Silicon Valley Is Bracing for a Permanent Underclass (29 minute read)
Silicon Valley insiders believe AI will soon displace millions of white-collar jobs and create a permanent underclass, but the industry has no coherent plan to address the disruption it's building.
What: An opinion essay based on interviews with AI company employees, executives, and policy experts reveals widespread concern in Silicon Valley that advanced AI will automate away knowledge work faster than society can adapt, concentrating wealth in AI companies while rendering ordinary workers economically irrelevant.
Why it matters: This signals a rare moment of class anxiety among privileged tech workers who are usually the automaters rather than the automated, potentially creating political momentum for policies that would normally be too progressive for mainstream voters to support.
Takeaway: Watch for AI economic impact policies in the 2028 election cycle, as this is becoming a top voter concern and creating rare political space for normally-unpopular progressive economic interventions.
Deep dive
- AI benchmarks now directly measure how well models can replace humans in specific jobs like investment banking, law, medicine, and consulting, with OpenAI's GDPVal showing over 80% win rates against human professionals within months of release
- Block CEO Jack Dorsey laid off nearly half his company in March 2026 citing AI coding agents, triggering a 25% stock surge and pressuring other executives to follow suit even before knowing if AI can actually replace those roles
- OpenAI's 2025 white paper proposes New Deal-style interventions including a 32-hour workweek, higher capital gains taxes, and a public wealth fund giving citizens equity stakes in AI companies, though the company hasn't committed to lobbying for specific legislation
- Anthropic CEO Dario Amodei predicts 50% of entry-level white-collar jobs could disappear by 2030 and warns that AI may create an unemployed underclass as it outpaces people with "lower intellectual ability," eventually encompassing more of the population
- AI-related investments accounted for 39% of US economic growth in the first three quarters of 2025, giving the federal government a vested interest in sustaining the AI boom regardless of worker impacts
- Research shows junior engineers using AI coding agents don't complete tasks faster and understand their work less when quizzed, suggesting AI tools may stunt skill development precisely when early-career workers face AI competition for jobs
- The "permanent underclass" meme posits people have limited time to build wealth before AI and robotics can fully replace human labor, freezing everyone in their current class positions
- Democratic pollster David Shor found 79% of voters worry about government lacking a plan to protect workers from AI, and policies like federal jobs guarantees poll better than universal basic income
- Anthropic's enterprise AI agents have driven annualized revenue to $30 billion (up from $9 billion at end of 2025), but the company hasn't released specific economic policies it supports beyond vague principles
- OpenAI removed its 100x profit cap in 2025 and a pro-AI super PAC funded partly by OpenAI's president spent $2 million against a congressional candidate who proposed AI safety regulation and taxation for direct payments
- Economists predict AI will look like accelerated deindustrialization, with companies outsourcing to AI agents instead of overseas workers, but compressed into two years rather than a decade
- Some AI employees are pre-committing billions in individual donations to nonprofits addressing AI harms, working 80-hour weeks while running multiple Claude Code agents overnight in anticipation of their own obsolescence
Decoder
- AGI (Artificial General Intelligence): AI systems capable of performing any intellectual task that humans can do, rather than specialized narrow AI
- Permanent underclass: A theory that once AI can do all human work, people will be frozen in their current economic class with the rich deploying AI and everyone else unemployable
- Labor share: The portion of economic output that goes to workers as wages versus owners as capital returns, which has been declining
- GDPVal benchmark: OpenAI's evaluation measuring AI performance across 44 occupations to assess real-world economic utility
- AI agents: Large language models that can autonomously complete multi-step tasks and work independently for hours, like Claude Code or OpenAI's Codex 5.3
- The China shock: Rapid deindustrialization when US manufacturing jobs moved overseas, now used as analogy for AI displacement happening faster
Original article
Many people in Silicon Valley believe that AI will soon surpass human capabilities. While this should produce tremendous growth and scientific achievement, it will displace millions of jobs, depress economic mobility, and exacerbate inequality. The technology will ferry power and wealth to the AI companies and existing owners of capital while ordinary people lose their economic leverage. It could create a permanent underclass as people are rendered useless and unemployable.