Massive-scale online collaboration - Luis von Ahn
how<00:00:15.519> many<00:00:15.679> of<00:00:15.759> you<00:00:15.839> have<00:00:15.920> had<00:00:16.000> to<00:00:16.080> fill<00:00:16.320> out
how many of you have had to fill out
how many of you have had to fill out
some<00:00:16.640> sort<00:00:16.800> of<00:00:16.880> web<00:00:17.039> form<00:00:17.199> where<00:00:17.359> you've<00:00:17.440> been
some sort of web form where you've been
some sort of web form where you've been
asked<00:00:17.680> to<00:00:17.760> read<00:00:17.920> a<00:00:18.000> distorted<00:00:18.400> sequence<00:00:18.720> of
asked to read a distorted sequence of
asked to read a distorted sequence of
characters<00:00:19.199> like<00:00:19.359> this<00:00:20.000> yeah<00:00:20.240> how<00:00:20.400> many<00:00:20.560> of
characters like this yeah how many of
characters like this yeah how many of
you<00:00:20.720> found<00:00:20.960> it<00:00:21.039> really<00:00:21.199> really<00:00:21.439> annoying
you found it really really annoying
you found it really really annoying
okay<00:00:23.359> outstanding<00:00:23.920> so<00:00:24.080> i<00:00:24.240> invented<00:00:24.720> that
or<00:00:27.920> i<00:00:28.000> was<00:00:28.160> one<00:00:28.320> of<00:00:28.400> the<00:00:28.480> people<00:00:28.640> who<00:00:28.800> did<00:00:28.960> it
or i was one of the people who did it
or i was one of the people who did it
that<00:00:29.199> thing<00:00:29.359> is<00:00:29.439> called<00:00:29.599> a<00:00:29.760> captcha<00:00:30.320> and<00:00:30.480> the
that thing is called a captcha and the
that thing is called a captcha and the
reason<00:00:30.880> is<00:00:30.960> there<00:00:31.119> is<00:00:31.199> to<00:00:31.359> make<00:00:31.439> sure<00:00:31.599> that<00:00:31.760> you
reason is there is to make sure that you
reason is there is to make sure that you
the<00:00:32.160> entity<00:00:32.399> filling<00:00:32.719> out<00:00:32.800> the<00:00:32.960> form<00:00:33.200> are
the entity filling out the form are
the entity filling out the form are
actually<00:00:33.520> a<00:00:33.600> human<00:00:34.160> and<00:00:34.239> not<00:00:34.399> some<00:00:34.559> sort<00:00:34.719> of
actually a human and not some sort of
actually a human and not some sort of
computer<00:00:35.200> program<00:00:35.520> that<00:00:35.680> was<00:00:35.840> written<00:00:36.079> to
computer program that was written to
computer program that was written to
submit<00:00:36.480> the<00:00:36.640> form<00:00:36.800> millions<00:00:37.120> and<00:00:37.200> millions<00:00:37.520> of
submit the form millions and millions of
submit the form millions and millions of
times<00:00:38.239> the<00:00:38.399> reason<00:00:38.640> it<00:00:38.800> works<00:00:39.120> is<00:00:39.280> because
times the reason it works is because
times the reason it works is because
humans<00:00:40.079> at<00:00:40.160> least<00:00:40.399> non-visually<00:00:40.960> impaired
humans at least non-visually impaired
humans at least non-visually impaired
humans<00:00:41.680> have<00:00:41.840> no<00:00:42.000> trouble<00:00:42.239> reading<00:00:42.480> these
humans have no trouble reading these
humans have no trouble reading these
distorted<00:00:43.040> squiggly<00:00:43.440> characters<00:00:43.840> whereas
distorted squiggly characters whereas
distorted squiggly characters whereas
computer<00:00:44.480> programs<00:00:45.200> simply<00:00:45.440> can't<00:00:45.600> do<00:00:45.760> it<00:00:45.920> as
computer programs simply can't do it as
computer programs simply can't do it as
well<00:00:46.160> yet<00:00:46.719> so<00:00:46.879> for<00:00:47.039> example<00:00:47.360> in<00:00:47.440> the<00:00:47.520> case<00:00:47.680> of
well yet so for example in the case of
well yet so for example in the case of
ticketmaster<00:00:48.399> the<00:00:48.559> reason<00:00:48.800> you<00:00:48.879> have<00:00:49.039> to<00:00:49.120> type
ticketmaster the reason you have to type
ticketmaster the reason you have to type
these<00:00:49.440> distorted<00:00:49.920> characters<00:00:50.320> is<00:00:50.480> to<00:00:50.559> prevent
these distorted characters is to prevent
these distorted characters is to prevent
scalpers<00:00:51.440> from<00:00:51.600> writing<00:00:51.920> a<00:00:52.000> program<00:00:52.320> that<00:00:52.480> can
scalpers from writing a program that can
scalpers from writing a program that can
buy<00:00:52.879> millions<00:00:53.280> of<00:00:53.360> tickets<00:00:53.680> two<00:00:53.840> at<00:00:53.920> a<00:00:54.000> time
buy millions of tickets two at a time
buy millions of tickets two at a time
you<00:00:54.800> know<00:00:54.960> captures<00:00:55.280> are<00:00:55.360> used<00:00:55.520> all<00:00:55.680> over<00:00:55.840> the
you know captures are used all over the
you know captures are used all over the
internet<00:00:56.399> and<00:00:56.480> since<00:00:56.719> they're<00:00:56.879> used<00:00:57.120> so<00:00:57.360> often
internet and since they're used so often
internet and since they're used so often
a lot<00:00:57.840> of<00:00:57.920> times<00:00:58.079> the<00:00:58.160> precise<00:00:58.480> sequence<00:00:58.879> of
a lot of times the precise sequence of
a lot of times the precise sequence of
random<00:00:59.359> characters<00:00:59.760> that<00:00:59.920> are<00:01:00.000> shown<00:01:00.239> to<00:01:00.399> the
random characters that are shown to the
random characters that are shown to the
user<00:01:01.039> is<00:01:01.199> not<00:01:01.359> so<00:01:01.600> fortunate<00:01:02.399> so<00:01:02.559> this<00:01:02.800> is<00:01:02.879> an
user is not so fortunate so this is an
user is not so fortunate so this is an
example<00:01:03.520> from<00:01:03.680> the<00:01:03.920> yahoo<00:01:04.320> registration<00:01:05.040> page
example from the yahoo registration page
example from the yahoo registration page
the<00:01:06.240> random<00:01:06.799> characters<00:01:07.280> that<00:01:07.360> happen<00:01:07.600> to<00:01:07.680> be
the random characters that happen to be
the random characters that happen to be
shown<00:01:08.000> to<00:01:08.080> the<00:01:08.240> user<00:01:08.479> were<00:01:08.640> w-a-i-t<00:01:09.680> which<00:01:09.920> of
shown to the user were w-a-i-t which of
shown to the user were w-a-i-t which of
course<00:01:10.240> spell<00:01:10.479> a<00:01:10.560> word<00:01:11.200> but<00:01:11.360> the<00:01:11.439> best<00:01:11.600> part<00:01:11.840> is
course spell a word but the best part is
course spell a word but the best part is
the
the
the
message<00:01:13.280> that<00:01:13.439> the<00:01:13.520> yahoo<00:01:13.840> helped<00:01:14.080> us<00:01:14.159> got
message that the yahoo helped us got
message that the yahoo helped us got
about<00:01:14.640> 20<00:01:14.880> minutes<00:01:15.200> later
this<00:01:23.920> person<00:01:24.320> thought<00:01:24.479> they<00:01:24.640> needed<00:01:24.880> to<00:01:24.960> wait
this person thought they needed to wait
this person thought they needed to wait
this<00:01:25.280> of<00:01:25.360> course<00:01:25.520> is<00:01:25.600> not<00:01:25.680> as<00:01:25.759> bad<00:01:25.920> as<00:01:26.080> this
this of course is not as bad as this
this of course is not as bad as this
poor<00:01:26.400> person<00:01:26.880> who<00:01:27.119> uh
okay<00:01:30.960> now<00:01:31.360> captcha<00:01:31.680> project<00:01:32.000> is<00:01:32.079> something
okay now captcha project is something
okay now captcha project is something
that<00:01:32.400> we<00:01:32.479> did<00:01:32.560> here<00:01:32.720> at<00:01:32.799> carnegie<00:01:33.119> mellon<00:01:33.360> over
that we did here at carnegie mellon over
that we did here at carnegie mellon over
10<00:01:33.680> years<00:01:33.840> ago<00:01:34.079> and<00:01:34.240> it's<00:01:34.400> been<00:01:34.560> used
10 years ago and it's been used
10 years ago and it's been used
everywhere<00:01:35.439> uh<00:01:35.759> let<00:01:35.920> me<00:01:36.000> not<00:01:36.079> tell<00:01:36.240> you<00:01:36.320> about
everywhere uh let me not tell you about
everywhere uh let me not tell you about
a<00:01:36.560> project<00:01:36.799> that<00:01:36.960> we<00:01:37.040> did<00:01:37.200> a<00:01:37.280> few<00:01:37.439> years<00:01:37.680> later
a project that we did a few years later
a project that we did a few years later
which<00:01:38.159> is<00:01:38.240> sort<00:01:38.400> of<00:01:38.479> the<00:01:38.960> next<00:01:39.200> evolution<00:01:39.680> of
which is sort of the next evolution of
which is sort of the next evolution of
captchas<00:01:40.400> just<00:01:40.560> a<00:01:40.640> project<00:01:40.960> that<00:01:41.040> we<00:01:41.200> call
captchas just a project that we call
captchas just a project that we call
recaptcha<00:01:42.560> which<00:01:42.720> is<00:01:42.880> something<00:01:43.119> that<00:01:43.200> we
recaptcha which is something that we
recaptcha which is something that we
started<00:01:43.600> here<00:01:43.759> at<00:01:43.840> carnegie<00:01:44.159> mellon<00:01:44.640> then<00:01:45.040> we
started here at carnegie mellon then we
started here at carnegie mellon then we
turned<00:01:45.759> it<00:01:45.840> into<00:01:46.079> a<00:01:46.159> startup<00:01:46.479> company<00:01:47.119> and
turned it into a startup company and
turned it into a startup company and
then<00:01:47.360> about<00:01:47.600> a<00:01:47.680> year<00:01:47.840> and<00:01:47.920> a<00:01:48.000> half<00:01:48.159> ago<00:01:48.399> google
then about a year and a half ago google
then about a year and a half ago google
actually<00:01:49.119> acquired<00:01:49.520> this<00:01:49.680> company<00:01:50.399> so<00:01:50.640> let<00:01:50.799> me
actually acquired this company so let me
actually acquired this company so let me
tell<00:01:51.040> you<00:01:51.119> what<00:01:51.280> this<00:01:51.600> project<00:01:52.000> started
tell you what this project started
tell you what this project started
okay<00:01:53.040> so<00:01:53.200> this<00:01:53.360> project<00:01:53.680> started<00:01:53.920> from<00:01:54.079> the
okay so this project started from the
okay so this project started from the
following<00:01:54.479> realization<00:01:55.280> it<00:01:55.439> turns<00:01:55.759> out<00:01:56.079> that
following realization it turns out that
following realization it turns out that
approximately<00:01:56.880> 200<00:01:57.280> million<00:01:57.600> captures<00:01:57.920> are
approximately 200 million captures are
approximately 200 million captures are
typed<00:01:58.320> every<00:01:58.399> day<00:01:58.560> by<00:01:58.719> people<00:01:58.880> around<00:01:59.040> the
typed every day by people around the
typed every day by people around the
world
world
world
okay<00:02:00.640> when<00:02:00.799> i<00:02:00.960> first<00:02:01.200> heard<00:02:01.360> this<00:02:01.600> i<00:02:01.680> was<00:02:01.840> quite
okay when i first heard this i was quite
okay when i first heard this i was quite
proud<00:02:02.159> of<00:02:02.320> myself<00:02:02.640> i<00:02:02.719> thought<00:02:02.880> look<00:02:03.040> at<00:02:03.119> the
proud of myself i thought look at the
proud of myself i thought look at the
impact<00:02:03.520> that<00:02:03.600> my<00:02:03.759> research<00:02:04.079> has<00:02:04.240> had
impact that my research has had
impact that my research has had
but<00:02:05.600> then<00:02:05.840> i<00:02:05.920> started<00:02:06.159> feeling<00:02:06.399> bad<00:02:06.719> so<00:02:06.880> here's
but then i started feeling bad so here's
but then i started feeling bad so here's
the<00:02:07.200> thing<00:02:07.439> each<00:02:07.600> time<00:02:07.759> you<00:02:07.920> type<00:02:08.080> a<00:02:08.160> captcha
the thing each time you type a captcha
the thing each time you type a captcha
essentially<00:02:09.599> you<00:02:09.759> waste<00:02:10.000> 10<00:02:10.239> seconds<00:02:10.560> of<00:02:10.640> your
essentially you waste 10 seconds of your
essentially you waste 10 seconds of your
time
time
time
and<00:02:12.000> if<00:02:12.160> you<00:02:12.239> multiply<00:02:12.560> that<00:02:12.720> by<00:02:12.800> 200<00:02:13.200> million
and if you multiply that by 200 million
and if you multiply that by 200 million
you<00:02:13.599> get<00:02:13.680> the<00:02:13.760> humanity<00:02:14.239> as<00:02:14.319> a<00:02:14.400> whole<00:02:14.640> is
you get the humanity as a whole is
you get the humanity as a whole is
wasting<00:02:15.120> about<00:02:15.280> 500<00:02:15.840> 000<00:02:16.160> hours<00:02:16.480> every<00:02:16.640> day
wasting about 500 000 hours every day
wasting about 500 000 hours every day
typing<00:02:17.120> these<00:02:17.280> annoying<00:02:17.599> captchas
typing these annoying captchas
typing these annoying captchas
so<00:02:19.120> then<00:02:19.360> i<00:02:19.440> started<00:02:19.680> feeling<00:02:20.000> bad
so then i started feeling bad
so then i started feeling bad
and<00:02:22.879> then<00:02:23.120> i<00:02:23.200> started<00:02:23.440> thinking<00:02:23.760> well<00:02:23.920> of
and then i started thinking well of
and then i started thinking well of
course<00:02:24.239> we<00:02:24.319> can't<00:02:24.480> just<00:02:24.640> get<00:02:24.720> rid<00:02:24.879> of<00:02:25.040> captchas
course we can't just get rid of captchas
course we can't just get rid of captchas
because<00:02:25.760> the<00:02:25.920> security<00:02:26.239> of<00:02:26.319> the<00:02:26.400> web<00:02:26.560> sort<00:02:26.720> of
because the security of the web sort of
because the security of the web sort of
depends<00:02:27.120> on<00:02:27.200> them<00:02:27.440> but<00:02:27.599> then<00:02:27.760> i<00:02:27.840> started
depends on them but then i started
depends on them but then i started
thinking<00:02:28.400> is<00:02:28.560> there<00:02:28.720> any<00:02:28.879> way<00:02:29.120> which<00:02:29.360> we<00:02:29.520> can
thinking is there any way which we can
thinking is there any way which we can
use<00:02:29.840> this<00:02:30.160> effort<00:02:30.640> for<00:02:30.720> something<00:02:31.040> that<00:02:31.120> is
use this effort for something that is
use this effort for something that is
good<00:02:31.440> for<00:02:31.520> humanity
good for humanity
good for humanity
so
so
so
see<00:02:34.000> here's<00:02:34.239> the<00:02:34.319> thing<00:02:34.480> while<00:02:34.640> you're<00:02:34.800> typing
see here's the thing while you're typing
see here's the thing while you're typing
a<00:02:35.120> captcha<00:02:35.440> during<00:02:35.680> those<00:02:35.840> 10<00:02:36.080> seconds<00:02:36.640> your
a captcha during those 10 seconds your
a captcha during those 10 seconds your
brain<00:02:37.120> is<00:02:37.280> doing<00:02:37.440> something<00:02:37.760> amazing<00:02:38.239> your
brain is doing something amazing your
brain is doing something amazing your
brain<00:02:38.879> is<00:02:39.040> doing<00:02:39.200> something<00:02:39.519> that<00:02:39.599> computers
brain is doing something that computers
brain is doing something that computers
cannot<00:02:40.400> yet<00:02:40.640> do<00:02:41.200> so<00:02:41.360> can<00:02:41.519> we<00:02:41.680> get<00:02:41.760> you<00:02:41.920> to<00:02:42.000> do
cannot yet do so can we get you to do
cannot yet do so can we get you to do
useful<00:02:43.040> work<00:02:43.280> for<00:02:43.440> those<00:02:43.599> 10<00:02:43.840> seconds<00:02:44.319> another
useful work for those 10 seconds another
useful work for those 10 seconds another
way<00:02:44.720> of<00:02:44.800> putting<00:02:45.040> it<00:02:45.200> is<00:02:45.280> there's<00:02:45.440> some
way of putting it is there's some
way of putting it is there's some
humongous<00:02:46.000> problem<00:02:46.319> that<00:02:46.400> we<00:02:46.480> cannot<00:02:46.800> get
humongous problem that we cannot get
humongous problem that we cannot get
computers<00:02:47.360> to<00:02:47.519> solve<00:02:47.840> but<00:02:48.000> that<00:02:48.160> somehow<00:02:48.480> we
computers to solve but that somehow we
computers to solve but that somehow we
can<00:02:48.800> split<00:02:49.200> into<00:02:49.440> tiny<00:02:49.760> 10<00:02:50.000> second<00:02:50.319> chunks
can split into tiny 10 second chunks
can split into tiny 10 second chunks
such<00:02:51.120> that<00:02:51.280> each<00:02:51.440> time<00:02:51.599> somebody<00:02:52.000> solves<00:02:52.319> a
such that each time somebody solves a
such that each time somebody solves a
capture<00:02:52.720> they<00:02:52.879> solve<00:02:53.120> a<00:02:53.200> little<00:02:53.360> bit<00:02:53.519> of<00:02:53.599> this
capture they solve a little bit of this
capture they solve a little bit of this
problem<00:02:54.480> and<00:02:54.560> the<00:02:54.720> answer<00:02:54.959> to<00:02:55.040> that<00:02:55.200> is<00:02:55.280> yes
problem and the answer to that is yes
problem and the answer to that is yes
and<00:02:55.599> this<00:02:55.760> is<00:02:55.840> what<00:02:56.000> we're<00:02:56.160> doing<00:02:56.400> now<00:02:56.720> so<00:02:56.879> what
and this is what we're doing now so what
and this is what we're doing now so what
you<00:02:57.120> may<00:02:57.280> not<00:02:57.440> know<00:02:57.680> is<00:02:57.760> that<00:02:57.840> nowadays<00:02:58.319> while
you may not know is that nowadays while
you may not know is that nowadays while
you're<00:02:58.560> typing<00:02:58.879> a<00:02:58.959> caption<00:02:59.360> not<00:02:59.519> only<00:02:59.680> are<00:02:59.760> you
you're typing a caption not only are you
you're typing a caption not only are you
authenticating<00:03:00.560> yourself<00:03:00.879> as<00:03:00.959> a<00:03:01.040> human<00:03:01.760> but
authenticating yourself as a human but
authenticating yourself as a human but
in<00:03:02.080> addition<00:03:02.400> you're<00:03:02.560> actually<00:03:02.800> helping<00:03:03.120> us
in addition you're actually helping us
in addition you're actually helping us
to<00:03:03.360> digitize<00:03:03.840> books<00:03:04.640> okay<00:03:04.879> so<00:03:05.040> let<00:03:05.200> me<00:03:05.280> explain
to digitize books okay so let me explain
to digitize books okay so let me explain
how<00:03:05.680> this<00:03:05.840> works<00:03:06.080> so<00:03:06.159> there's<00:03:06.319> a<00:03:06.400> lot<00:03:06.560> of
how this works so there's a lot of
how this works so there's a lot of
projects<00:03:06.879> out<00:03:07.040> there<00:03:07.120> trying<00:03:07.280> to<00:03:07.440> digitize
projects out there trying to digitize
projects out there trying to digitize
books<00:03:08.319> google<00:03:08.640> has<00:03:08.879> one<00:03:09.120> the<00:03:09.200> internet
books google has one the internet
books google has one the internet
archive<00:03:09.920> has<00:03:10.159> one<00:03:10.400> amazon<00:03:10.720> now<00:03:10.879> with<00:03:10.959> the
archive has one amazon now with the
archive has one amazon now with the
kindles<00:03:11.360> trying<00:03:11.599> to<00:03:11.680> digitize<00:03:12.080> books
kindles trying to digitize books
kindles trying to digitize books
basically<00:03:12.879> the<00:03:12.959> way<00:03:13.120> this<00:03:13.200> works<00:03:13.519> is<00:03:13.840> you
basically the way this works is you
basically the way this works is you
start<00:03:14.239> with<00:03:14.400> an<00:03:14.560> old<00:03:15.120> book<00:03:15.360> like<00:03:15.519> a<00:03:15.599> physical
start with an old book like a physical
start with an old book like a physical
thing<00:03:17.040> you've<00:03:17.200> seen<00:03:17.360> those<00:03:17.519> things<00:03:17.680> right
thing you've seen those things right
thing you've seen those things right
like
like
like
okay
okay
okay
so<00:03:21.200> you<00:03:21.360> start<00:03:21.519> with<00:03:21.680> a<00:03:21.760> book<00:03:22.159> and<00:03:22.239> then<00:03:22.560> you
so you start with a book and then you
so you start with a book and then you
scan<00:03:23.040> it<00:03:23.599> now<00:03:23.920> scanning<00:03:24.319> a<00:03:24.400> book<00:03:24.640> is<00:03:24.720> like
scan it now scanning a book is like
scan it now scanning a book is like
taking<00:03:25.200> a<00:03:25.280> digital<00:03:25.680> photograph<00:03:26.159> of<00:03:26.319> every
taking a digital photograph of every
taking a digital photograph of every
page<00:03:26.640> of<00:03:26.720> the<00:03:26.800> book<00:03:27.040> it<00:03:27.120> gives<00:03:27.280> you<00:03:27.360> an<00:03:27.599> image
page of the book it gives you an image
page of the book it gives you an image
for<00:03:28.080> every<00:03:28.239> page<00:03:28.480> of<00:03:28.560> the<00:03:28.640> book<00:03:28.800> this<00:03:28.879> is<00:03:28.959> an
for every page of the book this is an
for every page of the book this is an
image<00:03:29.360> with<00:03:29.599> text<00:03:30.159> for<00:03:30.319> every<00:03:30.480> page<00:03:30.640> of<00:03:30.720> the
image with text for every page of the
image with text for every page of the
book<00:03:31.280> the<00:03:31.360> next<00:03:31.599> step<00:03:31.760> in<00:03:31.920> the<00:03:32.000> process<00:03:32.319> is
book the next step in the process is
book the next step in the process is
that<00:03:32.560> the<00:03:32.640> computer<00:03:33.040> needs<00:03:33.200> to<00:03:33.280> be<00:03:33.360> able<00:03:33.519> to
that the computer needs to be able to
that the computer needs to be able to
decipher<00:03:34.239> all<00:03:34.400> of<00:03:34.480> the<00:03:34.560> words<00:03:34.879> in<00:03:34.959> this<00:03:35.200> image
decipher all of the words in this image
decipher all of the words in this image
that's<00:03:36.000> done<00:03:36.239> using<00:03:36.400> a<00:03:36.480> technology<00:03:36.959> called
that's done using a technology called
that's done using a technology called
ocr<00:03:37.680> for<00:03:37.920> optical<00:03:38.319> character<00:03:38.720> recognition
ocr for optical character recognition
ocr for optical character recognition
which<00:03:39.680> takes<00:03:40.000> a<00:03:40.080> picture<00:03:40.400> of<00:03:40.560> text<00:03:41.280> and<00:03:41.440> tries
which takes a picture of text and tries
which takes a picture of text and tries
to<00:03:41.760> figure<00:03:42.000> out<00:03:42.080> what<00:03:42.239> the<00:03:42.319> text<00:03:42.560> is<00:03:42.720> in<00:03:42.799> there
to figure out what the text is in there
to figure out what the text is in there
now<00:03:43.840> the<00:03:44.080> problem<00:03:44.799> is<00:03:44.879> that<00:03:45.120> ocr<00:03:45.519> is<00:03:45.599> not
now the problem is that ocr is not
now the problem is that ocr is not
perfect<00:03:46.319> especially<00:03:46.799> for<00:03:47.040> older<00:03:47.360> books<00:03:47.680> where
perfect especially for older books where
perfect especially for older books where
the<00:03:48.000> ink<00:03:48.239> has<00:03:48.480> faded<00:03:48.959> and<00:03:49.120> the<00:03:49.200> pages<00:03:49.519> have
the ink has faded and the pages have
the ink has faded and the pages have
turned<00:03:49.920> yellow
turned yellow
turned yellow
ocr<00:03:51.280> cannot<00:03:51.519> recognize<00:03:51.920> a<00:03:52.000> lot<00:03:52.159> of<00:03:52.239> the<00:03:52.319> words
ocr cannot recognize a lot of the words
ocr cannot recognize a lot of the words
for<00:03:52.799> example<00:03:53.040> for<00:03:53.200> things<00:03:53.360> that<00:03:53.440> were<00:03:53.599> in<00:03:53.680> more
for example for things that were in more
for example for things that were in more
than<00:03:54.000> 50<00:03:54.319> years<00:03:54.480> ago<00:03:54.959> the<00:03:55.120> computer<00:03:55.440> cannot
than 50 years ago the computer cannot
than 50 years ago the computer cannot
recognize<00:03:56.159> about<00:03:56.319> 30<00:03:56.799> of<00:03:56.879> the<00:03:56.959> words
recognize about 30 of the words
recognize about 30 of the words
so<00:03:58.000> what<00:03:58.159> we're<00:03:58.319> doing<00:03:58.560> now<00:03:58.720> is<00:03:58.799> we're<00:03:58.959> taking
so what we're doing now is we're taking
so what we're doing now is we're taking
all<00:03:59.439> the<00:03:59.519> words<00:03:59.760> that<00:03:59.840> the<00:03:59.920> computer<00:04:00.239> cannot
all the words that the computer cannot
all the words that the computer cannot
recognize<00:04:01.519> and<00:04:01.680> we're<00:04:01.840> getting<00:04:02.159> people<00:04:02.480> to
recognize and we're getting people to
recognize and we're getting people to
read<00:04:02.879> them<00:04:03.200> for<00:04:03.439> us<00:04:04.000> while<00:04:04.400> they're<00:04:04.640> typing<00:04:04.879> a
read them for us while they're typing a
read them for us while they're typing a
caption<00:04:05.280> on<00:04:05.360> the<00:04:05.439> internet<00:04:06.000> okay<00:04:06.239> so<00:04:06.400> next
caption on the internet okay so next
caption on the internet okay so next
time<00:04:06.720> you<00:04:06.799> type<00:04:06.959> a<00:04:07.040> captcha<00:04:07.360> these<00:04:07.599> words<00:04:07.920> that
time you type a captcha these words that
time you type a captcha these words that
you're<00:04:08.239> typing<00:04:08.799> are<00:04:08.959> actually<00:04:09.360> words<00:04:09.680> that
you're typing are actually words that
you're typing are actually words that
are<00:04:09.920> coming<00:04:10.159> from<00:04:10.319> books<00:04:10.560> that<00:04:10.720> are<00:04:10.799> being
are coming from books that are being
are coming from books that are being
digitized<00:04:11.599> that<00:04:11.760> the<00:04:11.920> computer<00:04:12.159> couldn't
digitized that the computer couldn't
digitized that the computer couldn't
recognize<00:04:13.439> and<00:04:13.519> now<00:04:13.840> the<00:04:14.000> reason<00:04:14.319> we<00:04:14.480> have<00:04:14.640> two
recognize and now the reason we have two
recognize and now the reason we have two
words<00:04:15.200> nowadays<00:04:15.680> instead<00:04:15.920> of<00:04:16.000> one<00:04:16.320> is<00:04:16.479> because
words nowadays instead of one is because
words nowadays instead of one is because
you<00:04:16.799> see<00:04:16.959> one<00:04:17.120> of<00:04:17.280> the<00:04:17.359> words<00:04:17.759> is<00:04:17.840> a<00:04:17.919> word<00:04:18.160> that
you see one of the words is a word that
you see one of the words is a word that
we<00:04:18.639> that<00:04:18.799> the<00:04:18.959> system<00:04:19.280> just<00:04:19.440> got<00:04:19.680> out<00:04:19.759> of<00:04:19.840> a
we that the system just got out of a
we that the system just got out of a
book<00:04:20.320> you<00:04:20.479> didn't<00:04:20.720> know<00:04:20.880> what<00:04:21.040> it<00:04:21.120> was<00:04:21.359> and
book you didn't know what it was and
book you didn't know what it was and
it's<00:04:21.600> going<00:04:21.680> to<00:04:21.759> present<00:04:22.160> it<00:04:22.240> to<00:04:22.400> you<00:04:22.800> but
it's going to present it to you but
it's going to present it to you but
since<00:04:23.199> it<00:04:23.280> doesn't<00:04:23.520> know<00:04:23.680> the<00:04:23.840> answer<00:04:24.160> for<00:04:24.320> it
since it doesn't know the answer for it
since it doesn't know the answer for it
it<00:04:24.560> cannot<00:04:24.880> grade<00:04:25.199> it<00:04:25.280> for<00:04:25.440> you<00:04:25.919> so<00:04:26.080> what<00:04:26.240> we<00:04:26.320> do
it cannot grade it for you so what we do
it cannot grade it for you so what we do
is<00:04:26.560> we<00:04:26.639> give<00:04:26.800> you<00:04:26.880> another<00:04:27.120> word<00:04:27.360> one<00:04:27.520> for
is we give you another word one for
is we give you another word one for
which<00:04:27.919> the<00:04:28.000> system<00:04:28.320> does<00:04:28.479> know<00:04:28.639> the<00:04:28.800> answer
which the system does know the answer
which the system does know the answer
okay<00:04:29.600> we<00:04:29.759> don't<00:04:29.840> tell<00:04:30.000> you<00:04:30.080> which<00:04:30.320> one's<00:04:30.560> which
okay we don't tell you which one's which
okay we don't tell you which one's which
and<00:04:30.880> we<00:04:30.960> say<00:04:31.199> please<00:04:31.360> type<00:04:31.600> both<00:04:32.320> and<00:04:32.479> if<00:04:32.560> you
and we say please type both and if you
and we say please type both and if you
type<00:04:32.960> the<00:04:33.040> correct<00:04:33.360> word<00:04:33.600> for<00:04:33.680> the<00:04:33.759> one<00:04:33.919> for
type the correct word for the one for
type the correct word for the one for
which<00:04:34.240> the<00:04:34.320> system<00:04:34.560> already<00:04:34.880> knows<00:04:35.040> the
which the system already knows the
which the system already knows the
answer<00:04:35.680> it<00:04:35.840> assumes<00:04:36.160> you're<00:04:36.320> human<00:04:36.800> and<00:04:36.960> it
answer it assumes you're human and it
answer it assumes you're human and it
also<00:04:37.280> gets<00:04:37.440> some<00:04:37.680> confidence<00:04:38.080> that<00:04:38.240> you<00:04:38.320> type
also gets some confidence that you type
also gets some confidence that you type
the<00:04:38.639> other<00:04:38.800> word<00:04:38.960> correctly
the other word correctly
the other word correctly
and<00:04:40.560> if<00:04:40.880> we<00:04:41.040> repeat<00:04:41.360> this<00:04:41.520> process<00:04:41.840> to<00:04:41.919> like<00:04:42.080> 10
and if we repeat this process to like 10
and if we repeat this process to like 10
different<00:04:42.560> people<00:04:42.880> and<00:04:43.040> all<00:04:43.199> of<00:04:43.280> them<00:04:43.520> agree
different people and all of them agree
different people and all of them agree
on<00:04:43.919> what<00:04:44.080> the<00:04:44.160> new<00:04:44.320> word<00:04:44.560> is<00:04:44.720> then<00:04:44.880> we<00:04:44.960> get<00:04:45.120> one
on what the new word is then we get one
on what the new word is then we get one
more<00:04:45.520> word<00:04:45.759> digitized<00:04:46.320> accurately<00:04:47.280> so<00:04:47.440> this
more word digitized accurately so this
more word digitized accurately so this
is<00:04:47.680> how<00:04:47.840> the<00:04:48.000> system<00:04:48.320> works<00:04:48.800> and<00:04:49.040> basically
is how the system works and basically
is how the system works and basically
since<00:04:49.759> we<00:04:49.919> released<00:04:50.240> it<00:04:50.400> about<00:04:50.560> three<00:04:50.800> or<00:04:50.880> four
since we released it about three or four
since we released it about three or four
years<00:04:51.280> ago<00:04:51.680> a<00:04:51.759> lot<00:04:52.000> of<00:04:52.080> websites<00:04:52.479> have<00:04:52.560> started
years ago a lot of websites have started
years ago a lot of websites have started
switching<00:04:53.120> from
switching from
switching from
the<00:04:54.160> old<00:04:54.320> capture<00:04:54.639> where<00:04:54.800> people<00:04:54.960> wasted<00:04:55.280> the
the old capture where people wasted the
the old capture where people wasted the
time<00:04:55.600> to<00:04:55.680> the<00:04:55.759> new<00:04:56.000> capture<00:04:56.400> where<00:04:56.560> people<00:04:56.720> are
time to the new capture where people are
time to the new capture where people are
helping<00:04:57.040> to<00:04:57.120> digitize<00:04:57.520> books<00:04:57.919> so<00:04:58.080> for<00:04:58.240> example
helping to digitize books so for example
helping to digitize books so for example
ticketmaster<00:04:59.280> so<00:04:59.440> every<00:04:59.600> time<00:04:59.759> you<00:04:59.840> buy
ticketmaster so every time you buy
ticketmaster so every time you buy
tickets<00:05:00.320> on<00:05:00.400> ticketmaster<00:05:01.040> you<00:05:01.199> have<00:05:01.280> to
tickets on ticketmaster you have to
tickets on ticketmaster you have to
digitize<00:05:01.840> a<00:05:01.919> book<00:05:02.400> uh<00:05:02.639> facebook<00:05:03.040> every<00:05:03.199> time
digitize a book uh facebook every time
digitize a book uh facebook every time
you<00:05:03.520> add<00:05:03.600> a<00:05:03.759> friend<00:05:03.919> or<00:05:04.080> poke<00:05:04.320> somebody<00:05:04.639> you
you add a friend or poke somebody you
you add a friend or poke somebody you
help<00:05:04.880> to<00:05:04.960> digitize<00:05:05.360> a<00:05:05.520> book<00:05:06.000> twitter<00:05:06.720> and
help to digitize a book twitter and
help to digitize a book twitter and
about<00:05:07.199> 350<00:05:08.000> 000<00:05:08.320> other<00:05:08.479> sites<00:05:08.720> are<00:05:08.880> all<00:05:08.960> using
about 350 000 other sites are all using
about 350 000 other sites are all using
recaptcha<00:05:09.919> and<00:05:10.080> in<00:05:10.160> fact<00:05:10.320> the<00:05:10.400> number<00:05:10.639> of
recaptcha and in fact the number of
recaptcha and in fact the number of
sites<00:05:10.960> that<00:05:11.199> are<00:05:11.280> using<00:05:11.520> recapture<00:05:11.919> so<00:05:12.080> high
sites that are using recapture so high
sites that are using recapture so high
that<00:05:12.400> the<00:05:12.560> number<00:05:12.800> of<00:05:13.039> words<00:05:13.280> that<00:05:13.440> we're
that the number of words that we're
that the number of words that we're
digitizing<00:05:14.080> per<00:05:14.240> day<00:05:14.720> is<00:05:14.880> really<00:05:15.120> really
digitizing per day is really really
digitizing per day is really really
large<00:05:15.840> it's<00:05:16.000> about<00:05:16.160> 100<00:05:16.479> million<00:05:16.800> a<00:05:16.880> day<00:05:17.199> which
large it's about 100 million a day which
large it's about 100 million a day which
is<00:05:17.520> the<00:05:17.680> equivalent<00:05:18.400> of<00:05:18.560> about<00:05:19.280> two<00:05:19.520> and<00:05:19.600> a
is the equivalent of about two and a
is the equivalent of about two and a
half<00:05:19.919> million<00:05:20.240> books<00:05:20.479> a<00:05:20.639> year<00:05:21.280> and<00:05:21.440> this<00:05:21.600> is
half million books a year and this is
half million books a year and this is
all<00:05:21.919> being<00:05:22.160> done<00:05:22.400> one<00:05:22.560> word<00:05:22.800> at<00:05:22.880> a<00:05:22.960> time<00:05:23.120> by
all being done one word at a time by
all being done one word at a time by
just<00:05:23.440> people<00:05:23.759> typing<00:05:24.000> captures<00:05:24.320> on<00:05:24.400> the
just people typing captures on the
just people typing captures on the
internet
okay<00:05:33.199> now<00:05:33.600> of<00:05:33.759> course<00:05:34.320> since<00:05:35.120> we're<00:05:35.280> doing<00:05:35.520> so
okay now of course since we're doing so
okay now of course since we're doing so
many<00:05:35.919> words<00:05:36.240> per<00:05:36.400> day
many words per day
many words per day
funny<00:05:37.759> things<00:05:38.000> can<00:05:38.160> happen<00:05:38.479> and<00:05:38.560> this<00:05:38.639> is
funny things can happen and this is
funny things can happen and this is
especially<00:05:39.120> true<00:05:39.280> because<00:05:39.520> see<00:05:39.680> now<00:05:39.840> we're
especially true because see now we're
especially true because see now we're
giving<00:05:40.160> people<00:05:40.400> two<00:05:40.880> randomly<00:05:41.440> chosen
giving people two randomly chosen
giving people two randomly chosen
english<00:05:42.400> words<00:05:42.800> next<00:05:43.039> to<00:05:43.199> each<00:05:43.360> other<00:05:43.759> okay<00:05:44.000> so
english words next to each other okay so
english words next to each other okay so
funny<00:05:44.320> things<00:05:44.479> can<00:05:44.639> happen<00:05:44.800> so<00:05:44.960> for<00:05:45.120> example
funny things can happen so for example
funny things can happen so for example
we<00:05:45.840> presented<00:05:46.400> this<00:05:46.560> word<00:05:46.880> it's<00:05:47.039> the<00:05:47.120> word
we presented this word it's the word
we presented this word it's the word
christians<00:05:48.080> there's<00:05:48.320> nothing<00:05:48.639> wrong<00:05:48.800> with<00:05:48.960> it
christians there's nothing wrong with it
christians there's nothing wrong with it
but<00:05:49.360> if<00:05:49.440> you<00:05:49.520> present<00:05:49.840> it<00:05:49.919> along<00:05:50.160> with<00:05:50.320> another
but if you present it along with another
but if you present it along with another
randomly<00:05:51.199> chosen<00:05:51.600> word<00:05:52.240> bad<00:05:52.479> things<00:05:52.720> can
randomly chosen word bad things can
randomly chosen word bad things can
happen
but<00:05:56.160> it's<00:05:56.320> even<00:05:56.560> worse<00:05:56.800> because<00:05:57.039> the
but it's even worse because the
but it's even worse because the
particular<00:05:57.600> website<00:05:58.000> where<00:05:58.160> we<00:05:58.320> show<00:05:58.479> this
particular website where we show this
particular website where we show this
actually<00:05:59.360> happened<00:05:59.680> to<00:05:59.759> be<00:06:00.000> called<00:06:00.639> the
actually happened to be called the
actually happened to be called the
embassy<00:06:01.280> of<00:06:01.360> the<00:06:01.440> kingdom<00:06:01.759> of<00:06:01.840> god
oops
yep<00:06:09.280> uh<00:06:09.440> here's<00:06:09.680> another<00:06:10.000> really<00:06:10.240> bad<00:06:10.479> one<00:06:10.720> at
yep uh here's another really bad one at
yep uh here's another really bad one at
john<00:06:11.600> edwards.com
okay<00:06:18.319> so<00:06:18.479> we<00:06:18.560> keep<00:06:18.800> on<00:06:18.960> insulting<00:06:19.360> people<00:06:19.600> left
okay so we keep on insulting people left
okay so we keep on insulting people left
and<00:06:19.919> right<00:06:20.080> every<00:06:20.319> day
and right every day
and right every day
now<00:06:21.280> of<00:06:21.360> course<00:06:21.520> we're<00:06:21.680> not<00:06:21.840> just<00:06:22.000> insulting
now of course we're not just insulting
now of course we're not just insulting
people<00:06:22.720> see<00:06:22.880> here's<00:06:23.120> the<00:06:23.199> thing<00:06:23.600> since<00:06:23.759> we're
people see here's the thing since we're
people see here's the thing since we're
presenting<00:06:24.319> two<00:06:24.479> randomly<00:06:24.880> chosen<00:06:25.199> words
presenting two randomly chosen words
presenting two randomly chosen words
just<00:06:26.400> interesting<00:06:26.880> things<00:06:27.120> can<00:06:27.280> happen<00:06:27.520> so
just interesting things can happen so
just interesting things can happen so
this<00:06:27.840> has<00:06:28.000> actually<00:06:28.240> has<00:06:28.479> given<00:06:28.960> rise<00:06:29.600> to<00:06:30.319> a
this has actually has given rise to a
this has actually has given rise to a
really
really
really
big<00:06:31.680> internet<00:06:32.160> meme<00:06:32.560> that<00:06:32.880> tens<00:06:33.120> of<00:06:33.280> thousands
big internet meme that tens of thousands
big internet meme that tens of thousands
of<00:06:33.600> people<00:06:33.759> have<00:06:33.919> participated<00:06:34.639> in<00:06:34.880> which<00:06:35.120> is
of people have participated in which is
of people have participated in which is
called<00:06:35.840> captcha<00:06:36.319> art<00:06:36.639> i'm<00:06:36.800> sure<00:06:36.960> some<00:06:37.120> of<00:06:37.280> you
called captcha art i'm sure some of you
called captcha art i'm sure some of you
have<00:06:37.520> heard<00:06:37.759> about<00:06:38.000> it<00:06:38.319> here's<00:06:38.800> here's<00:06:39.120> how<00:06:39.280> it
have heard about it here's here's how it
have heard about it here's here's how it
works<00:06:39.759> okay<00:06:40.000> imagine<00:06:40.240> you're<00:06:40.400> using<00:06:40.800> it
works okay imagine you're using it
works okay imagine you're using it
you're<00:06:41.039> using<00:06:41.280> the<00:06:41.360> internet<00:06:41.680> and<00:06:41.759> you<00:06:41.840> see<00:06:42.000> a
you're using the internet and you see a
you're using the internet and you see a
captcha<00:06:42.400> that<00:06:42.560> you<00:06:42.639> think<00:06:42.800> is<00:06:43.039> somewhat
captcha that you think is somewhat
captcha that you think is somewhat
peculiar<00:06:44.240> like
peculiar like
peculiar like
this<00:06:45.199> captcha
this captcha
this captcha
then<00:06:46.639> what<00:06:46.800> you're<00:06:46.960> supposed<00:06:47.280> to<00:06:47.360> do<00:06:47.520> is<00:06:47.600> you
then what you're supposed to do is you
then what you're supposed to do is you
take<00:06:47.919> a<00:06:48.000> screenshot<00:06:48.560> of<00:06:48.720> it<00:06:49.120> then
take a screenshot of it then
take a screenshot of it then
of<00:06:50.160> course<00:06:50.319> you<00:06:50.400> fill<00:06:50.560> out<00:06:50.639> the<00:06:50.720> capture
of course you fill out the capture
of course you fill out the capture
because<00:06:51.199> you<00:06:51.280> help<00:06:51.440> us<00:06:51.520> digitize<00:06:51.919> a<00:06:52.000> book<00:06:52.560> but
because you help us digitize a book but
because you help us digitize a book but
then
then
then
first<00:06:53.759> you<00:06:53.840> take<00:06:54.000> a<00:06:54.160> screenshot<00:06:54.639> and<00:06:54.800> then<00:06:55.199> you
first you take a screenshot and then you
first you take a screenshot and then you
draw<00:06:55.600> something<00:06:55.919> that<00:06:56.080> is<00:06:56.240> related<00:06:56.639> to<00:06:56.800> it
that's<00:06:59.280> how<00:06:59.440> it<00:06:59.520> works
that's how it works
that's how it works
there<00:07:01.840> are
there are
there are
tens<00:07:02.960> of<00:07:03.199> thousands<00:07:03.759> of<00:07:03.919> these<00:07:04.319> some<00:07:04.479> of<00:07:04.639> them
tens of thousands of these some of them
tens of thousands of these some of them
are<00:07:05.840> very<00:07:06.080> cute
some<00:07:09.759> of<00:07:09.840> them<00:07:10.000> are<00:07:10.160> funnier
and<00:07:16.960> some<00:07:17.120> of<00:07:17.280> them<00:07:17.759> like<00:07:18.280> paleontological
and some of them like paleontological
and some of them like paleontological
schwizzle
they<00:07:22.160> contain<00:07:22.560> snoop<00:07:22.800> dog
okay<00:07:27.039> so<00:07:27.360> this<00:07:27.599> is<00:07:27.680> my<00:07:27.759> favorite<00:07:28.080> number<00:07:28.319> of
okay so this is my favorite number of
okay so this is my favorite number of
recaptcha<00:07:29.039> so<00:07:29.120> this<00:07:29.280> is<00:07:29.520> the<00:07:29.599> favorite<00:07:29.919> thing
recaptcha so this is the favorite thing
recaptcha so this is the favorite thing
that<00:07:30.240> i<00:07:30.400> like<00:07:30.560> about<00:07:30.720> this<00:07:30.880> whole<00:07:31.039> project
that i like about this whole project
that i like about this whole project
this<00:07:31.840> is<00:07:32.000> the<00:07:32.080> number<00:07:32.400> of<00:07:32.560> distinct<00:07:33.039> people
this is the number of distinct people
this is the number of distinct people
that<00:07:33.440> have<00:07:33.599> helped<00:07:33.840> us<00:07:33.919> digitize<00:07:34.479> at<00:07:34.560> least
that have helped us digitize at least
that have helped us digitize at least
one<00:07:35.039> word<00:07:35.440> out<00:07:35.520> of<00:07:35.680> a<00:07:35.759> book<00:07:36.240> through<00:07:36.400> recaptcha
one word out of a book through recaptcha
one word out of a book through recaptcha
750<00:07:37.919> million<00:07:38.240> which<00:07:38.400> is<00:07:38.560> a<00:07:38.639> little<00:07:38.880> over<00:07:39.039> 10<00:07:39.440> of
750 million which is a little over 10 of
750 million which is a little over 10 of
the<00:07:39.599> world's<00:07:39.840> population<00:07:40.400> has<00:07:40.560> helped<00:07:40.880> us
the world's population has helped us
the world's population has helped us
digitize<00:07:41.599> human<00:07:41.919> knowledge<00:07:42.400> and<00:07:42.560> it<00:07:42.720> is<00:07:42.880> words
digitize human knowledge and it is words
digitize human knowledge and it is words
it<00:07:43.599> is<00:07:43.759> numbers<00:07:44.080> like<00:07:44.240> these<00:07:44.479> that<00:07:44.639> motivate
it is numbers like these that motivate
it is numbers like these that motivate
my<00:07:45.360> research<00:07:45.680> agenda<00:07:46.080> so<00:07:46.479> the<00:07:46.720> question<00:07:46.960> that
my research agenda so the question that
my research agenda so the question that
motivates<00:07:47.520> my<00:07:47.680> research<00:07:48.000> is<00:07:48.160> the<00:07:48.240> following
motivates my research is the following
motivates my research is the following
if<00:07:49.120> you<00:07:49.199> look<00:07:49.360> at<00:07:49.440> humanity's<00:07:50.240> large-scale
if you look at humanity's large-scale
if you look at humanity's large-scale
achievements<00:07:51.440> these<00:07:51.680> really<00:07:52.000> big<00:07:52.319> things
achievements these really big things
achievements these really big things
that<00:07:52.720> humanity<00:07:53.199> has<00:07:53.360> gotten<00:07:53.599> together<00:07:53.919> and
that humanity has gotten together and
that humanity has gotten together and
done<00:07:54.319> like<00:07:54.479> historically<00:07:55.360> like<00:07:55.520> for<00:07:55.680> example
done like historically like for example
done like historically like for example
uh<00:07:56.639> building<00:07:56.879> the<00:07:56.960> pyramids<00:07:57.360> of<00:07:57.520> egypt<00:07:58.080> or<00:07:58.479> the
uh building the pyramids of egypt or the
uh building the pyramids of egypt or the
panama<00:07:59.039> canal<00:07:59.520> or<00:07:59.840> putting<00:08:00.080> a<00:08:00.160> man<00:08:00.319> on<00:08:00.400> the
panama canal or putting a man on the
panama canal or putting a man on the
moon
moon
moon
there's<00:08:01.840> a<00:08:01.919> curious<00:08:02.240> fact<00:08:02.479> about<00:08:02.720> them<00:08:03.039> and<00:08:03.199> it
there's a curious fact about them and it
there's a curious fact about them and it
is<00:08:03.440> that<00:08:03.680> they<00:08:03.840> were<00:08:04.000> all<00:08:04.240> done<00:08:04.400> with<00:08:04.560> about
is that they were all done with about
is that they were all done with about
the<00:08:04.879> same<00:08:05.199> number<00:08:05.440> of<00:08:05.599> people<00:08:06.160> it's<00:08:06.319> weird
the same number of people it's weird
the same number of people it's weird
they're<00:08:07.039> all<00:08:07.120> done<00:08:07.360> with<00:08:07.520> about<00:08:07.759> a<00:08:07.840> hundred
they're all done with about a hundred
they're all done with about a hundred
thousand<00:08:08.400> people
thousand people
thousand people
and<00:08:09.440> the<00:08:09.520> reason<00:08:09.840> for<00:08:10.000> that<00:08:10.160> is<00:08:10.319> because
and the reason for that is because
and the reason for that is because
before<00:08:10.800> the<00:08:10.960> internet<00:08:11.840> coordinating<00:08:13.039> more
before the internet coordinating more
before the internet coordinating more
than<00:08:13.360> a<00:08:13.440> hundred<00:08:13.680> thousand<00:08:13.919> people<00:08:14.160> let<00:08:14.319> alone
than a hundred thousand people let alone
than a hundred thousand people let alone
paying<00:08:14.960> them<00:08:15.520> what's<00:08:15.759> essentially
paying them what's essentially
paying them what's essentially
impossible<00:08:16.960> but<00:08:17.120> see<00:08:17.280> now<00:08:17.440> with<00:08:17.520> the<00:08:17.680> internet
impossible but see now with the internet
impossible but see now with the internet
i've<00:08:18.080> just<00:08:18.240> shown<00:08:18.400> you<00:08:18.479> a<00:08:18.639> project<00:08:19.039> where
i've just shown you a project where
i've just shown you a project where
we've<00:08:19.360> gotten<00:08:19.759> 750<00:08:20.560> million<00:08:20.879> people<00:08:21.120> to<00:08:21.280> help
we've gotten 750 million people to help
we've gotten 750 million people to help
us<00:08:21.599> digitize<00:08:22.080> human<00:08:22.400> knowledge<00:08:23.039> so<00:08:23.199> the
us digitize human knowledge so the
us digitize human knowledge so the
question<00:08:23.680> that<00:08:23.759> motivates<00:08:24.160> my<00:08:24.319> research<00:08:24.800> is
question that motivates my research is
question that motivates my research is
if<00:08:25.199> we<00:08:25.360> can<00:08:25.440> put<00:08:25.680> a<00:08:25.759> man<00:08:26.000> on<00:08:26.160> the<00:08:26.240> moon<00:08:26.400> with<00:08:26.560> a
if we can put a man on the moon with a
if we can put a man on the moon with a
hundred<00:08:26.879> thousand<00:08:27.520> what<00:08:27.680> can<00:08:27.840> we<00:08:28.000> do<00:08:28.400> with<00:08:28.639> 100
hundred thousand what can we do with 100
hundred thousand what can we do with 100
million
million
million
so<00:08:30.240> based<00:08:30.479> on<00:08:30.560> this<00:08:30.800> question<00:08:31.039> we've<00:08:31.199> had<00:08:31.440> a
so based on this question we've had a
so based on this question we've had a
lot<00:08:31.680> of<00:08:32.000> different<00:08:32.320> projects<00:08:32.719> that<00:08:32.959> we've
lot of different projects that we've
lot of different projects that we've
been<00:08:33.279> working<00:08:33.519> on<00:08:33.680> let<00:08:33.839> me<00:08:33.919> tell<00:08:34.080> you<00:08:34.240> about
been working on let me tell you about
been working on let me tell you about
one<00:08:35.120> that<00:08:35.279> i'm<00:08:35.440> most<00:08:35.680> excited<00:08:36.080> about<00:08:36.320> this<00:08:36.479> is
one that i'm most excited about this is
one that i'm most excited about this is
something<00:08:36.800> that<00:08:36.880> we've<00:08:37.039> been<00:08:37.440> so
something that we've been so
something that we've been so
semi-quietly<00:08:38.640> working<00:08:38.880> on<00:08:39.039> for<00:08:39.200> the<00:08:39.279> last
semi-quietly working on for the last
semi-quietly working on for the last
year<00:08:39.680> and<00:08:39.839> a<00:08:39.839> half<00:08:40.080> or<00:08:40.159> so<00:08:40.640> it<00:08:40.719> hasn't<00:08:41.120> been<00:08:41.279> yet
year and a half or so it hasn't been yet
year and a half or so it hasn't been yet
been<00:08:41.760> launched<00:08:42.159> it's<00:08:42.320> called<00:08:42.560> duolingo<00:08:43.360> since
been launched it's called duolingo since
been launched it's called duolingo since
it<00:08:43.760> hasn't<00:08:44.000> been<00:08:44.159> launched
uh<00:08:46.959> oh<00:08:47.120> yeah<00:08:47.279> i<00:08:47.360> can<00:08:47.519> trust<00:08:47.760> you<00:08:47.920> with<00:08:48.160> that
uh oh yeah i can trust you with that
uh oh yeah i can trust you with that
okay<00:08:49.519> so<00:08:49.680> this<00:08:49.760> is<00:08:49.920> a<00:08:50.000> project<00:08:50.240> here's<00:08:50.480> how<00:08:50.560> it
okay so this is a project here's how it
okay so this is a project here's how it
started<00:08:51.120> it<00:08:51.279> started<00:08:51.519> with<00:08:51.680> me<00:08:51.920> posting<00:08:52.240> a
started it started with me posting a
started it started with me posting a
question<00:08:52.640> to<00:08:52.720> my<00:08:52.880> graduate<00:08:53.360> student<00:08:53.839> severan
question to my graduate student severan
question to my graduate student severan
hacker<00:08:54.880> okay<00:08:55.120> that's<00:08:55.360> evern<00:08:55.600> hacker
hacker okay that's evern hacker
hacker okay that's evern hacker
so<00:08:57.600> i<00:08:57.760> posed<00:08:58.000> a<00:08:58.080> question<00:08:58.320> to<00:08:58.480> my<00:08:58.560> graduate
so i posed a question to my graduate
so i posed a question to my graduate
student<00:08:59.279> by<00:08:59.360> the<00:08:59.440> way<00:08:59.680> he's<00:08:59.839> left<00:09:00.080> you<00:09:00.160> did
student by the way he's left you did
student by the way he's left you did
hear<00:09:00.560> me<00:09:00.720> correctly<00:09:00.959> his<00:09:01.120> last<00:09:01.279> name<00:09:01.519> is
hear me correctly his last name is
hear me correctly his last name is
hacker
hacker
hacker
so<00:09:03.440> i<00:09:03.519> posed<00:09:03.760> this<00:09:03.920> question<00:09:04.160> to<00:09:04.320> him<00:09:04.880> how<00:09:05.040> can
so i posed this question to him how can
so i posed this question to him how can
we<00:09:05.360> get<00:09:05.600> a<00:09:05.760> hundred<00:09:06.160> million<00:09:06.480> people
we get a hundred million people
we get a hundred million people
translating<00:09:07.519> the<00:09:07.600> web<00:09:08.000> into<00:09:08.240> every<00:09:08.480> major
translating the web into every major
translating the web into every major
language<00:09:09.279> for<00:09:09.440> free
language for free
language for free
okay<00:09:10.640> so<00:09:10.800> there's<00:09:10.959> a<00:09:11.040> lot<00:09:11.200> of<00:09:11.279> things<00:09:11.440> to<00:09:11.519> say
okay so there's a lot of things to say
okay so there's a lot of things to say
about<00:09:11.839> this<00:09:12.000> question<00:09:12.240> first<00:09:12.480> of<00:09:12.560> all
about this question first of all
about this question first of all
translating<00:09:13.279> the<00:09:13.440> web<00:09:13.920> so<00:09:14.240> right<00:09:14.399> now<00:09:14.560> the<00:09:14.640> web
translating the web so right now the web
translating the web so right now the web
is<00:09:15.200> partitioned<00:09:15.680> into<00:09:15.920> multiple<00:09:16.240> languages<00:09:16.880> a
is partitioned into multiple languages a
is partitioned into multiple languages a
large<00:09:17.200> fraction<00:09:17.519> of<00:09:17.680> it<00:09:17.760> is<00:09:17.839> in<00:09:18.000> english<00:09:18.240> if
large fraction of it is in english if
large fraction of it is in english if
you<00:09:18.480> don't<00:09:18.560> know<00:09:18.720> any<00:09:18.880> english<00:09:19.600> you<00:09:19.680> can't
you don't know any english you can't
you don't know any english you can't
access<00:09:20.320> it
access it
access it
but<00:09:21.120> there's<00:09:21.440> large<00:09:21.680> fractions<00:09:22.080> in<00:09:22.160> other
but there's large fractions in other
but there's large fractions in other
different<00:09:22.640> languages<00:09:23.200> and<00:09:23.279> if<00:09:23.440> you<00:09:23.519> don't
different languages and if you don't
different languages and if you don't
know<00:09:23.839> those<00:09:24.000> languages<00:09:24.640> you<00:09:24.800> can't<00:09:24.959> access<00:09:25.279> it
know those languages you can't access it
know those languages you can't access it
so<00:09:25.600> i<00:09:25.680> would<00:09:25.760> like<00:09:25.920> to<00:09:26.000> translate<00:09:26.399> all<00:09:26.560> of<00:09:26.640> the
so i would like to translate all of the
so i would like to translate all of the
web<00:09:27.040> or<00:09:27.200> at<00:09:27.279> least<00:09:27.600> most<00:09:27.839> of<00:09:28.000> the<00:09:28.080> web<00:09:28.640> into
web or at least most of the web into
web or at least most of the web into
every<00:09:29.040> major<00:09:29.360> language<00:09:29.920> okay<00:09:30.080> so<00:09:30.240> that's
every major language okay so that's
every major language okay so that's
that's<00:09:31.360> what<00:09:31.519> i<00:09:31.600> would<00:09:31.680> like<00:09:31.839> to<00:09:32.000> do<00:09:32.399> now<00:09:32.800> some
that's what i would like to do now some
that's what i would like to do now some
of<00:09:33.120> you<00:09:33.360> may<00:09:33.600> say<00:09:33.760> well<00:09:33.920> why<00:09:34.160> can't<00:09:34.320> we<00:09:34.399> use
of you may say well why can't we use
of you may say well why can't we use
computers<00:09:35.040> to<00:09:35.200> translate<00:09:35.839> why<00:09:36.000> can't<00:09:36.160> we<00:09:36.320> use
computers to translate why can't we use
computers to translate why can't we use
machine<00:09:36.720> translation<00:09:37.279> machine<00:09:37.519> translation
machine translation machine translation
machine translation machine translation
nowadays<00:09:38.480> starting<00:09:38.720> to<00:09:38.800> translate<00:09:39.200> some
nowadays starting to translate some
nowadays starting to translate some
sentences<00:09:39.760> here<00:09:39.920> and<00:09:40.000> there<00:09:40.320> why<00:09:40.480> can't<00:09:40.640> we
sentences here and there why can't we
sentences here and there why can't we
use<00:09:40.800> it<00:09:40.880> to<00:09:40.959> translate<00:09:41.279> the<00:09:41.360> whole<00:09:41.519> web<00:09:42.160> well
use it to translate the whole web well
use it to translate the whole web well
the<00:09:42.399> problem<00:09:42.640> with<00:09:42.800> that<00:09:42.959> is<00:09:43.040> that<00:09:43.120> it's<00:09:43.440> not
the problem with that is that it's not
the problem with that is that it's not
yet<00:09:43.760> good<00:09:43.920> enough<00:09:44.160> and<00:09:44.320> it<00:09:44.399> probably<00:09:44.720> won't<00:09:44.880> be
yet good enough and it probably won't be
yet good enough and it probably won't be
for<00:09:45.120> the<00:09:45.200> next<00:09:45.440> 15<00:09:45.680> to<00:09:45.839> 20<00:09:46.080> years<00:09:46.640> it<00:09:46.720> makes<00:09:46.959> a
for the next 15 to 20 years it makes a
for the next 15 to 20 years it makes a
lot<00:09:47.120> of<00:09:47.279> mistakes<00:09:48.080> even<00:09:48.320> when<00:09:48.480> it<00:09:48.560> doesn't
lot of mistakes even when it doesn't
lot of mistakes even when it doesn't
make<00:09:48.959> a<00:09:49.120> mistake<00:09:49.519> since<00:09:49.760> it<00:09:49.839> makes<00:09:50.000> so<00:09:50.160> many
make a mistake since it makes so many
make a mistake since it makes so many
mistakes<00:09:50.720> you<00:09:50.800> don't<00:09:50.880> know<00:09:50.959> whether<00:09:51.120> to<00:09:51.279> trust
mistakes you don't know whether to trust
mistakes you don't know whether to trust
it<00:09:51.600> or<00:09:51.760> not
it or not
it or not
so<00:09:53.040> let<00:09:53.200> me<00:09:53.279> show<00:09:53.440> you<00:09:53.600> an<00:09:53.760> example<00:09:54.240> of<00:09:54.640> uh
so let me show you an example of uh
so let me show you an example of uh
something<00:09:55.200> that<00:09:55.279> was<00:09:55.519> translated<00:09:56.000> with<00:09:56.160> a
something that was translated with a
something that was translated with a
machine<00:09:56.640> it's<00:09:56.880> actually<00:09:57.120> it<00:09:57.200> was<00:09:57.360> a<00:09:57.519> forum
machine it's actually it was a forum
machine it's actually it was a forum
post<00:09:58.560> it<00:09:58.640> was<00:09:58.800> somebody<00:09:59.120> who<00:09:59.200> was<00:09:59.279> trying<00:09:59.519> to
post it was somebody who was trying to
post it was somebody who was trying to
ask<00:09:59.839> a<00:10:00.000> question<00:10:00.880> about<00:10:01.200> javascript<00:10:01.920> it<00:10:02.000> was
ask a question about javascript it was
ask a question about javascript it was
translated<00:10:02.720> from<00:10:02.880> japanese<00:10:03.519> into<00:10:03.839> english<00:10:04.640> so
translated from japanese into english so
translated from japanese into english so
i'll<00:10:05.760> just<00:10:05.920> let<00:10:06.080> you<00:10:06.240> read<00:10:06.399> this<00:10:06.640> person
i'll just let you read this person
i'll just let you read this person
starts<00:10:07.360> apologizing<00:10:08.160> from<00:10:08.320> the<00:10:08.399> fact<00:10:08.640> that
starts apologizing from the fact that
starts apologizing from the fact that
it's
it's
it's
translated<00:10:10.000> with<00:10:10.080> a<00:10:10.160> computer<00:10:10.640> okay<00:10:10.880> so<00:10:11.279> the
translated with a computer okay so the
translated with a computer okay so the
next<00:10:11.680> sentence<00:10:12.080> is<00:10:12.240> going<00:10:12.399> to<00:10:12.480> be<00:10:13.120> the
next sentence is going to be the
next sentence is going to be the
preamble<00:10:13.920> to<00:10:14.079> the<00:10:14.160> question<00:10:14.399> so<00:10:14.560> he's<00:10:14.720> just
preamble to the question so he's just
preamble to the question so he's just
explaining<00:10:15.440> something<00:10:16.160> remember<00:10:16.560> it's<00:10:16.720> a
explaining something remember it's a
explaining something remember it's a
question<00:10:17.040> about<00:10:17.200> javascript
okay
okay
okay
then
then
then
then<00:10:27.920> comes<00:10:28.160> the<00:10:28.240> first<00:10:28.480> part<00:10:28.640> of<00:10:28.720> the
then comes the first part of the
then comes the first part of the
question
then<00:10:37.440> comes<00:10:37.680> my<00:10:37.839> favorite<00:10:38.160> part<00:10:38.320> of<00:10:38.399> the
then comes my favorite part of the
then comes my favorite part of the
question
and<00:10:45.200> then<00:10:45.440> comes<00:10:45.680> the<00:10:45.839> ending<00:10:46.079> which<00:10:46.320> is<00:10:46.399> my
and then comes the ending which is my
and then comes the ending which is my
favorite<00:10:46.880> part<00:10:47.040> of<00:10:47.120> the<00:10:47.200> whole<00:10:47.440> thing
okay<00:10:54.079> so<00:10:54.240> computer<00:10:54.640> translation<00:10:55.519> not<00:10:55.680> yet
okay so computer translation not yet
okay so computer translation not yet
good<00:10:56.079> enough<00:10:56.880> okay<00:10:57.360> so<00:10:57.519> back<00:10:57.680> to<00:10:57.839> the<00:10:57.920> question
good enough okay so back to the question
good enough okay so back to the question
so<00:10:58.320> we<00:10:58.399> need<00:10:58.640> people<00:10:58.959> to<00:10:59.120> translate<00:10:59.600> the<00:10:59.680> whole
so we need people to translate the whole
so we need people to translate the whole
well<00:11:00.560> okay<00:11:01.200> so<00:11:01.600> now<00:11:01.760> the<00:11:01.839> next<00:11:02.079> question<00:11:02.399> you
well okay so now the next question you
well okay so now the next question you
may<00:11:02.640> have<00:11:02.800> is<00:11:02.880> well<00:11:03.040> why<00:11:03.200> can't<00:11:03.360> we<00:11:03.519> just<00:11:03.680> pay
may have is well why can't we just pay
may have is well why can't we just pay
people<00:11:04.160> to<00:11:04.320> do<00:11:04.399> this<00:11:04.560> we<00:11:04.640> could<00:11:04.720> pay
people to do this we could pay
people to do this we could pay
professional<00:11:05.360> language<00:11:05.600> translators<00:11:06.079> to
professional language translators to
professional language translators to
translate<00:11:06.480> the<00:11:06.560> whole<00:11:06.720> web
translate the whole web
translate the whole web
we<00:11:08.000> could<00:11:08.160> do<00:11:08.320> that<00:11:08.959> um<00:11:09.279> unfortunately<00:11:09.760> it
we could do that um unfortunately it
we could do that um unfortunately it
would<00:11:09.920> be<00:11:10.079> extremely<00:11:10.480> expensive<00:11:11.360> for<00:11:11.519> example
would be extremely expensive for example
would be extremely expensive for example
translating<00:11:12.480> a<00:11:12.560> tiny<00:11:12.800> tiny<00:11:13.120> fraction<00:11:13.440> of<00:11:13.600> the
translating a tiny tiny fraction of the
translating a tiny tiny fraction of the
whole<00:11:13.839> web<00:11:14.399> wikipedia<00:11:15.279> into<00:11:15.600> one<00:11:15.760> other
whole web wikipedia into one other
whole web wikipedia into one other
language<00:11:16.720> spanish<00:11:17.680> okay<00:11:18.079> you<00:11:18.240> know<00:11:18.480> wikipedia
language spanish okay you know wikipedia
language spanish okay you know wikipedia
exists<00:11:19.200> in<00:11:19.279> spanish<00:11:19.760> but<00:11:19.920> it's<00:11:20.160> very<00:11:20.320> small
exists in spanish but it's very small
exists in spanish but it's very small
compared<00:11:20.880> to<00:11:20.959> the<00:11:21.040> size<00:11:21.200> of<00:11:21.360> english<00:11:21.600> it's
compared to the size of english it's
compared to the size of english it's
about<00:11:21.839> 20<00:11:22.320> as<00:11:22.399> the<00:11:22.480> size<00:11:22.640> of<00:11:22.720> english<00:11:23.120> if<00:11:23.200> we
about 20 as the size of english if we
about 20 as the size of english if we
wanted<00:11:23.440> to<00:11:23.600> translate<00:11:23.920> the<00:11:24.079> other<00:11:24.240> 80
wanted to translate the other 80
wanted to translate the other 80
into<00:11:25.760> spanish<00:11:26.399> it<00:11:26.560> would<00:11:26.720> cost<00:11:27.040> at<00:11:27.200> least<00:11:27.519> 50
into spanish it would cost at least 50
into spanish it would cost at least 50
million<00:11:28.160> dollars<00:11:28.480> and<00:11:28.560> this<00:11:28.720> is<00:11:28.880> even<00:11:29.040> at<00:11:29.120> the
million dollars and this is even at the
million dollars and this is even at the
most<00:11:29.519> exploited<00:11:30.320> outsourcing<00:11:30.880> country<00:11:31.279> out
most exploited outsourcing country out
most exploited outsourcing country out
there
there
there
okay
okay
okay
so<00:11:33.279> it'd<00:11:33.360> be<00:11:33.519> very<00:11:33.680> expensive<00:11:34.079> so<00:11:34.320> what<00:11:34.480> we
so it'd be very expensive so what we
so it'd be very expensive so what we
want<00:11:34.720> to<00:11:34.800> do<00:11:34.880> is<00:11:34.959> we<00:11:35.040> want<00:11:35.120> to<00:11:35.200> get<00:11:35.360> 100<00:11:35.519> million
want to do is we want to get 100 million
want to do is we want to get 100 million
people<00:11:36.000> translate<00:11:36.399> the<00:11:36.480> web<00:11:36.800> into<00:11:37.040> every
people translate the web into every
people translate the web into every
major<00:11:37.440> language<00:11:37.920> for<00:11:38.160> free
major language for free
major language for free
okay<00:11:39.920> now<00:11:40.160> if<00:11:40.240> this<00:11:40.399> is<00:11:40.480> what<00:11:40.560> you<00:11:40.720> want<00:11:40.800> to<00:11:40.880> do
okay now if this is what you want to do
okay now if this is what you want to do
you<00:11:41.120> pretty<00:11:41.279> quickly<00:11:41.600> realize<00:11:41.920> you're<00:11:42.079> going
you pretty quickly realize you're going
you pretty quickly realize you're going
to<00:11:42.160> run<00:11:42.320> into<00:11:42.560> two<00:11:42.800> pretty<00:11:43.120> big<00:11:43.360> hurdles<00:11:43.920> or
to run into two pretty big hurdles or
to run into two pretty big hurdles or
two<00:11:44.240> big<00:11:44.480> obstacles<00:11:45.360> okay<00:11:45.600> the<00:11:45.839> first<00:11:46.160> one<00:11:46.800> is
two big obstacles okay the first one is
two big obstacles okay the first one is
a<00:11:47.200> lack<00:11:47.440> of<00:11:47.600> bilinguals<00:11:48.560> okay<00:11:48.880> so
a lack of bilinguals okay so
a lack of bilinguals okay so
i<00:11:50.240> don't<00:11:50.399> even<00:11:50.640> know<00:11:50.959> if<00:11:51.120> there<00:11:51.360> exists<00:11:51.760> 100
i don't even know if there exists 100
i don't even know if there exists 100
million<00:11:52.399> people<00:11:52.720> out<00:11:52.880> there<00:11:53.040> using<00:11:53.279> the<00:11:53.440> web
million people out there using the web
million people out there using the web
who<00:11:53.760> are<00:11:53.839> bilingual<00:11:54.399> enough<00:11:54.640> to<00:11:54.720> help<00:11:54.959> us
who are bilingual enough to help us
who are bilingual enough to help us
translate<00:11:55.920> okay<00:11:56.160> that's<00:11:56.399> a<00:11:56.480> big<00:11:56.720> problem<00:11:57.120> the
translate okay that's a big problem the
translate okay that's a big problem the
other<00:11:57.440> problem<00:11:57.600> that<00:11:57.760> you're<00:11:57.839> going<00:11:57.920> to<00:11:58.000> run
other problem that you're going to run
other problem that you're going to run
into<00:11:58.399> is<00:11:58.480> the<00:11:58.560> lack<00:11:58.800> of<00:11:58.880> motivation<00:11:59.920> how<00:12:00.079> are
into is the lack of motivation how are
into is the lack of motivation how are
we<00:12:00.240> going<00:12:00.399> to<00:12:00.480> motivate<00:12:00.959> people<00:12:01.440> to<00:12:01.600> actually
we going to motivate people to actually
we going to motivate people to actually
translate<00:12:02.320> the<00:12:02.399> web<00:12:02.720> for<00:12:02.959> free
translate the web for free
translate the web for free
this<00:12:04.079> is<00:12:04.160> normally<00:12:04.800> you<00:12:04.880> have<00:12:05.040> to<00:12:05.120> pay<00:12:05.440> people
this is normally you have to pay people
this is normally you have to pay people
to<00:12:05.760> do<00:12:05.920> this<00:12:06.320> so<00:12:06.480> how<00:12:06.560> we're<00:12:06.720> going<00:12:06.720> to
to do this so how we're going to
to do this so how we're going to
motivate<00:12:07.200> them<00:12:07.279> to<00:12:07.360> do<00:12:07.519> it<00:12:07.600> for<00:12:07.760> free<00:12:08.480> now<00:12:09.120> when
motivate them to do it for free now when
motivate them to do it for free now when
we<00:12:09.519> were<00:12:09.760> starting<00:12:10.000> to<00:12:10.079> think<00:12:10.320> about<00:12:10.480> this<00:12:10.639> we
we were starting to think about this we
we were starting to think about this we
were<00:12:10.959> blocked<00:12:11.200> by<00:12:11.279> these<00:12:11.440> two<00:12:11.600> things<00:12:11.760> but
were blocked by these two things but
were blocked by these two things but
then<00:12:12.000> we<00:12:12.160> realized<00:12:12.480> there's<00:12:12.639> actually<00:12:12.880> a<00:12:12.959> way
then we realized there's actually a way
then we realized there's actually a way
to<00:12:13.200> solve<00:12:13.519> both<00:12:13.760> of<00:12:13.839> these<00:12:14.079> problems<00:12:14.399> with<00:12:14.560> the
to solve both of these problems with the
to solve both of these problems with the
same<00:12:14.880> solution<00:12:15.600> there's<00:12:15.839> a<00:12:15.920> way<00:12:16.000> to<00:12:16.160> kill<00:12:16.399> two
same solution there's a way to kill two
same solution there's a way to kill two
birds<00:12:16.959> with<00:12:17.040> one<00:12:17.279> stone
birds with one stone
birds with one stone
and<00:12:18.399> that<00:12:18.639> is<00:12:18.720> to<00:12:18.959> transform<00:12:19.600> language
and that is to transform language
and that is to transform language
translation<00:12:20.639> into<00:12:20.880> something<00:12:21.120> that<00:12:21.360> millions
translation into something that millions
translation into something that millions
of<00:12:21.839> people<00:12:22.160> want<00:12:22.560> to<00:12:22.720> do
of people want to do
of people want to do
and<00:12:24.240> that<00:12:24.399> also<00:12:24.639> helps<00:12:24.880> with<00:12:25.040> the<00:12:25.120> problem<00:12:25.360> of
and that also helps with the problem of
and that also helps with the problem of
lack<00:12:25.680> of<00:12:25.760> bilinguals<00:12:26.720> and<00:12:26.880> that<00:12:27.040> is<00:12:27.279> language
lack of bilinguals and that is language
lack of bilinguals and that is language
education
education
education
okay<00:12:29.279> so<00:12:29.839> it<00:12:30.000> turns<00:12:30.240> out<00:12:30.399> that<00:12:30.560> today<00:12:31.200> there
okay so it turns out that today there
okay so it turns out that today there
are<00:12:31.519> over<00:12:31.920> 1.2<00:12:32.720> billion<00:12:33.120> people<00:12:33.519> learning<00:12:33.760> a
are over 1.2 billion people learning a
are over 1.2 billion people learning a
foreign<00:12:34.079> language
foreign language
foreign language
people<00:12:35.680> really<00:12:35.920> really<00:12:36.160> want<00:12:36.320> to<00:12:36.399> learn<00:12:36.560> a
people really really want to learn a
people really really want to learn a
foreign<00:12:36.880> language<00:12:37.200> and<00:12:37.279> it's<00:12:37.440> not<00:12:37.519> just
foreign language and it's not just
foreign language and it's not just
because<00:12:37.920> they're<00:12:38.079> being<00:12:38.240> forced<00:12:38.560> to<00:12:38.639> do<00:12:38.800> so<00:12:39.040> in
because they're being forced to do so in
because they're being forced to do so in
school<00:12:39.680> for<00:12:39.839> example<00:12:40.079> in<00:12:40.240> the<00:12:40.320> united<00:12:40.639> states
school for example in the united states
school for example in the united states
alone<00:12:41.040> there<00:12:41.200> are<00:12:41.279> over<00:12:41.440> 5<00:12:41.680> million<00:12:42.000> people
alone there are over 5 million people
alone there are over 5 million people
who<00:12:42.399> have<00:12:42.480> paid<00:12:42.720> over<00:12:42.959> 500<00:12:43.680> for<00:12:43.839> software<00:12:44.480> to
who have paid over 500 for software to
who have paid over 500 for software to
learn<00:12:44.880> a<00:12:44.959> new<00:12:45.120> language<00:12:45.680> okay<00:12:45.839> so<00:12:46.000> people
learn a new language okay so people
learn a new language okay so people
really<00:12:46.480> really<00:12:46.720> want<00:12:46.880> to<00:12:46.959> learn<00:12:47.120> a<00:12:47.200> new
really really want to learn a new
really really want to learn a new
language<00:12:48.000> so<00:12:48.160> what<00:12:48.320> we've<00:12:48.480> been<00:12:48.639> working<00:12:48.959> on
language so what we've been working on
language so what we've been working on
for<00:12:49.200> the<00:12:49.279> last<00:12:49.440> year<00:12:49.600> and<00:12:49.760> a<00:12:49.760> half<00:12:50.000> is<00:12:50.079> a<00:12:50.160> new
for the last year and a half is a new
for the last year and a half is a new
website<00:12:50.720> it's<00:12:50.880> called<00:12:51.040> duolingo<00:12:52.000> where<00:12:52.160> the
website it's called duolingo where the
website it's called duolingo where the
basic<00:12:52.639> idea<00:12:53.040> is<00:12:53.279> people<00:12:53.760> learn<00:12:54.000> a<00:12:54.079> new
basic idea is people learn a new
basic idea is people learn a new
language<00:12:54.800> for<00:12:55.120> free<00:12:55.760> while<00:12:56.079> simultaneously
language for free while simultaneously
language for free while simultaneously
translating<00:12:57.440> the<00:12:57.519> web<00:12:58.079> and<00:12:58.240> so<00:12:58.399> basically
translating the web and so basically
translating the web and so basically
they're<00:12:58.959> learning<00:12:59.279> by<00:12:59.519> doing
they're learning by doing
they're learning by doing
okay<00:13:00.800> so<00:13:00.959> the<00:13:01.040> way<00:13:01.279> this<00:13:01.440> works<00:13:01.760> is<00:13:02.000> whenever
okay so the way this works is whenever
okay so the way this works is whenever
you're<00:13:02.639> just<00:13:02.880> a<00:13:02.959> beginner<00:13:03.279> we<00:13:03.440> give<00:13:03.600> you<00:13:03.760> very
you're just a beginner we give you very
you're just a beginner we give you very
very<00:13:04.240> simple<00:13:04.560> sentences<00:13:05.040> there's<00:13:05.200> of<00:13:05.279> course
very simple sentences there's of course
very simple sentences there's of course
a<00:13:05.600> lot<00:13:05.760> of<00:13:05.839> very<00:13:06.079> simple<00:13:06.320> sentences<00:13:06.800> on<00:13:06.880> the
a lot of very simple sentences on the
a lot of very simple sentences on the
web<00:13:07.360> we<00:13:07.519> give<00:13:07.680> you<00:13:07.839> very<00:13:08.000> very<00:13:08.160> simple
web we give you very very simple
web we give you very very simple
sentences<00:13:08.800> along<00:13:09.120> with<00:13:09.360> what<00:13:09.600> each<00:13:09.760> word
sentences along with what each word
sentences along with what each word
means
means
means
okay<00:13:11.200> and<00:13:11.360> as<00:13:11.519> you<00:13:11.680> translate<00:13:12.160> them<00:13:12.320> and<00:13:12.480> as
okay and as you translate them and as
okay and as you translate them and as
you<00:13:12.639> see<00:13:12.800> how<00:13:12.959> other<00:13:13.120> people<00:13:13.440> translate<00:13:13.839> them
you see how other people translate them
you see how other people translate them
you<00:13:14.160> start<00:13:14.399> learning<00:13:14.720> the<00:13:14.800> language<00:13:15.120> and<00:13:15.279> as
you start learning the language and as
you start learning the language and as
you<00:13:15.600> as<00:13:15.760> you<00:13:15.839> get<00:13:16.240> more<00:13:16.480> and<00:13:16.560> more<00:13:16.720> advanced<00:13:17.040> we
you as you get more and more advanced we
you as you get more and more advanced we
give<00:13:17.360> you<00:13:17.440> more<00:13:17.600> and<00:13:17.680> more<00:13:17.839> complex<00:13:18.240> sentences
give you more and more complex sentences
give you more and more complex sentences
to<00:13:18.800> translate<00:13:19.200> but<00:13:19.360> at<00:13:19.519> all<00:13:19.680> times<00:13:19.920> you're
to translate but at all times you're
to translate but at all times you're
learning<00:13:20.399> by<00:13:20.639> doing<00:13:21.360> okay<00:13:21.600> now<00:13:21.760> the<00:13:21.920> crazy
learning by doing okay now the crazy
learning by doing okay now the crazy
thing<00:13:22.399> about<00:13:22.560> this<00:13:22.720> method<00:13:23.200> method<00:13:23.519> is<00:13:23.680> that
thing about this method method is that
thing about this method method is that
it<00:13:24.000> actually<00:13:24.320> really<00:13:24.639> works<00:13:25.360> okay<00:13:25.600> first<00:13:25.839> of
it actually really works okay first of
it actually really works okay first of
all<00:13:26.000> people<00:13:26.240> are<00:13:26.320> really<00:13:26.560> really<00:13:26.800> learning<00:13:27.120> a
all people are really really learning a
all people are really really learning a
language<00:13:27.519> we're<00:13:27.680> mostly<00:13:28.000> done<00:13:28.240> building<00:13:28.480> it
language we're mostly done building it
language we're mostly done building it
and<00:13:28.639> now<00:13:28.800> we're<00:13:28.959> testing<00:13:29.360> it<00:13:29.680> people<00:13:29.920> really
and now we're testing it people really
and now we're testing it people really
can<00:13:30.399> learn<00:13:30.639> a<00:13:30.720> language<00:13:31.120> with<00:13:31.279> it<00:13:31.920> and<00:13:32.160> and
can learn a language with it and and
can learn a language with it and and
they<00:13:32.480> learn<00:13:32.720> it<00:13:32.800> about<00:13:33.040> as<00:13:33.120> well<00:13:33.279> as<00:13:33.440> the
they learn it about as well as the
they learn it about as well as the
leading<00:13:33.839> language<00:13:34.240> learning<00:13:34.480> software<00:13:34.880> so
leading language learning software so
leading language learning software so
people<00:13:35.279> really<00:13:35.440> do<00:13:35.600> learn<00:13:35.760> a<00:13:35.839> language<00:13:36.240> and
people really do learn a language and
people really do learn a language and
not<00:13:36.480> only<00:13:36.639> do<00:13:36.800> they<00:13:36.959> learn<00:13:37.120> it<00:13:37.279> as<00:13:37.440> well<00:13:37.920> but
not only do they learn it as well but
not only do they learn it as well but
actually<00:13:38.720> it's<00:13:39.040> way<00:13:39.279> more<00:13:39.519> interesting
actually it's way more interesting
actually it's way more interesting
because<00:13:40.480> you<00:13:40.639> see<00:13:40.959> with<00:13:41.120> duolingo<00:13:41.680> people<00:13:41.839> are
because you see with duolingo people are
because you see with duolingo people are
actually<00:13:42.160> learning<00:13:42.480> with<00:13:42.639> real<00:13:42.880> content<00:13:43.600> as
actually learning with real content as
actually learning with real content as
opposed<00:13:44.000> to<00:13:44.160> learning<00:13:44.399> with<00:13:44.560> made<00:13:44.800> up
opposed to learning with made up
opposed to learning with made up
sentences<00:13:45.360> people<00:13:45.600> are<00:13:45.680> learning<00:13:45.920> with<00:13:46.000> real
sentences people are learning with real
sentences people are learning with real
content<00:13:46.480> which<00:13:46.639> is<00:13:46.720> inherently<00:13:47.279> interesting
content which is inherently interesting
content which is inherently interesting
okay<00:13:48.399> the<00:13:48.639> other<00:13:49.279> thing<00:13:49.600> so<00:13:49.760> so<00:13:50.000> people<00:13:50.240> really
okay the other thing so so people really
okay the other thing so so people really
do<00:13:50.560> learn<00:13:50.720> a<00:13:50.800> language<00:13:51.120> but<00:13:51.279> perhaps<00:13:51.680> more
do learn a language but perhaps more
do learn a language but perhaps more
surprisingly<00:13:53.040> the<00:13:53.279> translations<00:13:53.839> that<00:13:54.000> we
surprisingly the translations that we
surprisingly the translations that we
get<00:13:54.800> from<00:13:54.959> people<00:13:55.200> using<00:13:55.440> the<00:13:55.519> site<00:13:55.839> even
get from people using the site even
get from people using the site even
though<00:13:56.079> they're<00:13:56.240> just<00:13:56.480> beginners
though they're just beginners
though they're just beginners
the<00:13:57.680> translations<00:13:58.160> that<00:13:58.240> we<00:13:58.399> get<00:13:58.639> are<00:13:58.800> as
the translations that we get are as
the translations that we get are as
accurate<00:13:59.360> as<00:13:59.519> those<00:13:59.760> professional<00:14:00.160> language
accurate as those professional language
accurate as those professional language
translators<00:14:01.360> which<00:14:01.600> is<00:14:01.760> very<00:14:01.920> surprising<00:14:02.320> so
translators which is very surprising so
translators which is very surprising so
let<00:14:02.560> me<00:14:02.639> show<00:14:02.800> you<00:14:03.279> one<00:14:03.519> example<00:14:04.079> this<00:14:04.240> is<00:14:04.320> a
let me show you one example this is a
let me show you one example this is a
sentence<00:14:04.720> that<00:14:04.800> was<00:14:04.880> translated<00:14:05.360> from<00:14:05.519> german
sentence that was translated from german
sentence that was translated from german
into<00:14:06.240> english<00:14:06.800> the<00:14:07.040> top<00:14:07.279> is<00:14:07.360> the<00:14:07.519> german<00:14:08.160> the
into english the top is the german the
into english the top is the german the
middle<00:14:08.880> is<00:14:09.279> an<00:14:09.519> english<00:14:09.839> translation<00:14:10.320> that
middle is an english translation that
middle is an english translation that
was<00:14:10.560> done<00:14:10.800> by<00:14:10.959> somebody<00:14:11.360> who<00:14:11.519> was<00:14:11.680> a
was done by somebody who was a
was done by somebody who was a
professional<00:14:12.160> language<00:14:12.480> translator<00:14:12.959> who
professional language translator who
professional language translator who
paid<00:14:13.360> 20<00:14:13.600> cents<00:14:13.839> a<00:14:13.920> word<00:14:14.240> for<00:14:14.399> this
paid 20 cents a word for this
paid 20 cents a word for this
translation<00:14:15.279> and<00:14:15.360> the<00:14:15.440> bottom<00:14:15.760> is<00:14:15.839> the
translation and the bottom is the
translation and the bottom is the
translation<00:14:16.560> by<00:14:16.800> users<00:14:17.120> of<00:14:17.199> duolingo<00:14:18.160> none<00:14:18.399> of
translation by users of duolingo none of
translation by users of duolingo none of
whom<00:14:18.800> knew<00:14:19.120> any<00:14:19.360> german<00:14:19.839> before<00:14:20.079> they<00:14:20.240> started
whom knew any german before they started
whom knew any german before they started
using<00:14:20.720> the<00:14:20.880> site
using the site
using the site
you<00:14:21.920> can<00:14:22.000> see<00:14:22.240> it's<00:14:22.399> pretty<00:14:22.639> much<00:14:22.800> perfect<00:14:23.600> now
you can see it's pretty much perfect now
you can see it's pretty much perfect now
of<00:14:23.920> course<00:14:24.160> we<00:14:24.320> play<00:14:24.480> a<00:14:24.560> trick<00:14:24.880> here<00:14:25.279> to<00:14:25.440> make
of course we play a trick here to make
of course we play a trick here to make
the<00:14:25.760> translations<00:14:26.399> as<00:14:26.560> good<00:14:26.800> as<00:14:26.880> professional
the translations as good as professional
the translations as good as professional
language<00:14:27.519> translators<00:14:27.920> we<00:14:28.079> combine<00:14:28.399> the
language translators we combine the
language translators we combine the
translations<00:14:29.040> of<00:14:29.120> multiple<00:14:29.920> beginners<00:14:30.800> to
translations of multiple beginners to
translations of multiple beginners to
get<00:14:31.120> the<00:14:31.279> quality<00:14:31.680> of<00:14:31.760> a<00:14:31.839> single<00:14:32.399> professional
get the quality of a single professional
get the quality of a single professional
translator<00:14:33.920> okay<00:14:34.160> now<00:14:34.639> even<00:14:34.880> though<00:14:35.040> we're
translator okay now even though we're
translator okay now even though we're
combining
combining
combining
the<00:14:37.440> translations
the translations
the translations
the<00:14:39.120> the<00:14:39.360> site<00:14:39.680> actually<00:14:40.000> can<00:14:40.160> translate
the the site actually can translate
the the site actually can translate
pretty<00:14:40.959> fast<00:14:41.360> so<00:14:41.519> let<00:14:41.680> me<00:14:41.839> show<00:14:42.000> you<00:14:42.320> this<00:14:42.480> is
pretty fast so let me show you this is
pretty fast so let me show you this is
our<00:14:42.800> estimates<00:14:43.440> of<00:14:43.600> how<00:14:43.839> fast<00:14:44.079> we<00:14:44.160> could
our estimates of how fast we could
our estimates of how fast we could
translate<00:14:44.720> wikipedia<00:14:45.279> from<00:14:45.440> english<00:14:45.680> into
translate wikipedia from english into
translate wikipedia from english into
spanish<00:14:46.800> remember<00:14:47.199> this<00:14:47.360> is<00:14:47.440> 50<00:14:47.680> million
spanish remember this is 50 million
spanish remember this is 50 million
dollars<00:14:48.320> worth<00:14:48.560> of<00:14:48.639> value
dollars worth of value
dollars worth of value
okay<00:14:49.920> so<00:14:50.160> if<00:14:50.320> we<00:14:50.399> wanted<00:14:50.560> to<00:14:50.639> translate<00:14:51.199> into
okay so if we wanted to translate into
okay so if we wanted to translate into
spanish<00:14:51.920> we<00:14:52.079> could<00:14:52.240> do<00:14:52.320> it<00:14:52.480> in<00:14:52.639> five<00:14:52.959> weeks
spanish we could do it in five weeks
spanish we could do it in five weeks
with<00:14:53.360> 100<00:14:53.680> 000<00:14:54.000> active<00:14:54.240> users<00:14:55.040> and<00:14:55.120> we<00:14:55.279> could
with 100 000 active users and we could
with 100 000 active users and we could
do<00:14:55.519> it<00:14:55.600> in<00:14:55.760> about<00:14:55.920> 80<00:14:56.240> hours<00:14:56.399> with<00:14:56.560> a<00:14:56.639> million
do it in about 80 hours with a million
do it in about 80 hours with a million
active<00:14:57.199> users<00:14:57.839> since<00:14:58.079> all<00:14:58.240> the<00:14:58.320> projects<00:14:58.639> that
active users since all the projects that
active users since all the projects that
my<00:14:58.959> group<00:14:59.120> has<00:14:59.279> worked<00:14:59.440> on<00:14:59.519> so<00:14:59.680> far<00:15:00.079> have
my group has worked on so far have
my group has worked on so far have
gotten<00:15:00.480> millions<00:15:00.800> of<00:15:00.880> users<00:15:01.199> we're<00:15:01.279> hopeful
gotten millions of users we're hopeful
gotten millions of users we're hopeful
that<00:15:01.680> we'll<00:15:01.839> be<00:15:02.000> able<00:15:02.079> to<00:15:02.240> translate
that we'll be able to translate
that we'll be able to translate
extremely<00:15:03.120> fast<00:15:03.360> with<00:15:03.680> this<00:15:04.000> project
extremely fast with this project
extremely fast with this project
now<00:15:05.279> the<00:15:05.440> thing<00:15:05.600> that<00:15:05.760> i'm<00:15:06.160> most<00:15:06.720> excited
now the thing that i'm most excited
now the thing that i'm most excited
about<00:15:07.440> with<00:15:07.600> duolingo<00:15:08.079> is<00:15:08.160> that<00:15:08.320> i<00:15:08.399> think<00:15:08.560> this
about with duolingo is that i think this
about with duolingo is that i think this
provides<00:15:09.120> a<00:15:09.279> fair<00:15:09.680> business<00:15:10.079> model<00:15:10.320> for
provides a fair business model for
provides a fair business model for
language<00:15:10.880> education<00:15:11.760> so<00:15:11.920> here's<00:15:12.160> the<00:15:12.240> thing
language education so here's the thing
language education so here's the thing
the<00:15:12.720> current<00:15:12.959> business<00:15:13.199> model<00:15:13.440> for<00:15:13.519> language
the current business model for language
the current business model for language
education<00:15:14.399> is<00:15:14.639> the<00:15:14.880> student<00:15:15.360> pace
education is the student pace
education is the student pace
in<00:15:16.959> particular<00:15:17.440> the<00:15:17.680> student<00:15:18.000> pays<00:15:18.320> rosetta
in particular the student pays rosetta
in particular the student pays rosetta
stone<00:15:19.040> 500
that's<00:15:21.920> the<00:15:22.000> current<00:15:22.240> business<00:15:22.560> model<00:15:22.959> the
that's the current business model the
that's the current business model the
problem<00:15:23.519> with<00:15:23.600> this<00:15:23.839> business<00:15:24.160> model<00:15:24.720> is<00:15:24.880> that
problem with this business model is that
problem with this business model is that
95<00:15:25.920> of<00:15:26.000> the<00:15:26.079> world's<00:15:26.320> population<00:15:26.880> doesn't
95 of the world's population doesn't
95 of the world's population doesn't
have<00:15:27.199> 500<00:15:28.079> so<00:15:28.240> it's<00:15:28.399> extremely<00:15:29.040> unfair
have 500 so it's extremely unfair
have 500 so it's extremely unfair
towards<00:15:30.560> the<00:15:30.800> poor<00:15:31.199> okay<00:15:31.360> this<00:15:31.519> is<00:15:31.680> totally
towards the poor okay this is totally
towards the poor okay this is totally
biased<00:15:32.320> towards<00:15:32.560> the<00:15:32.639> rich<00:15:32.959> now<00:15:33.279> see<00:15:33.680> in
biased towards the rich now see in
biased towards the rich now see in
duolingo
duolingo
duolingo
because<00:15:35.839> while<00:15:36.000> you<00:15:36.240> learn<00:15:36.480> you're<00:15:36.720> actually
because while you learn you're actually
because while you learn you're actually
creating<00:15:37.680> value<00:15:38.560> you<00:15:38.800> are<00:15:39.120> translating<00:15:39.680> stuff
creating value you are translating stuff
creating value you are translating stuff
which<00:15:40.160> for<00:15:40.320> example<00:15:40.639> we<00:15:40.800> could<00:15:40.959> charge
which for example we could charge
which for example we could charge
somebody<00:15:41.519> for<00:15:41.760> translations<00:15:42.399> so<00:15:42.560> this<00:15:42.720> is<00:15:42.800> how
somebody for translations so this is how
somebody for translations so this is how
we<00:15:43.040> could<00:15:43.199> monetize<00:15:43.760> this
we could monetize this
we could monetize this
since<00:15:44.959> people<00:15:45.120> are<00:15:45.279> creating<00:15:45.760> value<00:15:46.079> while
since people are creating value while
since people are creating value while
they're<00:15:46.720> learning<00:15:47.279> they<00:15:47.440> don't<00:15:47.600> have<00:15:47.760> to<00:15:47.839> pay
they're learning they don't have to pay
they're learning they don't have to pay
with<00:15:48.160> their<00:15:48.320> money<00:15:48.639> they<00:15:48.800> pay<00:15:49.040> with<00:15:49.199> their
with their money they pay with their
with their money they pay with their
time<00:15:50.000> but<00:15:50.160> the<00:15:50.320> magical<00:15:50.720> thing<00:15:50.880> here<00:15:51.120> is<00:15:51.199> that
time but the magical thing here is that
time but the magical thing here is that
they're<00:15:51.440> paying<00:15:51.680> with<00:15:51.839> their<00:15:52.000> time<00:15:52.320> but<00:15:52.480> that
they're paying with their time but that
they're paying with their time but that
is<00:15:52.800> time<00:15:53.040> that<00:15:53.199> would<00:15:53.440> have<00:15:53.600> to<00:15:53.759> be<00:15:54.000> spent
is time that would have to be spent
is time that would have to be spent
anyways<00:15:55.279> learning<00:15:55.600> the<00:15:55.680> language<00:15:56.560> okay<00:15:56.800> so
anyways learning the language okay so
anyways learning the language okay so
the<00:15:57.040> nice<00:15:57.199> thing<00:15:57.360> about<00:15:57.440> duolingo<00:15:57.839> is<00:15:57.920> i<00:15:58.000> think
the nice thing about duolingo is i think
the nice thing about duolingo is i think
it<00:15:58.320> provides<00:15:58.639> a<00:15:58.720> fair<00:15:58.959> business<00:15:59.199> model<00:15:59.440> one
it provides a fair business model one
it provides a fair business model one
that<00:15:59.759> doesn't<00:15:59.920> discriminate<00:16:00.399> against<00:16:00.720> poor
that doesn't discriminate against poor
that doesn't discriminate against poor
people
people
people
so<00:16:02.399> here's<00:16:02.639> the<00:16:02.800> site
so<00:16:13.279> here's<00:16:13.519> the<00:16:13.600> site<00:16:13.920> we<00:16:14.000> haven't<00:16:14.320> yet
so here's the site we haven't yet
so here's the site we haven't yet
launched<00:16:15.120> um<00:16:15.519> but<00:16:15.680> if<00:16:15.839> you<00:16:15.920> go<00:16:16.160> there<00:16:16.560> you<00:16:16.720> can
launched um but if you go there you can
launched um but if you go there you can
sign<00:16:17.360> up<00:16:17.519> to<00:16:17.600> be<00:16:17.759> part<00:16:17.920> of<00:16:18.079> our<00:16:18.240> private<00:16:18.560> beta
sign up to be part of our private beta
sign up to be part of our private beta
which<00:16:19.040> are<00:16:19.120> probably<00:16:19.440> going<00:16:19.519> to<00:16:19.680> start<00:16:19.839> in
which are probably going to start in
which are probably going to start in
about<00:16:20.160> three<00:16:20.399> or<00:16:20.480> four<00:16:20.639> weeks<00:16:21.199> we<00:16:21.360> haven't<00:16:21.519> yet
about three or four weeks we haven't yet
about three or four weeks we haven't yet
launched<00:16:21.839> this<00:16:21.920> duolingo<00:16:22.399> by<00:16:22.560> the<00:16:22.560> way<00:16:22.720> i'm
launched this duolingo by the way i'm
launched this duolingo by the way i'm
the<00:16:23.040> one<00:16:23.279> talking<00:16:23.680> here<00:16:24.000> but<00:16:24.240> actually
the one talking here but actually
the one talking here but actually
duolingo<00:16:24.880> is<00:16:24.959> the<00:16:25.040> work<00:16:25.199> of<00:16:25.360> a<00:16:25.440> really<00:16:25.680> awesome
duolingo is the work of a really awesome
duolingo is the work of a really awesome
team<00:16:26.160> some<00:16:26.320> of<00:16:26.480> whom<00:16:26.639> are<00:16:26.800> here<00:16:27.519> so<00:16:28.160> thank<00:16:28.399> you
you
Luis von Ahn, TED, TEDx, TEDxCMU, Duolingo, language, TED-Ed, TED, Ed, TEDEducation