What Happened
A new essay titled "The Last Rung" by writer Mridul draws a sharp historical comparison between today's AI automation wave and the labor disruptions of the early 20th century. The piece, which surfaced on Hacker News on March 7, 2026, contrasts the current moment with the 1912 Taylor hearings, when Congress conducted adversarial questioning of Frederick Taylor, the architect of scientific management.
Back then, labor unions provided organized opposition. Congress held hearings. Workers had institutional backing to challenge how productivity gains were distributed. Mridul argues that none of those structures exist today.
The essay draws on Harry Braverman's labor analysis to make a specific claim: AI, particularly through RLHF (reinforcement learning from human feedback) training, extracts human judgment and tacit knowledge at a depth previous automation never reached. The workers whose labor trained these models receive no ongoing compensation. Meanwhile, productivity gains concentrate among AI companies capturing licensing revenue, while SaaS firms competing on price compress margins and cut staff.
Why It Matters
If you use AI tools daily, you are on both sides of this equation. Your prompts, feedback, and usage patterns improve the models. Your employer may be evaluating whether your role can be partially or fully automated. The essay's most uncomfortable point is that the frameworks for thinking about this collectively have atrophied. As Mridul puts it, neoliberalism has made the self "the only available scale for measuring anything," turning structural problems into individual ones.
This matters for practitioners because the productivity conversation around AI tools almost always focuses on what you can do faster. It rarely addresses what happens when "faster" means fewer people are needed. The 2025-2026 wave of AI-driven layoffs in content, customer support, and junior development roles is already providing data points.
Our Take
The essay diagnoses a real gap. The AI industry's labor conversation is dominated by two camps: accelerationists who dismiss displacement concerns, and doomers focused on existential risk. Almost nobody is building institutions to handle the mundane but massive problem of workers losing leverage.
That said, the essay's weakness is in offering direction. Identifying the absence of collective frameworks is useful. Proposing alternatives would be more useful. The comparison to 1912 is instructive but incomplete. The Taylor hearings happened because workers organized first, then demanded congressional action. The essay doesn't address what today's equivalent of that organizing would look like.
For people working with AI tools right now, the practical takeaway is this: understand the full picture of what these tools are doing to labor markets, not just what they do for your personal output. The productivity gains are real. So are the costs. Pretending otherwise is a choice.