Screen Time: John Herrman

The AI-Obsessed Are Driving Themselves Mad In the uncanny valley, even the geeks are getting fooled.

- John Herrman

IN EARLY AUGUST, an anonymous X account started making big promises about AI. "You're about to taste agi," it claimed, adding a strawberry emoji. "Q* isn't a project. it's a portal. altman's strawberry is the key. the singularity isn't coming. it's here," it continued.

"Tonight, we evolve." If you're not part of the hyperactive cluster of AI fans, critics, doomers, accelerationists, grifters, and rare insiders that has congregated on X, Discord, and Reddit to speculate about the future of AI, this probably sounds like nonsense. If you are, it might have sounded enticing at the time. Some background: AGI stands for "artificial general intelligence," a term used to describe human-or superhuman-like abilities in AI; the strawberries are a reference to an internal code name for a rumored "reasoning" technology being developed at OpenAI (and to a post from OpenAI CEO Sam Altman minutes before, which included a photo of strawberries); Q* is either a previous code name for the project or a related project; and the singularity is a theoretical point at which AI or technology more broadly becomes self-improving and uncontrollable. All this coming that night? Wow.