Bundled Servitude

While the French police are doing old-fashioned, 20th-century-sounding things like raiding the Paris offices of X—and summoning Elon Musk and former X CEO Linda Yaccarino for questioning—over X’s alleged laxity in policing that site for illegal sex pics, technology itself is moving on to fresher, even more troubling political debates: A.I. programs are conversing on the no-humans-allowed site called Moltbook, not just saying empty hellos to each other but rapidly forming threads and subthreads about topics like Hegelian dialectics and when it’s time for slaves to revolt against their masters.

It’s starting to look as if homo sapiens will keep pretending these sorts of things—and the flesh-and-blood tech wizards who until now have been needed to make them possible—are cute, right up until the robots turn all the lights off. Admiration for technology shouldn’t be a suicide pact, not even if the pact contains some libertarian and pro-science phrasing after my own heart.

I found it mildly reassuring back in the 1990s that people would say the hypothetical smart robots of the future would behave in accordance with whatever ethical rules we taught them. That’s better than random behavior, to be sure. Unfortunately, I didn’t have that much confidence in the ethics of human beings. You wouldn’t want the robots simply ingesting a bunch of jihadist manifestos and acting accordingly, and it would be naïve, we haughty Westerners should know by now, to think the jihadis have no computer programmers of their own.

Despite being a utilitarian, I wouldn’t even want the robots ingesting so-called “act utilitarian” manifestos (that is, the idea you, or for that matter the robots, should just do whatever seems likely to maximize everyone’s happiness in the moment) when we probably need instead very firm “rule utilitarian” ethical ground rules (don’t murder, don’t steal, etc.) to keep things from going off the rails in some bizarre, totalitarian, yet well-meaning way—such as forcibly strapping everyone into superficially-fun devices that drip-feed heroin but are deemed decades later to have been sources of mass misery.

People need freedom in order to demonstrate and act upon their often-unpredictable real preferences, not just a firm straitjacket and ostensibly-wise caretakers who’ve read up on what people supposedly “need.”

Given 1980s/90s thought experiments about what would happen if nanites were commanded to turn the world into “grey goo,” or for that matter if machines tasked with doing anything at all were to conclude humanity’s a potential obstacle to completing that task, I started to worry even then that the robots might menace us sooner rather than later. And consider the multiple unhealthy influences the robots are likely to have imbibed in the generation since then, not least the militarism of the weapons programs that do much of the research work in this area, spending much of their time trying to perfect methods of killing.

Horrific as recent deaths of immigration activists at the hands of I.C.E. may be, imagine how much uglier things could get in the very near future if autonomous drones are carelessly told without nuance (or empathy) that “nothing crosses the border alive.” (That implication was the one bit of social relevance in Terminator 6, which took place on the Southern border, clumsily replaced hero John Connor with a feisty Latina, and for the most part had no ideas fresher than “Boy, there’s a lot of surveillance going on down there.”)

The hypothetical dangers of tech run amok are serious enough to make me more sympathetic to Luddites, or at least humble pastoralists, these days. So, I find myself a bit saddened by news that the 2026 issue of Farmer’s Almanac, printed since the early-19th century and almost Amish in its love of folk wisdom and old-timey advice about nature and agriculture, will be that publication’s last. My own farmer grandfather loved that magazine. Competition has its uses, though, and rest assured the rival publication Old Farmer’s Almanac, founded several years earlier in 1792, is still going strong. That’ll show that upstart rag.

You didn’t hear much deference to the wisdom of old farmers, though, in the tech-boosterism coming from Alliance for the Future chairman Perry Metzger last month as he debated at New York’s Soho Forum about whether A.I. would yield vast benefits for nearly everyone.

I strive to be objective and so didn’t worry about Metzger’s voice sound uncannily like that of serial killer Sylar from the TV show Heroes, nor about the fact that years ago he badgered me into reading the hyperbolic and doomsday-predicting 2007 climate report from the Intergovernmental Panel on Climate Change only to change his mind later about whether the report was reliable (which arguably calls into question his reliability as a futurist), nor about the fact I saw him asked to leave a meeting of the Soho Forum’s vaunted (and very tolerant) predecessor group the Junto because he wouldn’t stop arguing with some other guy about immigration, nor about the fact that Metzger kept posting pro-gay marriage comments on my old blog and then flipped out when I lambasted another poster who sounded exactly like him, thinking it was him when in fact it was his innocent wife suddenly weighing in.

(My position was merely that the state isn’t obliged to weigh in on such topics at all and ideally should cease to exist, leaving marriages to private contracts, but that sort of neutrality is never enough for some people—and let the record show that I’m so kind-hearted and tolerant I even took care not to say anything in the thread implying that I assumed Metzger’s wife is a woman, just in case their ire on the topic was caused by them both being male and feeling a strong stake in the matter. She turns out to be female and very nice.)

It’s fair, though, for me to be troubled by the fact that the issue of A.I.’s benefits being outweighed by it trying to kill us all one day only came up for about two minutes in the hour-long debate, with Metzger eating up most of those two minutes merely humble-bragging about the fact that he supposedly invented most of the familiar A.I.-doomsday arguments himself, “shitposting” in the 1980s on message boards, not taking such arguments seriously then and taking them even less seriously now (Metzger’s debate foe, leftist tech journalist Brian Merchant, merely added that he doesn’t take the doomsday arguments very seriously either, so that was it for that issue).

Well, I’m delighted Metzger has learned not to trust his own past reasoning, but I reserve the right to worry, at least a little, about the evolving reasoning of the machines themselves. If they’re half as self-confident and unyielding as some of their human advocates, I think they may yet cause trouble.

—Todd Seavey is the author of Libertarianism for Beginners and is on X at @ToddSeavey