Massive-scale online collaboration | Luis von Ahn

Translator: Lenka Tušar Reviewer: Klavdija Cernilogar

How many of you had to fill out a web form

Koliko vas je pri izpolnjevanju spletnih obrazcev

where you've been asked to read

moralo razbrati tak popačen niz znakov?

a distorted sequence of characters like this?

Kolikim se zdi to res zoprno?

How many of you found it really annoying?

Prav, krasno. To sem torej izumil jaz.

(Laughter)

OK, outstanding. So I invented that.

(Smeh)

(Laughter)

Bil sem eden od izumiteljev.

Or I was one of the people who did it.

Ta reč se imenuje CAPTCHA.

That thing is called a CAPTCHA.

Njen namen je, da zagotovi, da obrazec izpolnjuje

And it is there to make sure you, the entity filling out the form,

človek in ne kak računalniški program,

are a human and not a computer program

sprogramiran, da nepretrgoma oddaja obrazce.

that was written to submit the form millions of times.

Stvar deluje, ker ljudje,

The reason it works is because humans, at least non-visually-impaired humans,

vsaj tisti, ki niso slabovidni,

brez težav prepoznajo te popačene znake,

have no trouble reading these distorted characters,

računalniški programi pa tega še ne zmorejo tako dobro.

whereas programs can't do it as well yet.

In the case of Ticketmaster,

Na strani Ticketmaster

the reason you have to type these characters

z vpisom teh skrivljenih znakov preprečite

is to prevent scalpers from writing a program

preprodajalcem vstopnic, da bi ustvarili program,

that can buy millions of tickets, two at a time.

ki bi jih lahko po dve naenkrat kupil na tisoče.

CAPTCHAs are used all over the Internet.

CAPTCHE se uporabljajo širom spleta.

And since they're used so often,

Zaradi pogostosti pojavljanja

a lot of times the sequence of random characters shown to the user

je rezultat vrstnega reda znakov, ki se izpiše,

pogosto precej ponesrečen.

is not so fortunate.

So this is an example from the Yahoo registration page.

Tu je primer s strani za prijavo na Yahoo.

The random characters that happened to be shown to the user

Naključni znaki, ki jih je moral uporabnik vpisati,

so W, A, I, T, kar seveda tvori besedo (počakaj).

were W, A, I, T, which, of course, spell a word.

Še bolj zanimivo pa je sporočilo,

But the best part is the message

that the Yahoo help desk got about 20 minutes later.

ki so ga na Yahoojevi podpori uporabnikom prejeli 20 minut kasneje.

[Help! I've been waiting for over 20 minutes and nothing happens.]

Besedilo: "Na pomoč! Čakam že več kot 20 minut, a se ne zgodi nič."

(Laughter)

(Smeh)

This person thought they needed to wait.

Ta uporabnik je torej mislil, da mora čakati.

This, of course, is not as bad as this poor person.

Še slabše jo je odnesla ta oseba. ("Ponovno zaženi")

(Laughter)

(Smeh)

CAPTCHA Project is something that we did at Carnegie Melllon over 10 years ago,

S projektom CAPTCHA smo pri Carnegie Mellon začeli pred dobrim desetletjem

in odtlej je v splošni uporabi.

and it's been used everywhere.

Let me now tell you about a project that we did a few years later,

Želel pa bi vam predstaviti projekt, ki smo ga začeli nekaj let kasneje,

neke vrste nadgradnjo CAPTCHE.

which is sort of the next evolution of CAPTCHA.

Projekt se imenuje reCAPTCHA,

This is a project that we call reCAPTCHA,

z njim smo začeli pri Carnegie Mellon,

which is something that we started here at Carnegie Mellon,

nato pa je prerasel v mlado podjetje.

then we turned it into a start-up company.

Pred letom in pol

And then about a year and a half ago, Google actually acquired this company.

je Google podjetje tudi prevzel.

Naj vam predstavim projekt.

Let me tell you what this project started.

Osnova je bilo spoznanje,

This project started from the following realization:

da se približno 200 milijonov CAPTCH

It turns out that approximately 200 million CAPTCHAs

vsakodnevno vtipka po vsem svetu.

are typed everyday by people around the world.

When I first heard this, I was quite proud of myself.

Ko sem to slišal, sem bil ponosen nase.

Vesel sem bil, da je moje delo imelo tak vpliv.

I thought, look at the impact my research has had.

Nato pa sem se začel počutiti slabo.

But then I started feeling bad.

Vsakič ko vtipkate CAPTCHO,

Here's the thing: each time you type a CAPTCHA,

namreč izgubite 10 sekund svojega časa.

essentially, you waste 10 seconds of your time.

And if you multiply that by 200 million,

Če to pomnožite s 200 milijoni,

you get that humanity is wasting about 500,000 hours every day

vidite, da človeštvo vsak dan zapravi 500.000 ur

typing these annoying CAPTCHAs.

med tipkanjem teh zoprnih CAPTCH.

(Laughter)

Počutil sem se slabo.

So then I started feeling bad.

(Laughter)

(Smeh)

And then I started thinking, of course, we can't just get rid of CAPTCHAs,

A kljub temu jih ne moremo kar prenehati uporabljati,

saj na njih temelji varnost spleta.

because the security of the web depends on them.

Hotel sem najti način, kako bi ta trud uporabnikov

But then I started thinking, can we use this effort

for something that is good for humanity?

spremenili v nekaj dobrega za človeštvo.

So see, here's the thing.

Gre namreč za to.

While you're typing a CAPTCHA, during those 10 seconds,

Ko tipkate CAPTCHO, vaši možgani teh 10 sekund

počnejo nekaj neverjetnega.

your brain is doing something amazing.

Your brain is doing something that computers cannot yet do.

Nekaj, česar računalniki še niso sposobni.

Kako lahko vaš trud teh 10 sekund uporabimo za nekaj koristnega?

So can we get you to do useful work for those 10 seconds?

Z drugimi besedami,

Is there some humongous problem that we cannot yet get computers to solve,

ali obstaja kakšna ogromna težava, ki je računalniki še ne morejo rešiti,

lahko pa jo razdelimo na 10-sekundne delčke,

yet we can split into tiny 10-second chunks

such that each time somebody solves a CAPTCHA,

da lahko vsakdo, ki razbere in vtipka CAPTCHO,

hkrati reši delček te težave?

they solve a little bit of this problem?

Odgovor je "da" in s tem se trenutno ukvarjamo.

And the answer to that is "yes," and this is what we're doing now.

Morda še ne veste, da se dandanes ob vsakem vnosu CAPTCHE

Nowadays, while you're typing a CAPTCHA,

ne le istovetite kot ljudje,

not only are you authenticating yourself as a human,

pač pa tudi pomagate pri digitalizaciji knjig.

but in addition you're helping us to digitize books.

Naj razložim, kako deluje.

Let me explain how this works.

Veliko projektov se ubada z digitalizacijo knjig.

There's a lot of projects trying to digitize books.

Taka sta tudi Google ter Internet Archive.

Google has one. The Internet Archive has one.

Tudi Amazon preko Kindla dela na digitalizaciji knjig.

Amazon, with the Kindle, is trying to digitize books.

To poteka tako,

Basically, the way this works is you start with an old book.

da začnemo s staro knjigo.

You've seen those things, right?

Saj poznate te stvari, kajne? Knjige?

Like a book?

(Smeh)

(Laughter)

So you start with a book and then you scan it.

Vzamemo torej knjigo in jo skeniramo.

To izgleda tako,

Now, scanning a book

is like taking a digital photograph of every page.

kot bi posneli digitalno fotografijo vsake strani v knjigi.

Dobimo torej podobo vsake strani,

It gives you an image for every page.

podobo z besedilom celotne knjige.

This is an image with text for every page of the book.

Pri naslednjem koraku mora biti računalnik zmožen

The next step in the process is that the computer needs to be able

dešifrirati vse besede na tej podobi.

to decipher the words in this image.

Pri tem mu pomaga tehnologija OCR

That's using a technology called OCR, for optical character recognition,

za optično prepoznavanje znakov,

which takes a picture of text

ki posname sliko besedila

and tries to figure out what text is in there.

in ga poskuša razbrati.

Now, the problem is that OCR is not perfect.

A tehnologija OCR ni popolna.

Še posebej pri starejših knjigah

Especially for older books

where the ink has faded and the pages have turned yellow,

z zbledelim črnilom in porumenelimi stranmi

OCR cannot recognize a lot of the words.

je mnogo besed, ki jih OCR ne prepozna.

For things that were written more than 50 years ago,

Pri besedilih, starejših od 50 let,

računalnik ne razbere skoraj 30 odstotkov besed.

the computer cannot recognize about 30 percent of the words.

Naše delo je,

So now we're taking all of the words that the computer cannot recognize

da zberemo vse besede, ki jih računalnik ne prepozna,

and we're getting people to read them for us

in jih posredujemo ljudem, da jih preberejo

while they're typing a CAPTCHA on the Internet.

in vtipkajo preko CAPTCHE na spletu.

Ko boste naslednjič vnašali CAPTCHO,

So the next time you type a CAPTCHA, these words that you're typing

bo to beseda iz knjige v postopku digitalizacije,

are actually words from books that are being digitized

that the computer could not recognize.

ki je računalnik ni uspel razbrati.

The reason we have two words nowadays instead of one

Dandanes pa CAPTCHE tvorita dve besedi in ne ena,

in ena od njiju je beseda,

is because one of the words

is a word that the system just got out of a book,

ki jo je sistem potegnil iz knjige,

je ni prepoznal in jo posredoval vam.

it didn't know what it was and it's going to present it to you.

A ker je ne pozna, ne ve, ali je vaš vnos pravi.

But since it doesn't know the answer, it cannot grade it.

Zato vam damo še drugo besedo,

So we give you another word,

for which the system does know the answer.

za katero sistem pozna odgovor.

We don't tell you which one's which and we say, please type both.

Vi ne veste, katera je katera, in vnesti je potrebno obe.

In če prav vnesete besedo,

And if you type the correct word

na katero sistem že pozna pravilen odgovor,

for the one for which the system knows the answer,

sklepa, da ste človek,

it assumes you are human

and it also gets some confidence that you typed the other word correctly.

in vam zaupa, da boste pravilno vnesli tudi drugo.

Če ta postopek ponovimo pri 10 različnih ljudeh

And if we repeat this process to 10 different people

in vsi enako vnesejo pomen nove besede,

and they agree on what the new word is,

s tem dobimo novo pravilno digitalizirano besedo.

then we get one more word digitized accurately.

Tako torej deluje ta sistem.

So this is how the system works.

Odkar smo s tem začeli pred tremi ali štirimi leti,

And since we released it about three or four years ago,

je mnogo spletnih strani zamenjalo

a lot of websites have started switching from the old CAPTCHA,

stare CAPTCHE, ki so zapravljale čas ljudi,

where people wasted their time,

za nove CAPTCHE, s katerimi digitaliziramo knjige.

to the new CAPTCHA where people are helping to digitize books.

Tak je tudi Ticketmaster.

So every time you buy tickets on Ticketmaster,

Vsakič ko kupite vstopnice na Ticketmastru, pomagate pri digitalizaciji knjig.

you help to digitize a book.

Vsakič ko na Facebooku dodate ali dregnete prijatelja,

Facebook: Every time you add a friend or poke somebody,

digitalizirate knjige.

you help to digitize a book.

Twitter in še 350.000 drugih spletnih strani uporablja reCAPTCHE.

Twitter and about 350,000 other sites are all using reCAPTCHA.

Število strani z reCAPTCHO je že tako veliko,

And the number of sites that are using reCAPTCHA is so high

da je ogromno tudi število besed, ki se dnevno digitalizirajo,

that the number of words we're digitizing per day is really large.

in sicer okrog 100 milijonov na dan,

It's about 100 million a day,

kar na letni ravni znese okrog 2,5 milijona knjig.

which is the equivalent of about two and a half million books a year.

Do tega pride le z eno besedo naenkrat,

And this is all being done one word at a time

z vnašanjem CAPTCH na spletnih straneh.

by just people typing CAPTCHAs on the Internet.

(Aplavz)

(Applause)

Now, of course,

Seveda pa se s pojavom

since we're doing so many words per day,

tolikih besed v enem dnevu

dogajajo tudi čudne stvari.

funny things can happen.

This is especially true because now we're giving people

Še posebej zdaj, ko uporabnikom posredujemo

dve naključno izbrani angleški besedi hkrati.

two randomly chosen English words next to each other.

Včasih so kombinacije zelo zabavne.

So funny things can happen.

Lep primer je ta beseda.

For example, we presented this word.

Pomeni "kristjani" in do tu še ni posebnosti.

It's the word "Christians"; there's nothing wrong with it.

Če pa jo postavite zraven povsem naključno izbrane besede,

But if you present it along with another randomly chosen word,

je lahko rezultat ponesrečen.

bad things can happen.

Nastane lahko to. (Besedilo: slabi kristjani)

So we get this.

[bad Christians]

Še slabše postane, ko ugotovimo, da se je bila ta kombinacija pojavila

But it's even worse, because the website where we showed this

na strani Ambasade božjega kraljestva.

actually happened to be called The Embassy of the Kingdom of God.

(Smeh)

(Laughter)

Ups.

Oops.

(Laughter)

(Smeh)

Tu je še en tak primer.

Here's another really bad one.

JohnEdwards.com

JohnEdwards.com

[Damn liberal]

(Besedilo: Prekleti liberalec)

(Laughter)

(Smeh)

Na ta način dnevno žalimo leve in desne.

So we keep on insulting people left and right everyday.

Seveda pa ne gre le za žalitve.

Of course, we're not just insulting people.

Ker gre za kombinacije dveh naključno izbranih besed,

Here's the thing. Since we're presenting two randomly chosen words,

se dogajajo zanimive stvari.

interesting things can happen.

So this actually has given rise to a really big Internet meme

To je spodbudilo nastanek

obsežnega spletnega fenomena,

that tens of thousands of people have participated in,

pri katerem sodeluje na desettisoče ljudi,

imenovanega CAPTCHA umetnost.

which is called CAPTCHA art.

I'm sure some of you have heard about it.

Gotovo je kdo od vas že slišal zanjo.

Here's how it works.

Deluje pa tako.

Imagine you're using the Internet and you see a CAPTCHA

Zamislite si, da med uporabo spleta naletite na CAPTCHO,

that you think is somewhat peculiar,

ki se vam zdi zanimiva,

like this CAPTCHA.

kot je na primer ta. (Besedilo: nevidni opekač)

[invisible toaster]

What you're supposed to do is you take a screenshot of it.

Nato naredite posnetek zaslona, ki jo prikazuje.

Seveda sledi vnos CAPTCHE,

Then of course, you fill out the CAPTCHA because you help us digitize a book.

saj pomagate digitalizirati knjigo.

Najprej torej naredite posnetek zaslona,

But first you take a screenshot

and then you draw something that is related to it.

nato pa poskušate narisati njen pomen.

(Smeh)

(Laughter)

That's how it works.

Tako to poteka.

(Laughter)

There are tens of thousands of these.

Obstaja na desettisoče takih podob.

Some of them are very cute.

Nekatere so zelo ljubke. (Besedilo: stisnil sem)

[clenched it]

(Smeh)

(Laughter)

Nekatere so bolj zabavne.

Some of them are funnier.

(Besedilo: zadeti ustanovitelji)

[stoned Founders]

(Laughter)

(Smeh)

And some of them, like paleontological shvisle ...

Ali pa ta,

paleontološki shvisle,

(Laughter)

kjer nastopa Snoop Dogg.

they contain Snoop Dogg.

(Laughter)

(Smeh)

OK, so this is my favorite number of reCAPTCHA.

Ta številka mi je pri reCAPTCHI najbolj všeč

in me pri celotnem projektu najbolj navdušuje.

So this is the favorite thing that I like about this whole project.

To je število ljudi,

This is the number of distinct people

ki so nam preko reCAPTCHE pomagali digitalizirati vsaj eno besedo iz knjige,

that have helped us digitize at least one word out of a book through reCAPTCHA:

750 milijonov ljudi,

750 million, a little over 10 percent of the world's population,

dobrih 10 odstotkov svetovnega prebivalstva,

has helped us digitize human knowledge.

nam je pomagalo digitalizirati človeško znanje.

And it is numbers like these that motivate my research agenda.

Številke, kot je ta, spodbujajo moje raziskovalne načrte.

So the question that motivates my research is the following:

Vprašanje, ki vodi mojo raziskavo, je sledeče:

pri velikopoteznih dosežkih človeštva,

If you look at humanity's large-scale achievements,

resnično velikih stvareh,

these really big things

that humanity has gotten together and done historically --

ki so v zgodovini združevale človeštvo,

like, for example, building the pyramids of Egypt

kot je na primer gradnja piramid v Egiptu,

or the Panama Canal

Panamskega prekopa,

or putting a man on the Moon --

ali človeška odprava na Luno,

there is a curious fact about them,

je zanimivo, da je pri vseh sodelovalo

and it is that they were all done with about the same number of people.

približno enako število ljudi.

Zanimivo je, da so vse združevale okrog sto tisoč ljudi.

It's weird; they were all done with about 100,000 people.

Razlog za to je, da je bila v času pred spletom

And the reason for that is because, before the Internet,

uskladitev več kot sto tisoč ljudi,

coordinating more than 100,000 people,

kaj šele plač za njihovo delo, praktično nemogoča.

let alone paying them, was essentially impossible.

Z uporabo spleta pa lahko izvedemo projekte, kot je ta,

But now with the Internet, I've just shown you a project

ki združuje 750 milijonov ljudi

where we've gotten 750 million people to help us digitize human knowledge.

v postopku digitalizacije človeškega znanja.

Pri mojem delu me zanima,

So the question that motivates my research is,

če lahko sto tisoč ljudi pošlje človeka na Luno,

if we can put a man on the Moon with 100,000,

what can we do with 100 million?

kaj šele lahko stori sto milijonov?

So based on this question,

Na tem vprašanju temelji še mnogo

we've had a lot of different projects that we've been working on.

drugih projektov, s katerimi se ukvarjamo.

Predstavil vam bom projekt, ki me najbolj navdušuje.

Let me tell you about one that I'm most excited about.

This is something that we've been semiquietly working on

Na njem smo na pol potiho delali

zadnje leto in pol.

for the last year and a half or so.

Nismo ga še zagnali. Imenuje se Duolingo.

It hasn't yet been launched. It's called Duolingo.

Ker še ni zunaj, pšššt!

Since it hasn't been launched, shhh!

(Smeh)

(Laughter)

Yeah, I can trust you'll do that.

Kot da vam lahko zaupam.

To je ta projekt. Začel pa se je tako.

So this is the project. Here's how it started.

Zastavil sem vprašanje svojemu študentu,

It started with me posing a question to my graduate student, Severin Hacker.

Severinu Hackerju.

To je Severin Hacker.

OK, that's Severin Hacker.

Zastavil sem mu torej vprašanje.

So I posed the question to my graduate student.

Mimogrede, prav ste slišali,

By the way, you did hear me correctly; his last name is Hacker.

piše se Hacker (heker).

(Laughter)

Vprašanje se je glasilo:

So I posed this question to him: How can we get 100 million people

kako pripraviti sto milijonov ljudi,

translating the web into every major language for free?

da prevedejo spletne vsebine v vse velike jezike, in to brezplačno?

Pri tem vprašanju je treba povedati veliko stvari.

There's a lot of things to say about this question.

Prvič, prevajanje spleta.

First of all, translating the web.

Trenutno je splet razdeljen na veliko jezikov.

Right now, the web is partitioned into multiple languages.

Velik del je v angleščini.

A large fraction of it is in English.

A če angleščine ne znate, vam ne koristi.

If you don't know English, you can't access it.

Tu je še velik del spleta v drugih jezikih,

But there's large fractions in other different languages,

a če jih ne poznate, vam ne koristi.

and if you don't know them, you can't access it.

Želel bi torej prevesti cel ali vsaj večino spleta

So I would like to translate all of the web,

or at least most of it, into every major language.

v vsak veliki jezik.

To torej želim doseči.

That's what I would like to do.

Now, some of you may say, why can't we use computers to translate?

Nekateri boste rekli, zakaj ne morejo tega storiti računalniki?

Zakaj ne moremo uporabiti strojnih prevodov?

Machine translation is starting to translate

Strojno prevajanje dandanes že prevaja določene stavke.

some sentences here and there.

Zakaj ne bi mogli strojno prevesti celotnega spleta?

Why can't we use it to translate the web?

Težava je, da je tak način prevajanja še zelo pomanjkljiv

The problem with that is it's not yet good enough

in ne bo zadovoljiv še naslednjih 15 do 20 let.

and it probably won't be for the next 15 to 20 years.

Dela veliko napak.

It makes a lot of mistakes. Even when it doesn't,

In tudi če ne naredi napake,

since it makes so many mistakes, you don't know whether to trust it or not.

mu zaradi prejšnjih napak ne zaupamo več popolnoma.

Naj vam pokažem primer

So let me show you an example

of something that was translated with a machine.

strojnega prevoda.

Gre za objavo na forumu.

Actually, it was a forum post.

It was somebody who was trying to ask a question about JavaScript.

Nekdo je želel postaviti vprašanje o JavaScriptu.

It was translated from Japanese into English.

Gre za prevod iz japonščine v angleščino.

So I'll just let you read.

Kar preberite.

This person starts apologizing

Ta oseba se opraviči,

for the fact that it's translated with a computer.

da se je poslužila strojnega prevoda.

So the next sentence is going to be the preamble to the question.

Naslednji stavek bo uvod v vprašanje.

Tu nekaj razlaga.

So he's just explaining something.

Ne pozabite, vprašanje zadeva JavaScript.

Remember, it's a question about JavaScript.

[At often, the goat-time install a error is vomit.]

(Besedilo: Pogosto kozji čas namestitev napake je bruhanje.)

(Laughter)

(Smeh)

Then comes the first part of the question.

Nato pride prvi del vprašanja.

[How many times like the wind, a pole, and the dragon?]

(Besedilo: Kolikokrat kot veter, drog in zmaj?)

(Laughter)

(Smeh)

Sledi pa meni najljubši del vprašanja.

Then comes my favorite part of the question.

[This insult to father's stones?]

(Besedilo: Ta žalitev očetovih kamnov?)

(Laughter)

(Smeh)

Zadnji del pa je najboljši.

And then comes the ending,

which is my favorite part of the whole thing.

(Besedilo: Prosim oprosti za tvojo neumnost. Tam je veliko hvala.)

[Please apologize for your stupidity. There are a many thank you.]

(Laughter)

(Smeh)

OK, so computer translation, not yet good enough.

Strojno prevajanje torej še ni zadovoljivo.

Vrnimo se k vprašanju.

So back to the question.

Potrebujemo ljudi za prevod celotnega spleta.

So we need people to translate the whole web.

Morda se sprašujete,

So now the next question you may have is,

zakaj preprosto ne moremo najeti ljudi za to?

well, why can't we just pay people to do this?

Lahko bi plačali poklicne prevajalce, da bi prevedli splet.

We could pay professional translators to translate the whole web.

Lahko bi.

We could do that.

Unfortunately, it would be extremely expensive.

Na žalost bi bilo to nepredstavljivo drago.

For example, translating a tiny fraction of the whole web, Wikipedia,

Vzemimo prevod majcenega delčka celotnega spleta, Wikipedije,

v drug jezik, španščino.

into one other language, Spanish.

OK? Wikipedia exists in Spanish,

Wikipedija obstaja v španščini,

but it's very small compared to the size of English.

a je njen obseg majhen v primerjavi z angleško različico.

Predstavlja okrog 20 odstotkov angleške.

It's about 20 percent of the size of English.

Če bi hoteli prevesti ostalih 80 odstotkov v španščino,

If we wanted to translate the other 80 percent into Spanish,

bi to stalo najmanj 50 milijonov dolarjev,

it would cost at least 50 million dollars --

in še to v najbolj izkoriščevalski in zatiralski državi.

and this is even at the most exploited, outsourcing country out there.

Bilo bi torej izjemno drago.

So it would be very expensive.

Namesto tega želimo dobiti 100 milijonov ljudi,

So what we want to do is, we want to get 100 million people

ki bi splet prevajali v vsak večji jezik

translating the web into every major language for free.

brezplačno.

Pri takem cilju se moramo zelo hitro

If this is what you want to do, you quickly realize

soočiti z dvema velikima preprekama,

you're going to run into two big hurdles, two big obstacles.

dvema velikima ovirama.

The first one is a lack of bilinguals.

Prva je pomanjkanje dvojezičnih ljudi.

So I don't even know

Nisem prepričan,

if there exists 100 million people out there using the web

da obstaja 100 milijonov uporabnikov spleta

who are bilingual enough to help us translate.

z zadostnim dvojezičnim znanjem, da lahko prevajajo.

To je velika težava.

That's a big problem.

The other problem you're going to run into is a lack of motivation.

Druga pa je pomanjkanje spodbude.

Kako bomo spodbudili ljudi,

How are we going to motivate people to actually translate the web for free?

da brezplačno prevajajo splet?

Običajno tako delo zahteva plačilo.

Normally, you have to pay people to do this.

So how are we going to motivate them to do it for free?

Kako naj jih torej spodbudimo k neprofitnemu delu?

Ko smo začeli razmišljati o tem, sta nas ti težavi ovirali.

When we were starting to think about this, we were blocked by these two things.

Nato pa smo se domislili načina,

But then we realized, there's a way

kako ju odpraviti z le eno rešitvijo.

to solve both these problems with the same solution.

Načinom, kako ubiti dve muhi na en mah.

To kill two birds with one stone.

Prevajanje smo spremenili v nekaj,

And that is to transform language translation

into something that millions of people want to do

kar milijoni ljudi želijo delati,

and that also helps with the problem of lack of bilinguals,

in kar hkrati odpravlja pomanjkanje dvojezičnih ljudi,

and that is language education.

in to je učenje jezika.

So it turns out that today,

Izkazalo se je,

there are over 1.2 billion people learning a foreign language.

da se danes več kot 1,2 milijarde ljudi uči tujega jezika.

Ljudje se res želijo naučiti novih jezikov.

People really want to learn a foreign language.

In ne le zato, ker jih v to silijo v šoli.

And it's not just because they're being forced to do so in school.

Le v ZDA, na primer,

In the US alone, there are over five million people

je pet milijonov ljudi plačalo več kot 500 dolarjev

who have paid over $500 for software to learn a new language.

za nakup programa za učenje tujega jezika.

Ljudje se resnično želijo naučiti novega jezika.

So people really want to learn a new language.

Zadnje leto in pol smo torej delali na novi spletni strani,

So what we've been working on for the last year and a half

imenovani Duolingo,

is a new website -- it's called Duolingo --

vodilna zamisel pa je brezplačno učenje novega jezika,

where the basic idea is people learn a new language for free

while simultaneously translating the web.

ob katerem se hkrati prevaja splet.

Uporabniki se učijo s prakso.

And so basically, they're learning by doing.

Deluje tako,

So the way this works

is whenever you're a just a beginner, we give you very simple sentences.

da tudi začetniki dobijo izjemno preproste stavke,

ki jih je na spletu seveda obilo.

There's a lot of very simple sentences on the web.

Dobite torej zelo preproste stavke

We give you very simple sentences along with what each word means.

skupaj s pomeni za vsako besedo.

And as you translate them

Ko jih prevedete in preverite prevod drugih,

and as you see how other people translate them,

se s tem učite novega jezika.

you start learning the language.

Ko napredujete,

And as you get more advanced,

we give you more complex sentences to translate.

dobite v prevod bolj zapletene stavke.

But at all times, you're learning by doing.

V vsakem primeru pa se učite s prakso.

Now, the crazy thing about this method is that it actually really works.

Najbolj noro pri tej metodi pa je,

da dejansko deluje.

People are really learning a language.

Prvič, ljudje se resnično učijo jezika.

We're mostly done building it and now we're testing it.

Program smo že postavili, sedaj ga testiramo.

Ljudje se z njim resnično lahko naučijo jezika.

People really can learn a language with it.

In uspeh je primerljiv z uporabniki vodilnih programov za učenje.

And they learn it about as well as the leading language learning software.

Ljudje resnično osvojijo nov jezik.

So people really do learn a language.

A ne samo osvojijo,

And not only do they learn it as well, but actually it's more interesting.

pač pa postane še bolj zanimivo.

Pri Duolingu se namreč učimo na podlagi resničnih vsebin.

Because with Duolingo, people are learning with real content.

V nasprotju z umetno ustvarjenimi stavki

As opposed to learning with made-up sentences,

se tu ljudje učijo na resničnih stavkih, kar je bistveno bolj zanimivo.

people are learning with real content, which is inherently interesting.

Jezika se torej uspešno naučijo.

So people really do learn a language.

A morda je bolj presenetljivo,

But perhaps more surprisingly,

da so prevodi, ki jih dobimo od uporabnikov,

the translations that we get from people using the site,

even though they're just beginners,

čeprav gre za začetnike,

the translations that we get

po kakovosti primerljivi s prevodi poklicnih prevajalcev,

are as accurate as those of professional language translators,

kar je res presenetljivo.

which is very surprising.

Naj vam pokažem primer.

So let me show you one example.

This is a sentence that was translated from German into English.

Ta stavek je bil preveden iz nemščine v angleščino.

Zgornji odstavek je v nemščini.

The top is the German. The middle is an English translation

V sredini je angleški prevod

that was done by a professional translator

poklicnega prevajalca za angleščino,

who we paid 20 cents a word for this translation.

ki je računal 20 centov na besedo.

Spodaj pa je prevod uporabnikov Duolinga,

And the bottom is a translation by users of Duolingo,

od katerih nihče ni znal nemško,

none of whom knew any German before they started using the site.

preden so začeli uporabljati to stran.

If you can see, it's pretty much perfect.

Kot vidite, je praktično popoln.

Of course, we play a trick here

Mi pa z zvijačo naredimo prevode

to make the translations as good as professional language translators.

primerljive s prevodi poklicnih prevajalcev.

Prevode več začetnikov združimo

We combine the translations of multiple beginners

in tako dosežemo kakovost enega poklicnega prevajalca.

to get the quality of a single professional translator.

Now, even though we're combining the translations,

In čeprav združujemo prevode,

the site actually can translate pretty fast.

stran prevaja precej hitro.

Naj vam pokažem naše ocene,

So let me show you,

this is our estimates of how fast we could translate Wikipedia

kako hitro bi lahko prevedli Wikipedijo

iz angleščine v španščino.

from English into Spanish.

Remember, this is 50 million dollars' worth of value.

Ne pozabite, da govorimo o vrednosti 50 milijonov dolarjev.

So if we wanted to translate Wikipedia into Spanish,

Če bi torej želeli prevesti Wikipedijo v španščino,

bi v petih tednih to dosegli s 100.000 aktivnih uporabnikov,

we could do it in five weeks with 100,000 active users.

oz. v 80 urah z milijonom aktivnih uporabnikov.

And we could do it in about 80 hours with a million active users.

Ker so projekti naše ekipe doslej zbrali že milijone uporabnikov,

Since all the projects my group has worked on so far

have gotten millions of users,

upamo, da nam bo ta projekt omogočil

we're hopeful that we'll be able to translate extremely fast.

izjemno hitro prevajanje.

Pri Duolingu me najbolj navdušuje to,

Now, the thing that I'm most excited about with Duolingo

da predstavlja pravičen poslovni način učenja jezikov.

is I think this provides a fair business model for language education.

Gre namreč za to:

So here's the thing:

sedanje poslovne prakse učenja jezikov delujejo tako,

The current business model for language education

da učenec plača za uporabo,

is the student pays,

and in particular, the student pays Rosetta Stone 500 dollars.

in sicer 500 dolarjev za program Rosetta Stone.

(Smeh)

(Laughter)

Taka je trenutna poslovna praksa.

That's the current business model.

Pri tem pa je težava,

The problem with this business model

da 95 odstotkov svetovnega prebivalstva nima 500 dolarjev.

is that 95 percent of the world's population doesn't have 500 dollars.

Je torej izjemno nepravična do revnejših

So it's extremely unfair towards the poor.

in namenjena bogatim.

This is totally biased towards the rich.

Duolingo pa je zasnovan tako,

Now, see, in Duolingo,

because while you learn, you're actually creating value,

da med učenjem pravzaprav

ne koristite le sebi, pač pa tudi prevajate,

you're translating stuff --

which, for example, we could charge somebody for translations,

za kar bi nekdo drug računal.

so this is how we could monetize this.

S tem ovrednotimo vaše učenje.

Since people are creating value while they're learning,

Ker ima rezultat učenja za nas vrednost,

jim ne izstavimo računa, saj plačajo s svojim časom.

they don't have to pay with their money, they pay with their time.

Čudovito pri tem je, da unovčijo čas,

But the magical thing here

is that is time that would have had to have been spent anyways

ki bi ga itak porabili

za učenje jezika.

learning the language.

So the nice thing about Duolingo

Prednost Duolinga je, da predstavlja pravičen posloven način,

is, I think, it provides a fair business model --

ki ne zapostavlja revnejših ljudi.

one that doesn't discriminate against poor people.

To je ta stran. Hvala vam.

So here's the site. Thank you.

(Aplavz)

(Applause)

To je torej naslov strani.

We haven't yet launched,

Nismo je še zagnali,

but if you go there, you can sign up to be part of our private beta,

a kljub temu se lahko vpišete in se pridružite naši razvojni različici,

which is probably going to start in three or four weeks.

ki naj bi se zagnala v treh ali štirih tednih.

Duolingo torej še ni na voljo.

We haven't yet launched it.

By the way, I'm the one talking here,

Mimogrede, o njem govorim le jaz,

but Duolingo is the work of a really awesome team,

Duolingo pa je delo izjemne ekipe, katere del je z nami.

some of whom are here. So thank you.

Hvala še enkrat.

(Applause)

(Aplavz)