Most AI value will come from broad automation, not from R & D

Emotional response to techno‑optimism

• Many describe current AI/tech optimism as depressing or dystopian: it feels like hype detached from tangible social benefit and mostly about cost-cutting and surveillance.
• Others argue tech progress is historically net-positive, but concede that in the short/medium term it often worsens inequality and can feel like “snake oil” (crypto comparisons come up).

Who benefits from automation and AI?

• Strong concern that AI/automation will magnify wealth concentration: more output with fewer workers, higher profits, soaring asset prices (especially housing), and no built‑in mechanism to share gains.
• Counterpoint: productivity gains have historically raised median living standards (more goods, better health), even while inequality rose; both can be true at once.
• Several foresee “neo‑feudalism”: corporations owning robots, land, food, housing, and even breathable air, with most people as precarious tenants/consumers.

Work, jobs, and automation in practice

• Concrete impacts already visible in art, game assets, voice acting, and some back‑office tasks; less so in complex software or physical trades (plumbers, electricians, caregivers).
• Some engineers report large productivity gains using LLMs as advanced autocomplete, especially for boilerplate code, while others find hallucinations and unreliability negate the benefit.
• Widespread fear that “assistants” today are a stepping stone to job cuts tomorrow; several layoff stories are tied to management’s AI narrative.

Historical analogies and metrics

• Frequent comparison to agricultural and industrial revolutions: massive labor displacement, eventual new kinds of work, but requiring strong worker organization (unions, regulation) to avoid misery.
• Debate over whether productivity gains have truly reduced working hours or just shifted burdens (e.g., housing, education, healthcare costs).
• GDP is criticized as a misleading success metric that can rise even while unemployment, precarity, and inequality worsen.

Governance, regulation, and power

• Automation’s social outcomes are framed as political, not technical: “Star Trek vs Blade Runner” depends on property rights, labor power, and regulation.
• Many argue current governments are captured by capital, making “let the market handle it” or “government will fix distribution later” not credible.

R&D vs broad automation framing

• Several think the article’s R&D vs automation split is ill-posed: R&D underlies all automation; capital deepening doesn’t happen without prior research.
• Some call dismissals of R&D shortsighted and point out foundational researchers rarely capture a share of the value comparable to downstream corporations.

Technical side notes

• A few highlight constraint programming and deterministic automation as under-discussed alternatives/complements to stochastic LLMs for many “broad automation” tasks.