4 ms·

Oh man, this resonates quite a bit, as a perpetual "how can I find the perfect system that melds with my mind" sort of person. For me, one of the biggest lesso

by jrm4 5y ago

Oh man, this resonates quite a bit, as a perpetual "how can I find the perfect system that melds with my mind" sort of person. For me, one of the biggest lessons is really teasing out the differences between what the computer can do and is good at, and what I can do. It's VERY easy to confuse the two. In fact, I think this is why I think most AI is completely overhyped. Computers keep perfect records and can remind you of things. This is basically the only thing the computers can definitely outdo you, the human, on. They really can't think as good as you. I think a great example of this is the Mindmap/graph stuff prevalent in many note taking deals. I used to love these things (and they're fun) until I realized that my own brain is SO MUCH BETTER than this thing at this particular task. I make these connections far better than this thing ever can.

burrows 5y ago
> This is basically the only thing the computers can definitely outdo you, the human, on. They really can't think as good as you. Can you give some intuition about “thinking”? Is “thinking” what I do when playing chess? In my case probably not, but let’s say a grandmaster. Because if so, in that domain, whatever a computer is doing seems superior to “thinking”.
- jrm4 5y ago
  But for me that's a lot like "a car going faster than a person running." Machines are good at well defined tasks, but not good at determining what those tasks should be, or even how to do them, etc.
  - burrows 5y ago
    > But for me that's a lot like "a car going faster than a person running." Yeah, I’d love to have car-like locomotive capability built in to my bio-machine. Or, maybe I wouldn’t, I dunno what all the consequences would be. > Machines are good at well defined tasks, but not good at determining what those tasks should be, or even how to do them, etc. But as a human, I mostly just copy-paste other humans. I’m not out here Solving Life from first principles.
    - jrm4 5y ago
      Weirdly, I think your second bit (solving from first principles) is much closer to the truth of how human brains work than "copy-paste?" Which harmonizes quite well with my first point? As great as computers are, I think a big problem is how they've created this tendency for bad analogies when it comes to the details. Not really different from how Frankenstein came about from people watching the advent of mechanical stuff like gears and pistons and such, and then made the LEAP to "HUMAN LIKE MONSTER." I'm not saying it will never ever happen -- but I'm still comfortable saying, even today, computers are NOTHING AT ALL LIKE human brains.
      - burrows 5y ago
        > Weirdly, I think your second bit (solving from first principles) is much closer to the truth of how human brains work than "copy-paste?" Which harmonizes quite well with my first point? I suppose we just disagree about this. My own experiences indicate that mostly I ape other humans rather than “figure” things out. > As great as computers are, I think a big problem is how they've created this tendency for bad analogies when it comes to the details. Not really different from how Frankenstein came about from people watching the advent of mechanical stuff like gears and pistons and such, and then made the LEAP to "HUMAN LIKE MONSTER." But isn’t that happening in the real world? Machines intended to present as human are becoming more human-like overtime. > I'm not saying it will never ever happen -- but I'm still comfortable saying, even today, computers are NOTHING AT ALL LIKE human brains. I don’t think computers need to work like brains in order to do most of or all of the things that humans do. As you pointed out a car doesn’t need to work like a human’s locomotive system to do like human locomotion.
User23 5y ago
I would argue that computers cannot autonomously do anything, anymore than a hammer can swing itself. It's only when some user sets them to purpose that they become operative. And while a computer can be set to task by another computer, ultimately some person started it going.
- burrows 5y ago
  Human newborns can’t autonomously do shit either.
  - jimmyvalmer 5y ago
    Quite the contary; that's the only thing they can do by themselves.

by jrm4 5y ago

burrows 5y ago

> This is basically the only thing the computers can definitely outdo you, the human, on. They really can't think as good as you. Can you give some intuition about “thinking”? Is “thinking” what I do when playing chess? In my case probably not, but let’s say a grandmaster. Because if so, in that domain, whatever a computer is doing seems superior to “thinking”.

jrm4 5y ago
But for me that's a lot like "a car going faster than a person running." Machines are good at well defined tasks, but not good at determining what those tasks should be, or even how to do them, etc.
- burrows 5y ago
  > But for me that's a lot like "a car going faster than a person running." Yeah, I’d love to have car-like locomotive capability built in to my bio-machine. Or, maybe I wouldn’t, I dunno what all the consequences would be. > Machines are good at well defined tasks, but not good at determining what those tasks should be, or even how to do them, etc. But as a human, I mostly just copy-paste other humans. I’m not out here Solving Life from first principles.
  - jrm4 5y ago
    Weirdly, I think your second bit (solving from first principles) is much closer to the truth of how human brains work than "copy-paste?" Which harmonizes quite well with my first point? As great as computers are, I think a big problem is how they've created this tendency for bad analogies when it comes to the details. Not really different from how Frankenstein came about from people watching the advent of mechanical stuff like gears and pistons and such, and then made the LEAP to "HUMAN LIKE MONSTER." I'm not saying it will never ever happen -- but I'm still comfortable saying, even today, computers are NOTHING AT ALL LIKE human brains.
    - burrows 5y ago
      > Weirdly, I think your second bit (solving from first principles) is much closer to the truth of how human brains work than "copy-paste?" Which harmonizes quite well with my first point? I suppose we just disagree about this. My own experiences indicate that mostly I ape other humans rather than “figure” things out. > As great as computers are, I think a big problem is how they've created this tendency for bad analogies when it comes to the details. Not really different from how Frankenstein came about from people watching the advent of mechanical stuff like gears and pistons and such, and then made the LEAP to "HUMAN LIKE MONSTER." But isn’t that happening in the real world? Machines intended to present as human are becoming more human-like overtime. > I'm not saying it will never ever happen -- but I'm still comfortable saying, even today, computers are NOTHING AT ALL LIKE human brains. I don’t think computers need to work like brains in order to do most of or all of the things that humans do. As you pointed out a car doesn’t need to work like a human’s locomotive system to do like human locomotion.

User23 5y ago

I would argue that computers cannot autonomously do anything, anymore than a hammer can swing itself. It's only when some user sets them to purpose that they become operative. And while a computer can be set to task by another computer, ultimately some person started it going.

burrows 5y ago
Human newborns can’t autonomously do shit either.
- jimmyvalmer 5y ago
  Quite the contary; that's the only thing they can do by themselves.