Of Villains and Clowns
Minimalism vs. Maximalism: What is the right way to deal with Trump and AI?
by Tim Leberecht
I was cross-reading two articles this week: Adam Gopnik’s “How Alarmed Should We Be If Trump Wins Again?” in The New Yorker and Anthropic CEO Dario Amodei’s “Machines of Loving Grace,” an essay aiming to provide a non-alarmist, cautiously optimistic vision for AI.
There are parallels: Gopnik distinguishes between Trump minimalists, those who see him as an aberration of the system that will eventually be reined in by democratic institutions, and Trump maximalists–those who view him as a fascist with the potential to bring down the system.
AI divides along similar lines: Minimalists consider it a tool whose unquestionable power to enhance human flourishing depends on our human values and agendas. Maximalists view it as an “agent, not a tool” (to borrow a helpful distinction by Yuval Noah Harari) whose power lies in radically altering the mechanics of language, meaning, and reality-making beyond our control.
There are differences: Trump’s danger stems from being a villain who dresses up as a clown (as history tells us, most fascists weren’t taken seriously until “they no longer laughed at me!”). In contrast, AI’s danger lies in being neither a clown nor a villain. It is just an incredibly smart, fast-learning digital species with no humor and no morals, and therefore no mercy. Trump is predictably unpredictable. AI is unpredictably predictable. Which one is more dangerous?
Heisenberg’s uncertainty principle applies to both. “Prepared by experience to believe in institutions, mainstream liberals believe in their belief even as the institutions are degraded in front of their eyes,” Gopnik writes.
The same is true for AI. We cannot really imagine what a world with ever more powerful AI will look like because we are too close to the systems and stories we have created to grasp that non-human systems and stories might soon overwrite them. Amodei’s essay—while thoughtful and detailed—is therefore nothing more than a unilateral declaration of best intentions. We should not take them too seriously. Trump’s, however, no matter how clownesque and frivolous, we should take by the word.
Gopnik concludes: “Generations of cultural critics have warned that we are lost in a labyrinth and cannot tell real things from illusion. Yet the familiar passage from peril to parody now happens almost simultaneously. Events remain piercingly actual and threatening in their effects on real people, while also being duplicated in a fictive system that shows and spoofs them at the same time. One side of the highway is all cancer; the other side all crazy. Their confoundment is our confusion.”
Whether it’s Trump or AI—as I’ve argued before, the task ahead will not so much be to discern what is real, but to create a more powerful reality.
“How Alarmed Should We Be If Trump Wins Again?”
https://lnkd.in/eaXrRmcc
“Machines of Loving Grace”
https://lnkd.in/euAifm3Z
“Goodbye Reality, For Real!”
https://lnkd.in/efZQXPug