Leave Me Behind
A personal essay from an Android developer who feels alienated by AI-assisted coding prompts a wider examination of how large language models are reshaping software work, learning, and identity. Commenters are split between seeing LLMs as powerful tools that boost productivity and unlock new projects, and warning that they erode craftsmanship, deskill programmers, worsen software quality, and concentrate power in proprietary platforms. Many note that management pressure, economic incentives, and broader social effects — from job insecurity to reduced human connection — may matter as much as the technology’s raw capabilities.
Craft, Community, and Loss
- Many agree the author is expressing grief over losing a beloved craft and sense of community (pairing, labs, meetups), not just complaining about tools.
- Some argue this nostalgia is real and valid; others see it as an existential crisis or sunk-cost issue that ignores that tools have always changed crafts.
- Several note that even before LLMs, industrial practices, Jira, remote work, and career progression already eroded the “joyful, communal” phase of programming.
AI as Tool vs Threat to Learning
- One camp treats LLMs as “Iron Man suits”: accelerators for typing, boilerplate, search, refactoring, tests, and small maintenance, while humans still design and reason.
- Another camp fears “outsourcing thinking,” deskilling, and a drug‑like dependence where people stop doing things they once could do, leading to “cognitive surrender.”
- Some solo or geographically isolated devs say AI is the first “collaborator/mentor” they’ve ever had and is empowering rather than dehumanizing.
Code Quality, Maintenance, and “Slop”
- Critics say stochastic code generation lowers the floor: it enables cheap, disposable, low‑quality software and encourages “vibecoding” without understanding.
- Others report the opposite in their own projects: more tests, better CI, more refactoring, and AI code review catching subtle bugs.
- Several stress that quality depends on how tools are used: careful review vs. blindly accepting agent output. The long‑term effect on the median codebase is seen as unclear.
Jobs, Inequality, and Historical Analogies
- Repeated comparisons to machinists→CNC operators, artisans→mass production, and furniture makers→IKEA: AI may make “hand‑crafted” software a niche craft.
- Disagreement over whether this is “just another automation” (manual labor analogy) or fundamentally different because it targets cognition and learning.
- Some warn of job devaluation and token‑usage metrics; others argue skills in specification, review, and system design will remain central.
Human Intelligence, Distraction, and Meaning
- Several worry AI plus existing media tech accelerate distraction, lazy thinking, and societal “dumbing down,” even if aggregate knowledge and GDP rise.
- Others argue human heedlessness predates AI; these tools mostly expose and scale existing tendencies.
- Multiple comments stress that people derive meaning from work; large‑scale removal of meaningful craft and autonomy is seen as socially dangerous.
Workplace Pressures and Power Concentration
- Reports of management demanding maximum AI usage, PR throughput, and token metrics create a zero‑sum, fear‑driven environment.
- Concerns that proprietary AI centralizes power and rents, unlike tools like git or Postgres; calls for open models and skepticism about further corporate control.
- A minority is enthusiastic about entrepreneurial upside and personal experimentation; another group prefers to “opt out” or stick to non‑AI workflows.