3 ms·

I've yet to see any evidence that human-like intelligence and computation speed are related in any way. There's a lot more to general intelligence than how fast

I've yet to see any evidence that human-like intelligence and computation speed are related in any way. There's a lot more to general intelligence than how fast numbers can be multiplied together.

danieltillett 11y ago
Actually in humans raw computational speed (nerve impulse speed) and intelligence (g) are quite correlated. We have a rough idea of how much processing the human brain is capable of by looking at the eye and optic nerve. Our most powerful computers are many orders of magnitude below human level processing.
- vl 11y ago
  >Our most powerful computers are many orders of magnitude below human level processing. How can this be true? Human brain consumes just so much energy, our chips are already running close to single-electron level switching, and consume comparable amounts of energy (not even talking about computer clusters/supercomputers). May be layouts/programming are not good enough, but bare computing power is there.
  - danieltillett 11y ago
    Human brains are far more efficient than computers. We tend to forget how efficient human brains are. You might find this post interesting [1]. 1. http://chrisfwestbury.blogspot.co.il/2014/06/on-processing-speed-of-human-brain.html http://chrisfwestbury.blogspot.co.il/2014/06/on-processing-s...
    - vl 11y ago
      This post lost me at last * 1000 - number of connection should not be counted. I.e. number of units x frequency gives you FLOPS. With this correction it's just 20 gigaflops for the human brain. Modern deep neutral network demonstrate that they do comparable decisions with relatively little power. (i.e. modern speech and image recognition running on laptops, for example). So I say for modern computers it's all about correct programming.
      - danieltillett 11y ago
        Well I am not sure what to add. Rather than arguing about FLOPS we should really be looking at how well computers can meet or beat humans in some activity. On the things we are bad at computers are already out in front, but at the thing our brains are good at (vision for example) they have a long way to go.

I've yet to see any evidence that human-like intelligence and computation speed are related in any way. There's a lot more to general intelligence than how fast numbers can be multiplied together.

danieltillett 11y ago

Actually in humans raw computational speed (nerve impulse speed) and intelligence (g) are quite correlated. We have a rough idea of how much processing the human brain is capable of by looking at the eye and optic nerve. Our most powerful computers are many orders of magnitude below human level processing.

vl 11y ago
>Our most powerful computers are many orders of magnitude below human level processing. How can this be true? Human brain consumes just so much energy, our chips are already running close to single-electron level switching, and consume comparable amounts of energy (not even talking about computer clusters/supercomputers). May be layouts/programming are not good enough, but bare computing power is there.
- danieltillett 11y ago
  Human brains are far more efficient than computers. We tend to forget how efficient human brains are. You might find this post interesting [1]. 1. http://chrisfwestbury.blogspot.co.il/2014/06/on-processing-speed-of-human-brain.html http://chrisfwestbury.blogspot.co.il/2014/06/on-processing-s...
  - vl 11y ago
    This post lost me at last * 1000 - number of connection should not be counted. I.e. number of units x frequency gives you FLOPS. With this correction it's just 20 gigaflops for the human brain. Modern deep neutral network demonstrate that they do comparable decisions with relatively little power. (i.e. modern speech and image recognition running on laptops, for example). So I say for modern computers it's all about correct programming.
    - danieltillett 11y ago
      Well I am not sure what to add. Rather than arguing about FLOPS we should really be looking at how well computers can meet or beat humans in some activity. On the things we are bad at computers are already out in front, but at the thing our brains are good at (vision for example) they have a long way to go.