09:30
TED2014

Randall Munroe: Comics that ask "what if?"

Filmed:

Web cartoonist Randall Munroe answers simple what-if questions ("what if you hit a baseball moving at the speed of light?") using math, physics, logic and deadpan humor. In this charming talk, a reader’s question about Google's data warehouse leads Munroe down a circuitous path to a hilariously over-detailed answer — in which, shhh, you might actually learn something.

- Cartoonist
Randall Munroe sketches elegant and illuminating explanations of the weird science and math questions that keep geeks awake at night. Full bio

So, I have a feature on my website where every week
00:12
people submit hypothetical questions
00:15
for me to answer,
00:17
and I try to answer them using math, science
00:18
and comics.
00:21
So for example, one person asked,
00:22
what would happen if you tried to hit a baseball
00:25
pitched at 90 percent of the speed of light?
00:26
So I did some calculations.
00:29
Now, normally, when an object flies through the air,
00:32
the air will flow around the object,
00:34
but in this case, the ball would be going so fast
00:36
that the air molecules wouldn't have time
00:37
to move out of the way.
00:39
The ball would smash right into and through them,
00:41
and the collisions with these air molecules
00:44
would knock away the nitrogen,
00:45
carbon and hydrogen from the ball,
00:47
fragmenting it off into tiny particles,
00:49
and also triggering waves of thermonuclear fusion
00:51
in the air around it.
00:53
This would result in a flood of x-rays
00:55
that would spread out in a bubble
00:57
along with exotic particles,
00:59
plasma inside, centered on the pitcher's mound,
01:00
and that would move away from the pitcher's mound
01:03
slightly faster than the ball.
01:07
Now at this point, about 30 nanoseconds in,
01:09
the home plate is far enough away
01:12
that light hasn't had time to reach it,
01:14
which means the batter
01:16
still sees the pitcher about to throw
01:17
and has no idea that anything is wrong.
01:20
(Laughter)
01:22
Now, after 70 nanoseconds,
01:24
the ball will reach home plate,
01:27
or at least the cloud of expanding plasma
01:28
that used to be the ball,
01:31
and it will engulf the bat and the batter
01:33
and the plate and the catcher and the umpire
01:37
and start disintegrating them all
01:40
as it also starts to carry them backward
01:43
through the backstop, which
also starts to disintegrate.
01:46
So if you were watching this whole thing
01:50
from a hill,
01:51
ideally, far away,
01:53
what you'd see is a bright flash of light
01:56
that would fade over a few seconds,
01:57
followed by a blast wave spreading out,
01:59
shredding trees and houses
02:01
as it moves away from the stadium,
02:04
and then eventually a mushroom cloud
02:06
rising up over the ruined city. (Laughter)
02:09
So the Major League Baseball rules
02:12
are a little bit hazy,
02:13
but — (Laughter) — under rule 6.02 and 5.09,
02:16
I think that in this situation,
02:20
the batter would be considered hit by pitch
02:22
and would be eligible to take first base,
02:25
if it still existed.
02:27
So this is the kind of question I answer,
02:30
and I get people writing in with
02:32
a lot of other strange questions.
02:34
I've had someone write and say,
02:36
scientifically speaking, what is the best
02:39
and fastest way to hide a body?
02:41
Can you do this one soon?
02:43
And I had someone write in,
02:45
I've had people write in about,
02:47
can you prove whether or not you can find love again
02:49
after your heart's broken?
02:51
And I've had people send in
02:53
what are clearly homework questions
02:54
they're trying to get me to do for them.
02:56
But one week, a couple months ago,
03:00
I got a question that was actually about Google.
03:03
If all digital data in the world
were stored on punch cards,
03:06
how big would Google's data warehouse be?
03:09
Now, Google's pretty secretive
about their operations,
03:12
so no one really knows how much data Google has,
03:15
and in fact, no one really knows
how many data centers Google has,
03:18
except people at Google itself.
03:20
And I've tried, I've met them a few times,
03:23
tried asking them, and they
aren't revealing anything.
03:24
So I decided to try to figure this out myself.
03:29
There are a few things that I looked at here.
03:32
I started with money.
03:34
Google has to reveal how much they spend,
03:36
in general, and that lets you put some caps
03:38
on how many data centers could they be building,
03:40
because a big data center costs
a certain amount of money.
03:44
And you can also then put a cap on
03:46
how much of the world hard
drive market are they taking up,
03:49
which turns out, it's pretty sizable.
03:51
I read a calculation at one point,
03:53
I think Google has a drive failure
03:55
about every minute or two,
03:56
and they just throw out the hard drive
04:00
and swap in a new one.
04:01
So they go through a huge number of them.
04:02
And so by looking at money,
04:05
you can get an idea of how
many of these centers they have.
04:06
You can also look at power.
04:08
You can look at how much electricity they need,
04:10
because you need a certain amount
of electricity to run the servers,
04:14
and Google is more efficient than most,
04:16
but they still have some basic requirements,
04:18
and that lets you put a limit
04:21
on the number of servers that they have.
04:23
You can also look at square footage and see
04:25
of the data centers that you know,
04:29
how big are they?
04:31
How much room is that?
04:32
How many server racks could you fit in there?
04:33
And for some data centers,
04:35
you might get two of these pieces of information.
04:37
You know how much they spent,
04:39
and they also, say, because they had to contract
04:41
with the local government
04:43
to get the power provided,
04:44
you might know what they made a deal to buy,
04:46
so you know how much power it takes.
04:49
Then you can look at the ratios of those numbers,
04:50
and figure out for a data center
04:53
where you don't have that information,
04:54
you can figure out,
04:56
but maybe you only have one of those,
04:57
you know the square footage,
then you could figure out
04:59
well, maybe the power is proportional.
05:01
And you can do this same thing
with a lot of different quantities,
05:03
you know, with guesses about
the total amount of storage,
05:05
the number of servers, the
number of drives per server,
05:08
and in each case using what you know
05:10
to come up with a model that narrows down
05:13
your guesses for the things that you don't know.
05:16
It's sort of circling around the
number you're trying to get.
05:18
And this is a lot of fun.
05:20
The math is not all that advanced,
05:23
and really it's like nothing more than
05:25
solving a sudoku puzzle.
05:28
So what I did, I went through all of this information,
05:30
spent a day or two researching.
05:35
And there are some things I didn't look at.
05:37
You could always look at the Google
05:39
recruitment messages that they post.
05:42
That gives you an idea of where they have people.
05:44
Sometimes, when people visit a data center,
05:46
they'll take a cell-cam photo and post it,
05:47
and they aren't supposed to,
05:49
but you can learn things about
their hardware that way.
05:51
And in fact, you can just
look at pizza delivery drivers.
05:54
Turns out, they know where all
the Google data centers are,
05:56
at least the ones that have people in them.
05:59
But I came up with my estimate,
06:02
which I felt pretty good about,
06:04
that was about 10 exabytes of data
06:06
across all of Google's operations,
06:09
and then another maybe five exabytes or so
06:11
of offline storage in tape drives,
06:15
which it turns out Google is
06:17
about the world's largest consumer of.
06:18
So I came up with this estimate, and this is
06:21
a staggering amount of data.
06:25
It's quite a bit more than any other organization
06:26
in the world has, as far as we know.
06:29
There's a couple of other contenders,
06:30
especially everyone always thinks of the NSA.
06:32
But using some of these same methods,
06:35
we can look at the NSA's data centers,
06:36
and figure out, you know, we
don't know what's going on there,
06:38
but it's pretty clear that their operation
06:40
is not the size of Google's.
06:43
Adding all of this up, I came up with
06:44
the other thing that we can answer, which is,
06:46
how many punch cards would this take?
06:48
And so a punch card can hold
06:50
about 80 characters,
06:53
and you can fit about 2,000 or so cards into a box,
06:55
and you put them in, say,
06:58
my home region of New England,
07:00
it would cover the entire region
07:02
up to a depth of a little less than five kilometers,
07:04
which is about three times deeper
07:08
than the glaciers during the last ice age
07:09
about 20,000 years ago.
07:11
So this is impractical, but I think
07:14
that's about the best answer I could come up with.
07:16
And I posted it on my website. I wrote it up.
07:19
And I didn't expect to get an answer from Google,
07:21
because of course they've been so secretive,
07:25
they didn't answer of my questions,
07:26
and so I just put it up and said,
07:28
well, I guess we'll never know.
07:29
But then a little while later
07:31
I got a message, a couple weeks later, from Google,
07:33
saying, hey, someone here has an envelope for you.
07:35
So I go and get it, open it up,
07:39
and it's punch cards. (Laughter)
07:42
Google-branded punch cards.
07:44
And on these punch cards,
there are a bunch of holes,
07:47
and I said, thank you, thank you,
07:50
okay, so what's on here?
07:52
So I get some software and start reading it,
07:53
and scan them, and it turns out
07:55
it's a puzzle.
07:57
There's a bunch of code,
07:58
and I get some friends to help,
08:00
and we crack the code, and then
inside that is another code,
08:01
and then there are some equations,
08:04
and then we solve those equations,
08:05
and then finally out pops a message from Google
08:06
which is their official answer to my article,
08:10
and it said, "No comment."
08:13
(Laughter) (Applause)
08:15
And I love calculating these kinds of things,
08:26
and it's not that I love doing the math.
08:29
I do a lot of math,
08:31
but I don't really like math for its own sake.
08:32
What I love is that it lets you take
08:35
some things that you know,
08:37
and just by moving symbols
around on a piece of paper,
08:39
find out something that you didn't know
08:43
that's very surprising.
08:45
And I have a lot of stupid questions,
08:47
and I love that math gives the power
08:49
to answer them sometimes.
08:51
And sometimes not.
08:54
This is a question I got from a reader,
08:55
an anonymous reader,
08:58
and the subject line just said, "Urgent,"
08:59
and this was the entire email:
09:01
"If people had wheels and could fly,
09:03
how would we differentiate them from airplanes?"
09:06
Urgent. (Laughter)
09:09
And I think there are some questions
09:11
that math just cannot answer.
09:15
Thank you.
09:17
(Applause)
09:20

▲Back to top

About the Speaker:

Randall Munroe - Cartoonist
Randall Munroe sketches elegant and illuminating explanations of the weird science and math questions that keep geeks awake at night.

Why you should listen

One of a small group of professional web cartoonists, math obsessive and chronic explainer Randall Munroe dazzles the online world (and racks up millions of monthly page views) with the meaninglessly-named (and occasionally heartbreaking) webcomic xkcd.
 
Munroe’s blog What If? specializes in cunning answers to, as the Atlantic put it, "the kinds of of wonderful and fanciful hypotheticals that might arise when the nerdily inclined get together in bars," like “How fast can you hit a speed bump while driving and live?” or “What would happen if a hair dryer with continuous power was turned on and put in an airtight 1x1x1 meter box?” As he told Math Horizons, I really enjoy solving these kinds of things, and it’s a bonus if I realize that I can put boxes around it and make it a comic."

He is also the author of the book Thing Explainer: Complicated Stuff in Simple Words

More profile about the speaker
Randall Munroe | Speaker | TED.com