More about the automation revolution

In an article titled ‘Why The World Hates Silicon Valley‘, Newsweek makes some points that expand upon the Guardian’s six-part series on artificial intelligence that I mentioned yesterday.  Here’s an excerpt.

Artificial intelligence is a game-changing technology, much like cloud-based apps over the past five years. It will be the basis for inventions we can hardly imagine now. (How about an AI-driven tiny drone that learns to buzz around and keep an eye on a building, replacing security guards? It’s coming!) And 3-D printing will get good enough so that a company like Nike will no longer make shoes in Asia and ship them back to the U.S. Instead, it will “print” them in a network of thousands of small factories peppered throughout cities and towns—so you can pick up your ready-made sneakers locally. Blockchain—the complex technology behind bitcoin—is only beginning to remake the financial industry. Virtual reality will get good enough to reinvent stuff like tourism, sports and doctor’s office visits. Biotech, robotics—an incredible array of technology is ready to burst upon us.

The impact will be so dramatic, Hemant Taneja of General Catalyst Partners tells me we’re heading into a “global application rewrite.” We are about to take apart every product and service in the world and put it back together with data, AI and all this other new stuff.

. . .

If you pick up your mobile phone, you’ll see a lot on there that you used to pay for and now comes free or cheap. You have a camera and a flashlight, both of which you used to buy. News is free—no need to buy a newspaper. International calls are cheap on Skype. Music—free or cheap on Spotify.

That device is just one example of the impact of technology and globalization. It’s increasingly making more things cheap or free, in many ways lowering our cost of living. That works on physical goods too—tech and global manufacturing are why you can buy nice clothes at H&M for way less than similar items cost 20 years ago. Technology will only accelerate this trend. Mike Maples, partner at tech investment company Floodgate, tells me we’re heading into an age of abundance, when we’ll have access to much more for much less than ever before. We’ll live better lives on less money. Which seems quite good.

However, … that same dynamic crushes the middle class by killing jobs and shrinking salaries. If more stuff is free or cheap, fewer people can earn money making and selling things. Instead, when something gets reduced to a cloud-based app, relatively few people can make it and sell it around the planet—and rake in all the money. Consider maps. Lots of companies used to print them, and lots of stores sold them. Today, there’s one consumer map company that matters globally: Google, based in Mountain View, California. Google gets all the map money, and most of those map jobs are gone.

For much of the world outside of Silicon Valley, the bad is starting to feel worse than the good. We love our phones and apps and cheap things, but we don’t like feeling economically marginalized.

There’s much more at the link.

It’s certainly thought-provoking stuff.  I’ve been saying for years that a great many ‘conventional’ jobs are under threat from the rise of artificial intelligence, robotics and automation in general.  I fear that far too many people are simply shrugging their shoulders and ‘giving up’ because they don’t think there’s anything they can do about it.  Others are turning their faces away from the prospect and refusing to face it, pretending it won’t happen to them.

I fear it’s going to happen to a lot of us.  We need to be preparing our game plan now, so that we can re-train for jobs that won’t be automated out from under us, or make other plans for a sustainable future for ourselves . . . because nobody else will.



  1. One has to look no further than the Japanese car industry… The number of salaryman jobs lost to robotics now exceeds the number of actual employees remaining…

  2. I read a pretty astute comment about technology displacing workers. His point was that if no one was employed and gained no money from that, how would they be expected to afford the now low priced materials ?

  3. If automation displaces enough workers, the conditions might be ripe for another "industrial revolution" type event.

    Investing in undersea mining and hydroponics might work out if enough people need a new line of work…

  4. I guess I'm an outlier on this. I started programming over 30 years ago and have been hearing the promise of AI for that entire time. The biggest win in the AI department recently is AlphaGo vs. Lee Sedol. The impressive part is the machine learning that created the neural network didn't require human intervention to learn how to win.

    On a very restricted problem. On 1200 CPUs and 176 GPUs.

    In other words, to play against 1 human, it required a building-worth of computing equipment and it's own electrical substation, and unknown training time before the match.

    And Moore's law has ended. We will hit the atomic limit in the next few years, and transistors won't be getting smaller. So those numbers will not be going down geometrically.

    I'm not worried about the machine overlords anytime soon … or ever.

    Lastly, the idea that custom manufacturing via 3D printing will EVER be less than 2-3x cost of mass production is silly. The idea of printing sneakers instead of rolling them off a factory line will always be preposterous.

  5. This is simply creative destruction. It's nothing new, and it's no worse than it has been in the past. Yes, many fields are being rendered obsolete and their tasks automated. That's good. That's how you get productivity increases.

    It will be tough for people who have their career automated out from under them. It always is. However, assuming that it's better to have them continue to do work that could be more efficiently done by a robot is just another form of the broken window fallacy.

  6. Mark – agree and disagree. The advancement of hardware is slowing, no doubt. But hardware isn't the limiting factor in AI at this time. It's software. And that is still advancing rapidly on some fronts; it only takes a couple of breakthroughs to make a huge difference. Example: a friend of mine works at MS research. He was brought in to help with a problem they were working on. the details are irrelevant, but a problem they'd been working on with a team for a couple of years was stuck at something like 2.5 frames per second. He spent a week or so examining the problem, took a different approach, and whipped up a solution that on the same hardware offered a ~4,000 fps solution.

    It will not be overnight. It will be here and there, hit and miss. But once the functions and economics for one building or industry are proven, it'll sweep the industry rapidly. I expect fast food kiosks will be rolled out rapidly in the near future, for example. A long time coming, but with $15/hr min wage laws hitting the books spurring adoption, it'll spread rapidly from those centers.

    When the economic downturn and resulting protectionist legislation finally hits home and hard, a lot of industries will see the economics shift, and changes will happen fast.

  7. Mark has a lot right, AI has been "just around the corner" since the 70's

    manufacturing is getting less ridged than it has been in the past, but you still want to keep the machines busy 24×7, and the delivery issues still favour larger companies that can make the investments in the machines and the maintenance for them.

    What's changing is less that small companies are manufacturing things themselves and more that the large companies are able to customize what they manufacture more easily. This includes being able to use their assembly lines to manufacture small batches of things for small companies.

    3D printers are seldom the right thing to use for production parts (there are exceptions), where they revolutionize things is the turn-around for prototypes of custom designs.

    another area they work well is in printing prototypes that are then used to create molds to cast the final parts.

    The areas where 3D printers are producing production parts are where they are printing in metal and they are producing components that have internal structure that just cannot be machined by traditional methods. This is starting to have a significant effect in aerospace (the new GE jet uses 3D printed fuel injector nozzles for example)

  8. Folks on the left like to ignore that each person has an educational ceiling. Tech and Engineering people think increased productivity is the prime goal of work. Not everyone can learn to do high tech work, what will you do with them? I've thought for years that increasing minimum wage increases incentive for automation, and disqualifies people from the labor market. Liberals complain that Walmart and McDonald's and grocery stores don't pay a "living wage", and force the issue, which leads to automated checkouts and ordering, which allow these companies to employ fewer people. Now, I won't say corporatism isn't part of the problem, as management is always pressured to increase profits for shareholders.

    @Mark: As to the idea of custom printed shoes being preposterous, they're already driving the research cost down. Printed casts are coming. Even if they start more expensive, they'll drop with tech improvements.

Leave a comment

Your email address will not be published. Required fields are marked *