Chris Urmson: How a driverless car sees the road TED Talk Minuto 7.50 en adelante
Una de las grandes tensiones en la instalación de la democracia occidental (aparte de saber si se compatible con el capitalismo sea dicho de paso, pequeño intríngulis ) es poder conciliar la libertad con la igualdad. Se han escrito toneladas de bibliotecas sobre estos temas y seguimos en ayunas.
Como siempre los dipolos dividen las opiniones y resulta dificil argumentar en forma definitiva a favor de una opción o de la otra.
En esta próxima sesión de #dipolosmimaguen debatiremos si son bienvenidos (o no) los coches autónomos. Aunque en América Latina parecen una fantasía pronto una decena de estados en USA aprobará su existencia,
Pero lo que nos interesa de esta problemática no son los aspectos tecnológicos, ni siquiera los ergonómicos, sino los problemas que generarán las máquinas autónomas cuando tomen decisiones en nombre nuestro.
Como siempre ocurrirá en estos debates dipolo, el fenómeno en cuestión (fascinante en si mismo) es un test proyectivo acerca de representaciones sociales, acerca de turismo acerca del comportamiento humano, acerca de las construcciones sociales convertidas en naturaleza humana hecha y derecha.
Porque es mas que sabido que los humanos somos muy irresponsables al conducir: Mueren 33.000 norteamericanos por año, 15.000 mexicanos; 7.500 argentinos. Cada año se producen entre 1,5 y 2 millones de muertos por causa de accidentes de tráfico (el equivalente a que en cada día laboral se cayera un avión 737), y en muchos países desarrollados constituye la principal causa de muerte entre los menores de 25 años.
También que la congestión de trafico debido a los conductores humanos es insoportable en términos objetivos y subjetivos. En el mundo hay 1.200 millones de coche y si China e India siguen creciendo como lo hacían hace un par de años atras habría 2.000 millones en 2025.
Hoy en USA el tráfico anual despilfarra 12.000 millones de litros, 7.000 millones de horas y ocasiona una perdida anual de U$ 160.000 millones. Los coches están detenidos 23 horas por día y se gasta un 30% de su consumo total buscando estacionamiento.
Pero renunciar al auto es para muchos renunciar a la personalidad. Renunciar al auto, sinónimo de libertad y elección equivale a sacrificar la libertad y la autonomía.
La discusión pasa pues por muchos andariveles
– Los coches son potestad de la clases media y alta, ¿renunciarán sus poseedores a este símbolo de status?
– 8 de las mas grandes compañías publicas del mundo son automotrices facturando combinadamente U$ 1.3 billones ¿abandonarán el terreno tan fácilmente preocupadas por las muertes que ellas mismas causan, y que generalmente atribuyen a accidentes o “error” humano?
– Machismo y SUV van de la mano. ¿Dejarán los hombres de basar su identidad en la propiedad de los vehículos?
How GM Beat Tesla to the First True Mass-Market Electric Car
– ¿Qué pasará con nuestras ciudades si la enorme inversión en transporte individual pasa al colectivo o social?
Yendo mas a fondo en los temas del dipolo para la discusión presencial
– ¿Aceptaremos que los programas manejan mejor que nosotros?
– ¿Dejaremos nuestra vida a su cargo cuando sabemos que el software nunca es fiable del todo?
The ethical dilemma of self-driving cars – Patrick Lin
– Los coches autónomos son la avanzada de una liberación creciente de los objetos del control y de la manipulación humanas. La Internet de las cosas es mucho mas que el mero afichaje de un chip a los objetos. Es la conversión de la logística y de la ecología de muchas industrias analógicas (salud, transporte, bancos, turismo, etc) a la lógica (y la topología) de Internet.
Pero con la eficientización de las redes, todas las prácticas humanas se ven cuestionadas. Si la mayoría de nuestros comportamientos son algoritimizables y computables, un programa los hará mejor que nosotros. ¿Sacrificaremos nuestra libertad (para el error) en aras de una perfección (o suficiencia, o seguridad) de la lógica maquinal?
Sigamos sumando preguntas y trabajemos en red y bajo la metodología del design thinking este próximo martes para ahondar mas en estas cuestiones.
Algunas referencias
1. Google’s Self-Driving Car Caused Its First Crash
Google’s cars have driven more than 1.3 million miles since 2009. They can recognize hand signals from traffic officers and “think” at speeds no human can match. As of January, they had been involved in 17 crashes, all caused by human error. Google has previously predicted they’ll be road-ready by 2020.
This was a minor crash with serious implications
________
2. Pagina de Wired sobre coches autónomos
_______
3. The Case for Making Self-Driving Cars Think Like Humans
Car crashes kill more than a million people each year, and roughly 90 percent of them are the result of human error. That’s the strongest argument for developing self-driving cars: Humans are lousy drivers. Coolly logical robots, the thinking goes, will far exceed us.
The irony of that is the best way to program those robots to drive may be to program them to drive more like humans. The UK is studying this very idea as part of “Move-UK,” an $8 million, three-year project intended to bring self-driving cars to market, on the double.
__________
4. Maven, GM’s Car-Sharing Scheme, Is Really About a Driverless Future
General Motors is launching a car-sharing program. It’s called Maven, it’s available in exactly one city, and, frankly, it’s an unexciting riff on ZipCar. But GM isn’t really competing with ZipCar. It’s placing a bet on the future.
Maven can be the foundation for the self-driving car network GM wants to build.
GM claims some 5 million people worldwide use vehicle sharing services like Uber, and that number is expected to hit 25 million by 2020.
The program is free to join and charges as little as $6 an hour to use a car, which includes insurance and gas.
_______
5. Toyota Eyes Satellite Antennas for Data-Hungry Connected Cars
The thing about a 4G connection is that it’s not enough, at least not for much longer. As cars become more and more like living rooms on wheels, they’ll need more and more data. It’s not just the ability to stream Netflix and download videogames while on the move. It’s the ability to download high-def 3-D maps, likely to be crucial for autonomous driving applications. It’s access to real-time weather information and local gas prices. It’s over the air updates that make the car’s software better as it ages.
the “mTenna,” a satellite antenna that’s just half an inch thick
______
6. Feds Say They’ll Count Computers As Human Drivers
It’s a major moment in the effort to introduce autonomous driving to America’s roads, which is hampered as much by regulatory questions as by technological hurdles.
“If no human occupant of the vehicle can actually drive the vehicle, it is more reasonable to identify the ‘driver’ as whatever (as opposed to whoever) is doing the driving.”
_________
7. Welcome to the Metastructure: The New Internet of Transportation | WIRED
For me, a car is like a suit or a good exoskeleton
Driving itself is changing. Between electric and self-?driving vehicles, ubiquitous sensors, network connectivity, and new kinds of transportation companies, everything is in flux: cars, how we feel about them, even roads and cities
It’s a shift we need. The fact is, too many people own too many cars—an estimated 1.2 billion vehicles globally. Congestion in many cities is already untenable, and it’s only getting worse.
Individually, the new tools and technologies for moving around are interesting; put them together and you get something profound. Connect these new systems and individual networks to each other and they self-?assemble into a transportation super-?network. It’s decentralized, offers multiple routes from node to node, carries any kind of person or thing to any kind of place, and adjusts itself in real time
This decentralized approach to remapping our physical roads is fundamentally (and finally) changing everything about how we get around
This future has a history. General Motors’ Futurama exhibit at the 1939 World’s Fair, harbinger of modernity, touted broad highways and self-driving cars. Transit engineers have advocated for personal, on-demand pods since at least the 1960s. But the developers of so-called Intelligent Vehicle Highway Systems didn’t predict that the primary intelligence would reside in the cars instead of a control tower. And they didn’t know that people would carry GPS-equipped supercomputers in their pockets.
The size and shape of a city, the literature says, is limited to the range someone can travel in 45 minutes to an hour.
Third phase transition.
Google isn’t the only company building robot cars, of course. Six major automakers (and Tesla) have announced plans to build varying levels of autonomy into their products.
Could mean the end of car crashes. Combine that with the end of internal combustion, and cars will change shape.
It could also mean the end of congestion.
All that traffic costs 3 billion gallons of fuel and 7 billion hours of time. In 2014, that added up to $160 billion.
The first example of high-utilization autonomous driving was the New York taxicab,” says Dan Ammann, president of General Motors. “Uber improved on that. From a customer’s point of view, that’s also an autonomous experience.”
In 1960, Kevin Lynch, an urban planning professor at MIT, sent his students to interview people about their cities. The resulting book, The Image of the City, identified five features that people almost always saw: paths, the routes they took habitually; edges, the places beyond which they had no idea about; districts, where they knew their way around; nodes, like “home” or “work”; and landmarks.
Without knowing it, Lynch was describing a city as a network—but an inefficient one. The packets never varied their routes, hugging the same paths and nodes. Sure, they might glance wistfully at a tower—an Eiffel or a Sears—as they chugged home, but they never wondered what was on the other side of the park or behind that brutalist mall.
[What does the city’s form actually mean to the people who live there? What can the city planner do to make the city’s image more vivid and memorable to the city dweller? To answer these questions, Mr. Lynch, supported by studies of Los Angeles, Boston, and Jersey City, formulates a new criterion—imageability—and shows its potential value as a guide for the building and rebuilding of cities. The wide scope of this study leads to an original and vital method for the evaluation of city form. The architect, the planner, and certainly the city dweller will all want to read this book]
Today, cars are people’s second-largest household expenditure, and they sit unused 23 hours a day.
Autonomous vehicles, powered by smartphones and sensors and the Internet, interpret those limitations as damage and route around them—just like the Internet. The Lynchian edges crack open. Drivers are as solvable a computational problem on the road as bits are on the Internet, moving from node to node in the most efficient way.
An amazing thing happens when that information gets freed: The city opens up.
A 2015 study of Lisbon, Portugal, by the Organisation for Economic Cooperation and Development found that a fleet of just 26,000 TaxiBots—hypothetical on-demand, autonomous carpool vehicles—could replace every one of the city’s 203,000 cars. Think about that: the same population, 565,000 people, served by a tenth the number of cars.
The metastructure is changing how people live
If increased mobility extends the 45-minute range just as millennials start needing to buy car seats—and as housing costs in downtowns keep going up—maybe they all move even farther into suburbs and exurbs
The metastructure can’t be something available only to people with credit cards
The truth is, whatever we fear about these changes, the car ownership model that has dominated the past 65 years isn’t sustainable. We can’t just keep building more roads.
VIDEOS
1 En un comienzo circa 2011 en Long Beach. Google autonomous car
2. The First Self Driving Car is 500 Years Old
3 Google Ready to Turn Driverless Cars Into a Business
Los grandes números
1. One: They’re here.
Born-again evangelists of self-driving cars include some of the most venerable names in the business, such as William Clay Ford Jr., executive chairman of the company founded by his Model T–building great-grandfather Henry, and Toyota Motor Corp. president Akio Toyoda, whose great-grandfather was known as the “king of Japanese inventors.” (Toyoda, a racing buff, was adamantly opposed to self-drivers before reversing himself late last year.)
Four U.S. states have legalized self-driving cars, and at least 13 more are
mulling similar laws.
2 They’re superior drivers.
The computer is simply a better driver than a human. Better at keeping its eyes on other drivers; better at maintaining a steady cruising speed and thereby maximizing fuel efficiency; better at parsing GPS data, weather data, traffic data—any and all
kinds of data, really—and better at making rapid-fire adjustments.
3. Three: They’re going to change everything.
Because the gulf between human and machine is so vast—and growing—the next step
after making driverless cars legal will be making them mandatory.
That connection between cars and drivers is nothing like the feeling we had for typewriters or landlines or any of a thousand technologies overthrown by computers and smartphones. That was utility; this is love.
* There are about 6 million car accidents—incidents serious enough to be reported to lawenforcement— each year in the U.S.
* About 33,000 Americans die annually as a result, with an additional 2 million or
so injured.
* (Worldwide, there are about 1.3 million traffic fatalities every year)
Some 94% of road accidents are the fault of drivers,
The price tag for this mayhem, by one estimate, runs $836 billion.
Other statistics tell of lesser forms of wastage.
The average American spends 42 hours per year stuck in traffic. In the country’s most congested areas Washington, D.C.; Los Angeles; and New York City— that figure climbs as high as 82 hours.
Just 12 years ago, when the U.S. government funded the first
international competition for self-piloting vehicles, not one of the challengers finished the 150-mile (240 km) desert course set out for them.
The most successful robocar covered a little more than 7miles (11 km) before stupidly getting itself stuck. (Its wheels also caught fire.)
The following year, only five of the 23 vehicles in the competition made it to the finish line, with the fastest one averaging a poky 19 m.p.h. (30 km/h).
One of the finishers weighed 30,000 lb. (13,600 kg)—roughly 10 Toyota Priuses—
Today Google’s autonomous test cars have logged more than 1.4 million miles (2.25 million km) on their odometers on public roads—equivalent to about 100 years of driving for the average individual. Total accidents: 17, all caused by human pilots.
Converting just 10% of the U.S. vehicle fleet to self-driving cars would educe the number of accidents each year by 211,000 and save 1,100 lives. In this modest scenario, the costs of human clumsiness would be cut by $25.5 billion.
If, somewhere down the road, the share of self-driving vehicles rises to 90%, the number of accidents avoided could reach 4.2 million per year, with 21,700 lives saved. Self-driving technology
In all the adoption of driverless cars in the U.S. could save $1.3 trillion a year, according to —including $158 billion in fuel costs, productivity increases of $507 billion
and $488 billion in accident-related savings. Total worldwide savings: $5.6 trillion.
For every car in the U.S. there are between two and three parking spaces—one at home, one at work and fractions at the mall, airport and stadium.
Together, these amount to about 500 million spaces in all, or a total area of more than 3,000 sq. mi. (7,770 sq km), some 2million acres (810,000
hectares).
30% of drivers in certain metropolitan business districts are basically driving in circles at any given moment, searching for an open spot. But every Eden has its serpent, the driverless autopia
The revolution will destroy a lot. The $198 billion auto-insurance industry, the $100 billion parking industry and the $300 billion auto-aftermarket business
The auto-insurance industry could shrink to less than 40% its current size over the next 25 years, just because of smarter cars. People will lose jobs. There are about 3 million truck drivers in the U.S., 200,000 cabbies, 170,000 auto-body and glassrepair
technicians.
Tesla founder Musk has pegged the driverless-car transition to begin around
2023, a date closer to us than 9/11. “You can’t have a person driving a 2-ton death machine,”
Prevalence in the next decade. Industry analysts roughly think 2035 to 2050.
What’s certain is that like all technological revolutions, this one will have a self-compounding effect: more and more driverless cars on the roadwill result in more and more machine-centric street designs.
4. Porqué la inteligencia humana no sera reemplazada nunca
Sé el primero en comentar