A roundup of the most engaging recent pieces on the web read by the Pittsburgh Review of Books this week.
“Punctuation: A Generational Divide” from The Atlantic
“The rules of punctuation have been codified since then, but they have never stopped evolving. Indeed, thanks to the advent of technologies that have us typing more and more, the rules have been hyper-evolving.”
“Why Is Digital Freedom Making Us Exhausted and Sad?” from The New Atlantis
“For more than two decades, Han has shaped a form of writing equal to his concerns: brief, concentrated books that think in movements rather than arguments. Their brevity feels deliberate, as if composed for the attention economy yet guided by another sense of time.”
“From A.I. to the Deep State, Michel Foucault Foresaw It All” from The New York Times
“Now, of course, with the spread of generative A.I., language without an author is everywhere. But this new era of authorless discourse is hardly proving to be the liberation for people that Foucault fantasized about.”
“Reading for Sins” from Liberties
“Is the only kind of magic left to literature dark magic — the power to harm with words?”
“Why Mourn Lindsey Graham and Not His Victims?” from Current Affairs
“Graham did not think much of the laws of war or civil liberties.”
“America 250” from Granta
“This was Fan Fest, a two-day event in Washington DC for UFC Freedom 250, advertised as ‘a once-in-a-generation celebration of the American fighting spirit’”.
“A City Bigger Than a Bomb: Dispatch from Beirut” from The Paris Review
“For almost a year after October 7, Israel’s war with Lebanon remained an oddly prim, choreographed conflict on the border, leaving Beirut largely untouched. When it blew across the whole country the following September, friends and family abroad could not understand why we decided to stay.”
“Empire of Signs” from The Los Angeles Review of Books
“The language model performs this operation at a near-incomprehensible scale. Trained on a corpus that is itself the sediment of particular people writing in particular places under particular pressures, it produces output that arrives smoothed, deracinated, generic, a kind of fluent average. Nowhere does the model disclose that its discourse on any subject arrives preformed by the structural biases of its training data.”
