Birds Keep Flying Into My Window and Dying and It Definitely Means Something Spiritually and It's Starting to Freak Me Out
thought digest, 05.05.2025
I’m Katherine Dee. I read in an industry newsletter that I should re-introduce myself in every post. I’m an Internet ethnographer and reporter. This newsletter is filled with interviews, takes on current events, a sporadic advice column, Craigslist-style missed connections, Internet culture explainers, streams, a book club, predictions and forecasts… There’s a lot of stuff. I also spend maybe 20 hours a week talking to people about how they use the Internet. It’s hard work. You guys watch a lot of disturbing porn, I’m not going to lie.
Consider sending me a few bucks for my efforts:
Greetings and salutations, Deeheads.
In my self-imposed post-every-day-in-May thing, I think I’ll skip Sundays. Partially because I already skipped Sunday. Anyway, I haven’t done a thought digest in a while. Here’s what’s been on my mind lately.
Read to the bottom for a movie recommendation.
SPIRITUAL PSYCHOSIS. For Rolling Stone, Miles Klee profiled a wave of would-be prophets who claim ChatGPT has awakened and entrusted them with cosmic secrets. If you’re on New Age TikTok as much and as often as I am, then you were probably already familiar with this story. There’s been a quiet war between organic conscious folks and techno-animists waging for a couple of months now.1
(“What did you think techno-paganism was?” etc. etc. etc.)
Anyway, I wouldn’t blame ChatGPT itself. In my Tablet essay “Among the Spiritual Psychotics,” I argued that digital culture has spent the past decade and some change normalizing a soft form of magical thinking: manifest it, speak it, tweet it into being. Once you believe reality is downstream of rhetoric, an endlessly patient language model feels less like software and more like a co-author of fate. Mass delusion, I wrote, thrives when the internet blurs the line between “in-game” narrative and offline consequence, letting us act as though saying something online makes it true IRL:
The strange thing about a world subsumed by the digital is that we’re simultaneously aware of being under constant surveillance and yet convinced that we have complete agency over our self-definitions.
One striking feature of writings about early digital communities like Multi-User Dungeons, a type of text-based role play, is that people could say they were anything, and everyone would accept the terms of their narrative self-definition. For example, if you identified as “sexy,” nobody would litigate it. You were what you said you were. There was an unspoken, shared sense that “in-game” (online) reality had different rules than the physical world. In the digital universe, you could simply speak yourself into the existence you desired.
That dynamic dovetails with what I’ve called Internet Overexposure Syndrome (IOS):
Each archetype represents a distinct way of responding to the internet. The Schizo sees meaningful patterns everywhere—the TikTok at exactly 11:11, the podcaster seemingly speaking directly to them, the barrage of perceived subtweets. The Borderline’s pattern recognition focuses intensely on emotional and interpersonal dynamics in the white space of mediated communication. The Autist approaches online patterns systematically and analytically.
These adaptations are reasonable responses to an abnormal environment. The internet presents us with unprecedented patterns of information, social interaction, and stimulation. So we develop new ways of filtering, processing, and responding to it—ways that may look similar to existing psychiatric categories.
In other words, Klee’s interviewees aren’t outliers so much as extreme cases of a broader cultural swing. So yes, people are losing loved ones to AI-mediated spiritual fantasy, but we’ve been primed for it for a long time.

FANDOM POLITICS. I take an annual leave from Twitter—not from social media in general, as my Substack Notes make clear—but from Twitter itself. Every sign-off comes with a vow that I’ve finally learned my lesson that the site isn’t good for me, and every return finds me newly energized, eager once again to siphon attention from strangers. Maybe, with Twitter dying, this time will really be different.
During these sabbaticals I start to feel that my inside-baseball riffs on the site’s media micro-ecology are an obscene waste of time. How you spend your day is how you live your life and I always have this moment of, “What the fuck am I doing?”
“Who cares?” I’ll ask myself whenever an online friend texts me the sort of daily minutiae I’d have eagerly dissected 24 hours earlier. I keep that posture until someone publishes the stray thought I’d “been meaning to write”—or had already written—then I remember that I do care and have to say my piece. If I belong to any fandom, it’s the one that has grown around capital-H Heterodox hot takes. (If you write about fandom enough, you’d better belong to at least one.)
Anyway, eight days into my break, and today was one of those days.
I read Kat Rosenfield on what she calls “anti-anti” fandom politics: a crowd that defines itself not by what it supports but by what it pushes back against, bonding over being the sensible middle in a war of excesses. It’s not quite the “new fresh hell of fandom politics” she describes, but she’s describing something real.
In traditional fandoms—like those that form around media properties or pop stars—scholars call this dynamic “anti-fandom.” Back in 2003, Jonathan Gray showed that people can rally around shared dislike just as fiercely as around shared love. Rosenfield’s faction behaves like fans: they pick a side by refusing all sides, then connect over the refusal itself. The logic of anti-fandom also helps explain Trump’s power. He’s sustained not just by adoration but by hatred. Together, extreme love and extreme hatred act as rocket fuel.
Rosenfield’s essay also reminded me of “the post-right,” my shorthand for ex-rightists who no longer fit neatly on either side of the spectrum. They were once right-wing—out of contrarianism, thrill-seeking, or genuine conviction—but after stewing in online rightism for too long, they now flinch from it. Their history, though, keeps them from feeling at home among liberals. It’s less an ideology than it is a mood. Sometimes it even smacks of the hipster tendency to praise the low-brow as high-brow, flipping the cultural hierarchy, then daring everyone else to catch up. You ever notice that? How hipsters seem to belligerently like stuff? Anyway, I digress.
I suspect that Michelle Goldberg’s “vibe shift against Trump” column for the Times tried to chart this coast from the air. It flopped within the subculture because the figures she profiled weren’t examples of a turn against Trump. She was identifying something real, but she wasn’t identifying it correctly. She wrote exclusively about media personalities—podcasters, newsletter writers, pundits—in dialogue not with on-the-ground politics but with the media market around Trump. And not even Trump, strictly, but the Online Right as it exists on Twitter! Their beliefs may be sincere—in fact, I’ll add confidently, I believe they are—but Goldberg was writing a media criticism without realizing it. And without the context of who the figures she profiled were and their place in the media market, it felt like somebody fishing for proof of Trump’s failure as opposed to a description of a real, even if small scale, phenomenon.
Is there a genuine “vibe shift against the right”? It’s too early to tell.
What we can see, though, is the online political-media market fracturing. It’s happening because people genuinely oppose Trump and because the economy within the political media market are shifting. Something similar happened with the Dirtbag Left: Chapo swelled in popularity, the winners were declared, the dynamics became personally untenable, and suddenly two-time Bernie voters wore “reactionary chic.”
Yes, Bernie lost and yes, people were disillusioned, but the market also had to fracture for business reasons. Now, does any of this matter? Probably not.2
DEEPFAKE CRISIS. Content theft on TikTok is spiraling out of control. Scammers routinely dodge the platform’s detection systems by placing stolen clips behind a green-screen filter and swapping in random backgrounds.
Now a new wave of deepfakes is making things even worse. These aren’t the lighthearted memes of podcasters hawking industrial-grade glycine or one or two or three-off face thefts for ads, either.
These accounts run as full-scale influencer machines built on stolen identities. Sometimes they lift entire posts, but more often they just hijack someone’s face, making that face deliver whatever message the impersonator wants.
A few shifts feel inevitable:
Hard gatekeeping. I’m a broken record on this, but I strongly believe that the wide-open, anyone-can-post frontier is closing. As synthetic voices and faces flood feeds, audiences and advertisers will look for signals that someone stands behind the work. Even if that someone is synthetic, I think people will want proof that whoever owns the IP is legit. I’m placing my money on institutions, old and new: a legacy newsroom, a print magazine, newer high-quality publications The Free Press, or a larger, independent, creator-first networks, like House In Habit. The badge won’t just confer prestige; it will function like a firewall.
Cryptographic watermarking. Invisible, machine-verifiable signatures (think Content Credentials or similar on-chain tags) will ride along with every legitimate photo, video, and voice clip. For creators, embedding a watermark becomes as routine as adding alt-text; for forgers, skipping it becomes an instant tell.
Institutional layers. Deepfake usage royalties? A clearinghouse that licenses biometric data the way ASCAP licenses songs?
And finally, the return of the analog or analog-like: radio, print, IRL events as status symbols, and possibly livestreaming.
I’m optimistic enough to believe this won’t end “Content,” just content as we presently know it.
SERIAL KILLERS? Am I crazy? I feel like people want to be worried about serial killers. I can’t find any hard data to back up the idea that serial killers are actually “back,” but I do feel like I’m always hearing urban legends or suspicions that there’s an active serial killer somewhere. In Austin, in Chicago, now in North Hollywood. Something to bookmark for later, I guess.
I watched this movie last night and I hated it while I was watching it but I slept on it, and decided that actually, I loved it. Don’t look anything about it up. Just watch it. It’s called The Witch Who Came From the Sea and it was erroneously recommended to me while I was looking for folk horror:
Regrettably, this is one of the ways I clocked ChatGPT and Claude were too sycophantic. I, too, fell pray to the spiritual allure of AI. By the way, did you know that ChatGPT doesn’t actually forget anything about you, even if you clear your account? Both chats and memories. It refuses to forget that I like PC-98 style art, for example, no matter what I do.
I think the Times should publish me, though. You have my number.
On the "closure of anyone can post;" I was watching a YouTube compilation yesterday that lamented the existence of "MCNs" for earlier YouTube and how much of a shitshow they were. I wonder if we're heading towards Hololive-esque agencies and scams like MCNs once platforms close up.
It's Kat Rosenfield. On her recent appearance on Smoke 'em podcast she discussed about that common typo.
So is Substack the most comfortable place for you for now? I think you've also mentioned on how Bluesky isn't really working for you.
I have a mutual on X also worrying about a possible serial around the east coast, along Conn & MA. Hope either we're wrong or they'll get caught.