I'm not worried about AI job loss

Original Article ↗ Hacker News Discussion ↗

Fear vs Optimism about AI and Jobs

  • Many commenters argue some “healthy fear” is rational, especially for people without savings or elite networks; optimism is seen as a luxury of the insulated.
  • Others say online “doom” is overblown and mostly an internet phenomenon; real life and markets don’t yet reflect a civilization-scale collapse.
  • Several note that belief by executives that AI can replace people may matter more than what AI can actually do.

Viral Essay, Hype, and Authenticity

  • The referenced “80–100M views” essay is widely criticized as marketing “slop,” hype-driven and possibly inauthentic as a personal story.
  • Some see it as fear-stoking advertorial for an AI product, with platform metrics overstating real engagement.
  • Others found its factual claims basically plausible but question its timelines and breathless tone.

Labor Substitution, Bottlenecks, and Comparative Advantage

  • Strong debate over whether AI will simply augment workers or directly substitute for them.
  • One side: automation historically shifts work toward higher-value tasks; Jevons-style effects mean more demand, not fewer jobs.
  • Other side: even 80% task automation can justify cutting most of a department; demand is not infinite, and many industries are bounded (e.g., food consumption).
  • Physical/robotic automation is framed as far less economically viable than software or call-center automation.

White-Collar vs Blue-Collar and “Ordinary People”

  • Many think computer-based, sequence-of-tasks roles (customer support, bookkeeping, much software work) are at higher risk than physical trades, though trades have training, risk, and pay issues.
  • Others note that “ordinary people” are already struggling; even if jobs remain, wages and security may erode and unemployment spikes could destabilize housing, banks, and social order.

Software Engineering, AI Tools, and Memory Limits

  • Experienced developers report using tools like Claude/Codex to generate most boilerplate while they handle architecture, debugging, and judgment; juniors often flail or ship dangerous code.
  • Some say 0% of their backlog can be fully automated; others claim nearly 100% could be, given good specs and agent frameworks—yet few see real-world products shipping 10x faster.
  • A recurring theme: current LLMs struggle with long-term context, large messy codebases, ambiguous tickets, and business nuance; “memory” is partially patched with notes, vector indexes, and scaffolding, but not solved.

Inequality, Social Risk, and Policy Blind Spots

  • Multiple commenters argue the real threat is not zero jobs but intensified inequality: owners capture AI gains while labor faces stagnant or falling wages.
  • Fears include mass white-collar unemployment, political instability, and potential “French Revolution 2.0” scenarios if millions of educated workers are sidelined.
  • Skepticism is widespread that governments will proactively manage the transition; many expect delayed, crisis-driven intervention at best.