Which is faster

[Goku1814]

Me and my brother had a heated argument about which is faster and which would be faster in the future. He said that computers would be able to in the near future be faster than a human brain and I said super computers are no where near to being able to faster than a human brain. So which one of us is correct?

[Rybo]

Define faster.

[minibois]

23 minutes ago, Goku1814 said:

Me and my brother had a heated argument about which is faster and which would be faster in the future. He said that computers would be able to in the near future be faster than a human brain and I said super computers are no where near to being able to faster than a human brain. So which one of us is correct?

At some point computers will be able to do everything better than humans. 

At the moment though, humans excel in more unpredictable scenarios, where as a computer can be woefully unprepared for anything it wasn't explicitly programmed to do.

 

But thing like math (especially in very large numbers) can be done much faster by computers than (most) humans.

[gloop]

9 minutes ago, Rybo said:

Define faster.

^ this.

[Goku1814]

7 minutes ago, gloop said:

^ this.

Which can process something faster. Like being able process something that you see or something that changes randomly.

Edited July 19, 2020 by Goku1814

[Rybo]

Just now, Goku1814 said:

Which can process something faster.

Computers are faster and have always been so: they process information at the speed of light, whereas the brain transmits it at the speed of chemical reactions.  For example, think of your reaction time.  When you see something and click your mouse, the most 1337 operator in the world is like 150 milliseconds -- that's the time it takes for a signal to go in through your eye, for your brain to decide what to do with it, and for the decision to make it to your finger.  The signal then goes from the mouse, through all of the computer's computations, then back out to the screen in...what, less than 10 milliseconds?  The computer is orders of magnitude faster from the perspective of reaction times (i.e., my interpretation of "raw speed").  

 

[Goku1814]

19 minutes ago, Rybo said:

Computers are faster and have always been so: they process information at the speed of light, whereas the brain transmits it at the speed of chemical reactions.  For example, think of your reaction time.  When you see something and click your mouse, the most 1337 operator in the world is like 150 milliseconds -- that's the time it takes for a signal to go in through your eye, for your brain to decide what to do with it, and for the decision to make it to your finger.  The signal then goes from the mouse, through all of the computer's computations, then back out to the screen in...what, less than 10 milliseconds?  The computer is orders of magnitude faster from the perspective of reaction times (i.e., my interpretation of "raw speed").  

 

But there's also some stuff that a supercomputer can't do faster than a brain. Like being able to process stuff like your brain telling your heart make blood and other stuff your heart does, to tell your ears to hear something etc. Stuff like that

[gloop]

3 minutes ago, Goku1814 said:

But there's also some stuff that a supercomputer can't do faster than a brain. Like being able to process stuff like your brain telling your heart make blood and other stuff your heart does, to tell your ears to hear something etc. Stuff like that

They pretty much can do that iirc. 

[Goku1814]

2 minutes ago, gloop said:

They pretty much can do that iirc. 

But no supercomputer is faster than a human nervous system. That's the kinda speed difference I'm talking about.

Edited July 19, 2020 by Goku1814

[Rybo]

1 minute ago, Goku1814 said:

But there's also some stuff that a supercomputer can't do faster than a brain. Like being able to process stuff like your brain telling your heart make blood and other stuff your heart does, to tell your ears to hear something etc. Stuff like that

It's not a function of speed, though -- it's a software problem.  The brain is purpose built to be able to sense things and reason about things.  When we purpose build computers to do things, they basically make humans irrelevant in the field.  Our only problem is that we're not good enough programmers yet to be able to tell computers how to do things that we're good at, like reasoning about things and walking and using our fingers and the like. One could make the argument that our computers don't have enough processing power to even do those tasks yet, which is probably true.  But that's also not a function of speed, necessarily: that's more a function of bandwidth: computers aren't nearly as dense as the human brain.  

[Goku1814]

26 minutes ago, Rybo said:

It's not a function of speed, though -- it's a software problem.  The brain is purpose built to be able to sense things and reason about things.  When we purpose build computers to do things, they basically make humans irrelevant in the field.  Our only problem is that we're not good enough programmers yet to be able to tell computers how to do things that we're good at, like reasoning about things and walking and using our fingers and the like. One could make the argument that our computers don't have enough processing power to even do those tasks yet, which is probably true.  But that's also not a function of speed, necessarily: that's more a function of bandwidth: computers aren't nearly as dense as the human brain.  

Ik that its a game but in the fallout series how were they able to use a brain as a computer? How have we not been able to do that in real life? I'm genuinely wondering.

Edited July 19, 2020 by Goku1814

[Guest]

Humans have always been faster. Computers are tools for a solution, designed by humans. Humans must have built the appropriate programming, so humans found the solution to give to computers to solve.

[Rybo]

33 minutes ago, Goku1814 said:

Ik that its a game but in the fallout series how were they able to use a brain as a computer? How have we not been able to do that in real life? I'm genuinely wondering.

Because we don't know our bodies nearly as well as you might think: we can't even keep a brain alive outside the body for long enough to hook it up to machinery.  The funny thing is that the earliest implementations of this scheme (if we ultimately go that route) will be to wire the brain up to some system and let it learn how to use those new pathways, not necessarily to hook up, for example, our biological arm pathways to the robot's arm.  

This is pretty similar to existing prosthetic technology.  People who lose parts of their arms can get a prosthetic with a sensor on it that detects when a now vestigial muscle flexes or relaxes: when the sensor is activated, it activates a function on the prosthetic -- such as closing the hand.  The brain is then tasked with associating "close hand" with flexing that muscle.  We may eventually be able to isolate the nerve impulses that the brain sends when it wants your arm to close its hand, but we ain't there yet.  We leave the job of making that connection to the brain.

 

All that is to say that the brain's software is so much more advanced than our own that when we eventually hook our primitive sticks up to it, we'll be depending on the brain to learn how to use it. This ties into why the brain can do things that our robots can't do yet.  The brain has been developed over billions of years to perform these tasks -- it has all kinds of programming optimizations and shortcuts in place to let it learn to do fairly complex things -- like walking and using hands.  It's so advanced that we can hook other things up to it -- like prosthetics, musical instruments, and cars, and it can, relatively quickly, learn how to use those things as though they were a part of it.  Compare that to our own efforts at making robots that can walk.  We've had companies doing it for something like 30 years now and they can kinda, sorta walk in relatively controlled environments.  

 

It may be that we have computers today that are, pound for pound, more capable of computation than the human brain.  The problem of figuring that out is twofold: 1) Computers and brains are so different that we can't directly compare them. And 2) the programs that we can stuff into the computer are so laughably simplistic that we can't use them to compare computers and brains.  The brain can literally write a program (i.e., learn how to) walk, use hands, dance, etc, largely on its own, whereas our programs can barely do the simplest of those things with decades of human effort into the project.  

[HM-2]

1 hour ago, Goku1814 said:

But no supercomputer is faster than a human nervous system. That's the kinda speed difference I'm talking about.

Nerve impulse is multiple orders of magnitude slower than electrical signals. The fastest nerve impulses are about 120m/s. Electrical signals travel circa 300,000,000m/s.

[butre]

computers can do addition and subtraction.  nothing more, nothing less.  sometimes it uses that basic math in interesting ways, such as right here, right now, displaying this post on this page, but all it can do is addition and subtraction.  it is significantly faster at addition and subtraction than we are, but as soon as you abstract it, or worse yet introduce an element of randomness, humans become significantly faster.

[butre]

hop on a linux machine, type cat /dev/random

close your eyes and bash the keyboard for 30 seconds.

see which method nets you the most characters in that timeframe

 

e: intended to hit the edit button.  my bad