‘A.I.-Washing’ Layoffs? + Why L.L.M.s Can’t Write Well + Tokenmaxxing

Synopsis

Tech layoffs at Atlassian, Block, and Meta are being branded as AI-driven restructuring, but the evidence suggests a messier truth: companies with cratering stock prices and bloated pandemic-era headcounts are using AI as a convenient narrative to satisfy investors — spending $68M on a Jay-Z event five months before cutting 4,000 jobs is not an AI story. Meanwhile, writer Jasmine Sun makes a counterintuitive technical case that LLMs have actually *regressed* as creative writers since GPT-2, because post-training processes optimized for helpfulness and factuality have stripped out the unpredictability that makes prose compelling — and since verifying what makes writing *good* is impossible to reduce to a rubric, the models get trained to produce corporate email, not literature. For any professional navigating workforce decisions, AI investment pitches, or content strategy, this episode delivers the clearest-eyed breakdown yet of where AI hype ends and genuine capability begins.

Speakers

Episode Breakdown