0:00:02.058,0:00:04.882
Hello. Well, let me start[br]by asking you a question:

0:00:05.537,0:00:08.905
How many of you had to fill out[br]some sort of web form

0:00:08.929,0:00:13.318
where you've been asked to read[br]a distorted sequence of characters?

0:00:16.163,0:00:18.097
How many of you found it really annoying?

0:00:18.611,0:00:20.761
Okay, outstanding. So I invented that.

0:00:20.761,0:00:22.373
(Laughter)

0:00:22.397,0:00:25.848
(Applause)

0:00:29.954,0:00:31.789
That thing is called a CAPTCHA.

0:00:31.789,0:00:35.837
And it is there to make sure[br]the entity filling out the form

0:00:35.861,0:00:38.713
is actually a human and not[br]some sort of computer program

0:00:38.737,0:00:43.084
that was written to submit the form[br]millions and millions of times.

0:00:43.108,0:00:45.014
The reason it works is because humans

0:00:45.038,0:00:47.443
have no trouble reading[br]these squiggly characters,

0:00:47.467,0:00:50.308
whereas computer programs[br]simply can't do it as well yet.

0:00:50.332,0:00:54.276
For example, when you're buying tickets[br]online for attending a concert

0:00:54.276,0:00:57.494
the reason you have to type

0:00:57.518,0:01:02.183
these distorted characters is to prevent

0:01:02.207,0:01:04.155
scalpers from writing a program

0:01:04.179,0:01:07.004
that can buy millions of tickets,[br]two at a time.

0:01:07.028,0:01:08.956
CAPTCHAs are used all over the Internet.

0:01:08.980,0:01:10.511
And since they're used so often,

0:01:10.535,0:01:15.822
a lot of times the precise sequence[br]of random characters shown to the user

0:01:15.846,0:01:17.422
is not so fortunate.

0:01:17.446,0:01:20.923
So this is an example from Yahoo.

0:01:20.947,0:01:23.314
The random characters that happened

0:01:23.338,0:01:27.306
to be shown to the user were W, A, I, T

0:01:27.330,0:01:29.621
which spells a word.

0:01:29.645,0:01:32.564
But the best part is the message[br]that the Yahoo help desk

0:01:32.588,0:01:35.634
got about 20 minutes later.

0:01:35.658,0:01:38.358
["Help! I've been waiting[br]for over 20 minutes,

0:01:38.382,0:01:41.800
and nothing happens."][br](Laughter)

0:01:42.390,0:01:45.734
This of course, is not as bad[br]as this poor person.

0:01:45.758,0:01:49.017
[REBOOT][br](Laughter)

0:01:49.041,0:01:51.549
I can tell funny stories[br]about captchas for hours

0:01:51.573,0:01:53.409
but since I cannot do that

0:01:53.433,0:01:56.218
let me tell you about a project[br]that we did afterwards

0:01:56.242,0:01:58.463
which is sort of the next[br]evolution of CAPTCHA.

0:01:58.488,0:01:59.845
We call it reCAPTCHA,

0:01:59.869,0:02:03.251
which is something that[br]we started at the University,

0:02:03.275,0:02:05.431
and then we turned it[br]into a startup company.

0:02:05.455,0:02:07.374
And then Google acquired this company.

0:02:07.399,0:02:09.917
so, all what I'm going to say[br]for the next 5 minutes

0:02:09.941,0:02:11.505
is owned by Google.

0:02:11.529,0:02:14.751
So, please, do not spread the word.

0:02:14.775,0:02:18.633
So let me tell you[br]how this project started.

0:02:18.657,0:02:23.127
It turns out that about 200 million[br]CAPTCHAs are typed everyday.

0:02:23.151,0:02:26.239
When I first heard this,[br]I was quite proud of myself.

0:02:26.263,0:02:29.266
I thought, "look at the impact[br]that my research has had."

0:02:29.290,0:02:31.205
But then I started feeling bad.

0:02:31.229,0:02:33.108
They are not only obnoxious, but also

0:02:33.133,0:02:36.577
each time you type a CAPTCHA

0:02:36.601,0:02:38.933
essentially you waste[br]10 seconds of your time.

0:02:38.957,0:02:42.684
And if you multiply that[br]by 200 million you get that

0:02:42.708,0:02:46.722
humanity as a whole is wasting[br]about 500,000 hours every day

0:02:46.746,0:02:48.949
typing these annoying CAPTCHAs.

0:02:48.973,0:02:51.417
So then I started feeling bad.

0:02:51.441,0:02:54.469
And then I started thinking,[br]is there any way

0:02:54.493,0:02:58.341
we can use this effort for something[br]that is good for humanity?

0:02:58.365,0:03:03.695
While you're typing a CAPTCHA,[br]during those 10 seconds,

0:03:03.719,0:03:05.844
your brain is doing something amazing.

0:03:05.868,0:03:09.576
Your brain is doing something[br]that computers cannot yet do.

0:03:09.600,0:03:11.432
So can we get you to do some

0:03:11.456,0:03:13.166
useful work to mankind?

0:03:13.190,0:03:14.371
Putting it differently,

0:03:14.396,0:03:16.977
is there some humongous problem[br]that we cannot yet get

0:03:17.001,0:03:18.523
computers to solve,

0:03:18.547,0:03:20.675
yet we can split into tiny chunks

0:03:20.699,0:03:22.856
such that each time[br]somebody solves a CAPTCHA

0:03:22.881,0:03:24.934
they solve a little bit of this problem?

0:03:24.957,0:03:27.868
The answer to that is "yes,"[br]and this is what we're doing now.

0:03:27.893,0:03:32.591
So what you may not know is that[br]nowadays while you're typing a CAPTCHA,

0:03:32.615,0:03:35.772
not only are you authenticating[br]yourself as a human,

0:03:35.796,0:03:38.676
but in addition you're actually[br]helping us to digitize books.

0:03:38.827,0:03:40.616
So let me explain how this works.

0:03:40.640,0:03:44.297
So there's a lot of projects out there[br]trying to digitize the existing books.

0:03:44.322,0:03:45.749
Google is digitizing books.

0:03:45.773,0:03:48.418
Amazon, with the Kindle,[br]is digitizing books.

0:03:48.443,0:03:51.116
The way this works[br]is you start with an old book.

0:03:51.141,0:03:53.142
You've seen those things, right?

0:03:53.166,0:03:54.204
Like a book?

0:03:54.229,0:03:55.557
(Laughter)

0:03:55.582,0:03:58.063
So you start with a book,[br]and then you scan it.

0:03:58.087,0:04:03.549
Now scanning a book is like taking[br]a digital photograph of every page.

0:04:07.567,0:04:09.883
The next step in the process[br]is that the computer

0:04:09.908,0:04:15.128
needs to be able to decipher[br]all of the words in this image.

0:04:15.152,0:04:18.815
Now the problem is that for older books[br]that were written several years ago

0:04:18.839,0:04:21.291
the computer cannot recognize[br]a lot of the words

0:04:21.315,0:04:25.119
because the ink has faded[br]and the pages have turned yellow.

0:04:25.143,0:04:26.953
Thus the words look a bit different

0:04:26.977,0:04:29.803
and the computer cannot recognize them.

0:04:29.827,0:04:32.468
So, for books that were written[br]more than 50 years ago,

0:04:32.493,0:04:35.895
the computer cannot recognize[br]about 30 percent of the words.

0:04:35.919,0:04:37.084
So what we're doing now

0:04:37.109,0:04:40.400
is we're taking all of the words[br]that the computer cannot recognize

0:04:40.425,0:04:43.824
and we're getting people to read them[br]for us while they're typing

0:04:43.848,0:04:45.311
a CAPTCHA on the Internet.

0:04:45.335,0:04:47.727
So, the next time you type a CAPTCHA -

0:04:47.751,0:04:54.322
(Applause)

0:04:54.346,0:04:57.535
these words that you're typing

0:04:57.559,0:05:00.927
are actually words that are coming[br]from books that are being digitized

0:05:00.951,0:05:03.145
that the computer could not recognize.

0:05:03.169,0:05:06.958
And now the reason we have[br]two words nowadays instead of one

0:05:06.982,0:05:12.167
is because we need to verify[br]if the answer is correct.

0:05:13.588,0:05:17.682
Because one of the words is such[br]that the system knows what it was,

0:05:17.706,0:05:20.728
and the other is a word that[br]the system just got out of a book,

0:05:20.753,0:05:23.538
it didn't know what it was,[br]and it's presented to you.

0:05:24.014,0:05:27.069
We're going to ask you to type both words.

0:05:27.093,0:05:29.072
And we won't tell you which one's which.

0:05:29.096,0:05:30.682
And if you type the correct word

0:05:30.707,0:05:33.483
for the one for which the system[br]already knows the answer,

0:05:33.508,0:05:35.069
it assumes you are human,

0:05:35.093,0:05:39.600
and it also gets some confidence[br]that you typed the other word correctly.

0:05:39.624,0:05:43.264
And if we repeat this process[br]to like 10 different people

0:05:43.288,0:05:45.582
and all of them agree[br]on what the new word is,

0:05:45.606,0:05:47.884
we are very confident[br]that this new word

0:05:47.908,0:05:49.850
was accurately digitized.

0:05:49.874,0:05:51.920
So this is how the system works.

0:05:51.944,0:05:54.396
The good thing is that[br]it has been very successful.

0:05:54.428,0:05:57.825
We're digitizing about[br]100 million words a day,

0:05:57.849,0:06:01.513
which is the equivalent of[br]about two million books a year.

0:06:01.537,0:06:04.436
And this is all being done[br]one word at a time

0:06:04.460,0:06:07.016
by just people typing CAPTCHAs[br]on the Internet.

0:06:07.040,0:06:10.672
Now, since we're doing[br]so many words per day,

0:06:10.696,0:06:14.724
funny things can happen.

0:06:14.748,0:06:18.286
And this is especially true[br]because now we're giving people

0:06:18.310,0:06:21.676
two randomly chosen English words[br]next to each other.

0:06:21.700,0:06:23.976
So funny things can happen.

0:06:24.000,0:06:25.976
For example, we presented this word.

0:06:26.000,0:06:30.284
It's the word "Christians";[br]there's nothing wrong with it.

0:06:30.308,0:06:33.976
But if you present it along with[br]another randomly chosen word,

0:06:34.000,0:06:35.276
bad things can happen.

0:06:35.300,0:06:38.128
So we get this.[br][bad Christians]

0:06:38.152,0:06:39.317
(Laughter)

0:06:39.341,0:06:40.860
It's quite funny.

0:06:40.884,0:06:43.316
But it's even worse,[br]because the particular website

0:06:43.340,0:06:45.251
where we showed this

0:06:45.275,0:06:50.080
actually happened to be called[br]The Embassy of the Kingdom of God.

0:06:50.104,0:06:51.841
(Laughter)

0:06:51.865,0:06:52.959
Oops!

0:06:52.984,0:06:54.882
Here's another really bad one.

0:06:54.906,0:06:59.417
American politician, JohnEdwards.com[br][Damn liberal]

0:06:59.441,0:07:03.691
(Laughter)

0:07:03.715,0:07:06.836
So we keep on insulting people everyday.

0:07:06.860,0:07:08.971
Now, we're not just insulting people.

0:07:08.995,0:07:11.851
Quite often,[br]interesting things can happen.

0:07:11.875,0:07:15.009
So this actually has given rise[br]to an Internet meme

0:07:15.033,0:07:17.367
that millions of people[br]have participated in,

0:07:17.391,0:07:20.173
which is called CAPTCHA art.

0:07:20.197,0:07:21.759
Here's how it works.

0:07:21.783,0:07:25.505
Imagine you're using the[br]Internet and you see a CAPTCHA

0:07:25.529,0:07:27.917
that you think is somewhat peculiar,

0:07:27.941,0:07:30.204
like this CAPTCHA.

0:07:30.228,0:07:33.484
Then what you're supposed to do[br]is you take a screen shot of it.

0:07:33.508,0:07:35.395
Then of course,[br]you fill out the CAPTCHA

0:07:35.420,0:07:38.119
because you help us[br]digitize a book, please.

0:07:38.143,0:07:40.160
But then, first you take a screen shot,

0:07:40.184,0:07:44.389
and then you draw something[br]that is related to it, like this.

0:07:44.413,0:07:46.082
[invisible toaster]

0:07:46.106,0:07:47.136
(Laughter)

0:07:47.161,0:07:50.639
It's just an example of CAPTCHA art.

0:07:50.662,0:07:54.359
There are tens of thousands of these.[br]Some of them are interesting.

0:07:54.383,0:07:57.672
Some of them are very cute.[br][clenched it!]

0:07:57.696,0:08:01.190
Some of them are funnier.

0:08:01.214,0:08:07.551
[stoned founders][br](Laughter)

0:08:07.575,0:08:11.761
This is my favorite number[br]of this whole project: 900 millions.

0:08:11.785,0:08:14.921
This is the number of distinct people

0:08:14.945,0:08:17.307
that have helped us digitize[br]at least one word

0:08:17.331,0:08:19.067
out of a book through reCAPTCHA.

0:08:19.091,0:08:21.186
A little over 10%[br]of the world's population,

0:08:21.211,0:08:22.946
has helped digitize human knowledge.

0:08:22.970,0:08:25.992
And it is numbers like these[br]that motivate my research agenda.

0:08:26.016,0:08:29.553
So the question that motivates[br]my research is the following:

0:08:29.577,0:08:32.626
If you look at humanity's[br]large-scale achievements,

0:08:32.652,0:08:35.371
these really big things that humanity[br]has gotten together

0:08:35.397,0:08:40.099
like building the pyramids of Egypt[br]or the Panama Canal

0:08:40.123,0:08:42.357
or putting a man on the Moon --

0:08:42.381,0:08:44.813
there is a curious fact about them,

0:08:44.837,0:08:48.268
and it is that they were all done[br]with about the same number of people.

0:08:48.292,0:08:50.836
They were all done with[br]about 100,000 people.

0:08:50.860,0:08:53.231
We can ask ourselves[br]why is that all of them used

0:08:53.256,0:08:54.842
about the same number of people.

0:08:54.867,0:08:57.368
The reason for that is because,[br]before the Internet,

0:08:57.393,0:09:00.797
coordinating more than[br]100,000 people was impossible.

0:09:00.821,0:09:03.535
But now with the Internet,[br]I've just shown you a project

0:09:03.559,0:09:06.433
where we've coordinated[br]900 million people.

0:09:06.457,0:09:08.734
So the question that[br]motivates my research is,

0:09:08.758,0:09:11.238
if we can put a man on the Moon[br]with 100,000 people,

0:09:11.262,0:09:14.219
what can we do with[br]100 million people?

0:09:14.243,0:09:17.371
Based on this question,[br]we've been working on a lot of projects.

0:09:17.395,0:09:19.397
I will not tell you[br]about all we have done.

0:09:19.421,0:09:22.446
But, let me tell you about[br]one that we are working on now.

0:09:22.470,0:09:25.479
We've been working on this[br]for about two years now.

0:09:25.503,0:09:29.737
And we're going to launch it[br]in about 30 days.

0:09:29.761,0:09:32.511
It's called Duolingo.

0:09:32.535,0:09:35.484
This project started asking[br]the following question:

0:09:35.508,0:09:39.798
How can we get 100 million people

0:09:39.822,0:09:45.173
translating the Web into[br]every major language for free?

0:09:45.197,0:09:47.736
So there's a lot of things[br]to say about this question.

0:09:47.760,0:09:49.441
First of all, translating the Web.

0:09:49.466,0:09:51.960
Right now it is partitioned[br]into multiple languages.

0:09:51.984,0:09:54.065
A large fraction of it is in English.

0:09:54.089,0:09:56.524
If you don't know any English,[br]you can't access it.

0:09:56.548,0:09:58.618
But large fractions are[br]in other languages,

0:09:58.643,0:10:01.443
and if you don't know the languages[br]you can't access them.

0:10:01.468,0:10:05.026
I would like to translate all[br]of the Web into every major language.

0:10:05.310,0:10:08.209
Now some of you may say,

0:10:08.233,0:10:10.686
why can't we use computers to translate?

0:10:10.710,0:10:14.677
Machine translation nowadays is starting[br]to translate some sentences.

0:10:14.701,0:10:17.079
Well the problem with that is that

0:10:17.103,0:10:18.797
it's not yet good enough,

0:10:18.821,0:10:23.275
and it probably won't be[br]for the next 20 to 30 years.

0:10:23.299,0:10:26.531
So let me show you an example of something

0:10:26.555,0:10:28.240
that was translated by a machine.

0:10:28.264,0:10:32.834
Actually it was a forum[br]about programming questions.

0:10:37.614,0:10:40.208
It was a programming question[br]translated from Japanese

0:10:40.233,0:10:44.667
into English and from then into Spanish,[br]though my translation is good.

0:10:44.691,0:10:46.941
The other one is bad. You'll see.

0:10:46.965,0:10:50.447
So I'll just let you read.

0:10:50.471,0:10:53.308
This person starts apologizing for[br]the machine translation.

0:10:53.332,0:10:55.954
Indeed, this was done with[br]the best translation program

0:10:55.978,0:10:57.421
from Japanese into English.

0:10:57.445,0:11:03.475
Remember, it's a question[br]about computer programming.

0:11:03.499,0:11:06.076
So here you are the preamble[br]to the question.

0:11:06.100,0:11:11.893
[At often, the goat-time install[br]a error is vomit.] (Laughter)

0:11:11.917,0:11:13.995
Then comes the first part of the question.

0:11:14.020,0:11:19.726
[How many times like the wind, a pole,[br]and the dragon?] (Laughter)

0:11:19.750,0:11:21.801
Then comes my favorite part[br]of the question.

0:11:21.825,0:11:25.939
[This insult to father's stones?][br](Laughter)

0:11:25.963,0:11:28.444
And then comes my favorite[br]part of the whole thing.

0:11:28.468,0:11:32.451
[Please apologize for your stupidity.[br]There are a many thank you.] (Laughter)

0:11:32.475,0:11:34.950
Okay, so computer translation[br]isn't yet good enough.

0:11:34.974,0:11:36.594
We need people to translate.

0:11:36.618,0:11:39.332
So what I want is to get[br]100 million people

0:11:39.356,0:11:42.640
translating the Web into[br]every major language for free.

0:11:42.665,0:11:45.600
I couldn't afford paying[br]100 million people for the job,

0:11:45.625,0:11:47.202
so I want them to do it for free.

0:11:47.227,0:11:48.940
Now if this is what you want to do,

0:11:48.965,0:11:51.485
you pretty quickly realize[br]you're going to run into

0:11:51.509,0:11:55.949
two pretty big obstacles,[br]needing to be hurdled.

0:11:55.973,0:12:00.028
The first one is a lack of bilinguals.

0:12:00.052,0:12:03.159
So I don't even know if there[br]exists 100 million people out there

0:12:03.183,0:12:06.501
using the Web who are bilingual[br]enough to help us translate.

0:12:06.525,0:12:07.570
That's a big problem.

0:12:07.594,0:12:09.708
The other problem is a lack of motivation.

0:12:09.732,0:12:13.720
How are we going to motivate people[br]to actually translate the Web for free?

0:12:13.744,0:12:18.844
After thinking about this for months,

0:12:18.868,0:12:20.864
we realized there's actually a way

0:12:20.888,0:12:23.610
to solve both these problems[br]with the same solution.

0:12:23.634,0:12:26.720
We realized that there's a way[br]to kill two birds with one stone.

0:12:26.745,0:12:30.892
And that is to transform[br]language translation

0:12:30.916,0:12:33.986
into something that millions[br]of people want to do,

0:12:34.010,0:12:37.882
and that also helps with[br]the problem of lack of bilinguals,

0:12:37.906,0:12:40.416
and that is language education.

0:12:40.441,0:12:44.061
It turns out there are millions of people[br]wanting to learn other languages.

0:12:44.085,0:12:49.119
Today there are over 1.2 billion people[br]learning a foreign language.

0:12:49.143,0:12:52.326
And it's not just because[br]they're being forced to do so in school.

0:12:52.350,0:12:54.936
For example, in the United States alone,[br]there are over

0:12:54.961,0:12:58.936
5 million people who have paid[br]over $500 for software

0:12:58.961,0:13:00.459
to learn a new language.

0:13:00.483,0:13:03.364
Many people want to learn a new language.

0:13:03.388,0:13:06.667
So what we've been working on[br]for the last two years

0:13:06.691,0:13:08.824
is a new website called Duolingo,

0:13:08.848,0:13:11.816
where the basic idea is[br]people learn a new language

0:13:11.840,0:13:17.188
for free, while simultaneously[br]translating the Web.

0:13:17.916,0:13:19.977
And so they're learning by doing.

0:13:20.001,0:13:21.634
So this is how it works.

0:13:21.658,0:13:25.528
The way this works is whenever[br]you're a just a beginner,

0:13:25.552,0:13:28.036
we give you very,[br]very simple sentences on the Web.

0:13:28.060,0:13:31.716
And if you don't know a word[br]we'll tell you what each word means

0:13:31.740,0:13:34.159
though you are asked[br]to "translate this sentence".

0:13:34.184,0:13:36.058
And it turns out that it really works.

0:13:36.082,0:13:38.358
Even though people know nothing[br]of the language

0:13:38.383,0:13:41.635
if we explain what each word means,[br]they'll be able to translate it.

0:13:41.660,0:13:43.432
As you see how other people translate

0:13:43.456,0:13:45.842
the same sentence,[br]you start learning the language.

0:13:45.866,0:13:47.857
And as you get more and more advanced,

0:13:47.881,0:13:51.078
we give you more and more[br]complex sentences to translate.

0:13:51.102,0:13:53.431
This is how[br]you are going to help us translate.

0:13:53.455,0:13:55.104
This is how the site works.

0:13:55.128,0:13:57.118
We're mostly done building it,

0:13:57.142,0:13:59.415
and now we're testing it.

0:13:59.439,0:14:01.184
When we started working on this

0:14:01.208,0:14:03.653
I didn't think it could work, really.

0:14:03.677,0:14:06.183
But it turns out that[br]it works, indeed. It's amazing.

0:14:06.207,0:14:08.988
First, people really[br]can learn a language with it.

0:14:09.012,0:14:12.369
In this case we are testing it with people

0:14:12.393,0:14:14.751
knowing English,[br]wanting to learn Spanish,

0:14:14.775,0:14:16.049
and vice versa.

0:14:16.073,0:14:18.309
So people really do learn a language.

0:14:18.333,0:14:19.920
And they learn it about as well

0:14:19.945,0:14:22.077
as the leading language learning software,

0:14:22.102,0:14:24.498
which is very good,[br]but perhaps more surprisingly,

0:14:24.523,0:14:28.924
the translations that we get from people[br]using the site are very good.

0:14:28.948,0:14:30.767
They are as accurate as those

0:14:30.791,0:14:33.291
of professional language translators.

0:14:33.315,0:14:36.407
Now of course,[br]we play a trick here and it is that

0:14:36.431,0:14:40.748
we combine the translations[br]of multiple beginners, several students,

0:14:40.772,0:14:42.490
and choose the best.

0:14:42.514,0:14:44.742
But it turns out[br]that that best translation

0:14:44.767,0:14:47.925
is as good as those of[br]professional language translators.

0:14:47.949,0:14:53.553
Now even though we're combining[br]multiple translations,

0:14:54.857,0:14:57.073
another good thing about Duolingo is that

0:14:57.097,0:15:00.188
the site actually[br]can translate pretty fast.

0:15:00.212,0:15:03.531
So let me show you an estimate[br]of how fast we could translate.

0:15:03.555,0:15:06.579
If we wanted to translate Wikipedia[br]from English into Spanish --

0:15:06.604,0:15:09.319
of course, Wikipedia exists in Spanish[br]but is much smaller

0:15:09.343,0:15:12.098
than its English counterpart,[br]about 20 percent of it --

0:15:12.122,0:15:15.711
If we wanted to translate Wikipedia[br]from English into Spanish using Duolingo

0:15:15.735,0:15:20.922
we could do it in five weeks[br]with 100,000 active users

0:15:20.946,0:15:22.500
learning English with Duolingo.

0:15:22.524,0:15:25.623
And we could do it in about 80 hours[br]with a million active users.

0:15:25.647,0:15:28.333
Since all the projects that[br]my group has worked on so far

0:15:28.357,0:15:29.994
have gotten millions of users,

0:15:30.018,0:15:33.452
we're hopeful that we'll be able[br]to translate the Web for free.

0:15:33.476,0:15:35.693
We haven't yet launched Duolingo,

0:15:35.717,0:15:39.113
(Applause)

0:15:45.159,0:15:46.895
I'd like to leave you with --

0:15:47.871,0:15:49.941
we haven't yet launched Duolingo[br]we plan to do so in 30 days.

0:15:49.966,0:15:52.230
If you visit Duolingo.com, you can sign up

0:15:52.254,0:15:56.547
to be part of our private beta[br]in about 30 days.

0:15:56.572,0:15:57.658
Help us.

0:15:57.683,0:15:58.833
Thank you.

0:15:58.858,0:16:00.373
(Applause)