Last year, Dave Eggers walked into OpenAI's offices and told the staff directly: ChatGPT was silencing an entire generation. This week, Min Jin Lee published a short story in The New Yorker about a boy who has spent three years sleeping, eating, and studying for a single exam. These two things are not coincidentally adjacent. They are the same story told from opposite ends of the pressure system.
The Suneung and the Prompt Box: Two Machines for Erasing Personhood
The Korean Suneung exam that structures Lee's story is, like ChatGPT, a system optimized for a single kind of output. Both promise access to opportunity. Both operate by narrowing the range of acceptable expression until what emerges is technically correct and spiritually empty. Eggers' charge, that AI is silencing a generation, lands differently when you hold it next to Lee's portrait of a student who has already silenced himself in pursuit of a number. Kyle Chayka's analysis of algorithmic homogenization is the theoretical scaffolding here: systems that optimize for legibility kill the peripheral, the weird, the alive.
What AI Companies Are Actually Selling
Eggers was invited to speak by Sam Altman, which is either an act of genuine intellectual humility or the most expensive therapy session in Silicon Valley history. The more interesting question is what Anthropic and its upstart competitor Thinking Machines are actually selling when they pitch themselves as intellectual, creative platforms. Fast Company's Rebecca Heilweil reports that Thinking Machines is explicitly coming for Anthropic's intellectual vibe, as if cultural prestige were a positioning variable rather than an outcome. Meanwhile Min Jin Lee's boy sits alone in a room, preparing to be judged by a system that has already decided what a correct answer looks like.