Source: arXiv

Paper

Who Will Become the Next Senior? How Generative AI Erodes the Development Pathway in Software Engineering — Sumin Yu, Taesup Moon (submitted 19 Jul 2026)
ID: arXiv:2607.17067

What it claims

Macro charts show junior hiring stress; lab studies show AI changes task speed. This paper asks how GenAI reshapes early-career development in real orgs and classrooms. Through 14 semi-structured interviews with juniors at the SE entry threshold and senior engineers in South Korea, analyzed with Reflexive Thematic Analysis, the authors name a pattern of Absorption: GenAI redirects entry-level work into senior–AI workflows.

Three consequences follow: (1) juniors lose the productive struggle that once built expertise; (2) university classrooms collectively normalize GenAI use, structurally reproducing that loss; (3) seniors and juniors see the problem asymmetrically, so neither side corrects it alone. The claim is institutional: GenAI may absorb not only tasks but parts of the pathway that produces the next seniors — fixing it needs design across classrooms, workplaces, and junior evaluation criteria.

The breakdown

The interviews sit in South Korea, but the mechanism is portable. Absorption means entry-level tickets that once forced juniors through debugging, ownership, and messy specs now collapse into a senior prompting a model. Classroom normalization then ships graduates who already expect that collapse. The perceptual asymmetry — seniors and juniors disagreeing about what is being lost — is why the pathway can erode without an alarm. Macro “junior hiring is down” charts do not show this; only workplace and university practice do.

Why readers outside the lab should care

If you hire Korean CS grads, run a Seoul/Busan engineering office, or manage a remote Korea pod, “juniors with Copilot ship faster” can hide a pipeline crash two years later. Multinationals that treat Korea as a cost-efficient junior farm need to redesign apprenticeship — or they will import seniors who never practiced the hard reps. Overseas students training in Korea should notice when AI completes the struggle that used to be the curriculum.

HR and engineering managers outside Korea can use the Absorption pattern as a diagnostic checklist even if their interview sample is local.

What travelers and expats should watch

• Do ask hiring managers what juniors still own end-to-end without AI autocomplete.
• Do design onboarding tasks that keep productive struggle (reviews, debugging, incident ownership).
• Don’t treat classroom GenAI norms as proof workplace apprenticeship is healthy.
• Expect interview study limits: n=14 in Korea — directional evidence, not a global census.
• Do review how your Korean university partners grade AI-assisted coursework — classroom norms feed the same Absorption loop.

Context

Read this as labor-pipeline risk, not as anti-AI nostalgia. Korelay frame: the Korea fieldwork is a warning about Absorption — seniors + models eating the apprenticeship layer that used to mint seniors. If your Korea office still reports “velocity up,” ask what skill still gets practiced without autocomplete — and write that into junior evaluation before the next hiring cycle.

Source

arXiv:2607.17067 — abstract and framing cited; open the OA PDF for themes and interview design. Do not republish the PDF.