Much hand-wringing over AI refuses to acknowledge the past. I have said before, and I repeat here, that AI is very likely to be just like email: a “time-saving” “productivity tool” that helps us “work more efficiently”. All the same promises, all the same kind of rollout from slow to rapid. Except in the 90s and early 00s, the rapid adoption of email was tied to expensive infrastructure. Cables in the ground, computer workstations, and staff training. AI has none of those roadblocks.
Like email, AI is the “time-saving” “productivity tool” that helps us “work more efficiently.” All the same hype, advertising, and process improvements as email. And what did we end up with? An endless inbox full of more work. More effort, more hours, and, I think, more stress and more irritation at our degraded quality of life. Great for business, bad for people.
We’ve only recently started to think about turning a corner on setting limits to our evening and weekend email use. This was hardly a problem in 1950 or even 1990. You could only reply to people based on the number of memos you could carry. You could call people, but if you couldn’t get them on the line, you’d just move on. “Oh, they’re not there.” And no one played email hot potato by just shunting “Thoughts?” messages at you.
AI is likely to lead us straight to this same problem. So you can deploy three apps in a day. So what? Now I just have expectations that you can make three good apps in a day. So you can respond to email faster. So what? Now I expect you can reply to my email even faster!. Businesses and customers will come to expect all this and more to be done fast, cheap, and easy. They might even do it themselves.
But the iron law of all economic things is that nothing is ever truly fast, cheap, and easy. These things always come with high costs, whether you bear the costs of those things or not. Cheap jeans are sewn by sweatshop labor. Cheap food is hardly food at all. Fast frequently means unsafe, shoddy, or poorly done. Easy is always relative, and usually means something was barely thought about.
Our entire world runs on solving complex problems. “AI is here to solve complex problems in a fraction of the time!” That’s great, but sometimes things are a little too complex. AI is not going to magically help fund every unmet childhood need, classroom expense, or hunger. AI is not going to help people decide what to build (or not) in their neighborhoods. AI is not going to help you get over a bad breakup or death.
But many people think it will.
This brings me to the heart of this post: just what exactly are we optimizing for?
The Amish and Mennonites are a fascinating case study. We think of them as allergic to technology and, perhaps, backward. But that’s far from true.
Amish communities optimize for community. Anything that potentially harms or degrades their community bonds is rejected. This is why they may have a telephone, or even a small tablet at an Amish country store to handle payment processing. But they have the one, in a defined place, in a defined role, and it does its one job.
We’re not nearly as good at setting these kinds of limits because we don’t optimize for community as they do.
Amish communities investigate new technology all the time. Heck, wouldn’t surprise me if they investigate AI someday soon! They’re not completely oblivious to the world happening around them. Designated members of the community, usually mature adult men, will try new technologies for a while to see how they work, evaluate the value to their community, and investigate risks to the community. Cars, for instance, allow people to spread farther and farther apart. This is bad for building a tight-knit community, so after some thought, they rejected cars.
Amish farmers reject tractors and machines in favor of horse-drawn tools. These tools are “slower”, “less efficient”, and therefore “produce less” than mechanical tools. But when you’re always concerned foremost with the health of the community, you realize the obvious downsides of mechanized tools. They require immense sums of money and debt, damage the land and wildlife (no horse-drawn team ever plowed over fawns), erode topsoil, and lead to row crops (big tractors can’t do much else) that quickly become monocultures because you have to pay the debt with the highest-earning crops. This ultimately reduces variation and forces most people to never let fields remain fallow (“unproductive”), leaving the ground unenriched. “Going into a lot of debt and harming the land that feeds us” is, fair to say, bad for a community. Just look at every non-Amish rural community in America.
We have no North Star in modern Western society. Our civilization is only oriented around money.
You can like that, I guess. I think most people reasonably don’t. There ought to be some things that deserve higher priority, like our community, health and well-being, and future generations.
AI enables none of this. Unless you are one of those people who believe AI will enable people to somehow do no work and yet make an income. That “employers and job creators” are going to share their profits to ensure some kind of Universal Basic Income. People who believe that have paid no attention to the United States and our policy of embracing “creative destruction”. Please don’t argue that we’ll all get “retraining” and be featured in political campaign ads, just like the coal miners, construction workers, and farmers.
I am neither an AI skeptic nor an AI optimist. This stuff is just this year’s telegram/automobile/telephone/workstation/internet thing. But like all of our prior “world-changing” industries, so much is lost because we optimize only for money.
We rebuilt our cities entirely around cars, and to the detriment of non-car users. We went from “An affordable way to get to church in our Sunday best,” to, “A way to sit in traffic for an hour to get to work while driving a 5-year loan repayment plan.”
We reoriented how we communicate based on, at one time or another, party lines, “free nights and weekends”, and now Zoom. And it’s become worse in some ways. We went from “We can have a call with grandma on Tuesday,” to also, “The staff can meet with people in any time zone all the time!”
Some of this is surely fine. It’s good that we can talk to people at long distances. But it also means we’ve separated those people from our community. We’ve “globalized”, a term that can’t possibly be good. We rightly get upset when our congressional districts get gerrymandered to include people from “way out there” in “that other place”, but somehow we’re thrilled to have Zoom calls with people in another country?
It can be good that we can travel long distances with trucks. We can now ship medicine and books to every town in the country year-round. And, evidently, also ship strawberries everywhere in the middle of winter. They taste terrible, cost a lot, and require immense amounts of chemicals and mechanization at a terrible economic cost. But, hey, $5 for a pint in February in Michigan!
We have to get our heads around that anything optimized for money, and certainly anything pitched by a corporation or a government, is going to optimize for things they like. It’s not going to be good for ordinary people. Just as the Amish rightly realized that cars, constantly carrying pocket computers, video games, and loads of other stuff were probably not ideal for their communities and relationships.
We see this with the EU’s regulation of U.S. tech companies. The very weird defense by Americans is some riff on, “Gosh, look at their piddly economic growth in the E.U. Their economy is so small compared to the U.S.! They should deregulate and let everything flourish!” Meanwhile, it seems to me Europeans have just made their own choices on what to optimize for. Things like defending against the whims of a few corporations and a foreign government, personal time, child leave, and other things Americans can’t understand all seem to make them pretty happy and well-adjusted. No ordinary person wakes up and thinks, “Gosh, I hope our GDP grows by 2.9% this year.” So in some sense, we’re all already optimizing for this reality. It’s just that our government and government-by-corporations certainly aren’t.
I hope you don’t think AI is going to bring new wealth to you or your family. It is not. It will, like email before it, bring either more work and a faster pace to an already unnatural pace of life at best or, at worst, unemployment.
It will also certainly bring “growth”, so I guess we can all be thankful for that.
