In a recent episode of *Decoder*, host Nilay Patel unpacked a concept he's dubbed "software brain" - a worldview that reduces everything to algorithms, databases, and loops. It's the kind of thinking that gave us Zillow (a database of houses), Uber (a database of cars and riders), and YouTube (a database of videos). But as AI supercharges this mindset, the gap between tech's excitement and the public's growing disdain has become a chasm you could drive a data center through.
The polling is brutal. An NBC News poll found AI has worse favorability than ICE and sits just above "the war in Iran" and "the Democrats generally," even though nearly two-thirds of respondents used ChatGPT or Copilot in the past month. Quinnipiac found over half of Americans think AI will do more harm than good, with more than 80 percent at least somewhat concerned. Only 35 percent are excited. Gen Z, the heaviest AI users, are the most sour: a Gallup poll shows only 18 percent are hopeful (down from 27 percent last year), while 31 percent are angry (up from 22 percent).
Tech execs see the numbers but keep misdiagnosing the problem as a marketing issue. OpenAI just spent $200 million on a podcast deal because Sam Altman thinks AI needs better branding. "If AI were a political candidate," Altman said, "it would be the least popular political candidate in history." Patel's retort: "AI doesn't have a marketing problem. People experience these tools every single day." ChatGPT has 900 million weekly users, trending toward a billion, and everyone has seen AI Overviews in Google Search and the ensuing slop. "You can't advertise people out of reacting to their own experiences."
The core issue, Patel argues, is that software brain sees the world as a series of controllable databases - but people aren't computers. They don't want to be flattened into legible data points for AI's convenience. The tech industry's ask - "make yourself legible to the AI" - is a fundamental misunderstanding of human nature. As Ezra Klein noted after visiting Silicon Valley, AI types are "racing one another to fully integrate AI into their lives" by giving it access to everything: files, email, calendar, messages. But for most people, that's not an opportunity; it's surveillance.
Patel points to the violent backlash: politicians who supported data centers have had their houses shot at, and Sam Altman's home was targeted with Molotov cocktails. He condemns the violence but notes it springs from a sense of helplessness that tech leaders have fostered by casually predicting AI will wipe out all the jobs. Anthropic CEO Dario Amodei, for instance, warns of a "serious employment crisis" as entry-level white-collar work is replaced. "The people do not yearn for automation," Patel says. "Not everything is a loop. The entire human experience cannot be captured in a database." So while the industry plows ahead - spending enormous sums on energy, emissions, and RAM - it misses the obvious: asking people to adapt to computers is a doomed idea. Computers should adapt to people. And no amount of slick marketing or executive haircuts is going to fix that.