The Peak and the Trough
I have a new piece up at Discourse looking at what just might be the real productive promise of artificial intelligence, and it’s not what you probably think—or at least, it’s not what we have mostly been discussing for the past year or so.
Writing fake articles and substituting for stock photos (which people used to use in their blog posts instead of AI-generated illustrations) is replacing work that is already relatively low-paid and not, alas, central to the economy….
The comment making the rounds on the internet, in various forms, is that AI should be doing tedious tasks for creative people, but instead it’s doing creative tasks for tedious people.
With that in mind, perhaps the really transformative use of AI, the one that would free up large amounts of creative effort, is one that is largely being overlooked. What we need is the old dream of the household robot servant—an AI Jeeves who will fold your laundry and make sure coffee and breakfast are waiting for you in the morning.
I use this as an excuse to talk about a report I came across recently which describes and quantifies the revolution in household appliances that transformed the 20th Century.
According to a recent study by the National Bureau of Economic Research, “Back in 1900, without household appliances, the average US household spent 58 hours per week on meal preparation, laundry, and cleaning.” As they used to say, every mother is a working mother. Despite the rosy conservative view of the traditional family, you can see why most women didn’t work outside the home. It wasn’t a luxury made possible by their husbands’ well-paying factory jobs. (In fact, most families enjoy a substantially higher standard of living today.) Rather, it was a necessity due to the sheer volume of menial labor required inside the home.
The number of hours required to maintain a household was reduced in 20th Century from 58 to 18. The different is a full 40-hour week—which explains the migration of women into the paying workforce.
I also look at an interesting study about why Western Europeans work fewer hours than Americans. It’s not because they’re living la dolce vita. It’s because they spend more time on “home production”—housework—because they can’t afford as many labor-saving appliances.
But the really interesting statistic is that this revolution in household work leveled out in the developed world by the 1990s.
Household machines have taken away or simplified much of the brute physical work. What remains is tending to the machines: gathering the dishes, sorting the clothes, folding them when they’re dry, preparing the ingredients to be cooked and so on. It’s not a huge amount of labor, but it is a significant expenditure of valuable time, and it can’t be automated mechanically. It requires intelligence and judgment—or something like human intelligence.
The idea of an artificially intelligent robot butler—a Jeeves to our Wooster—is still a long way off. What we have now are the first incremental bits, like a robot vacuum that will help keep your floors clean but is just learning how not to smear around dog vomit. There are some early attempts to devise a robot that can fold laundry, but it can’t yet match up your socks….
Interestingly, a group of executives from Cruise, one of the robotaxi companies, has just raised money for a new startup devoted to “household robots.” It is unlikely we will get the robot butler of science fiction, a humanoid automaton that does all our tedious domestic tasks itself. But that’s as unimaginative as expecting our robot butler to operate an old-fashioned hand-crank clothes wringer. We tend to make the mistake of assuming machines will do things the way we do them, rather than the way that is easiest for a machine to do them. It is more likely we will gradually get a variety of appliances or specialized robots that perform specific tasks. One to gather, sort, clean, fold and return your laundry; another to gather and clean dishes and put them back on the shelves; another to take ingredients from the fridge and make breakfast; another to mow the lawn and weed the flower beds. And so on.
I acknowledge that this is a long way off, “perhaps as far away as our contemporary household appliances were in 1900.” But it’s worth speculating what the results would be, both in terms of the other work it would free up, and perhaps more important in terms of the direct gains to our well-being and lifestyle, a sort of “Downton Abbey Effect.”
The Peak and the Trough
In some respects, this is not the best time for excited expectations about the promise of artificial intelligence, because the whole field seems to be sliding from last year’s Peak of Inflated Expectations down into the Trough of Disillusionment. I’ve been seeing a lot of people on social media making fun of the extravagant and obvious errors produced by ChatGPT and other AI chatbots, and I point out in my article the palpable disappointment that we still don’t have widely available self-driving cars.
An Axios report names the phenomenon.
Grumbles about generative AI's shortcomings are coalescing into a "trough of disillusionment" after a year and a half of hype about ChatGPT and other bots…. The "trough of disillusionment" was first named and defined by consulting firm Gartner in 1995 as part of its theory of hype cycles in tech.
On expert quips, “"No one wants to build a product on a model that makes things up," while another gets more specific about the economic: “The valuations anticipate trillion dollar markets, but the actual current revenues from generative AI are rumored to be in the hundreds of millions.”
This is all a demonstration of the difference I explained in depth some time ago between actual consciousness and the ability of artificial intelligence to process data.
Not being able to tell the difference led both to exaggerated fears of AI and inflated hopes.
But remember that the next stage of the hype cycle is the Plateau of Productivity, when people begin to identify the most useful applications of a new technology and optimize it for those purposes. So it’s still worth spending some time to anticipate what those applications might be.