http://en.wikipedia.org/wiki/Fuzzy_electronics

        Fuzzy logic:
        http://en.wikipedia.org/wiki/Fuzzy_logic
        • thumb

          James Wallis Martin

          • 0
          Dec 10 2011: I earn a living working with fuzzy logic. The question of preference for Chocolate Ice Cream over Vanilla Ice Cream does not have an algorithm.
        • thumb

          Mariusz K

          • 0
          Dec 10 2011: It could choose between chocolate ice cream and vanilla ice cream by asking virtual organism simulation for it needs (which is analogy to our likes / desires). If you would add hormones' mechanics, organism's mechanics (that influence our human thoughts) then you could get to somethign no different than human intelligence.

          Are we not just AI influenced by organism's needs / desires?
      • thumb

        Grant Robertson

        • +2
        Dec 10 2011: James, in regards to your statement: "The question of preference for Chocolate Ice Cream over Vanilla Ice Cream does not have an algorithm."

        I don't think that was exactly the question. The original question was "Will we discover an algorithm for generating true human intelligence?" Now, we could pick some nits with the wording of this question. For instance, algorithms are designed not discovered. And the only "true" human intelligence would, by definition, be only in a human. But there is one key nit that I believe you missed which creates an important distinction in this context. Whether intentionally or not, Logan did not ask if we will develop an algorithm that itself has human intelligence. He asked if we would develop an algorithm which would in turn be able to GENERATE human intelligence.

        Genetic algorithms can be used to GENERATE other algorithms and/or hardware which quickly become more complex than even the creator of the original algorithm can understand. I believe the right algorithm, running on the right hardware and making modifications to the right mix of hardware and software can and will eventually generate human intelligence. I feel the problem many researchers in this field have is that they limit themselves to what they can do with the kinds of binary, sequential processing, instruction following, register shifting kind of logic which they have available to them now. Remember, massively parallel processors are still a finite number of processors which each execute instructions sequentially.

        Think of the one major difference between a brain and ANY current processor.... nerves don't have a clock signal. Nerves just react in their own time. This is more akin to how digital electronics worked before timing clocks were invented and the signals just cascaded throughout the circuit. The clock was invented to tame the difficulty of getting all the timings correct throughout the circuit. Imagine using those timings to your advantage instead.
        • thumb

          Logan Crouse

          • 0
          Dec 10 2011: Yes sir, that's exactly what I had in mind when I worded the question. You were right in your other nits, as well; algorithms are designed, not discovered, and, of course, the only true "human" intelligence would be in a human.

          I am glad that you moved beyond those nits to grasp the essence of what I was attempting to ask, though, and you said it quite eloquently in the subsequent paragraphs.
        • thumb

          James Wallis Martin

          • 0
          Dec 11 2011: My point is there is more to true human intelligence that just intellect (which in humans is flawed - e.g. we forget) there is emotional intelligence and some would argue spiritual intelligence. We have both nurture and nature that govern our experiences and shape our human intelligence. Star Trek: The Next Generation dealt with the topic throughout all of the series with Data and even in the original series with Spock. In order for the algorithm to be "true human" would have to be a human. What you are arguing is what we do when we procreate.

    • Richard Fravel

      • +1
      Dec 9 2011: If we think we have we would spend a millinium tearing it apart because too any of us wouldn't believe it or would theorize it to death to the point we would never trust it.
      • thumb

        Logan Crouse

        • 0
        Dec 10 2011: Hehehe. You mean---like we do with our fellow human beings?

Showing single comment thread. View the full conversation.