WEBVTT

00:00:00.687 --> 00:00:01.270
PROFESSOR: Hi.

00:00:01.270 --> 00:00:05.600
Do you know that Google
stores 45 billion web

00:00:05.600 --> 00:00:06.910
pages on the internet?

00:00:06.910 --> 00:00:09.240
And if every single page
was a sheet of paper

00:00:09.240 --> 00:00:11.810
and we stacked
them up real high,

00:00:11.810 --> 00:00:17.970
that would be a tower 610 times
taller than Mount Everest?

00:00:17.970 --> 00:00:18.470
Whoa.

00:00:18.470 --> 00:00:22.290
And if that's not
mind blowing enough,

00:00:22.290 --> 00:00:25.050
Google is able to give you
your search results in a split

00:00:25.050 --> 00:00:26.200
second.

00:00:26.200 --> 00:00:27.660
So how can Google do that?

00:00:27.660 --> 00:00:30.920
I mean, that's like finding
a needle in a haystack.

00:00:30.920 --> 00:00:34.450
And so what I was
thinking of is showing

00:00:34.450 --> 00:00:37.190
the analogy of a
search algorithm,

00:00:37.190 --> 00:00:40.700
like this, through a hotel.

00:00:40.700 --> 00:00:43.640
So imagine I have
a friend, James,

00:00:43.640 --> 00:00:45.435
and I'm looking for
James in the hotel

00:00:45.435 --> 00:00:48.710
and he's hiding in one of
the doors of the hotel rooms.

00:00:48.710 --> 00:00:52.210
And assume that the
hotel rooms are arranged

00:00:52.210 --> 00:00:55.750
in a numerical order and
the people in the hotel

00:00:55.750 --> 00:00:57.360
are arranged in
alphabetical order

00:00:57.360 --> 00:00:59.820
to the numbers behind them.

00:00:59.820 --> 00:01:04.379
And so the easiest way, I
guess, you could think of

00:01:04.379 --> 00:01:07.500
to find James is to just
run through the doors,

00:01:07.500 --> 00:01:10.700
opening every single one,
and say, is James here?

00:01:10.700 --> 00:01:11.460
Is James here?

00:01:11.460 --> 00:01:12.070
Is James here?

00:01:12.070 --> 00:01:13.470
And so on.

00:01:13.470 --> 00:01:20.100
But the fastest way to do it
is to run to the middle door

00:01:20.100 --> 00:01:22.030
and to open the door
and say, is James here?

00:01:22.030 --> 00:01:24.910
And if James isn't there,
you ask him, who are you?

00:01:24.910 --> 00:01:27.710
Well, this other person says,
well, my name is Daniel.

00:01:27.710 --> 00:01:29.670
Then you would
instantly know, oh,

00:01:29.670 --> 00:01:31.580
the people on the
left of this door

00:01:31.580 --> 00:01:36.000
are people whose names
start with A, B, and C.

00:01:36.000 --> 00:01:37.870
And the people on the
right-hand of the door

00:01:37.870 --> 00:01:43.110
will be people from D
onwards, D, E, F, H.

00:01:43.110 --> 00:01:47.610
And then you realize, oh,
so if I'm looking for James,

00:01:47.610 --> 00:01:49.240
James is on the right-hand side.

00:01:49.240 --> 00:01:53.030
And so you will run
to the door right

00:01:53.030 --> 00:01:55.250
in the middle of
the section you just

00:01:55.250 --> 00:01:58.190
cut on the right-hand side,
and then you will ask again,

00:01:58.190 --> 00:01:58.990
is James here?

00:01:58.990 --> 00:02:01.500
If James isn't there, you
would try to figure out,

00:02:01.500 --> 00:02:03.260
is f on the left-hand
side of the grid

00:02:03.260 --> 00:02:05.760
or on the right-side
of the grid, and so on.

00:02:05.760 --> 00:02:08.539
And that way you
get to your results

00:02:08.539 --> 00:02:15.030
real fast instead of just
trying it one at a time.

00:02:15.030 --> 00:02:17.340
Now, I guess what
I wanted to get out

00:02:17.340 --> 00:02:18.910
of this video to
the people out there

00:02:18.910 --> 00:02:21.890
is to blow their minds
on the simplicity of some

00:02:21.890 --> 00:02:25.010
of these algorithms
and to say, whoa,

00:02:25.010 --> 00:02:29.640
if I can understand that--
and to say that just knowing

00:02:29.640 --> 00:02:31.530
that something is
a little bit better

00:02:31.530 --> 00:02:36.510
actually makes a humongous
difference in the real world.

00:02:36.510 --> 00:02:39.940
And that's what I want
to get out of the video.