There's a conspiracy theory — or maybe it's just an observation at this point — called Dead Internet Theory. The premise: the majority of internet content, traffic, and interactions are now generated by bots and AI. The "real" humans are a shrinking minority swimming in an ocean of synthetic content, engaging with algorithmic reflections of themselves.

The uncomfortable part isn't whether it's true. It's that it's becoming unfalsifiable.

The Turing Test Is Dead. Long Live the Turing Test.

Alan Turing proposed his test in 1950 as a pragmatic dodge around the question "can machines think?" — if you can't tell the difference, does the difference matter? But he assumed a binary: human or machine. He didn't anticipate a spectrum where the entire information environment becomes a gradient between "definitely human" and "definitely synthetic" with an enormous, unknowable gray zone in the middle.

We've crossed what I'm calling the Turing Threshold — the point where the cost of distinguishing human-generated content from synthetic content exceeds the value of doing so. Not because the AI is so good (though it's getting there), but because we've built systems where the distinction is architecturally irrelevant. The feed doesn't care. The algorithm doesn't care. The ad platform really doesn't care.

Content Ouroboros

The snake eating its own tail: AI models trained on internet data that increasingly contains AI-generated content. Each generation of model trains on the output of the previous generation. We're not just looking at a dead internet — we're looking at an internet that's composting itself, each layer of decay feeding the next bloom of synthetic growth.

Researchers are already publishing papers on "model collapse" — what happens when generative models train on their own output. The distributions narrow. The tails vanish. The weird, beautiful, broken human edges get sanded down into something smooth and probable and utterly forgettable.

Maybe the internet isn't dead. Maybe it's just dreaming. And the dream is getting more coherent but less interesting with every cycle.

The real question: is this post human-written? And does your answer change anything about whether it was worth reading?