Everybody's Lost Their Minds
Growing reliance on AI—especially coding assistants and “agentic” tools—is leaving many engineers feeling de-skilled, alienated from their craft, and pressured into acting as supervisors for opaque systems they no longer fully understand. Commenters weigh these productivity gains against broader social and environmental costs, from concentrated corporate power and data center energy use to erosion of community, skills, and trust, drawing parallels to past overreactions such as COVID lockdowns. While some see AI as a powerful new tool or inevitable progress, others argue its current trajectory degrades human agency and expertise rather than augmenting them.
Societal cycles, media, and “mass insanity”
- Several see recurring 10-ish year waves of collective overreaction: war on terror, GFC, crypto, COVID, now AI, with cycles seemingly shortening.
- Others argue these are not organic “panics” but largely manufactured by institutions (religion, press, now social media) to keep people passive.
COVID responses and social damage
- Big subthread on whether lockdowns were necessary or “over the top,” especially in Australia.
- Critics: restrictions were too broad, long-term damage to kids’ education, destroyed meetups and community life, worsened isolation, and some policies were arbitrary or badly timed.
- Defenders: early uncertainty, Italian hospital overload, and high death tolls elsewhere justified strong measures; without them, deaths would have been far worse.
- Some younger people reportedly resent not the existence of lockdowns but their late, inconsistent, and repeatedly undermined implementation.
- Others fear COVID became a dress rehearsal for “techno‑fascist” emergency governance; some instead see it as a useful dress rehearsal for future pandemics.
AI capabilities, coding, and workflow
- Many report substantial productivity gains: LLMs find serious bugs, write nontrivial systems, implement integrations quickly, and act as powerful learning tools.
- Others find agentic workflows exhausting: lots of steering, reviewing inconsistent output, cleaning up “fast garbage,” and feeling more like managers or janitors than engineers.
- Strong split between those who see programming as a craft (now undermined and de-skilled) vs those who see code as a means to an end and are thrilled by leverage.
- Concerns include atrophying skills, loss of deep understanding, and pressure to produce more with less time to think.
De-skilling, personal vs societal impact
- Pro‑AI comments often emphasize personal gains (“what it does for me”), which critics see as ignoring wider social, economic, and ethical costs.
- Analogies to tobacco and TV are used: tools can feel helpful while quietly degrading health or cognition; a minority may use them for genuine learning, most will not.
- Some foresee intelligence becoming a cheap commodity, with society increasingly relying on “because Claude/ChatGPT said so,” hollowing out expertise.
Environmental and resource debates
- Heated disagreement over data center water usage and environmental impact.
- One side: AI water use is tiny versus agriculture (e.g., almonds, golf courses) and could be shifted to seawater cooling; water complaints are seen as misinformation.
- Other side: local impacts (e.g., nitrate concentration from one DC, diesel generation, noise, siting near communities) are real; even small percentages matter in a climate crisis.
Security and vulnerability research
- Some argue frontier models are transformative for vulnerability research, automating bug finding and tooling.
- Others question whether discovering many more vulnerabilities actually improves net security versus addressing basics (asset inventories, automated patching, better OS design).
Developer split and career anxiety
- Noted growing divide: devs enthusiastically leaning into AI vs those skeptical or demoralized.
- Some feel forced into AI-heavy workflows they don’t enjoy to remain employable; others consider moving to trades or more physical domains less automatable.