Thinking will become a hobby
Claims that AI will soon handle most “cognitive work” and relegate human thinking to a hobby provoke sharp disagreement. Many argue that current LLMs don’t truly think and actually increase human cognitive load, while others believe rapid progress could eventually automate both mental and physical labor, forcing society to rethink work, education, democracy, and how people derive purpose. Across viewpoints, there is broad anxiety about economic disruption, concentration of power, and the risk that critical thinking and literacy—already fragile—could further atrophy if humans increasingly offload judgment and understanding to machines.
Nature of LLM “Thinking”
- Many argue current LLMs don’t truly “think”; they’re seen as powerful autocomplete trained on human output.
- Others respond that the internal process is irrelevant if outputs match human reasoning for practical purposes.
- Several note “jagged intelligence”: models are superhuman on some tasks (e.g., formal math) and very weak on others.
- Some point to RL and latent/abstract internal representations as evidence that future systems might reason in less word-like ways.
Impact on Cognitive Work and Jobs
- Some developers report higher cognitive load: AI writes more code, but humans must constantly review, integrate, and correct it.
- A recurring view: humans are being “promoted” to spec-setting and middle-management of agents, not removed from the loop.
- Others believe progress is so rapid that large swaths of cognitive labor, and later physical labor via “physical LLMs,” will be automated.
- Skeptics reply that robotics is far behind: real-world edge cases, physical constraints, maintenance, and costs still favor humans.
Economy, Inequality, and Democracy
- One camp foresees near‑post‑scarcity: automated food, housing, and services, with painful transition but eventual abundance.
- Critics say history shows capital does not share gains without being forced; absent major reforms, mass unemployment and pauperization are likely.
- If economic value is fully decoupled from human labor, some fear human political leverage will collapse, threatening liberal democracy.
- Others argue we must explicitly decouple human worth from economic output via new social contracts (e.g., UBI or similar).
Is Thinking Already a Hobby?
- Many claim most people already outsource thinking to apps, institutions, and algorithms; deep reflection is a niche activity.
- Historically, extended “thinking” was often reserved for elites with surplus time; some see AI pushing us back toward that pattern.
Human Roles and Values
- Even optimistic users say AI increases the amount and level of thinking they must do: system design, judgment, and responsibility remain human.
- Suggested enduring human niches: value-setting (“what and why”), high-level judgment, arts and creativity, social and caring professions, and political decision‑making.
Tech Trajectory and Uncertainty
- Some extrapolate recent AI progress as exponential and inevitable; others warn of sigmoid slowdowns and an “AI winter.”
- Several emphasize that outcomes depend on collective choices and policy, not technology alone.