AI may be making us think and write more alike
Large language models are increasingly used to draft emails, reviews and plans, leading many to worry that human writing styles and even thought patterns are being flattened into a bland, “average” corporate dialect. Commenters describe workplaces where managers outsource communication and evaluation to AI, making it harder to see what people actually know, and raising fears about loss of individuality, degraded skills, and overreliance on machine judgement. Others argue this homogenization echoes earlier shifts like the printing press or social media, suggesting that while AI will change how we write and reason, social norms and competitive pressures may eventually push back toward more distinctive human expression.
Workplace Communication and Authenticity
- Many describe managers and leads who communicate almost exclusively via LLM outputs (emails, reviews, tickets, plans).
- This feels dehumanizing and “proxy-like”: harder to gauge intent, negotiate, or push back, because the real person is not reasoning.
- Some are job-hunting specifically to escape AI-mediated bosses, but expect this pattern to be widespread.
- LLM use is said to bloat slide decks and specs with slick but noisy, redundant content, adding overhead for those doing the real work.
- Others distinguish acceptable uses (grammar/tone checks, idea validation) from full delegation of writing, which they see as crossing a line.
Cognitive and Cultural Effects
- Multiple comments report people unconsciously mimicking LLM phrases and tone (“you’re absolutely right”, “I hallucinated that”).
- Concern: offloading “stuck” thinking to LLMs may atrophy problem‑solving and originality; some fear “LLM-brain” or mild psychosis‑like detachment.
- Others argue human communication is highly resilient and novelty‑driven; they predict LLM‑speak will eventually become unfashionable, like past fads.
Creativity, Secrecy, and a New “Dark Age”
- A long, widely discussed comment envisions a “second dark age”: people hide techniques from LLM training, open communities shrink, and trusted human‑only circles rise.
- Some are already withholding code, art, and patterns they would once have open‑sourced, believing training “dilutes” their advantage.
- Debate: one side sees this as selfish or “dog in the manger”; the other claims LLM benefits are net‑negative so far (spam, slop, surveillance, skill loss).
Education, Skills, and Productivity
- Teachers and mentors report that student essays and junior code now look same-y, making it harder to assess actual understanding.
- Others note that to use LLMs safely you must finally do long‑neglected work: detailed specs, tests, docs. This may cancel much of the supposed productivity gain.
Average‑ness, Safety, and Homogenization
- Many emphasize that LLMs, by design, gravitate toward the “average” and then are further tuned for inoffensive, corporate‑bland safety.
- Some like the precise vocabulary and clear grammar they pick up from LLMs; others see the tone as neutered and impersonal.
- There is disagreement over whether this convergence is just another wave of normalization (like printing presses and standardized spelling) or a qualitatively worse flattening of thought.