Yes, and
Rapid advances in AI coding tools are forcing people to reassess whether it still makes sense to study computer science or become a programmer. Commenters contrast “vibe coding” with LLMs against deliberate, fundamentals-based software engineering, debating whether code-reading, architecture, and domain expertise will grow in value or whether AI will eventually subsume most of that work. Underneath are broader anxieties about a brutal junior job market, long‑term career viability, and the possibility that programming itself may follow the path of other crafts that were largely industrialized.
Future of programming jobs with AI
- Many expect LLMs to cut the number of traditional dev roles, especially juniors; some report they’re already gone.
- Others argue cheaper coding increases demand for software, so programmers (or programmer‑adjacent roles) remain valuable, though proportions and tasks will change.
- One view: devs become spec writers, architectural decision‑makers, and AI supervisors; a more extreme view says AI will eventually take those over too, like any screen‑based job.
Learning to code vs “translator” roles
- Some say deep programming is a bad investment; better to become a domain expert (construction, mining, healthcare, etc.) who uses AI to build tools.
- Pushback: you can’t reliably use AI to generate and assess code if you don’t understand code; domain + programming is portrayed as a “superpower.”
- There’s broad support for “X + CS”: accountants, analysts, and managers who can code are seen as 10x in their roles.
Reading vs writing code, fundamentals
- Disagreement whether reading and writing code are the same skill; several say reading large, messy codebases is distinct and much harder.
- Some fear LLMs will atrophy code‑reading and deep thinking skills; others note LLMs can already find subtle bugs better than they can.
- Many still value fundamentals (data structures, low‑level concepts, complexity control) as the durable part of CS education.
Vibe coding, complexity, and architecture
- “Vibe coding” via LLMs is praised for speed but criticized for generating “slop”: redundant, overly complex, or inconsistent systems.
- Concerns: rapid feature output without understanding, accumulating code faster than humans can grok, and agents struggling with brownfield/legacy systems.
- Several argue architecture, refactoring, and complexity management gain importance as AI makes code volume cheap.
LLMs vs compilers and determinism
- Major thread argues prompting→code is unlike high‑level→assembly: you can reason about compiler output; LLM behavior must be probed experimentally.
- Some label LLMs “chaotic”: tiny prompt changes can yield large, unpredictable differences, even if temperature is 0.
- Others counter that real‑world compilers and systems are already unpredictable enough that most devs just “poke until it works” anyway.
Education, struggle, and job market
- Debate over forcing students to hand‑code: some see necessary “character‑building” struggle; others say no one will forgo an AI “do it” button.
- AI as “great TA” is a recurring theme when used for explanations, not turnkey solutions.
- Multiple comments describe the current junior job market as exceptionally bad, with networking (broadly defined) and non‑pure‑programmer paths suggested as survival strategies.