Mark Zuckerberg doesn’t understand how to live
A recent opinion piece in The Verge opens with a small, telling scene: a rock climber proudly describing an AI-generated motivational poster he had hung on his bedroom wall - a bear walking a slackline over a canyon, holding a sign reading "Do cool shit." The reaction from those around him had been consistently muted. The writer uses this moment as a lens through which to examine something larger: the growing habit of reaching for generative AI to produce objects of personal meaning, and what that habit costs.
The piece is framed partly as a response to Mark Zuckerberg's public statements about AI - what the author characterizes as a dim and telling vision of human life. Zuckerberg has spoken enthusiastically about AI companions, AI-generated content, and AI as a tool for self-improvement and connection. The critique here is not purely technical. It is about what these framings assume: that the value of a motivational poster, a piece of writing, or a creative act lies in the output itself, rather than in the effort, judgment, and personal history that went into making it.
This kind of argument sits in an interesting position relative to the generative AI industry. Tools like Midjourney, DALL-E, and Adobe Firefly have made image generation fast and accessible, and there is genuine utility in that. But the poster anecdote points at something the capability benchmarks do not measure - whether a thing made effortlessly by a model carries the same weight for its owner as something made with difficulty and intention. The climber's poster said something he believed. The silence of everyone he told it to suggested the message had not traveled.
The broader implication is that AI companies, and Zuckerberg's Meta in particular, may be building products optimized for output volume rather than for what makes creative and social acts meaningful to people. That is not a new criticism, but the framing here - grounded in the ordinary awkwardness of a conversation on a climbing wall - makes it land with some force. The article is less a technical assessment than a reminder that the question of what generative AI should be for is still very much open.
