How algorithms shape our world | Kevin Slavin

Translator: Toni Beltrán Reviewer: Fran Ontanaya

This is a photograph

Aquesta fotografia és

by the artist Michael Najjar,

de l'artista Michael Najjar,

and it's real,

i en cert sentit és real,

in the sense that he went there to Argentina

doncs l'artista va anar a Argentina

to take the photo.

a fer la foto.

But it's also a fiction. There's a lot of work that went into it after that.

Però també és ficció. Hi ha un munt de feina al darrere.

And what he's done

El que l'artista ha fet,

is he's actually reshaped, digitally,

en realitat, és modificar, digitalment,

all of the contours of the mountains

tots els contorns de les muntanyes

to follow the vicissitudes of the Dow Jones index.

per a fer-les encaixar amb les fluctuacions de l'índex Dow Jones

So what you see,

Així que, això que veieu,

that precipice, that high precipice with the valley,

aquest gran precipici al costat de la vall,

is the 2008 financial crisis.

és la crisi financera del 2008.

The photo was made

La foto es va fer

when we were deep in the valley over there.

quan érem a les profunditats de la vall, allà.

I don't know where we are now.

No sé on som ara.

This is the Hang Seng index

Això és l'índex Hang Seng

for Hong Kong.

de Hong Kong.

And similar topography.

Té una topografia similar,

I wonder why.

i em pregunto perquè.

And this is art. This is metaphor.

I això és art, és metàfora.

But I think the point is

Però el més important, penso jo,

that this is metaphor with teeth,

és que és una metàfora amb força.

and it's with those teeth that I want to propose today

I és amb aquesta força que vull proposar, avui,

that we rethink a little bit

que pensem una mica

about the role of contemporary math --

en el paper de les matemàtiques modernes --

not just financial math, but math in general.

no només matemàtiques financeres, sinó matemàtiques en general

That its transition

En la seva transició

from being something that we extract and derive from the world

de quelcom que extraiem i derivem del món

to something that actually starts to shape it --

cap a quelcom que, en realitat, comença a donar-li forma --

the world around us and the world inside us.

tant al món que ens envolta com al nostre món interior.

And it's specifically algorithms,

En particular, els algoritmes

which are basically the math

són les matemàtiques

that computers use to decide stuff.

que els ordinadors usen per prendre decisions.

They acquire the sensibility of truth

Adquireixen sensibilitat a la veritat

because they repeat over and over again,

a base de repetir una vegada rere una altra.

and they ossify and calcify,

I es van ossificant i calcificant

and they become real.

i es transformen en realitat.

And I was thinking about this, of all places,

Estava pensant en això, ara fa dos anys

on a transatlantic flight a couple of years ago,

en un vol transatlàntic

because I happened to be seated

perquè resulta que seia

next to a Hungarian physicist about my age

al costat d'un físic hongarès de la meva edat

and we were talking

i estàvem xerrant

about what life was like during the Cold War

de com era la vida durant la guerra freda

for physicists in Hungary.

pels físics hongaresos

And I said, "So what were you doing?"

I li vaig dir, "I doncs, què fèieu?"

And he said, "Well we were mostly breaking stealth."

i ell va dir, "Bé, bàsicament, trencàvem l'Stealth."

And I said, "That's a good job. That's interesting.

I jo, "És una bona tasca. Interessant.

How does that work?"

Com funciona?"

And to understand that,

Per entendre això,

you have to understand a little bit about how stealth works.

cal que entengueu una mica com funciona l'Stealth.

And so -- this is an over-simplification --

Això és una simplificació,

but basically, it's not like

però bé, bàsicament,

you can just pass a radar signal

no pots simplement travessar una senyal de radar

right through 156 tons of steel in the sky.

amb 156 tones de metall al cel.

It's not just going to disappear.

No pot simplement desaparèixer.

But if you can take this big, massive thing,

Però si poguessis agafar aquesta gran cosa,

and you could turn it into

i la poguessis transformar en

a million little things --

un milió de petites coses --

something like a flock of birds --

com un esbart d'ocells --

well then the radar that's looking for that

aleshores el radar que està buscant

has to be able to see

hauria de ser capaç de veure

every flock of birds in the sky.

tots els esbarts d'ocells del cel.

And if you're a radar, that's a really bad job.

I, si ets un radar, no és massa bona idea.

And he said, "Yeah." He said, "But that's if you're a radar.

I va dir: "Si". Va dir: "Això si ets un radar.

So we didn't use a radar;

De manera que no vam fer servir un radar;

we built a black box that was looking for electrical signals,

vam construir una caixa negra que buscava senyals elèctriques,

electronic communication.

comunicacions electròniques.

And whenever we saw a flock of birds that had electronic communication,

I quan veiem un esbart d'ocells que tenia comunicacions electròniques,

we thought, 'Probably has something to do with the Americans.'"

sabíem que segurament tenia a veure amb els americans."

And I said, "Yeah.

I jo vaig dir, "Sí,

That's good.

molt bé.

So you've effectively negated

Per tant, acabes de destrossar

60 years of aeronautic research.

60 anys de recerca aeronàutica

What's your act two?

i després, què?

What do you do when you grow up?"

Què vas fer quan vas créixer?

And he said,

I va dir,

"Well, financial services."

"Serveis financers"

And I said, "Oh."

I jo vaig dir, "Oh."

Because those had been in the news lately.

Perquè feia poc havia sortit a les notícies.

And I said, "How does that work?"

I li vaig dir: "Com funciona?"

And he said, "Well there's 2,000 physicists on Wall Street now,

I ell em va dir, "Bé, ara mateix hi ha uns 2000 físics a Wall Street,

and I'm one of them."

i jo sóc un d'ells."

And I said, "What's the black box for Wall Street?"

I jo vaig dir: "I com és la caixa negra a Wall Street?"

And he said, "It's funny you ask that,

I va dir: "És graciós que em preguntis això,

because it's actually called black box trading.

perquè, en realitat, es parla de comerç de caixa negra,

And it's also sometimes called algo trading,

o també comerç algo,

algorithmic trading."

comerç algorítmic."

And algorithmic trading evolved in part

I el comerç algorítmic ha evolucionat, en part,

because institutional traders have the same problems

perquè els comerciants institucionals tenen els mateixos problemes

that the United States Air Force had,

que les Forces Armades dels Estats Units tenien,

which is that they're moving these positions --

és a dir, que estan movent aquestes accions --

whether it's Proctor & Gamble or Accenture, whatever --

ja sigui "Proctor & Gamble" o "Accenture", és igual --

they're moving a million shares of something

estan movent un milió d'accions d'alguna cosa

through the market.

a través dels mercats.

And if they do that all at once,

I si les mouen totes de cop,

it's like playing poker and going all in right away.

passa com al pòquer: apostar-ho tot d'entrada.

You just tip your hand.

És com ensenyar les teves cartes.

And so they have to find a way --

Per tant necessiten una manera --

and they use algorithms to do this --

és a dir, algoritmes --

to break up that big thing

per dividir aquell gran moviment

into a million little transactions.

en un milió de petites transaccions.

And the magic and the horror of that

I la màgia, o l'horror, de la qüestió

is that the same math

es que es poden fer servir

that you use to break up the big thing

les mateixes matemàtiques tant per trencar

into a million little things

la gran cosa en un milió de petites coses

can be used to find a million little things

com per buscar un milió de petites coses

and sew them back together

i recomposar una gran cosa i,

and figure out what's actually happening in the market.

d'aquesta manera, saber que està passant realment als mercats.

So if you need to have some image

Així que si us voleu fer una idea

of what's happening in the stock market right now,

del que passa a la borsa ara mateix,

what you can picture is a bunch of algorithms

podeu pensar en una colla d'algoritmes

that are basically programmed to hide,

pensats bàsicament per amagar,

and a bunch of algorithms that are programmed to go find them and act.

i una altra colla d'algoritmes pensats per trobar i actuar.

And all of that's great, and it's fine.

I tot això està molt bé

And that's 70 percent

i representa el 70 %

of the United States stock market,

de la borsa dels Estats Units

70 percent of the operating system

el 70 % del sistema operatiu

formerly known as your pension,

altrament dit la teva pensió,

your mortgage.

la teva hipoteca,

And what could go wrong?

Què podria fallar?

What could go wrong

Podria passar tal com

is that a year ago,

fa un any, quan

nine percent of the entire market just disappears in five minutes,

el 9% del total del mercat simplement va desaparèixer durant 5 minuts,

and they called it the Flash Crash of 2:45.

i n'hi van dir el flash crash de les 2:45.

All of a sudden, nine percent just goes away,

Tot de cop, el 9% simplement desapareix,

and nobody to this day

i fins ara, ningú

can even agree on what happened

ha pogut aclarir que va passar,

because nobody ordered it, nobody asked for it.

perquè ningú ho va manar, ningú ho va demanar.

Nobody had any control over what was actually happening.

De fet, ningú tenia cap control sobre el que realment succeía.

All they had

Tot el que tenien era, simplement

was just a monitor in front of them

una pantalla davant d'ells

that had the numbers on it

amb un munt de números,

and just a red button

i un botó vermell

that said, "Stop."

que posava "Stop."

And that's the thing,

El que està passant,

is that we're writing things,

es que escrivim coses,

we're writing these things that we can no longer read.

coses que ja no podem llegir

And we've rendered something

Hem generat quelcom

illegible,

il·legible

and we've lost the sense

I ja no sabem

of what's actually happening

realment què està passant

in this world that we've made.

en aquest món que hem construït.

And we're starting to make our way.

I això és el començament.

There's a company in Boston called Nanex,

Hi ha una companyia a Boston, Nanex,

and they use math and magic

utilitza matemàtiques i màgia

and I don't know what,

no sé exactament com,

and they reach into all the market data

però agafen dades dels mercats

and they find, actually sometimes, some of these algorithms.

i troben, de vegades, alguns d'aquests algoritmes.

And when they find them they pull them out

Quan els troben, els extreuen

and they pin them to the wall like butterflies.

i els claven a la paret com papallones.

And they do what we've always done

I aleshores fan allò que tots hem fet

when confronted with huge amounts of data that we don't understand --

quan ens hem trobat davant de grans quantitats de dades que no enteniem--

which is that they give them a name

simplement, els hi posen un nom

and a story.

i una història.

So this is one that they found,

A aquest que van trobar,

they called the Knife,

li van posar "Knife",

the Carnival,

el "Carnival".

the Boston Shuffler,

el "Boston Shuffler"

Twilight.

"Twilight".

And the gag is

El més graciós de tot és

that, of course, these aren't just running through the market.

que, òbviament, aquests algoritmes no funcionen només als mercats.

You can find these kinds of things wherever you look,

Aquest tipus de coses es poden trobar arreu,

once you learn how to look for them.

un cop se sap com buscar-les.

You can find it here: this book about flies

Per exemple, aquí: aquest llibre sobre mosques

that you may have been looking at on Amazon.

que potser us heu estat mirant per Amazon.

You may have noticed it

Potser el vàreu veure

when its price started at 1.7 million dollars.

quan costava 1.700.000 dòlars

It's out of print -- still ...

No s'imprimeix -- encara...

(Laughter)

(Rialles)

If you had bought it at 1.7, it would have been a bargain.

Si l'haguéssiu comprat a 1.700.000 hauria sigut una ganga,

A few hours later, it had gone up

doncs al cap d'unes hores havia pujat

to 23.6 million dollars,

fins als 23.600.000 dòlars.

plus shipping and handling.

costs d'enviament apart.

And the question is:

La pregunta és:

Nobody was buying or selling anything; what was happening?

Si ningú estava comprant ni venent res; què estava passant?

And you see this behavior on Amazon

Aquests tipus de comportament s'observen tant a Amazon

as surely as you see it on Wall Street.

com a Wall Street.

And when you see this kind of behavior,

I el que aquests comportaments són

what you see is the evidence

una conseqüència

of algorithms in conflict,

d'algoritmes en conflicte,

algorithms locked in loops with each other,

algoritmes bloquejats en bucles,

without any human oversight,

sense cap supervisió humana,

without any adult supervision

sense la supervisió d'un adult

to say, "Actually, 1.7 million is plenty."

que pugui dir: "Home, 1.700.000 dòlars ja és bastant."

(Laughter)

(Rialles)

And as with Amazon, so it is with Netflix.

Passa el mateix amb Netflix.

And so Netflix has gone through

Netflix ha fet servir

several different algorithms over the years.

diferents algoritmes al llarg dels anys:

They started with Cinematch, and they've tried a bunch of others --

van començar amb Cinematch, i després n'han anat provant d'altres.

there's Dinosaur Planet; there's Gravity.

com "Dinosaur Planet". "Gravity".

They're using Pragmatic Chaos now.

i, actualment, "Pragmatic Chaos".

Pragmatic Chaos is, like all of Netflix algorithms,

El que aquest algoritme intenta, igual tots els altres,

trying to do the same thing.

es el següent:

It's trying to get a grasp on you,

Tracta d'entendre't a tu,

on the firmware inside the human skull,

el que hi ha a dins teu,

so that it can recommend what movie

per a ser capaç de recomanar-te

you might want to watch next --

quina pel·lícula t'agradaria veure --

which is a very, very difficult problem.

i això és una tasca molt, molt difícil.

But the difficulty of the problem

Però que la tasca sigui difícil,

and the fact that we don't really quite have it down,

i el fet que encara no sabem realment com resoldre-la

it doesn't take away

no elimina

from the effects Pragmatic Chaos has.

les conseqüències d'usar "Pragmatic Chaos".

Pragmatic Chaos, like all Netflix algorithms,

Al cap i a la fi, aquest algoritme, i tots els altres

determines, in the end,

que Netflix ha usat, determina

60 percent

el 60%

of what movies end up being rented.

de les pel·lícules que els clients acaben llogant.

So one piece of code

És a dir, que un tros de codi,

with one idea about you

que conté una sola idea sobre tu,

is responsible for 60 percent of those movies.

és el responsable del 60% de les vendes.

But what if you could rate those movies

Ara bé, què passaria si poguéssim puntuar aquestes pel·lícules

before they get made?

abans de que existissin?

Wouldn't that be handy?

Oi que seria molt pràctic?

Well, a few data scientists from the U.K. are in Hollywood,

Bé, alguns científics del Regne Unit, experts en dades, estan a Hollywood,

and they have "story algorithms" --

desenvolupant algoritmes per guions --

a company called Epagogix.

Hi ha una empresa anomenada Epagogix

And you can run your script through there,

que, a partir del teu guió

and they can tell you, quantifiably,

et poden dir, quantitativament,

that that's a 30 million dollar movie

si serà una pel·lícula de 30 milions de dòlars

or a 200 million dollar movie.

o una de 200 milions.

And the thing is, is that this isn't Google.

I això no és Google,

This isn't information.

no és informació,

These aren't financial stats; this is culture.

ni estadística financera: això és cultura.

And what you see here,

I el que heu vist avui aquí,

or what you don't really see normally,

o el que normalment no veieu

is that these are the physics of culture.

és la física de la cultura.

And if these algorithms,

I si aquests algoritmes,

like the algorithms on Wall Street,

igual que els de Wall Street,

just crashed one day and went awry,

produissin un crash,

how would we know?

com ho sabríem?

What would it look like?

quina pinta faria?

And they're in your house. They're in your house.

També són a casa teva.

These are two algorithms competing for your living room.

Això són dos algoritmes competint pel teu saló

These are two different cleaning robots

Són dos robots netejadors diferents

that have very different ideas about what clean means.

i tenen idees força diferents sobre que vol dir netejar

And you can see it

Això es pot veure

if you slow it down and attach lights to them,

posant càmera lenta i engantxant-los unes llums.

and they're sort of like secret architects in your bedroom.

Són com una mena d'arquitectes secrets a la teva habitació.

And the idea that architecture itself

I la idea que la pròpia arquitectura

is somehow subject to algorithmic optimization

està subjecte a l'optimització algorítmica

is not far-fetched.

no és gens descabellada.

It's super-real and it's happening around you.

És ben real, i està passant al teu voltant.

You feel it most

Ho notareu, per exemple,

when you're in a sealed metal box,

si esteu en una d'aquestes caixes metàl·liques hermètiques

a new-style elevator;

un d'aquests ascensors d'última generació

they're called destination-control elevators.

ascensor amb control de destinació, se'n diuen.

These are the ones where you have to press what floor you're going to go to

Són aquests on has de prémer el botó del pis on vas

before you get in the elevator.

abans d'entrar a l'ascensor.

And it uses what's called a bin-packing algorithm.

Utilitzen un algoritme d'empaquetament per grups.

So none of this mishegas

És a dir, que res de la bogeria

of letting everybody go into whatever car they want.

de deixar que cadascú es fiqui a l'ascensor que vol:

Everybody who wants to go to the 10th floor goes into car two,

Tothom qui va al desè pis, ha d'agafar el segon ascensor,

and everybody who wants to go to the third floor goes into car five.

i tothom qui va al tercer pis ha d'agafar el cinquè ascensor.

And the problem with that

El problema, però,

is that people freak out.

és que la gent s'espanta.

People panic.

S'esgarrifa.

And you see why. You see why.

I no és difícil veure perquè passa això.

It's because the elevator

Passa perquè a l'ascensor

is missing some important instrumentation, like the buttons.

li falten certs elements bàsics, com els botons.

(Laughter)

(rialles)

Like the things that people use.

Coses que la gent fa servir.

All it has

Tot el que té

is just the number that moves up or down

és l'indicador del pis

and that red button that says, "Stop."

i el boto vermell que diu: "STOP".

And this is what we're designing for.

I això és el que estem creant.

We're designing

Estem creant

for this machine dialect.

aquesta comunicació amb les màquines

And how far can you take that? How far can you take it?

I fins on podem arribar, estirant aquest concepte?

You can take it really, really far.

Doncs bastant, bastant lluny

So let me take it back to Wall Street.

Deixeu-me que torni a Wall Street.

Because the algorithms of Wall Street

Els algoritmes de Wall Street

are dependent on one quality above all else,

depenent, per sobre de tot, d'una característica:

which is speed.

la velocitat.

And they operate on milliseconds and microseconds.

Operen en milisegons i microsegons

And just to give you a sense of what microseconds are,

Per que us en feu una idea,

it takes you 500,000 microseconds

es triguen uns 500.000 microsegons

just to click a mouse.

a fer clic amb el ratolí.

But if you're a Wall Street algorithm

Però si ets un algoritme de Wall Street

and you're five microseconds behind,

i vas 5 microsegons enrere

you're a loser.

estàs perdut.

So if you were an algorithm,

Per tant, si fossis un algoritme

you'd look for an architect like the one that I met in Frankfurt

buscaries un arquitecte com el que jo vaig conèixer a Frankfurt

who was hollowing out a skyscraper --

que estava buidant un gratacels --

throwing out all the furniture, all the infrastructure for human use,

llençant tots els mobles i tot allò que les persones usen,

and just running steel on the floors

i posant acer al terra

to get ready for the stacks of servers to go in --

per a poder posar-hi piles de servidors --

all so an algorithm

tot per tal que un algoritme

could get close to the Internet.

pogués estar més a prop d'Internet.

And you think of the Internet as this kind of distributed system.

Potser penseu que Internet es un sistema distribuït.

And of course, it is, but it's distributed from places.

És distribuït, però distribuït des de certs llocs.

In New York, this is where it's distributed from:

A Nova York, per exemple, es distribueix des d'aquí:

the Carrier Hotel

l'Hotel Carrier

located on Hudson Street.

al carrer Hudson.

And this is really where the wires come right up into the city.

Aquí és de on realment surten els cables cap a la ciutat

And the reality is that the further away you are from that,

I el que passa és que si estàs més lluny d'aquí,

you're a few microseconds behind every time.

estàs uns microsegons enrere.

These guys down on Wall Street,

I aquesta gent de Wall Street,

Marco Polo and Cherokee Nation,

Marco Polo i Cherokee Nation,

they're eight microseconds

van 8 microsegons endarrere

behind all these guys

respecte tota aquesta gent

going into the empty buildings being hollowed out

que van als edificis que es buiden

up around the Carrier Hotel.

al voltant de l'hotel Carrier.

And that's going to keep happening.

I això seguirà passant.

We're going to keep hollowing them out,

Seguirem buidant edificis,

because you, inch for inch

ja que, centímetre a centímetre,

and pound for pound and dollar for dollar,

lliura a lliura i dòlar a dòlar

none of you could squeeze revenue out of that space

cap de nosaltres podria treure més profit d'aquest espai

like the Boston Shuffler could.

que el Boston Shuffler.

But if you zoom out,

Però si ens allunyem,

if you zoom out,

si ens allunyem,

you would see an 825-mile trench

veiem una rasa de 1300 quilòmetres de llargada

between New York City and Chicago

entre Nova York i Chicago

that's been built over the last few years

que ha estat construïda darrerament

by a company called Spread Networks.

per una empresa anomenada Spread Networks.

This is a fiber optic cable

Això és un cable de fibra òptica

that was laid between those two cities

estès entre aquestes dues ciutats.

to just be able to traffic one signal

només per transmetre un senyal

37 times faster than you can click a mouse --

37 vegades més ràpid que el clic d'un ratolí,

just for these algorithms,

només per aquests algoritmes:

just for the Carnival and the Knife.

el Carnival i el Knife.

And when you think about this,

I quan pensem això,

that we're running through the United States

que estem foradant els Estats Units

with dynamite and rock saws

amb dinamita i serres de roca

so that an algorithm can close the deal

per a que un algoritme pugui tancar un contracte

three microseconds faster,

tres microsegons més ràpid,

all for a communications framework

tot en un marc de comunicacions

that no human will ever know,

que cap humà arribarà a conèixer,

that's a kind of manifest destiny;

és una mena de destí manifest

and we'll always look for a new frontier.

que sempre cercarà una frontera nova.

Unfortunately, we have our work cut out for us.

Però encara tenim molta feina a fer.

This is just theoretical.

Tot això es teoria

This is some mathematicians at MIT.

d'uns matemàtics del MIT.

And the truth is I don't really understand

I, la veritat, és que no entenc

a lot of what they're talking about.

gaire del que parlen.

It involves light cones and quantum entanglement,

Es tracta de cons lluminosos y connexions quàntiques.

and I don't really understand any of that.

i el cert és que no comprenc res d'això.

But I can read this map,

Això sí, puc llegir aquest mapa.

and what this map says

que diu que

is that, if you're trying to make money on the markets where the red dots are,

si intentem guanyar diners als mercats a on es troben els punts vermells,

that's where people are, where the cities are,

és a dir, a on es troba la gent, a on són les ciutats,

you're going to have to put the servers where the blue dots are

haurem de posar els servidors a on estan els punts blaus

to do that most effectively.

per tal d'obtenir la màxima eficiència.

And the thing that you might have noticed about those blue dots

I potser heu notat que la majoria de punts blaus

is that a lot of them are in the middle of the ocean.

es troben al mig de l'oceà.

So that's what we'll do: we'll build bubbles or something,

Així doncs, haurem de construir bombolles

or platforms.

o plataformes.

We'll actually part the water

En realitat, anem a compartir l'aigua

to pull money out of the air,

per tal d'extreure diners a l'aire

because it's a bright future

perquè allí hi ha un futur brillant

if you're an algorithm.

si som algoritmes.

(Laughter)

(Rialles)

And it's not the money that's so interesting actually.

En realitat, els diners no són el que ens interessa més,

It's what the money motivates,

sinó més aviat la motivació que porten els diners.

that we're actually terraforming

El fet de transformar

the Earth itself

el mateix planeta

with this kind of algorithmic efficiency.

amb aquesta mena d'eficiència algorítmica.

And in that light,

Sota aquesta perspectiva

you go back

tornem a veure

and you look at Michael Najjar's photographs,

les fotografies de Michael Najjar

and you realize that they're not metaphor, they're prophecy.

i ens adonem de que no són metafòriques, són profètiques.

They're prophecy

S'anticipen

for the kind of seismic, terrestrial effects

als efectes sísmics, terrestres

of the math that we're making.

de les matemàtiques que fem.

And the landscape was always made

I el paisatge sempre ha estat configurat

by this sort of weird, uneasy collaboration

per aquesta mena de col·laboració, estranya i difícil,

between nature and man.

entre la natura i el home.

But now there's this third co-evolutionary force: algorithms --

Però ara existeix aquesta tercera força coevolutiva: els algoritmes;

the Boston Shuffler, the Carnival.

el "Boston Shuffler", el "Carnival".

And we will have to understand those as nature,

I haurem de considerar-los com una part de la natura.

and in a way, they are.

I d'alguna, ho són.

Thank you.

Gràcies.

(Applause)

(Aplaudiment)