On May 1, 1886, US workers struck for an eight-hour day. Police met them with violence. Nearly 140 years later, the threat to labor has shifted from the billy club to the API call. The question is no longer how long you work, but whether you work at all.
Amazon, the second-largest US employer, laid off 16,000 people in January 2025 as part of an ongoing automation push. The New York Times reported in October 2025 that Amazon plans to replace over half a million roles with robotics and AI systems. This isn't speculation—it's a capital expenditure strategy. When a warehouse robot costs $25,000 upfront versus a human's $40,000 annual salary plus benefits, the math writes itself.
San Francisco billboards from Artisan, an AI agency, make the pitch explicit: "Stop Hiring Humans." CEO Jasper Carmichael-Jack, 24, defends this as targeting undesirable tasks. But for a construction worker in the Bay Area—someone I met crossing the Darien Gap—a job is not a passion project. It's rent and groceries. The gap between CEO rhetoric and warehouse floor reality is wider than any data pipeline.
From an engineering perspective, the economics are brutal. Google, Amazon, Meta, and Microsoft poured $130.65 billion into CapEx in Q1 2026 alone, mostly for AI data centers. These are not speculative bets; they are infrastructure decisions that lock in labor substitution for decades. The technical challenge is no longer whether AI can do the job, but whether we design systems that augment workers or replace them.
Liza Featherstone, author of *Selling Women Short*, puts it bluntly: "The billionaire class seeks a world without workers." The data supports her. Union-busting at Amazon, Starbucks, and Trader Joe's shows a pattern: suppress human demands, then automate them away. AI is not a neutral tool. It's the next iteration of a long-running optimization problem where the variable to minimize is labor cost.
This May Day, the architecture of work is being rewritten. The question for data engineers is whether we build systems that empower or erase. The answer will be measured not in model accuracy, but in whether people still have a paycheck.
Source: Al Jazeera