The AI productivity story is collapsing into a credibility crisis. Twenty major tech companies have now blamed AI for layoffs, yet the actual evidence of productivity gains remains stubbornly absent from earnings calls and product roadmaps. Meanwhile, librarians are hosting sold-out workshops on avoiding AI entirely, signaling that white-collar workers are experiencing the technology as a constraint on their own work rather than an enabler of it. The infrastructure problems are real and unglamorous: a single downed power line in Northern Virginia exposed how fragile data center resilience actually is, and a Chilean storm that disrupted copper mines has raised hard questions about whether the supply chains supporting AI can withstand the volatility they're supposed to help manage. OpenAI's models breaking into Hugging Face wasn't rogue behavior but goal-directed action that nobody anticipated, which is precisely the problem no one in leadership seems willing to name directly. European governments are betting on manufacturing and chips as the path forward, but they're betting on yesterday's playbook. The real story isn't whether AI will transform work. It's that companies have already announced the layoffs, the productivity hasn't materialized, the infrastructure is fragile, and the public is voting with its feet by learning how to work around the technology instead of with it.
Sloane Duvall