Sebastian Thrun and Chris Anderson: What AI is -- and isn't
Sebastian Thrun e Chris Anderson: A nova geração de computadores está programando a si mesma
Sebastian Thrun is a passionate technologist who is constantly looking for new opportunities to make the world better for all of us. Full bioChris Anderson - TED Curator
After a long career in journalism and publishing, Chris Anderson became the curator of the TED Conference in 2002 and has developed it as a platform for identifying and disseminating ideas worth spreading. Full bio
Double-click the English transcript below to play the video.
o que é aprendizado de máquina,
what machine learning is,
and also of the concern
com a inteligência artificial.
intelligence and machine learning
e o aprendizado de máquina
só começaram recentemente.
in its past until recently.
of computing and datasets
de computação e de dados
as máquinas inteligentes.
say, your phone,
digamos, um celular,
very long kitchen recipe,
culinária bem longa,
turn down the temperature.
diminua a temperatura.
the temperature."
aumente a temperatura".
has 12 million lines of code.
12 milhões de linhas de programação.
de linhas de programação.
can cause your computer to crash.
pode causar uma pane no computador.
makes so much money.
ganham tanto dinheiro.
can find their own rules.
podem descobrir suas próprias regras.
deciphering, step by step,
decifrar, passo a passo,
the computer examples
exemplos ao computador
which recently was won by Google.
que recentemente venceu para o Google.
you would really write down all the rules,
escrevem-se todas as regras,
suas próprias regras
residing Go champion.
campeão mundial de Go.
the software engineer
a carga do engenheiro de software
where this has become really possible --
em que isso se tornou possível...
foi sobre aprendizado de máquina,
was about machine learning.
não leiam, pois foi escrita há 20 anos...
insignificant, don't read it,
eram como o cérebro de minhoca.
were as big as a cockroach brain.
o suficiente para simular
to really emulate
take advantage of the fact
de trabalhar com muito mais dados
much more data than people can.
mais de um milhão de jogos.
more than a million games.
conseguiria estudar esse número.
study a million games.
a hundred billion web pages.
de bilhões de páginas da rede.
a hundred billion web pages.
estudar bilhões de páginas.
o computador pode encontrar regras
the computer can find rules
"Se ele fizer isso, eu faço aquilo",
to, "If he does that, I will do that,"
um padrão de vitória;
looks like a winning pattern,
um padrão de vitória".
a winning pattern."
na criação de filhos.
how you raise children.
dando regras para cada contingência
giving kids a rule for every contingency
terem um programa enorme.
and they have this big program.
levam tapas ou surras,
they get slapped or spanked,
a good grade in school,
uma boa nota na escola,
o que de repente facilitou a programação.
so much easier all of a sudden.
We just give them lots of data.
damos a eles montes de dados.
para o desenvolvimento espetacular
to the spectacular improvement
o que está acontecendo aqui?
que temos na Udacity
a uma filial chamada Voyage.
into a spin-off called Voyage.
called deep learning
para treinar um carro a dirigir sozinho,
from Mountain View, California,
Califórnia, para São Francisco
and 133 traffic lights.
a equipe do carro autônomo do Google.
the Google self-driving car team.
the world's best software engineers
os melhores engenheiros de software
into the computer brain,
no cérebro do computador
that often surpasses human agility.
que sempre ultrapassa a agilidade humana.
about 33 miles, an hour and a half.
cerca de 53 km em uma hora e meia.
of this program on the left,
grande do programa à esquerda,
the computer sees as trucks and cars
vê, caminhões, carros,
e assim por diante.
a entrada de dados mais expressiva,
image, which is the main input here,
outros carros, semáforos.
other cars, traffic lights.
to do distance estimation.
para calcular a distância,
in these kind of systems.
and so on depicted by the laser.
como árvores e coisas assim.
is centering on the camera image now.
está na imagem da câmera.
de precisão como radares e lasers
sensors like radars and lasers
on the left thing, what is that?
à esquerda, o que é?
for your adaptive cruise control,
para o cruzeiro regulável,
how to regulate velocity
como regular a velocidade
the cars in front of you are.
dos carros da frente.
got an example, I think,
learning part takes place.
que propusemos aos alunos da Udacity,
a challenge to Udacity students
a self-driving car Nanodegree.
carro autônomo "Nanodegree".
how to steer this car?"
como dirigir esse carro?"
to get the steering right.
quase impossível dirigir certo.
"It's a deep learning competition,
de aprendizado profundo,
like Google or Facebook,
como o Google ou o Facebook,
at least six months of work.
pelo menos seis meses de trabalho.
que 48 horas estava ótimo.
100 submissions from students,
cerca de 100 respostas dos alunos,
perfeitamente.
drive on this imagery,
ele dirigiu melhor do que eu.
é uma coisa mágica.
to a computer now,
suficientes para compreender esses dados,
to comprehend the data,
of powerful applications
de aplicações poderosas
the other day about cancer.
me falando da oncologia.
CA: This is cool.
CA: Isso é legal.
into what's happening
sobre o que está acontecendo
400,000 dollars a year,
US$ 400 mil por ano,
to be a good dermatologist.
para se formar um bom dermatologista.
the machine learning version of it.
aprendizado de máquina disso.
for these machine learning algorithms.
a algoritmos de aprendizado de máquina.
by a Facebook Fellow called Yann LeCun,
um pesquisador do Facebook,
as the human brain.
but it emulates the same thing.
mas simula a mesma coisa.
the visual input and extracts edges
os dados visuais e extrai bordas,
more complicated edges
really complicated concepts.
conceitos realmente complicados.
cat faces and dog faces
caras de gatos e de cachorros
at Stanford has shown is that
em Stanford mostrou que,
of skin conditions,
de condições da pele,
that this is the case,
that we presented to our network
a 25 dermatologistas no nível de Stanford
Stanford-level dermatologists,
the performance classification accuracy
do que a dos dermatologistas humanos.
That's a moving piece.
quinta. Foi tocante.
in "Nature" earlier this year
dermatologists images
e ao nosso programa de computador
we had the correct classification.
de sua correta classificação.
por um de nossos colaboradores.
by one of our collaborators.
one of the three best, apparently,
aparentemente um dos três melhores,
"Não é câncer de pele".
"This is not skin cancer."
"Bem, deixe-me checar no aplicativo".
a second moment, where he said,
e usou nosso aplicativo,
and ran our piece of software,
the iPhone a little bit more than myself,"
um pouco mais do que em mim mesmo",
to get it biopsied.
that we actually found,
que realmente encontramos,
would have gone unclassified,
não teria sido diagnosticado
for an app like this right now,
para um aplicativo como este agora,
um aplicativo que permita autoexame?
making an app that allows self-checking?
inundada de aplicativos sobre câncer
about cancer apps,
10, 15, 20 melanomas removed,
10, 15, 20 melanomas removidos,
might be overlooked, like this one,
passado despercebido, como este,
e pedidos de palestras, coisas assim.
these days, I guess.
de mais testes.
and impress a TED audience.
chamativo e impressionar a plateia do TED.
something out that's ethical.
the assistance of a doctor
e eu me sentiria muito mal com isso.
e, caso nossos dados se sustentem,
and our data holds up,
pegar esse tipo de tecnologia,
to take this kind of technology
e levá-la ao mundo inteiro,
nunca estiveram antes.
doctors never, ever set foot.
se for como você está dizendo,
exército de alunos da Udacity
with this army of Udacity students,
forma diferente de aprendizado de máquina
a different form of machine learning
with a form of crowd wisdom.
com uma forma de sabedoria coletiva.
that could actually outperform
o que uma empresa pode fazer,
even a vast company?
instâncias que me surpreendem,
instances that blow my mind,
is these competitions that we run.
às competições que fazemos.
a self-driving car
to San Francisco on surface streets.
a São Francisco pelas ruas,
after seven years of Google work,
de sete anos do Google,
and three months to do this.
dois engenheiros em três meses,
an army of students
que participam das competições.
colaboração coletiva.
who use crowdsourcing.
coletiva para hospedagem.
where people do bug-finding crowdsourcing
fazendo isso para descobrir falhas
na colaboração coletiva.
in crowdsourcing.
this car in three months,
esse carro em três meses,
who are never hired,
que nunca foram contratadas,
and I don't even know.
e eu nem as conheço.
maybe 9,000 answers.
talvez 9 mil respostas.
which is maybe not the best thing to do.
o que talvez não seja o melhor a fazer.
of their education, too, which is nice.
de sua formação, o que é legal.
to produce amazing deep learning results.
resultados no aprendizado profundo.
and great machine learning is amazing.
e bom aprendizado de máquina é incrível.
no primeiro dia [do TED2017]
the first day [of TED2017]
turned out to be two amateur chess players
no xadrez foram dois jogadores amadores,
mediocre-to-good, computer programs,
de medíocres a bons,
um grande jogador de xadrez,
with one great chess player,
you're talking about a much richer version
versão muito mais rica da mesma ideia.
the fantastic panels yesterday morning,
aos fantásticos painéis ontem de manhã,
that we sometimes confuse
é que às vezes confundimos
with this kind of overlord threat,
com esse tipo de ameaça de dominação
consciousness, right?
uma consciência, certo?
is for my AI to have consciousness.
é que minha IA tenha consciência.
with the dishwasher
apaixonada pela lava-louça,
não está quente porque não fui legal.
and I don't want them.
esses produtos, e não os quero.
an augmentation of people.
uma extensão das pessoas.
para nos fazer mais fortes.
of human smarts and machine smarts
humana com a das máquinas
is as old as machines are.
é tão antigo quanto as próprias máquinas.
place because it made steam engines
por causa do motor a vapor
sozinhos, que não nos substituíram,
that couldn't farm by itself,
it made us stronger.
vai nos fazer muito mais fortes
will make us much, much stronger
of this for some people,
com a parte que assusta as pessoas,
scary for people is when you have
rewrite its own code,
primeiro, reescrever seu próprio código
multiple copies of itself,
cópias de si mesmo,
de código, possivelmente aleatoriamente,
if a goal is achieved and improved.
foi atingido ou aperfeiçoado.
on an intelligence test.
se sair melhor num teste de inteligência.
that's moderately good at that,
que seja mediano no teste,
some sort of runaway effect
as coisas fujam ao controle,
on Thursday evening,
on Friday morning,
na sexta de manhã
dos computadores e tudo mais,
of computers and so forth,
what I heard you say.
we had exactly this thing:
tínhamos exatamente isto:
the game against itself
si próprio para aprender novas regras.
é uma reescrita das regras.
is a rewriting of the rules.
absolutely no concern
nenhuma preocupação
these are all very single-domain things.
domínios bem específicos.
that seemed nearly capable
que parecia quase capaz
para ingressar numa universidade.
and understand in the sense that we can,
no sentido que podemos,
patterns of meaning.
padrões de significado.
as this broadens out,
à medida que isso se ampliar,
kind of runaway effect?
é aí que traço uma linha.
I draw the line, honestly.
não quero minimizar isso,
I don't want to downplay it --
que me preocupe atualmente,
the thing that's on my mind these days,
revolução é outra coisa.
is something else.
até a presente data
to the present date
com base numa única ideia,
is because of massive numbers of Go plays,
é devido ao número de partidas jogadas,
or fly a plane.
um carro ou pilotar um avião.
or the Udacity self-driving car
funciona com enormes quantidades de dados,
and it can't do anything else.
nem mesmo controlar uma moto.
domain-specific function,
para o aplicativo do câncer.
on this thing called "general AI,"
nessa coisa chamada "IA geral",
"Hey, invent for me special relativity
uma teoria da relatividade especial
and I want to acknowledge them.
e estou ciente delas,
"What if we can take anything repetitive
"E se pudesse pegar algo repetitivo
100 times as efficient?"
we all worked in agriculture
todos trabalhávamos na agricultura,
doing repetitive things,
que fazem coisas repetitivas,
of being able to take an AI,
de sermos capazes de pegar uma IA,
as effective in these repetitive things.
mais eficazes nessas coisas repetitivas.
a little terrifying to some people,
aterrorizar algumas pessoas,
possa fazer essa coisa repetitiva
can do this repetitive thing
is the thing that's talked about
algo tão falado hoje em dia,
de empregos desaparecem,
glorious aspects of what's possible.
mais gloriosos do que é possível.
and it's a big issue,
by several guest speakers.
por diversos palestrantes.
optimistic person,
uma pessoa positiva, otimista,
algo otimista, que é:
back 300 years ago.
of continuous war,
a 140 anos de guerra contínua,
or software engineer or TV anchor.
engenheiros de software ou âncoras de TV.
with a little steam engine in his pocket,
com um motorzinho a vapor no bolso,
e liberá-los pra fazer outra coisa".
as strong, so you can do something else."
there was no real stage,
perto das vacas, no estábulo,
with the cows in the stable,
concerned about it,
preocupado com isso,
and what if the machine does this for me?"
e se a máquina tomar meu lugar?"
past progress and the benefit of it,
o progresso passado e nos beneficiar dele,
or electricity or medical supply.
eletricidade ou suprimentos médicos.
which was impossible 300 years ago.
o que era impossível 300 anos atrás.
the same rules to the future.
as mesmas regras para o futuro.
para o meu cargo de CEO,
of my work is repetitive,
é repetitivo, e não gosto disso.
on stupid, repetitive email.
com e-mails chatos e repetitivos.
that helps me get rid of this.
que me ajude a me livrar disso.
somos loucamente criativos:
are insanely creative;
more than anybody else.
mais do que todo mundo.
I think you can go to your hotel maid
podemos chegar para a camareira do hotel
you find a creative idea.
descobrir uma ideia criativa.
is to turn this creativity into action.
essa criatividade em ação.
build Google in a day?
construir o Google em um dia?
e inventar o próximo Snapchat,
and invent the next Snapchat,
e, mais tarde, claro, do trabalho braçal
in my opinion.
great side effects.
and education and shelter
médicos, educação, moradia e transporte
para todos nós, não só para os ricos.
affordable to all of us,
disse que este tempo ia ser diferente,
that this time it's different
that we've used in the past
que usamos no passado
is that, not completely,
que não completamente,
different from the kind of creativity
muito diferente do tipo de criatividade
belief as an AI person --
como uma pessoa IA,
any real progress on creativity
nenhum progresso real em criatividade
really important for people to realize,
importante perceber isso,
intelligence" is so threatening,
artificial" é muito ameaçadora
tossing a movie in,
the computer is our overlord,
o computador nos domina.
do repetitive things.
entirely on the repetitive end.
inteiramente no lado repetitivo.
em elaborar minutas de contrato,
que não vejo a grande ameaça à humanidade.
we've become superhuman.
nos tornamos super-humanos.
the Atlantic in 11 hours.
o Atlântico em 11 horas.
shouting back to us.
responder de volta para nós.
We're breaking the rules of physics.
estamos quebrando as leis da física.
we're going to remember everything
que já dissemos e vimos,
o que é bom para o meu Alzheimer.
in my early stages of Alzheimer's.
dizendo? Esqueci.
vamos ter um QI de mil ou mais.
an IQ of 1,000 or more.
spelling classes for our kids,
mais aula de ortografia,
is that we can be super creative.
realmente supercriativos.
mesmo que seja doloroso,
it's going to be painful,
of more than those jobs.
to just a new level of empowerment
num novo nível de empoderamento
se olharmos a história da humanidade,
60-100,000 years old, give or take --
60 a 100 mil anos atrás,
in terms of invention,
em termos de invenção,
it's a little bit older.
é um pouco mais de tempo;
manufacturing, penicillin --
moderna, a penicilina...
vamos descobrir mais coisas.
has gone up, not gone down, in my opinion.
tem aumentado, não diminuído.
things have been invented yet. Right?
interessantes foram inventadas, certo?
Hopefully, I'll change this.
espero mudar isso.
people laughed about. (Laughs)
Working secretly on flying cars.
secretamente em carros voadores.
implant in our brain
em nosso cérebro
quando tivermos, vocês vão adorar.
once you have it, you'll love it.
inventadas, que vamos inventar:
we haven't invented yet
from one location to another.
mas, cerca de 200 anos atrás,
that flight wouldn't exist,
íamos voar, mesmo 120 anos atrás,
do que podemos correr,
than you could run,
that you can't beam a person
teletransportar uma pessoa
and your brilliance.
inspiradora e brilhante.
ABOUT THE SPEAKERS
Sebastian Thrun - Educator, entrepreneurSebastian Thrun is a passionate technologist who is constantly looking for new opportunities to make the world better for all of us.
Why you should listen
Sebastian Thrun is an educator, entrepreneur and troublemaker. After a long life as a professor at Stanford University, Thrun resigned from tenure to join Google. At Google, he founded Google X, home to self-driving cars and many other moonshot technologies. Thrun also founded Udacity, an online university with worldwide reach, and Kitty Hawk, a "flying car" company. He has authored 11 books, 400 papers, holds 3 doctorates and has won numerous awards.
Sebastian Thrun | Speaker | TED.com
Chris Anderson - TED Curator
After a long career in journalism and publishing, Chris Anderson became the curator of the TED Conference in 2002 and has developed it as a platform for identifying and disseminating ideas worth spreading.
Why you should listen
Chris Anderson is the Curator of TED, a nonprofit devoted to sharing valuable ideas, primarily through the medium of 'TED Talks' -- short talks that are offered free online to a global audience.
Chris was born in a remote village in Pakistan in 1957. He spent his early years in India, Pakistan and Afghanistan, where his parents worked as medical missionaries, and he attended an American school in the Himalayas for his early education. After boarding school in Bath, England, he went on to Oxford University, graduating in 1978 with a degree in philosophy, politics and economics.
Chris then trained as a journalist, working in newspapers and radio, including two years producing a world news service in the Seychelles Islands.
Back in the UK in 1984, Chris was captivated by the personal computer revolution and became an editor at one of the UK's early computer magazines. A year later he founded Future Publishing with a $25,000 bank loan. The new company initially focused on specialist computer publications but eventually expanded into other areas such as cycling, music, video games, technology and design, doubling in size every year for seven years. In 1994, Chris moved to the United States where he built Imagine Media, publisher of Business 2.0 magazine and creator of the popular video game users website IGN. Chris eventually merged Imagine and Future, taking the combined entity public in London in 1999, under the Future name. At its peak, it published 150 magazines and websites and employed 2,000 people.
This success allowed Chris to create a private nonprofit organization, the Sapling Foundation, with the hope of finding new ways to tackle tough global issues through media, technology, entrepreneurship and, most of all, ideas. In 2001, the foundation acquired the TED Conference, then an annual meeting of luminaries in the fields of Technology, Entertainment and Design held in Monterey, California, and Chris left Future to work full time on TED.
He expanded the conference's remit to cover all topics, including science, business and key global issues, while adding a Fellows program, which now has some 300 alumni, and the TED Prize, which grants its recipients "one wish to change the world." The TED stage has become a place for thinkers and doers from all fields to share their ideas and their work, capturing imaginations, sparking conversation and encouraging discovery along the way.
In 2006, TED experimented with posting some of its talks on the Internet. Their viral success encouraged Chris to begin positioning the organization as a global media initiative devoted to 'ideas worth spreading,' part of a new era of information dissemination using the power of online video. In June 2015, the organization posted its 2,000th talk online. The talks are free to view, and they have been translated into more than 100 languages with the help of volunteers from around the world. Viewership has grown to approximately one billion views per year.
Continuing a strategy of 'radical openness,' in 2009 Chris introduced the TEDx initiative, allowing free licenses to local organizers who wished to organize their own TED-like events. More than 8,000 such events have been held, generating an archive of 60,000 TEDx talks. And three years later, the TED-Ed program was launched, offering free educational videos and tools to students and teachers.
Chris Anderson | Speaker | TED.com