(No. 20, a ±07 minute read)
Maybe after all it really is people vs. the robots.
If you read this regularly you know that I was in San Francisco in November. The city that was my home for many years and the city that is shorthand stand-in for both the triumphs and excesses of tech and its bros. And, I suppose, shorthand also for the American Right’s understanding of liberal “lawlessness.” Oh well.
I arrived too late and met a dear friend for dinner in the Haight where she lives. We sat at a corner table, in a restaurant on a corner, which afforded me a view of an intersection where a stunning number of cars cruised by without occupants—robots operated on city streets by Google. It was uncanny and I could not take my eyes off of them when they passed by.
Flippantly, but not wrongly, my friend dismissed them as the most recent manifestation of disruptive tech that San Francisco residents have been forced to accept as merely the latest tech fait accompli in their lives. Famously, Cruise (owned by Chevrolet) has pulled its cars after the catastrophic dragging of a pedestrian at Fifth and Mission, my old neighborhood. According a source speaking with the NY Times each Chevy Cruise outfitted for autonomous driving costs between $150,000 to $200,000 dollars and requires a staff of one and a half people when it’s on the road.1 Cruise had 400 such cars in its fleet, half in San Francisco.
That’s a lot of scratch that could be paying a lot of cabbies who would likely stop after driving over someone at five miles-per-hour while pulling to the curb.
But that’s not how it works. Move fast and break things has been tech’s mantra. This is another way of describing anti-social behavior; just gussied up in piratical business-speak; a language all about renaming ordinary things, intending to obfuscate, in the service of justifying an often predatory white collar workforce’s existence and influence. Finance has been doing it for decades.
Last week Google abruptly caused its AI engine, called Gemini, previously Bard, to cease rendering people. This was done after a scalawag asked it to generate pictures of German World War II “solidier” to get around its refusal to do so when spelled correctly.
Oops. Is 25% accuracy enough to allow for “breaking things” while “moving fast?” I suppose that depends on how bad the public outcry might be? The tech industry hasn’t always had the best ability to gauge that, as anyone who lived through either of the dot-com booms in the city can attest. A heap of money thrown at people with maybe poor senses of history, moral reasoning, social structures, and maybe chock-a-block full of greed—what could go wrong?
As an aside it is worth noting, by way of cadging a sense of tech’s self-image, that this AI engine was initially named “Bard",” presumably after the West’s most famous literary figure and ultimate chronicler of human compulsion gone sideways. As if giving an AI engine that, at best, cobbles together sentences that are legible, a name evoking Shakespeare isn’t arrogance enough changing it to Gemini, either to evoke the remarkable work of our second manned space program or the constellation associated most closely with the twins Castor and Pollux may be telling. (Pollux, through no fault of his own, was fathered by Zues who, masquerading as a swan, raped Leda.) Both mythological figures had great careers of slaving, theft, and kidnapping (of their wives-to-be), and became enshrined as beacons of immortality and death, shining from the Gemini constellation.
It seems that at some level the folks at Google must be in on the joke that is being had at our expense with that name change—or are they just not up on Greek mythology? Or unable to google things? Capitalism, as it has been entrenched today with shareholder value preeminent over all other considerations, is by its nature anti-social. I suppose one can’t fault Google for either comparing its AI engine to Shakespeare, for all the hubris that might evoke, or for renaming it after Castor and Pollux once it left prose behind and became able to envision equal-opportunity Nazis.
It’s not novel to list tech’s casualties but it is worth reminding us: local news papers fell initially to Craigslist’s online classified ads, cabs were prey to unregulated ride-sharing, cohesive local neighborhoods in tourist destination towns were consumed by short-term vacation rentals, civil discourse (as much as we had it) was exploded by anonymity on the internet and social media—which fed off the carcasses of newspapers and magazines. People who study such things tell us social media is also consuming our and our children’s mental well-being. These are the big casualties.
At what point in the race to ultimate AI is it worth engaging the precautionary principal? Among the professionally global warming-concerned, Jevon’s Paradox2 is never far off when discussing power efficiencies or increases in sustainable power generation—should some incarnation of it be applied to AI and its ability to potentially eat white collar and creative labor for breakfast? The “better” AI becomes the more it will be wielded and disrupt social systems for the worse, and against societal interests?
Remember when modernity’s efficiencies—phones, computers, industrial robots, etc—were going to bring us increased leisure time? How’s that working for you?
Me too.
My sister is a SAG member who was on strike for a contract that prevented studios from using her likeness to generate, with AI, characters created from her likeness, depriving her of work. I am a journalist who makes images and writes prose that, on the surface, can be duplicated by AI engines. These are occupations that at some level are under threat, as were factory workers when automated assembly protocols began to be phased in and more recently cabbies facing the rise of Uber and Lyft. There are many such examples of older paradigms of work being upended dating back to the Luddites smashing looms.
To be sure not all such upendings are bad—welding robots can make better welds than people, remove people from hazardous environments, and don’t get repetitive stress injuries. But what of those welders ousted from their means of making a living? We left them out in the cold, and, frankly, this has contributed to the rise of populism in the U.S. and around the world. Why should we “knowledge workers” get any special treatment that a welder did not see as merelythe latest batch of people perhaps on the threshold of labor disenfranchisement?
We shouldn’t, all of us displaced by technology need to not be left to our own devises. But I don’t see that happening. That would take some measure of real debate with actionable outcomes facing tech’s full-bore ahead, damn the torpedoes thinking and in light of the societal effects of their advances.
Unfortunately I don’t see a lot of that debate happening at a policy level. One thing that gives me a little measure of hope though is that we are storytelling animals and our stories, whether those bearing witness, or those which we build our individual relationships with, cannot be divorced from us as organisms. Telling plausible stories about our habitation of the earth is not something that I see AI becoming capable of. At lest not with any real veracity.
This is another rough newsletter, I know, and to soften the blow I offer this photo of Leimomi Ho, a storyteller, captured as well as I could, demonstrating the human element of storytelling that no AI will master. The stories of Hawaiʻi that Ho and dancers like her tell (with the aid of the Kilohana Serenaders here) are not something AI will master—told as they are with wisdom, love, and occasional bawdy humor.
Let’s figure out how to do this throughout human endeavor, if I may be allowed a moment of truly hippie shit. We certainly need it in the face of tech’s profit-seeking juggernaut. After all we are people, not machines. And I can tell you, sitting at a red light while a car empty of driver bears down on you, wondering if it will stop, is no way to live. Not for people.
https://www.nytimes.com/2023/11/03/technology/cruise-general-motors-self-driving-cars.html
This is the idea that increasing efficiencies are consumed by increasing demand. In the world of power generation this means no net decrease, and typically increases in energy consumption following efficiency gains. Those efficiency gains are generally realized as cheaper power production or uses of power being made less wasteful. Jevon’s Paradox bodes poorly for less costly green power—as it gets cheaper than carbon-based energy we will find more uses for that power, creating more or bigger gadgets which use dwindling resources and still more energy to produce, and perhaps not netting us a gain in the fight against a warming atmosphere. In other words our profligate nature as a commodity-oriented modern species may well and truly fuck us even in the face of great gains made in generating renewable energy. Or we will figure out how to live within our planets resource-means.