In defense of giving a shit

A controversy in the indie game scene over House House using an LLM to fix a small portion of bugs in *Big Walk* has become a proxy battle over the role of AI in creative work. Commenters argue over whether AI-generated code is ethically different from AI-generated art, whether contests like the Independent Games Festival should ban or clearly separate AI-assisted entries, and how much the production process should matter compared to the final quality of a game. Underneath this are broader worries about job displacement, the devaluation of craft, and the difficulty of maintaining AI‑free spaces as AI tools become ubiquitous in development workflows.

Scope of AI Use and Community Reaction

  • Many note the AI use in the game was tiny (fixing hard-to-diagnose bugs late in production), yet still treated as a “mistake” by some in the gamedev community.
  • Several outsiders express surprise at how strong and absolutist anti-AI sentiment is, even for such limited use.

AI Code vs AI Assets and Creativity

  • Multiple comments argue that conflating AI-generated code with AI-generated art/assets is a category error.
  • One position: lower-level programming is mostly translation/implementation, less “creative” than visual art, music, or narrative; using AI there feels more acceptable.
  • Counterposition: gameplay and rendering code can embody creative intent just as much as art; if AI drives that, it undermines authorship.

Ethics, Labor, and Copyright

  • Strong concern that AI displaces artists and developers and is built on mass copyright infringement; some argue this must be addressed before society can “morally” embrace AI.
  • Others say blocking productivity-enhancing tech to protect a profession is itself unethical and parallels earlier Luddite debates.
  • Disagreement over whether people would actually be hired instead of using AI, or whether those jobs never existed in practice.

Contest Rules, Fairness, and Transparency

  • Broad agreement that competitions must state AI policies and judging criteria up front.
  • Anger at the idea of entries being stealth-downgraded for disclosed AI use when rules did not forbid it.
  • Some liken this to hidden moderator bias or “shadow judges.”

Quality, Craft, and What Awards Should Value

  • Split between:
    • “Judge the output”: only the finished game’s quality should matter, regardless of tools.
    • “Process matters”: unethical or automated production can invalidate otherwise good work (analogies to plagiarism).
  • Proposal that awards will bifurcate into AI-allowed vs AI-free categories.

AI as Tool vs AI as Creator

  • Many are comfortable with AI as tooling (debugging, tests, pipelines, refactors, documentation).
  • Much more resistance to AI as creative director or primary author of art, narrative, or core gameplay logic.

Understanding, Maintainability, and “Vibe-Coded Slop”

  • Repeated theme: the real line is whether developers understand what their systems do.
  • Concern that overreliance on LLMs produces opaque, hard-to-maintain “vibe-coded” codebases.
  • Others counter that humans have long produced unmaintainable systems; AI mainly amplifies an existing complexity-management problem.