Thoughts on thinking

Impact on Learning and Education

• Many see LLMs as a “negative crutch” that bypasses struggle, undermining deep learning, critical thinking, and writing skills, especially for kids.
• Schools’ reactions vary: some “hardcore ban” AI (in-class handwritten work, honor codes, anti-cheating rules); others are urged to treat it like calculators or the internet—teach how and when to use it, not just forbid it.
• Active vs passive learning is a core split: critics say LLMs push passive consumption of finished answers; defenders say they can enable active learning if used to interrogate texts, ask follow-up questions, and explain confusing concepts.
• There is concern that only intrinsically motivated or curious students will benefit; for the majority, AI makes it even easier to avoid thinking.
• Oral exams, live Q&A, process logs, and “proof of process” are proposed as better assessments than AI-vulnerable take‑home essays.

Thinking as Exercise vs Tool Use

• A recurring analogy compares thinking to lifting weights: calculators and LLMs are like machines that do the work for you—useful for production, harmful if you skip all the “mental gym” work.
• Some argue “manual thinking” will become rare and valuable, something you deliberately practice (chess, mental arithmetic, writing, languages) even when not strictly needed.
• Others counter that tools have always displaced certain mental skills (e.g., hand square-root algorithms) without making people broadly “dumber,” as long as fundamentals are learned first.

Creativity, Originality, and Motivation

• The article’s core anxiety—“why create when AI can do it better?”—resonates with many in coding, writing, and drawing, who feel pride and meaning eroding when outputs can be replicated by prompts.
• Critics argue this reveals an unhealthy fixation on outperforming others or being “first”; they emphasize process, personal expression, and unique human experience as the real value.
• Many contest the premise that LLMs already produce superior thought or art, describing outputs as polished, average, formulaic, or shallow—especially for poetry, serious essays, and nontrivial code.
• There’s worry that auto-regressive models reinforce existing norms and “average” ideas, potentially discouraging off‑norm, breakthrough thinking.

Work, Economics, and Identity

• Several comments express fear that AGI will steadily devalue knowledge workers, leading to existential crises about purpose, skill, and livelihood.
• Others suggest the real problem is societal structure: capitalism will deploy AI to extract more value from labor, not to liberate people, unless wealth and power are intentionally restructured.
• A different camp views AI as a productivity amplifier: it frees them from drudge work and lets them attempt more ambitious projects (software systems, hardware builds, multidisciplinary hobbies).

Using LLMs Well vs Poorly

• Productive patterns: using LLMs as research assistants, explainers, hypothesis checkers, language partners, or brainstorming “sparring partners” that push back rather than just agree.
• Harmful patterns: letting LLMs draft essays, code, or ideas wholesale and merely “approving” them—this feels like sedation rather than augmentation and leads to observable skill atrophy in some developers.
• Several advocate explicit constraints: first think or write your own attempt, then use AI for verification, refinement, or alternative perspectives.

Cultural and Social Shifts

• Some foresee conversations and collaboration degenerating into meta‑discussions about prompts, with groups effectively channeling their chosen LLMs instead of themselves.
• Others report early signs of backlash: friend groups or creative communities lose meaning when AI-generated content floods them, prompting a renewed appreciation for clearly human, live, or handmade work.
• A meta‑concern: if AI absorbs and regurgitates most written thought, human incentives to contribute new, carefully crafted work may erode, especially if credit and traffic increasingly bypass original creators.