Eleven years in, self-driving cars are still a ways off

Transportation has changed in the last decade but self-driving cars will still take some more time:

The boldest bid to remake transportation with tech was also among the earliest, and so far, the most disappointing. In 2009, Google cofounder Larry Page tapped computer scientist Sebastian Thrun to build a self-driving car. Make a vehicle that moves people safely and efficiently, Page said (in Thrun’s telling), and you could have a business as big as Google itself. The resulting effort, now known as Waymo, helped trigger a global race for autonomy, one that many predicted would bear fruit by the decade’s end. Tesla CEO Elon Musk said a Tesla would drive itself across the country in 2017. General Motors promised to launch a robo-taxi service in 2019. Nissan targeted 2020 for the market debut of its self-driving car. Former Waymo lead Chris Urmson said he hoped his sons would never need to learn how to drive.

But billions of dollars and thousands of engineers haven’t produced a robot that can match, let alone eclipse, the ability of the human driver. AV developers have retreated to quiet suburbs and simple interstates, hoping they can master at least some corner of a profoundly complex world. GM pushed back its debut date indefinitely. Nissan has stopped talking about self-driving. Waymo is just starting to take the human backups out of its cars in the Phoenix suburbs. Musk never made his road trip.

Reading this brief overview, two things struck me:

  1. Having a computer do all that is needed to drive is a monumental task. There is a lot of information to take in from behind the wheel and the environment keeps changing. This makes human drivers look pretty good. Even with all the accidents and deaths that occur every year, that humans can handle all of this at 60 mph or higher is remarkable.
  2. All the money and effort that has gone into this simply reinforces the car as the primary agent of transportation in the United States. While having no human driver could be a game changer, all this effort does little to displace the car as center of social life, work, urban planning, and sprawl. Perhaps it would be too much to ask Americans to give up cars but this could be viewed by future Americans as a missed opportunity to reorganize society.

Even if the next decade features truly autonomous vehicles, it will take more time for these vehicles to work their way through the system. Since I have also seen lists of the new laws and regulations going into effect January 1, is it far-fetched to imagine a new rule starting in early 2025 that all new vehicles purchased must be fully autonomous?

Claim: “The physical environment feels depressingly finished”

As Derek Thompson of The Atlantic considers innovation and Silicon Valley, he includes this paragraph regarding innovation in the physical and urban realm:

And if you look up from your smartphone, progress becomes harder to see. The physical world of the city—the glow of electric-powered lights, the rumble of automobiles, the roar of airplanes overhead and subways below—is a product of late-19th-century and early-20th-century invention. The physical environment feels depressingly finished. The bulk of innovation has been shunted into the invisible realm of bytes and code.

There are several pieces that can be pulled out of this an examined:

1. Has innovation in cities and urban areas slowed? Many of the major changes may have already happened – think the modern skyscraper, the car and all the roads to go with them – but I’m guessing there are some lesser-known changes in the last few decades that have made a major difference. (For better or worst, one would be the global shift toward and innovations in capitalism, neoliberalism, and the finance industry that has had large effects on numerous cities and neighborhoods.)

2. If “the physical environment feels depressingly finished,” does this mean a change in aesthetics or style could alter this? Science-fiction films and shows tend to depict cities as white, gleaming, and move curved than they are today. Think Her which merges city life and technological change. Or, find images of cities from researchers, activists, and architects who imagine much greener cities full of plants and life rather than hard surfaces and cars. Perhaps the problem is not innovation as it is described in this article; one issue is that the look of big cities has not changed much in the fifty years or so (even as some individual buildings or projects might stand out).

3. If the look and feel of cities has not changed as much recently, could “the invisible realm of bytes and code” bring significant changes to the physical environment in the next few decades? In contrast to #2, perhaps future innovation in spaces will be less about collective experiences and aesthetics and more about changed private experiences. Imagine Virtual Reality in cities that allows pedestrians to see or overlay different information over their immediate surroundings. Or, easier access to Big Data in urban settings that will help individuals/consumers make choices.

Tech jobs continue to congregate in particular metropolitan regions

A new analysis looks at where tech jobs located between 2005 and 2017:

Researchers from the Brookings Institution and the Information Technology and Innovation Fund, a tech-industry-backed think tank, arrived at their conclusion by looking at a fairly narrow slice of jobs—13 industries that involve the highest rate of research and development spending and STEM degrees per worker. That includes much of the software industry, as well as jobs in areas like pharmaceuticals and aerospace. The researchers found that, between 2005 and 2017, five metro areas—San Jose, San Francisco, Seattle, San Diego, and Boston— not only added lots of jobs, they were also becoming more dominant in those industries overall.

TechJobsWired2005to2017

In part, that’s due to changes in what businesses need, says Enrico Moretti, an economist at UC Berkeley who wasn’t involved in the study. The enduring dominance of some tech hubs is somewhat counterintuitive. Technology was supposed to be a democratizing force—the internet and iPhone would make it possible to do innovative work from just about anywhere. But instead, high-tech industries became about proximity to your fellow high-tech workers. Businesses clustered around hubs of investment, in places where skilled workers could stick around after school, hop between jobs, and stay in touch with contacts. That plays out on an individual level too, Moretti says. In recent research tracking the patent activity of scientists as they moved in and out of places like the Bay Area, Moretti found that they were far more productive in those innovative hubs…

The researchers’ point is that it’s hard to build hubs of innovation from scratch—in places where the economy is really struggling, and where there’s little existing tech talent. Instead, you want to start with places that are already buzzing, and through a mix of investment—in things like R&D, education fellowships, and financing for small businesses—and tax incentives to encourage new business, nudge them to become innovation hubs. In other words, those places are already fertile ground for high-tech companies, but they need a little more fertilizer to get there. The researchers prefer federal investment to local subsidies that try to attract individual businesses—an often fruitless effort for smaller communities, as incidents like the downsized Foxconn factory in Wisconsin and Amazon’s HQ2 search demonstrate.

How exactly these centers of industry arise, thrive, and consolidate (and then maybe fade away or die?) is a good subject of academic study. Through a series of decisions, conditions, and good circumstances, agglomerations start. Inertia can carry them for a long time. As noted in the last paragraph, it can be difficult to introduce competition from other centers or create new centers once the main locations are well-established. Tech center do not just happen; they are the result of multiple social processes, interactions, and decisions.

Additionally, it is interesting to see that there is still a lot of value of actual physical locations near other businesses or organizations – even in a field that can render spatial and time distances less relevant. Being close to other people, being able to actually stop by or talk to them, still matters. All of this can add up to a location with a collection of similar organizations being more than the sum of its parts.

Yea! The Internet enables American workers to work more

A working paper links the number of hours American white-collar employees put in and the Internet:

In a new working paper, the economists Edward E. Leamer, of UCLA, and J. Rodrigo Fuentes, of Pontificia Universidad Católica de Chile, studied data about working hours from the American Community Survey. They found that hours worked since 1980 increased nearly 10 percent for Americans with bachelor’s and advanced degrees. Leamer told me that he believes this is because computing has shifted much of the economy from manufacturing to neurofacturing, Leamer’s term for intellectually intensive white-collar labor that is often connected to the internet, such as software programming, marketing, advertising, consulting, and publishing.

Neurofacturing jobs lend themselves to long hours for several reasons, Leamer said. They’re less physically arduous, as it’s easier to sit and type than to assemble engine parts. What’s more, the internet makes every hour of the day a potential working hour…

As Leamer and Fuentes write in the paper, “The innovations in personal computing and internet-based communications have allowed individual workers the freedom to choose weekly work hours well in excess of the usual 40.”

The internet has also supercharged global competition and forced international firms to outwork rivals many thousands of miles away. This has created a winner-take-all dynamic that’s trickled down to the workforce. In their 2006 study, “Why High Earners Work Longer Hours,” the economists Peter Kuhn and Fernando Lozano found that the premium paid for longer workweeks has increased since 1980 for educated workers, but not for less educated workers. Their theory is that at the most competitive firms, ambitious workers putting in super-long hours are sending a clear message to the boss: Promote me! And the boss isn’t just getting the message; he’s actively soliciting it. At many firms, insanely long hours are the skeleton key to the C-suite and the partner track. Thus, overwork becomes a kind of arms race among similarly talented workers, exacerbated by the ability to never stop working, even at home. It’s mutually assured exhaustion.

“Mutually assured exhaustion” is the result of zealous workers, managers asking more of employees, or the product of a unique work ethic in the United States?

This could lead to a basic question that I ask myself from time to time: has the Internet made life better? Is humanity thriving more, feeling better, doing more good, and experiencing a better life because of the Internet? The personal comparison is harder in that I was much younger when the Internet was not available but I can still imagine the comparisons. How might my academic work be different? My family life? My leisure time? And so on.

Additionally, the study also seems ripe for a comparison to other countries around the world that also have the Internet. Is the Internet the driver here or a tool that the American economic and social system utilizes to push a particular kind of work and approach to life? The Internet is not all powerful and cultural and social decisions in other societies seem to provide room for pushing against the possibility of working all day that the Internet allows.

Palaces for the People, Part 4: Facebook community versus physical community

I recently read Eric Klinenberg’s 2018 book Palaces for the People. Today, I highlight the last of four passages from the book that make some interesting connections regarding physical places.

Returning to social media toward the end of the book and specifically discussing Facebook, Klinenberg suggests online interaction is not a good substitute for interaction in real space in real time:

But no matter how the site’s designers tweak Facebook content, the human connections we need to escape danger, establish trust, and rebuild society require recurrent social interaction in physical places, not pokes and likes with “friends” online. (212)

This is a regular theme in the book: social media interaction cannot match social interaction that takes place face-to-face in real places.

I would guess social media platforms will try really hard in the next few years to change their platforms to encourage more positive social interactions. Some users already work hard to avoid negative interactions. Facebook, for example, is pushing community groups more. Instagram is hiding likes. Twitter is allowing people to hide responses to their posts. Will this all work? Possibly. But, Klinenberg argues that all of these efforts can only go so far. Humans will still need physical places that encourage interaction, trust, and new ideas.

Imagine social media in ten years that is primarily made up of positive interactions. Perhaps then it will be criticized for largely hiding negative emotions or conflict. Perhaps it be dull in the way that endlessly cheery stories might be. Or perhaps it will be seen as a supplement to offline relationships rather than competition for them.

Another way to think about Klinenberg’s ideas: what do public spaces need to be in order to entice people away from social media? There are ingredients that make public spaces more interesting such as a regular flow of people, a variety of activity, a human scale, and perceived safety. Do we have enough of these to truly people people way from their smartphones and if not, how much work would it take to develop spaces like this all over the country?

Palaces for the People, Part 1: building relationships in physical proximity

I recently read Eric Klinenberg’s 2018 book Palaces for the People. In the next few days, I will highlight a few short passages from the book that make some interesting connections regarding physical places.

In a discussion of relationships and social media in Chapter One, Klinenberg concludes:

Building real connections requires a shared physical environment – a social infrastructure. (41)

Research on social media tends to back this up: meaningful or lasting relationships on social media are often grounded in offline interactions and relationships. Social media may be particularly good at helping people maintain connections over time but many social media relationships have roots in or also take place offline. These deeper connections take place in particular settings. Physical spaces can help foster social interaction and togetherness.

This reminds me of Herbert Gans’ conclusions about the lives of teenagers in an early Levittown: there was nowhere for them to go. If there are not tangible physical spaces for young adults to gather (a role formerly played by the shopping mall), then the smartphone and social media look more attractive. Communities may struggle to find places for teenagers to go and be welcome – for example, even shopping malls did not necessarily want them – but the alternative may be worse.

Could AI ever replace diplomacy?

A thought from watching the impeachment proceedings: the relationship between major countries hinges on the personal interactions of a relatively small set of people on each side. The interests of the United States, a global power with more than 327 million people, come down to personal interactions between diplomats. However deep the Deep State might be, a small set of relationships matter for all countries in how they get along with other countries.

In the world of the Internet and computers, does it seem feasible to replace person-to-person diplomacy with Artificial Intelligence? Two humorous examples suggest this could be very hard:

1. The diplomacy built in to the computer game Civilization that never seems to work that well.

2. The arduous negotiations that can occur in the board game Diplomacy over relatively simple moves.

But, imagine the possibilities. A much reduced diplomatic staff! Quicker negotiations! Being able to blame an algorithm for mistakes rather than people!

Ultimately, would governments trust artificial intelligence to put their diplomatic fate in its hands?