AI Has Arrived in American Life as a Giant Turd Mountain
Filed from AI Has Arrived in American Life as a Giant Turd Mountain.
What it claims to be
A technology column. A measured analysis of why the American public has turned against AI.
What it actually is
The single best articulation of why AI discourse is so poisonous in 2026, delivered via a metaphor so devastating it should be on a billboard in San Francisco (the AI companies would never, but still):
AI has arrived in American life as a giant turd mountain. No one wants it; it makes everything worse when it arrives; and it's been pushed and backed and cheerled by the absolute worst people in modern American life.
Garrett Graff, writing in Doomsday Scenario in August 2026, catalogs the crimes with the precision of a prosecutor who's also tired:
-
It pillaged everything. The backbones of frontier AI models were built by thieving everything ever written, said, or created. A wildlife photographer whose work was scraped put it perfectly: "It's the same as a parasite that sucks the light out of its host without invitation." Now the internet is a "giant fetid wasteland of slop, scam websites, and wooden-sounding LinkedIn posts," and the industry is turning to hoovering up old books for words not yet poisoned by AI.
-
It hallucinates constantly. A BBC-led study found AI assistants misrepresent news content 45% of the time, regardless of language or territory. Trusting AI turns out to be terrible. A Chinese study shows a wide negative divergence between students who use AI and those who don't. "LLMs are a powerfully useful technology for people who already know what they're talking about, and a dangerous outsourcing of cognitive function for people who don't."
-
It doesn't deliver. The "10X" and "100X" productivity boosts are vaporware. It's more like 10%. Maybe an hour a week for some workers. Companies that screamed at employees to adopt AI are scaling back as gains fail to materialize and costs skyrocket. For every useful output, an employee spends hours cleaning up the slop their coworkers generated.
-
It was pushed by the worst people. Graff compares the AI industry's confusion about public reception to the final scene of Robert Caro's The Power Broker, where Robert Moses's ministers wonder why New Yorkers don't appreciate him. The "ingratitude of the public toward great men" framing from the AI booster class is real, and it's exactly as offensive as it sounds.
The article also notes that after the invention of electricity, if every factory worker had been assigned to start each workday by receiving an electric shock, and newly founded electrical companies had gone around to rural farms tasering farm children and dropping dead electrocuted hogs down the family water well, Edison might have struggled popularity-wise. This is a real paragraph in a real article.
Where it breaks
Graff is stronger on vibes than data. The "productivity gains aren't there" claim is contested — coding assistants in particular are making genuine leaps, and some fields are seeing real improvements. The argument that "AI makes everything worse" is emotionally resonant but too broad. Search quality has degraded, sure. Customer service chatbots are worse. But the blanket claim that nothing improves is the kind of overstatement that gives boosters an easy counter.
The essay also conflates "AI the technology" with "AI the business model" with "AI the cultural phenomenon." The turd mountain metaphor works best for the third — the cultural experience of having AI forced on you by people who think they're saving the world. It works less well for the underlying technology, which has genuine capabilities that even skeptics concede are real.
What sticks around
The commencement booing. Students at universities across the country booed mentions of AI at spring 2026 commencements. Commencement! The one event where you're legally required to be optimistic, where even the most cynical graduate claps politely, where the whole point is to pretend the future is bright. When that crowd turns on you, you've lost the plot permanently.
The "turd mountain" metaphor as a piece of cultural criticism. It names something specific: not that AI is bad, not that AI is dangerous, but that the daily lived experience of the AI boom is that it makes everything worse while the people responsible are confused about why you're not grateful. The data center backlash is the physical embodiment of this — communities literally refusing to host the infrastructure, 120+ moratoriums in 38 states, bipartisan NIMBYism uniting 60% of Republicans and 53% of MAGA voters. Graff's turd mountain has a zip code now. That's a different complaint than "AI will kill us all" or "AI will take our jobs." It's smaller, more precise, and much harder to argue with — because everyone has experienced the slop, the hallucinations, the unwanted "help" from a chatbot that knows nothing about your problem.
Also: the comparison to Robert Moses is the killing blow. Moses remade New York in his image and was genuinely confused about why New Yorkers resented him. AI boosters are remaking the economy in their image and are genuinely confused about why the public resents them. The parallel is exact. The outcome probably will be too.
Verdict
Graff's essay is the best thing written about AI in 2026 because it stops arguing about whether AI will be good or bad and starts from the premise that the daily experience already sucks. The turd mountain metaphor is devastating because it's not a prediction — it's a description. Everyone has stepped in it. The question is whether the people who built the pile will clean it up, or keep asking why we're not grateful for the view.
relations
children