conv.

All stories
AIQuiet 43d · day 43

LLMs breed false confidence in non-experts, sparking debate on skill erosion

Developers warn that AI tools remove the productive friction that signals the need to learn a domain, enabling incompetence at scale.

What to know

  • LLMs lower barriers to entry so steeply that non-experts now tackle specialized work without realizing their incompetence, enabled by AI that masks rather than solves knowledge gaps.
  • The core concern: friction—the difficulty of learning a domain—is essential signal, not noise; removing it creates false confidence without corresponding competence.
  • The pattern echoes existing problems in trades and manufacturing, where automation has eroded craft knowledge and left fewer people able to judge tool output.

The dispute No significant disagreement; all voices treat LLM-enabled incompetence as a problem, differing only in emphasis—whether the focus is skill erosion or the psychological mechanism of false confidence. · positions read across 7 posts and comments

most voices

LLMs enable false competence by removing the friction that signals the need to learn a domain; this erodes real expertise.

  • “The thing about LLMs is that they can remove an enormous amount of that friction. That friction is not noise, it is signal…Removal of this friction is a very, very bad thing.”

    jwcph · Mastodon
some voices

LLMs breed unearned confidence, mimicking the overconfidence of those who lack self-awareness about their own limits.

  • “LLMs give everyone the confidence of a mediocre middle-aged straight white guy.”

    gvwilson (via spouse) · Mastodon

“Verstärkter Einsatz von immer besseren Maschinen führt zu Verlust von Fertigkeiten und Kenntnissen.”

arthurdent, Mastodon commenter · Mastodon · Aug 17, 4:43 AM

Elizafox DeveloperA CEO (unnamed company) Subject of anecdote

The record 1 articles and posts · last 30 days

  1. RE: https:// social.treehouse.systems/@Eliz afox/117108843517245521 My spouse's response to this is, "LLMs give… post · Mastodon · gvwilson · 42d ago · 102▲ · 3 comments

What people are saying 6 voices from 1 site · best of 7 · verbatim