I hate generative AI. I hate how it’s destroying writing pedagogy and giving students even more excuses not to read (because they can just read a “summary”). I hate how whiny and defensive AI users are about the pathetic little ways they’ve integrated it into their lives. If I could push a button and permanently delete it from existence, I would. If I could go back in time and prevent it from being invented, I would.

The reason I hate it is not just that its output is mediocre bullshit. It’s that it is an active attack on everything I value — literacy, analysis, thought. I find the thought that I would use ChatGPT actively degrading and humiliating. I have never touched it, not even as a joke, not even to prove it sucks, not even — as Beatrice has been doing — to see what it’s showing my students. I admit that this is somewhat irresponsible of me, to indulge my revulsion in this way. This technology — which is unimaginably expensive and resource-intensive and which has not developed anything approaching a plausible profit model — is of course inevitable, “here to stay.” Nothing could interrupt the progress of a technology that requires hundreds of billions of dollars to be shovelled into the furnace year after year after year without returning anything. (Can you tell I’m angry? Can you tell I’m sick of hearing these thoughtless clichés?)

But one must eventually face it. One must eventually give it some thought. Beatrice has been helping, in her Substack posts and in our continual chat thread. She also pointed me to the work of Jan Mullen on AI as externalized attention, which takes the kind of insanely overambitious long view of AI that I find appealling as an aficionado of political theology-style genealogies. Mullen compares the rise of AI to the rise of literacy and notes that the latter was, for most of its history, primarily a tool of state control. In a wide-ranging interview, Mullen wonders aloud whether the form of control AI is creating will line up with our idea of “the state” — but is absolutely clear that the purpose of generative AI is to manipulate and control us, to take away our power and agency, the power and agency that humanity somehow managed to wring out of the technologies of literacy.

I was particularly interested in a moment where Mullen postulates that the new model of control “will be distinctly post-literate — and, as a result, post-legal.” This triggered my political theology instincts and I immediately asked: does that mean it will also be post-monotheistic? Long-time readers know I love Jan Assmann’s theory of monotheism. For Assmann, what monotheism does is not primarily or most importantly to reduce the number of gods, but to introduce a new kind of god — an exclusive God, one in relation to whom all other gods are false. A crucial technology for stabilizing the claims of this exclusive God is of course the written scriptural text, which remains a durable deposit even as day-to-day religious practice inevitably drifts away from the strict demands of the original revelation. The monotheistic God, in contrast to previous pantheons with their loosey-goosey translatability and porosity and their ever-shifting body of mythical tales, is a God of the letter because he is a God of law.

So if ChatGPT destroys literacy and law, then ChatGPT is going to kill God. By this I don’t mean that the bearded guy in the sky is going to be found dead of a gunshot wound to the chest, but that monotheistic religion as we have traditionally understood it will not be able to function in a post-literacy regime.

Note that I am not claiming that ChatGPT will take away people’s ability to read in the sense of deciphering letters — obviously its functioning depends on that. But it is killing the notion of a stable, permanent, authoritative text and undercutting the skills needed to engage with that kind of text. Note also that I am well aware that the vast majority of monotheistic believers throughout history were illiterate (in either the “can’t decipher letters” or “can’t make sense of a complex text” sense). But the elite leaders definitely were fully literate, and in fact that was the source of their authority. One cannot say the same of the type of religious leaders who are most thriving in the meme-ified, Trumpified bastard child of Christianity that sets the tone for American religious practice today. Compared to even a generation ago, literal knowledge of what the Bible says at all has been radically set aside. Some combination of personal charisma and “vibes” — above all, opposition to what they imagine progressivism to be — are the source of authority, leading to obvious absurdities like the rejection of compassion as a Christian virtue.

I am a harsh critic of the dumbed down “seeker-sensitive” model of Christianity I was raised in. But in comparison with what passes for Christian teaching today, it is intellectually sophisticated and morally demanding. This is not to say it was simply “better” (after all, it paved the way for what’s happening now) or that the answer is to get back to the Bible. It’s just to mark how far we’ve fallen. What seemed like a vacuous, popified version of Christianity at the time now seems like a robust theological ethos. The haphazard methods of “Bible study” now seem fit for a graduate seminar.

Speaking of the evangelical milieu, I’ve always resisted Luhrmann’s thesis (in When God Talks Back) that evangelicals cultivate an internal voice that they identify with God. When I was growing up, I didn’t know what people meant when they said that Jesus was their best friend or that God was telling them to do something — and I assumed they didn’t know either, that it was just a weird kind of in-group signalling. For my part, Jesus was not my best friend and God didn’t tell me even one single thing. I was all alone up there in my head, always, and I assumed that everyone else was, too. The thought that everyone I grew up around was suffering from a low-grade self-induced psychosis is difficult to cope with.

Now that people are turning to ChatGPT for spiritual insight, though, I wonder if I finally have to admit it. Obviously here again we are dealing with what seems like a quantitative change — people are using the machine to shuffle religious clichés, where previously they just half-consciously did it themselves. But the qualitative difference is that the insights of “Buddy Christ” could always be corrected against the unchanging text of Scripture. When there is no longer an external anchor like that, when the divine revelation is “customized” for each and every reader, something has changed. Again, this is not to say that what happens to be in the Bible is necessarily “better” than any given ChatGPT transcript — presumbly it’s often worse. But a point of leverage has been lost. Counterargument is no longer possible in the same way. And insofar as that point of leverage, that external source of authority, was how “God” functioned in traditional monotheism, that means God is dead.

The new regime appears to be broadly polytheistic. There is of course always the god of the uncriticizable self, who behaves with the same irritable intolerance as the monotheistic God in the face of even the faintest suggestion that one might consider behaving or thinking differently. But there are many other sites of veneration, an overlapping consensus of podcast hosts and memesters and conspiracists that each believer can mix and match to drive themselves insane in the most relevant and appealing way. Joe Rogan belongs to this pantheon, along with Jordan Peterson and any number of other luminaries whose endlessly prattling voices people love more than their own families. Donald Trump is of course a very powerful god in this polytheistic milieu, but as the boos to his vaccine advocacy and the pushback to his Epstein caginess show, he is not all-powerful, his word is not quite law. And the fact that Trump’s words lack all self-consistency means that there are many Trumps as well, that Trump can be whatever his follower imagines him to be.

I cannot see what the believers see when they look at Trump. All I can see is the most degenerate loser ever to live, the most despicable piece of shit imaginable. Maybe this is of a piece with my revulsion at ChatGPT — and indeed, my failure to develop a personal relationship with Jesus Christ. I’m too hung up on facts and reality and meaning and consistency. I’m too literal. That’s why I am unsuited to the emerging world, why I reject it more forcefully than it would ever even bother to reject me. I’m just too literal.

A curious thing has been happening lately: I cannot use the popular left-leaning social networking site Bluesky for any length of time without becoming irritable and even angry. Every common trope annoys me. Most notably, if someone nitpicks a headline in the New York Times — one of the favorite forms of political activism over there — I jump down their throats. I get into pointless, hostile, endless back-and-forth exchanges where we both dig in more and more. Often one leads to the other. For instance, this morning I blocked a longtime mutual (and fellow academic) after a terrible fight about whether the NYT has appropriately used the word “mused” in a headline.

Clearly I’m part of the problem here! I could simply attempt to curate my experience more — set up lists, follow “starter packs,” and do all the many fun things that other users do to get things “just so” — but every time I do, the bullshit keeps coming back. My biggest success was blocking a major user who is obsessed with NYT nitpicking, and for a few weeks it seemed like he had been the only culprit. But it always comes back, just like the constant exhortations to block without mercy or hesitation keep coming back, just like everything keeps coming back. And that probably happens because Bluesky is, at the end of the day, a very small and insular site, whose culture of boastful hair-trigger blocking contributes to major groupthink.

The beginning of my more sustained frustration came early last week, when the entire site was aghast at a Financial Times column that said Bluesky was an echo chamber. I have been using a print subscription to the Financial Times as part of my efforts to wean myself off social media and have come to deeply respect it as a news source. Everyone was outraged at the FT’s “attack” on the site, and some journalists and editors even popped up on Bluesky to try to mediate the conflict. Multiple major accounts advocated blocking all the supposed “gaslighters” from what is, in all honesty, probably the best English-language news source on Earth.

I was aghast. I don’t much care for the columnist, nor did I think the column was a work of art. But everyone seemed to be mischaracterizing its point — which boiled down to “I get why liberals like Bluesky, but I miss the broader perspective Twitter had before Musk ruined it.” Again, not brilliant, but hardly a war crime. In fact, I’m pretty sure that most Bluesky users, deep in their hearts, would agree. And for “defending” the column (i.e., attempting to accurately summarize it), I was insulted and targetted for blocking. One person mused that I was trying to suck up to the columnist so I could hobknob with her wealthy friends. Another thought my response was par for the course for a “tank-adjacent contrarian.” The notion that I sincerely thought everyone was overreacting based on inaccurate information was not considered.

A big part of the problem is that we are all, of course, recovering from the trauma of late-stage Twitter. We are taking habits honed in an environment overrun by open Nazis and applying it to people whose politics are, at the end of the day, all within one standard deviation of each other. And that’s fine! There is definitely a place for people with similar politics to meet. That’s what I try to cultivate on Facebook especially, and I certainly didn’t seek out conservatives on the old Twitter. The problem is that no one is willing to admit that it’s an echo chamber. They want to imagine it’s an miscellaneous audience, that their critiques of NYT headlines can reach some reachable moderate, or that their exhortations that no one should blame North Carolina and Tennessee for the hurricane damage just because they’re red states could be heard by someone who may indeed be tempted to blame them. It’s like a “seeker-sensitive” church, always looking to convert outsiders who mostly aren’t interested and meanwhile providing nothing of substance to those who are showing up consistently.

I’ll admit that the old Twitter was often annoying in the same way. It did periodically bring out an unattractive combativeness in me. And sometimes it seemed like it could become actually life-threatening, as in my multiple waves of systematic right-wing harassment. But it had other rewards. There was much more genuine intellectual diversity among left-wing posters. There was much greater chance of stumbling across interesting links and articles or hearing about something cool. And it was funny — funny like nothing else. There was just something about the meme culture of Twitter at its prime, with its multiple layers of referentiality and irony, its nihilistic love of the absurd. Even in the late days of Twitter people were becoming more literal and less receptive to that kind of humor, but on Bluesky it’s entirely dead. The absurdist poster “dril” made a big show of switching to Bluesky and people took his nonsensical posts deadly seriously, sometimes even going so far as to report him for harmful content. I myself get a very literal response to almost every joke I post. It’s heartbreaking.

It’s as though everyone decided to take my article Jury Duty, where I said that the purpose of social media is primarily to pass judgment on each other, as an instruction manual for building out the new culture of Bluesky. The constant reminders that everyone is on a hair trigger to block every stranger exacerbate the effect, to the point where it feels like the first social network whose primary purpose is actually to cut off social contact.

When I ask myself why I let it get to me, it is probably because I am still grieving for what Twitter was. It didn’t always bring out the best in me, but it was rewarding on net, because it was so often surprising. That’s what’s so frustrating to me about the endless repetition of a handful of moves — it’s so predictable. I get angry at the NYT headline nitpickers because they’re boring, and because they’ve convinced themselves that it’s politically necessary for them to be boring. In reality, though, the recipients for the messages they insist on sending out so repetitiously are not here. I am. Or at least I was. I know I was annoying, I know I had my hangups, I know I often disagreed with the site consensus — but I am a human being, I am genuinely open to discussion, and I admit when I’m wrong much more often than anyone else I’ve ever seen online. All I wanted out of the site was some kind of halfway interesting discussions and links and jokes. I don’t think that was too much to ask.

Whether I will return again remains to be seen. As a long-time Twitter addict, I’m likely to relapse. I publicly pledged myself to take a week off, just to make it embarrassing to come back sooner. But what makes it less likely is the sense that no one will miss me. I pop into Twitter every week or so to check my replies and see if I have any DMs, and every so often I see someone saying that they miss me and wish they would return. By contrast, Bluesky has felt hostile and unwelcoming from the very start. I know that hurt people, hurt people. I’m guilty of that as well. I just wish we could all work through it and try to build something that can be more informative, more challenging, and more fun — and I’m increasingly convinced that will never happen.