Re: TI-H: mind vs. computer

To: <ti-hardware@lists.ticalc.org>
Subject: Re: TI-H: mind vs. computer
From: "David Knaack" <dknaack@geocities.com>
Date: Thu, 17 Dec 1998 21:34:02 -0600
Delivered-To: ti-hardware-outgoing@towerguard.unix.edu.sollentuna.se
Reply-To: ti-hardware@lists.ticalc.org

From: John Teffer <jdteffer@wcf.net>
>David Knaack wrote:
>>
>> From: Dan <danti@applecyber.dyndns.com>
>> >Nah... you can't measure it.. cells die, and therefore memories die...
>>
>> The neurons that you are born with are the same ones that you
>> die with.  If they were to die regularly it is unlikely that we could
>> continue to function for very long.
>>
>
>It's estimated that we lose 50,000 neurons a day, out of a total of 100
>billion.

Thats 5/10,000ths of a percent, or less than 0.1% over a lifetime,
not a significant loss (I should have said 'if a significant number were
to die...).

>The important concepts are stored with such massive redundancy
>that a couple hundred thousand neurons is nothing.

Also taking into account the adaptability of the neurons when faced
with new concepts, the loss of a few can probably easily be compensated
for with minor changes to configuration.

>In reference to the capacity and speed of the brain, I'm looking at book
>from 1990 that puts the capacity of the average brain at 100 trillion bits
>(sorry, to lazy to convert to terabytes), and that's assuming each neuron
>can hold 1,000 bits.

Hmm, I don't know where they get their numbers, but if they are calculating
assuming that every neuron in the brain is associated with storing memories
I'd say their estimate is probably high.

>Neurons perform an analog computation in ~5 milliseconds.  The brain has
>about 100 trillion neuron connections, if one percent of the connections
>are active, that gives us 200 trillion computations a second. (but I cant
>seem to convert that to mhz in my head at the moment, so is my four bit
>pocket calculator more powerful than my brain?  No, don't answer that.)

Difficult to compare, since processors run (mostly) sequentially, and brains
run massively parallel.  A brain given a task such as image recognition will
easily perform the compare (which lends itself to parallel computation) in
just a fraction of a second.  A sufficently fast processor may be able to
keep
up, but at current technology thats unlikely.  Given a sequential
calculation,
say, determining if a 100 digit number is prime, the brain will likely take
vastly longer to do the calculation, since it is not designed to do such
calculations with any kind of speed.

Its kind of like comparing the merits of an operating system, the strength
of the system will depend on the task it is assigned.

DK