🤖 The return of automation anxiety
Also: 5 Quick Questions for . . . Scott Kupor of venture capital firm Andreessen Horowitz
“Technology is a gift of God. After the gift of life it is perhaps the greatest of God’s gifts. It is the mother of civilizations, of arts and of sciences.” - Freeman Dyson
In This Issue:
Long Read: The return of automation anxiety
5QQ: 5 Quick Questions for . . . Scott Kupor of venture capital firm Andreessen Horowitz
Micro Reads: Cellular rejuvenation, autonomous air taxis, fashion in the metaverse, and more . . .
Long Read
🤖 The return of automation anxiety
In the Before Times, there was a strange surge in automation anxiety. Lots of books and think pieces about “robots taking all the jobs.” Those concerns were spurred partly by numerous reports about big advances in machine learning and autonomous driving — not to mention numerous Boston Dynamics videos of automatons doing parkour. Recall Andrew Yang using the automation issue to make a splash in the 2020 Democratic primaries. As Yang predicted to The New York Times: “All you need is self-driving cars to destabilize society. That one innovation will be enough to create riots in the street. And we’re about to do the same thing to retail workers, call center workers, fast-food workers, insurance companies, accounting firms.”
At the same time such fears were gaining purchase, however, the US labor market was gaining momentum. By 2015, the long post–financial crisis expansion was slowly increasing wage growth — especially at the bottom of the income distribution — and boosting prime-age labor participation. By February 2020, the unemployment rate was at its lowest level since 1969. And employment rates for the workers with the least education had risen higher above their pre-crisis levels than those for college graduates. (For more, check out the wonderful and highly readable paper “The Recovery From The Great Recession: A Long, Evolving Expansion” by Jay C. Shambaugh of George Washington University and Michael R. Strain, my colleague at the American Enterprise Institute.)
But maybe automation anxiety wasn’t wrong, just early. Even as our focus has been consumed by the Covid-19 pandemic, AI and robotics have kept getting better. Deere & Co. recently announced that it had developed a fully autonomous tractor designed for large-scale farming, and that it plans to sell the machine later this year. Robot umpires will be calling balls and strikes this year in Triple-A, the highest level of the baseball’s minor leagues, with a promotion to the majors seemingly inevitable.
Then there’s all the kiosks. The retail sector had five fewer cashiers and salespeople per hundred workers in June 2021 versus the average in 2019, according to a new Goldman Sachs analysis. The leisure sector has moved in a similar direction, GS adds, “in part reflecting the rise of mobile ordering, online check-in/check-out, and housekeeping service on request.”
And it’s not over. Going forward, the bank’s economic forecasts assume a significant substitution of in-person sales and support staff with remote or automated services. “These business model changes have outlasted — or in the case of the leisure sector, occurred after — the spring 2020 lockdowns, and they have not reversed in 2021 despite a sharp rebound in retail traffic and dining, hotel, and recreation services consumption. . . . Specifically, we assume 1.4 million fewer in-person sales staff in the private sector economy. . . .”
I’m not going to spend time arguing that robots won’t take all the jobs. (If you’re looking for that kind of thing, check out this short/sharp analysis by Sam Korus over at Ark Invest.) But I would point to the December 2021 paper, “What Are the Labor and Product Market Effects of Automation? New Evidence from France” on the subject. It investigates the effects of automation technologies on employment, wages, prices, sales, and profits at French manufacturing companies between 1995 and 2017.
At all levels of analysis — plant, firm and industry — the relationship between automation and employment is positive, indicating that in practice the productivity effect tends to outweigh the displacement effects. . . . These patterns can be explained by a simple consumer demand substitution channel. After adopting automation technologies, firm owners increase their profits but pass through some of the productivity gains to consumers, inducing scale effects. Automation can thus lead to higher firm profits, lower consumer prices, increased consumer demand, and in turn to increased firm and industry scale, higher labor demand and higher domestic employment at the expense of foreign competitors.
Anyway, there are concerns about automation other than mass joblessness. A recent New York Times piece by reporter Steve Lohr, “Economists Pin More Blame on Tech for Rising Inequality,” points to inequality as a reason for concern:
Half or more of the increasing gap in wages among American workers over the last 40 years is attributable to the automation of tasks formerly done by human workers, especially men without college degrees, according to some of his recent research. . . . So-so technologies replace workers but do not yield big gains in productivity. As examples, [economist Daron] Acemoglu cites self-checkout kiosks in grocery stores and automated customer service over the phone. Today, he sees too much investment in such so-so technologies, which helps explain the sluggish productivity growth in the economy. By contrast, truly significant technologies create new jobs elsewhere, lifting employment and wages.
This isn’t an anti-innovation argument. Rather, the NYT piece is suggesting that the US economy is generating too much innovation of the sort that replaces workers and not enough that creates new things for them to do. Digital examples of the latter cited by Acemoglu include databases, spreadsheets, search engines, and digital assistants.
Which brings us to the public policy question: Should government try to correct this supposed innovation imbalance? One possible action would be to alter how the tax code treats hiring a human versus buying a machine. Here’s Acemoglu in a podcast chat with me back in 2017:
We implicitly subsidize production with machines relative to production with labor. If you buy a machine, you don’t have to pay payroll taxes, you can debt finance it, and that’s going to get a subsidy from the government. And the capital income that is generated is going to be taxed at a lower rate. If you hire a worker, you have to pay payroll taxes, they are going to be taxed at a much higher rate, you have to put up with lots of other costs coming from regulations, and other things you have to do when you’re employing workers. So that creates a very non-level playing field, so we’re essentially subsidizing firms implicitly for using machines rather than labor. It’s probably a good idea to start thinking about how we can even that playing field a little bit [although] I am definitely not suggesting a robot tax so that we can raise revenues.
This is a serious and thoughtful proposal that I’ve heard echoed by other economists, as well. But it’s not my favorite for several reasons:
There’s a lot of gray area and overlap between job-displacing and job-enabling technologies. Both replace workers and both make them productive. Although ATMs, really just glorified kiosks, have meant fewer bank tellers per bank branch (though they’ve also resulted in more branches), the job of a teller has changed into one that’s more of a sales consultant and financial products adviser.
Moreover, who knows exactly how a technology will evolve over time, and how those changes will affect its job-displacing and job-enabling attributes? The need for that sort of foresight reminds me why I’m skeptical of many industrial policy ideas.
Another way this time-horizon factor plays out is with the displaced workers themselves. They all don’t stay displaced forever, sitting at home. And making sure they re-enter the labor market is where policy has a stronger role rather than in tinkering with the tax code. I’m thinking here of earnings subsidies, relocation assistance, and work-based learning programs where market needs determine skills rather than government bureaucrats.
Finally, if business isn’t investing enough in awesome tech rather than so-so tech, maybe we should think about policies that would address that issue, from greater federal R&D funding to rolling back regulation that makes it hard to innovate in atoms rather than bits. I worry about a policy response where business ends up investing less in machines overall — something that our uninspiring productivity stats suggest would be a bad outcome.
5QQ
❓❓❓❓❓ 5 Quick Questions for . . . Scott Kupor of venture capital firm Andreessen Horowitz
Scott Kupor is an Investing Partner at the venture capital firm Andreessen Horowitz, where he is also responsible for all operational aspects of running the firm. In addition, Kupor is the author of The Wall Street Journal bestselling book Secrets of Sand Hill Road: Venture Capital and How to Get It. He recently wrote an essay, “On Supercities, Economic Growth, and Income Inequality in a Post-COVID World,” that I really liked and wanted to ask a few questions about.
1/ Why do you think that pandemic/remote work has undermined the power of location enough to create the potential for significant innovation hubs in places other than the “supercities” — especially in the Rust Belt or “left behind” America?
I am not sure we know enough to say this is a permanent shift away from the innovation hubs, but for the immediate term, we have already seen the ex-migration. Simply put, I think Covid forced us to reconcile our theory on workforce participation (that distributed workforces could not succeed) with the reality — we figured it out over the last two years and have realized that our theories do not match reality. In fact, in some ways, the reality is superior — workers feel more in control of their lives and employers benefit from alleviating their geographic constraints on recruiting.
2/ There’s a strong economic consensus about the difficulty of building innovation clusters. Do we really understand how to mix those ingredients (startups + entrepreneurs + institutions + policies + universities + government R&D + VC funding + favorable legal conditions) properly? After all, Silicon Valley was a happy accident in many ways.
I don't think we fully understand the mix of ingredients yet, but we do understand network effects. We know that building networks requires a balance of supply (workers with the appropriate skills and "centers of excellence") with demand (employers in need of that confluence of skills). Thus, the easier way to create a network effect will be to focus on specialization, or comparative advantage, versus trying to go too broad. If innovation clusters figure out what they are experts in and build funding, R&D and educational centers around this particular domain, we are more likely to create successful clusters.
3/ The notion of government as venture capitalist and investing in applied research, or even young companies, has gained some momentum in Washington. Given your experience in the VC business, what are the odds of government success here?
The government should not play venture capitalist. Rather, the government should focus on the "R" in R&D and leave it to the private sector to work on the "D." This means broadening funding to fundamental science, but also doing so in conjunction with the idea of creating research (and ultimately) innovation hubs. Thus, rather than sprinkling basic science funding across all geos, the government should effectively function as a kingmaker by matching specific geos with specific research mandates.
4/ How much should we really worry about countries such as China developing “centralized mechanisms to invest in core enabling technologies (e.g., artificial intelligence)” given the long history of governments being bad at developing frontier technologies from the top down?
You're right about the history of top-down frontier technology development — although one might argue that the Kennedy-sponsored space race was one positive counter-example that ultimately did result in success! I believe that what matters most, however, is whether a country is the most attractive place for researchers and entrepreneurs — in terms of workforce, resources, economic incentives — a place that the US has occupied for a very long period of time. We are threatening that supremacy with, among other things, our immigration policies and increasing regulatory scrutiny on entrepreneurship. If we continue down that path, I think the risk is that capital — both human and financial — flows to countries that become more attractive.
5/ What do you see happening in Washington that gives you confidence in its ability to execute on its part of the formula you outline in the piece — or is it even open to that formula?
It's hard to have too much confidence these days given the challenges in getting any legislation done in Washington these days, but, ultimately, politicians (and all others) respond to incentives. I think we are finally at a point where most politicians recognize that income and opportunity inequality must be addressed in part by creating opportunities for economic growth — broadly across geographies traditionally under-served. Pork barrel politics is of course not new; but perhaps now the crisis of inequality may be an impetus toward constructive pork barrel politics.
Bonus: If there is no New Roaring Twenties of faster growth, what probably went wrong?
We have every opportunity in front of us to enjoy the fruits of economic growth; this game is ours to lose. If we in fact lose, it will be of our own doing — because we failed to maintain our supremacy as the most attractive place for smart, innovative, risk-taking individuals who want to better themselves and others by creating important, lasting and high-growth enterprises. That could be through a combination of restrictive immigration policies that limit our ability to attract and retain innovators globally and through the imposition of other regulatory policies that entrench incumbents and make it harder for startups to challenge the status quo.
Micro Reads
🧬 Billionaires Bankroll Cell Rejuvenation Tech as the Latest Gambit to Slow Aging - Michael Eisenstein, Scientific American | Yamanaka factors, which enable researchers to reprogram a regular somatic cell into a pluripotent stem cell, show promise for rolling back the clock on cellular aging. Several startups, funded by billionaires such as Jeff Bezos, are racing to leverage Yamanaka factors for anti-aging therapy. And breakthrough innovations in other fields are aiding the search for cellular rejuvenation: Using machine learning, Shift Bioscience hopes to identify genes that can turn back aging without taking cells all the way back to pluripotency. “This message needs to be clear — it’s not about extending the lifespan,” the piece quotes Vittorio Sebastiano of Stanford University. “What we care about is increasing the healthspan of people . . . and that you don’t have to live for a long time in a condition of frailty.”
🚁 Boeing pumps extra $450m into self-flying air taxi start-up - Sylvia Pfeifer, Financial Times |
“Autonomy is the key to unlocking scale across all advanced air mobility applications, from passenger to cargo and beyond. That’s why straight-to-autonomy is a core first principle,” said Marc Allen, chief strategy officer at Boeing. Rival plane makers Airbus and Embraer, alongside a host of start-ups, are developing their own electric air taxis. It has been one of the hottest markets, despite most start-ups not yet having a flying prototype able to carry passengers Investors last year poured more than $7bn into companies focused on all types of future solutions for air mobility, according to analysis by McKinsey, more than doubling the funding of the previous 10 years in the sector. Overall funding has reached $12.7bn since 2010.
🩸 What Elizabeth Holmes and Theranos Reveal About Venture Capitalism - Sebastian Mallaby, The New York Times | The fall of Theranos and the fraud convictions of founder Elizabeth Holmes has catalyzed criticism of venture capitalists as incompetent gamblers. But we’d sure miss VCers if they were gone. Mallaby points out that the fewer than 1 percent of US firms that receive venture funding account for 47 percent of publicly traded, nonfinancial companies and 89 percent of R&D investment by firms on the stock exchange. Venture capitalists expect most of their investments to fail and a few bets to pay off in a big way, but these aren't blind gambles. VC firms serve an important market function by identifying coming trends and investing in startups well-positioned to take advantage of the new wave.
🔬 The Nanotechnology Revolution Is Here — We Just Haven’t Noticed Yet - Christopher Mims, The Wall Street Journal | We don’t yet have molecule-size machines that “patrol our bloodstream and repair damage” or “microscopic factories capable of churning out endless copies of themselves.” But the present reality of nanomachines is still pretty cool. Moore's Law has given rise to ever-smaller components of technologies we consider mundane, from inkjet printers to airbags. What's next? Mims: "Just around the bend, nanotechnology could enable ultra-tiny cameras, as well as a dizzying array of other kinds of sensors, able to detect everything from air pollution and black ice to hacking attempts and skin cancer."
🛰 Can Starlink Save the American Mother? - Katherine Boyle, The Rambler | Starlink, a satellite internet service operated by SpaceX, is poised to unleash opportunity on the developing world where the lack of broadband infrastructure inhibits connection to the global economy. But Boyle argues that Starlink could also have big implications for the US. Fast, reliable internet in rural parts of America's heartland, combined with the rise of remote work, could unleash an outpouring of elites from America's crowded and expensive coastal cities with implications for federalism, housing, and even family formation and fertility.
🥽 What to Wear in the Metaverse - Vanessa Friedman, The New York Times | Friedman raises the question of how the metaverse will resolve the tension between the limitless creativity of virtual apparel and the need to physically represent yourself in the virtual world. (Well, maybe no tension for Meta boss Mark Zuckerberg who is sticking with jeans and a t-shirt.) Your metaverse avatar is more than a customizable video game character. It is, in the virtual world, actually you. So get ready to be judged, just like in the real world. Friedman writes: “Just because a virtual space is free from certain real life limitations does not mean it is free from the preconceptions everyone brings to parsing character. In fact, absent other clues, like profession, a virtual space makes clothes even more important.”
For customers I would say that 75% of the technology used to assist consumers causes to to three times the time and effort to resolve an issue than talking to a human being. So how is efficiency measured. Sure it is more efficient for the seller of a product or service. However for the consumer it is a frustrating waste of time. Holistically much of the technology reducing jobs simply shifts the “ cost” to the ultimate user and as service and product options shrink due to lack of competition Americans are frustrated and mad as they attempt to accomplish compel tasks.