“Turing Test with Numbers – Part II”
Note: You should read Free Will #2 before you
read this!
In my
previous entry, I proposed the following test to illustrate something uniquely
human that computers cannot do (and I suggest, will never be able to do). I wondered if it would be possible to create
a series of numbers – digits – that are uniquely human? I came up with a series that I thought might
fit the bill. I listed six series of
digits, twenty digits each series, and one of the
six series I claim is uniquely human. Could you figure it out?
Here are the number series:
A.
34123034323639550691
B.
12345678901234567890
C. 98832863158718824883
D.
14916253649648110012
E. 71828182845904523536
F.
04656464621583251630
What’s the
answer? Well, first let’s eliminate the
two easy ones:
B – That just the digits 0-9 in order, starting with 1. If you struggled with that one, then you
really are human!!
Before I go
on to the rest, did you try “googling” these number series? I’m assuming that most advanced computers
would do just that. The amount of
information you might gain is astounding and, at least to some extent, quite
decipherable by a computer with an intelligently (i.e. by a human!) programmed
algorithm. Well, if you did google these,
you’d find the following (at least this is what I found…of course if this blog
gets indexed, then when you search, you might find this blog and be able to go
directly to the answer. This is a major
problem that I discuss later in this entry.):
A. No match found for
this number.
B. Lots of hits. Not sure that this would be helpful to a
computer or a human. But again, this one
was really obvious. If you can’t pick
out this pattern, what pattern can you pick out? Again, the inability to identify this one
would certainly be uniquely human.
C. Google identified
this as a possible Fedex package tracking number. It is not.
Looks like no real matches were found.
D. I found one
matching page when I searched this.
That’s hard to do. I remember
there used to be a game where you tried to come up with a term that only
matched on a single page. The problem is
that as soon as you post it, it is no longer on just one page! Anyway, if you could read Japanese…or was it
Chinese (?) … it might have helped you.
But again, this should have been a pretty simple series (like B).
E. Of all the
entries, googling this one should have provided some help in figuring it
out. This is one you either know
immediately because you are a real math geek, or you would have a really hard
time figuring it out without the assistance of google (or a friendly math geek
to ask).
F. No entries found.
OK, what do
we learn?
D. Maybe not quite as
easy as B, but this one would have been simple for any computer, and I think if
you thought about it for a bit, you would get it. 1, 4, 9, 16, 25…it’s just the square of the
digits starting with 1. And, since these
series only have 20 digits, the square of 11 is truncated in this series. I wonder if that would throw off a
computer? Probably not.
E. Well, maybe google
helped you out here. Or maybe you took
the hint from the body of my previous entry when I talked about pi. Now
do you know it? These are the first
twenty decimal digits of the value of “e”.
If you don’t know what e is, don’t worry about it – you probably learned
about it in algebra, but you have long forgotten. Computers don’t forget. I think this would be easy for a computer to
identify. If you are a real math geek,
and decided that memorizing the digits of pi was too blasé, then you probably
memorized this sometime in your past and you picked it up faster than even the
computer could! As for me, my memory
never allowed me to memorize pi or e. I
would never have figured it out without google (except, of course, that I made
up the test so I already knew).
This leaves
A, C, and F. Google wasn’t much help for
any of these, except that C happens to start with digits that must look like a
FedEx number. Are FedEx numbers random? They must not be.
I told you
that two of the series are random, so you know that two of A, C, and F are
random and the remaining one must be the “uniquely human” series. Oddly, I will tell you that one of the random
number sequences was taken directly from a random number table published on
some website, so I was expecting that google would find it, but when I searched
for those 20 digits, it did not come up.
There are probably more sophisticated searches that would find it. For the other random number series, I just used
a random number generator and came up with one number at a time and wrote it
down, so that particular series would not necessarily be anywhere on the
internet. I couldn’t find it. But you kind of wonder – at some point in the
future, will every random series of 20 digits be on the internet and identified
as being random? Maybe – it’s feasible.
You might be
tempted to say that since C might be a FedEx package number, that would qualify
as the “uniquely human” number. Well,
FedEx has nothing to do with it being right, but you would have a lucky guess. Being lucky might also be uniquely
human! But, for the record, A and F are
random number sequences, and unless I accidentally hit the
one-in-whatever-astronomical-odds lottery, it has no pattern that could be
identified by either man or computer. They
are both random.
So, that
leaves C: 98832863158718824883
C has a human
pattern that requires creative thinking to understand. Actually, for some of you, if I just tell you
to “think outside the box”, you might get it rather quickly. Can a computer think creatively? If, by “creative thinking”, we mean the idea
of coming up with a thought that has never been thought before, or a linkage
that has never been made before, then I suggest that computers can’t do
that. I think computers could be programmed
to “think outside the box”, as long as some intelligent person programs them to
do so, but in that case computers would just be thinking in some bigger
box. I think computers cannot think
truly creatively, and I also think that humans can. Many of you would disagree with me on both
counts!
OK, I guess I
have to tell you what C is. I hate to do
it, because doing so will completely void this version of the “Turing Number
Test”. That’s the problem – once the
answer is known, then it is no longer uniquely human because of two
things: 1) it’s now searchable on
google; and 2) it’s now known by any number of programmers, who can easily
modify any computer to “think” this way and recognize what C is.
C is a
sentence. It is the sentence “Neat test
of a sweet feat” with no punctuation. It
just uses the first letter of each digit when it is spelled out. So, the digit “9”, spelled “nine”, is
equivalent to the letter “n”. I then
took the further step of saying that if the pronunciation of a letter made the
sound of a different letter, then that digit could also stand for the
pronounced letter. Thus, the digit “8”
can be “e” (for “eight”) or it can be “a” because it is pronounced starting
with the “a” sound (in English). The
digit “1” can therefore be “o” or “w”.
Anyway, if you work all that out, picking the letter associated with
each digit as it makes sense, you get “NEATTESTOFASWEETFEAT”. Actually, the last word is ambiguous, because
it could also be “feet”. But, in the
context of the sentence and the test, would that makes sense??? Not to a human, I wouldn’t think.
So, there you
have it. I don’t think a computer, told
to evaluate a number sequence for a pattern (or lack of pattern in the case of
a random number) would be able to identify series C as anything but random,
whereas a human being can. Would all
humans figure this one out? No. And that doesn’t make them non-human. My proposition is this: if the entity you are communicating with can
identify sequence C as a sentence, then the entity you are dealing with is
human. I would probably further propose
that if the entity you are communicating with can’t identify sequence B as a pattern, then the entity you are
communicating with is human. You can’t
use this test to confirm a computer, but you can’t use it to deny a computer. I believe any human could pretend to be a
computer, so just stating that C is a random number sequence doesn’t mean you
are a computer. But the point of the
Turing Test concept is for a computer to fool a human into thinking it is
human. If a computer figured it out,
then that would destroy my proposition, or at least destroy my version of this
test.
This type of
thinking is largely what the MENSA testing evaluates. Creativity.
How is it possible for computers to be truly creative? I know that a computer could generate art or
music and could, in a sense, generate a piece of art that would be unique in
the sense that there would be no existing artwork exactly like it. The same could be said of a two-year old
human. But a computer will generate art
that is within a framework of rules and concepts, all introduced by a human. Can a computer think up a new concept that has
never been thought of before? Can a
computer, given knowledge of number sequences, decide to consider the starting
letter and sound of the written digits to generate a puzzle? Well, yes, of course if the human who
programmed it gives it that ability. Or
gives it the ability to learn and find examples where humans have done just
that sort of thing. But I don’t consider
that the same as original creativity.
You might argue that the first human
who thought of the idea of using the first letters of the words of the digits to
create a puzzle just based it on some of their past experience, and thus would not be that different from a computer, who would have learned it from the past experience of other humans. But doesn’t that line of thinking assume then
that the idea has always been around since the beginning of time? Where?
Encoded in what or who? I guess you
would argue that it is an extension of previous things – so the idea itself has
not always existed, but it was an extension of previous concepts (such as “thinking
out of the box”) that previously existed – those things themselves being
extensions of previous thoughts – back to some simple and obvious basic thought
that started it all. To me, that doesn’t
solve the problem, it just makes it seem simpler – but if you start with no
ideas, then the first idea, no matter how simple, is a huge step change in
thinking. To come up with the first “idea”
is true creativity!
All I can tell you is that
somewhere back in history, a human being was the first to think “I could
scratch a picture into a rock face, and thus record my experiences for the
future.” That is the ultimate in
creativity, in my opinion. The idea of
writing, having never had any concept of writing before, is just mind-boggling. And, to give you a taste of where I intend to
go with this line of thinking: I
consider human creativity to be an “uncaused cause”, which makes them a true act
of original creation! But that is a
discussion for the future.
No comments:
Post a Comment