Tinkering Won’t Prevent an AI Underclass

Individual early adoption of AI tools (learning prompts, building automations, experimenting with assistants) can produce temporary advantage, but rapid product and platform change erodes that edge and leaves systemic outcomes driven by policy, corporate strategy, and labor markets. The public debate should therefore shift from personal self‑help to political choices about training, redistribution, and platform power. — This reframing shifts responsibility from individuals to institutions, changing what solutions (regulation, collective bargaining, public training) are seen as legitimate and urgent.