Matthew Lillard and the Dead Internet

On what it means that we all feel like the last human being left online

Something exceptionally strange is happening in the replies to a tweet about Matthew Lillard and his children. The tweet, from the verified, purposeless content-aggregating account “DiscussingFilm,” is as close to the platonic late 2023 tweet as possible—incoherent engagement-bait on a site that has come to monetize engagement, an insignificant “news hit” from the type of account that has come to dominate both Twitter feeds and the internet writ large. “Matthew Lillard,” the tweet goes, “says his ‘middle kid called me crying because they were so happy for me when “FIVE NIGHTS AT FREDDYS” came out.’

‘They went to the movie with all their friends, and I think they were very proud of me being a part of it.’”

The tweet garnered 209 replies, each almost without exception forming a body of responses so unambiguously AI-generated and syntactically nonsensical that it veers away from the mildly amusing and into the realm of uncanny valley, paranoiac dread. “That’s heartwarming! Matthew Lillard’s kid’s reaction to ‘Five Nights at Freddy’s’ is just precious. *star emoji* *clap emoji*” goes one common reply. “GOAT........WOW love him,” goes another. In the replies to an engagement-bait tweet about the star of a widely panned film adaptation of a jumpscare-based video game, those inclined to look can watch the internet turning into a wasteland of AI-generated shitposts before their eyes. Any regular Twitter user will recognize this experience as one that happens every single day.

---

The dead internet theory, birthed in early 2021 and mainstreamed through an Atlantic piece in late 2021, posits that some time around 2016 the internet was functionally abandoned by human beings; from that point onward, the theory goes, nearly all content, interaction, and engagement on the internet was powered by a network of artificial intelligence bots employed by the government to “gaslight” the population. The theory, like many a unifying conspiratorial perspective, is both obviously false and also speaks to something overwhelmingly true about the world around us. Dead internet’s central insight is not undermined by the fact that the internet has become a dystopian, inhuman wasteland through monetization, consolidation and rampant short-cutting, rather than some globohomo mind-control scheme. Like the people who use it, the internet is not on its way to being destroyed out of a need for control, but because it is perceived as a carcass to be picked at for profit.

To scroll the internet in 2023 is to feel, more often than not, like you are the last person alive. As Max Read wrote back in 2018, an overwhelming portion of the internet content apparatus has been fake for quite some time. Read paints the picture of a highly monetized house of cards: advertisers paying sites for engagement that is completely astroturfed in click farms and by bots, online businesses price-gouging and hawking fake goods, algorithmic and artificial intelligence churning spam content out into the void. In the New Yorker, Kyle Chayka writes about the way in which the internet feels emptier, less fun, and more artificial than ever. On the macro level, the situation is grim: blogs die or are consumed, writers are fired, re-hired, and fired ad infinitum, searches are harder to parse, human beings are nearly impossible to find. The individual experience is just as dire: the internet, to which our lives have been migrated without our say, feels like an empty, fun-house suburban mall. It’s all mannequins and haunted music; the rare human being you encounter wears the mannequin’s clothes.

Bots, especially since Elon Musks’s purchase of Twitter, are everywhere. In their chase for engagement, they flood the internet with a nearly illegible, obviously non-human torrent of slop. The few humans who continue to post, whether out of some duty, compulsion, or pathology, affect a homogeneity of their own; as I’ve written in the Letterboxd context, the overwhelming majority of human-generated internet posts hew to a set of codes, memes, and take-formats that render their speech something like an imitation of AI, rather than the other way around. The pattern, however regrettable, is remarkably hard to break. Social network algorithms and private equity-induced layoffs have decimated the distribution of the written word; recorded music breaks through to audiences at the whims of the TikTok terms of virality; if finding one’s own voice amidst the sea of AI vernacular is hard, sharing that voice has become increasingly impossible. The dead internet theory being untrue does not change the fact that the internet is dying.

Scrolling the internet today, then, threatens to be the ultimate act of solipsism. It is easy, surrounded by bots trying to sound like humans and humans trying to sound like bots, to feel like the last person alive. The other day, I spent what became hours scrolling through Twitter, surveying the horrific wreckage of that day’s continued onslaught of Gaza’s. I saw calls for justice, for a ceasefire, for liberation, because these are the sort of calls that my algorithm understands that I want to see. I also saw dismissals, justifications, gleeful gloating, cynical counterarguments, people convinced that the suffering was faked, worth celebrating, or both. These responses provoked the type of horror in me that any internet user knows well, and so I could not stop myself from scrolling to see more. I gorged on callousness, so appalled at the lack of humanity I saw that I had no choice but to convince myself that the people responding were not people at all, but AI. Whether or not these accounts were run by human beings or source code meant nothing to me; they had to be robots for me to be able to continue to trust the human capacity for empathy and moral certitude. Had I responded, I’m sure that I would have been considered a bot by somebody else in turn.

The point, as best as I can understand it, is that there will be real consequences for a generation reared and socialized on the internet when the internet passes the point of plausibility, when the platform that nominally existed to connect people to other human beings, human knowledge, and human endeavors becomes a mass facsimile. There will always be rhetorical safety in the glib, meme-y response—log off, “touch grass”—that presupposes that life before the internet is something that society, en masse, might easily return to. But for those more skeptical or discerning, the prospects are bleak: either continue to post into the abyss, feeding the next generation of artificial intelligence grist with which they might more closely resemble humans, or stumble in the darkness to find whether something new can come into being.