A piece in the New York Times Andy Grove's Warning To Silicon Valley points to Groves's essay from 2010: How America Can Create Jobs. Groves, who served as Intel CEO for many years, saw a big problem in the attitudes in American business toward manufacturing. He saw the shift of factories abroad as robbing the United States of much needed expertise and competitive advantage.
This downgrading of the importance of manufacturing has fallen most heavily on the less intellectually able. As I've pointed out previously, the employment:population ratio spans over 29 percentage points between high school drop-outs and those with at least a bachelor's degree in college. While the high school drop-outs are around 42-43% employed (unemployment rates are misleading) the bachelors and up are around 72% employed. The high school grads are around 55% while those with some college courses are around 63-64%.
What we are seeing is cognitive stratification of the labor market where the future is still bright for the brightest minds but increasingly dimmer at lower levels of intellectual octane. The factory jobs for the low skilled are getting automated and/or moved abroad. Even if businesses were to heed Groves' warning (unlikely) this would only slow the rate of decline of the prospects of most potential workers. Automation seems set to eliminate more routine jobs and boost the cognitive demands of the average job.
What I'm left wondering: will robots work for the poor unemployed masses?
Are you ready to deal with a computer interface to order a burger and fries? The online survey at the MarketWatch site has 73% of the readers saying yes. The numbers vary quite a bit by state with Oregon having only 45% in favor of skipping the human order takers but 77% in Virginia in favor. The age 18-24 year old people are most in favor followed by 25-34.
This could be rolled out first in high volume restaurants with both human and computer order taking. Throw in a downloadable app so that people can order while en route. People will opt for the fast way whenever there is a line to talk to a human worker.
Using tablet interfaces will be faster. No waiting in line. You won't have to repeat yourself when misunderstood. Even better: be able to order and pay online before you get to the store. Then the order will be ready sooner after you walk in the door. I've done this with Domino's Pizza several times.
Start-up Eatsa has introduced auto-ordering in San Francisco. This starts to bring back the Horn & Hardart Automat of the 1930s and 1940s.
Seriously, what's your favorite fast food chain? Would you still go and order if it did not have a human order taker?
My guess: People do self-serve gasoline just fine. They use self-serve ATMs. They order a rising fraction of their purchases online. They buy from vending machines. I think most people will shift to self-serve restaurant ordering just fine. Some fast food chains could just switch over entirely. They might lose some customers while gaining others.
At lunch time restaurants could even gain customers (especially with pre-order) due to the need for employees to get their food quickly and get back to work. The ability of a work group to all order from their desks and then walk to a restaurant down the street, pick up food immediately, and start eating would be a big plus.
Tyler Cowen reviews the latest research: Silicon Valley Has Not Saved Us From A Productivity Slowdown. Seems persuasive to me. But then I already believed the argument.
Why I already believed the argument: An opposing viewpoint is that the benefits from computers are underappreciated due to quality and convenience improvements that do not get captured in productivity measures. But that opposing viewpoint does not explain why the computer revolution hasn't just boosted productivity growth even higher. So why wouldn't the older forms of innovation for productivity growth also continue while the computer innovations got put on top of them? Granted, some of the older ways to, say, improve a factory got replaced by computers that control what more purely mechanical systems would have done. But that sort of improvement by computer should have shown up in the productivity stats just as improving old mechanical systems would have.
The benefit to one's personal life so far seems oversold. Certainly I get a lot of benefit out of the internet in a variety of ways. I no longer go to stores like Target or Wal-Mart or even the post office. I read far more and can get far more information at my fingertips. But in the rest of my life there's been less change. Airplanes haven't gotten any faster. Rather, travel has become a bigger hassle. Reliability in cars went thru a big improvement in the 80s (bye bye points and condensers) and 90s (Japanese makers raised the bar), but seemingly not so much since. Most remaining household chores aren't going away any time soon. We aren't living in the sci fi future yet.
Looks like driving is going to get automated in 10-15 years time. That'll wipe out truck driving, taxi driving, and other driving jobs. Lots of the remaining agriculture jobs will get automated by autonomous tractors and autonomous harvesting equipment. But it is less clear when medicine will go thru an automation revolution.
So what's your sense of it? Is productivity soaring much faster than measured and there is no problem? Or is there a problem but it will get addressed in the next 10-15 years by machine learning models doing more and more things for us? Or will the slow down last a long time?