← Back to digest
HN Threads

Stanford: AI is hitting entry-level jobs hardest

What's this about?

A Stanford study found employment effects concentrated at the entry level rather than spread across seniority — the jobs most exposed are the ones people use to enter a field.

What exactly does "entry-level exposure" mean?

Junior roles consist disproportionately of well-specified, reviewable tasks, which is precisely what these models do well. Senior roles are more about judgement, context and responsibility. So the automation lands on the rung people climb first — which is a different problem from net job loss, because it breaks the pipeline that produces senior people later.

What's the community saying?

HN 135 points, 160 comments
  1. The dominant thread connects this directly to the coding-expertise item above: if juniors cannot get in, the expertise pipeline breaks a decade from now.
  2. Methodological pushback on separating AI effects from the general hiring slowdown.
  3. Anecdotal confirmation from people hiring, describing junior postings quietly disappearing.
← Anthropic appears to be A/B testing reduced effort levels in Claude CodeOpenAI restores 5-hour Codex and Work limits for ChatGPT Plus →