PR spam today looks like email spam in the early 2000s

Original Article ↗ Hacker News Discussion ↗

Motivations for AI‑Generated PRs

  • Many think AI PR spam is driven by career incentives: padding GitHub profiles, meeting school/job requirements, or “brand building,” not helping OSS.
  • Others argue many contributors genuinely believe they’re helping or finally able to contribute using LLMs, even if they lack skills.
  • Several note mixed motives: altruism and self‑interest often coexist; people choose helpful activities that also benefit their résumé or ego.

Impact on Open Source Projects

  • Maintainers report being overwhelmed by low‑quality, AI‑authored “slop” PRs, often from people who don’t understand the code.
  • Some now close AI‑generated PRs on sight, disable public PRs, or only accept contributions from known/trusted people.
  • Contributors who try hard to submit high‑quality fixes feel discouraged when they’re treated as noise and ignored.

Altruism, Reputation, and Incentives

  • Discussion explores whether any contribution framed as altruistic is really self‑interested (feeling good, social status, “doing something visible”).
  • Examples like inefficient food banks are used to argue that people often prefer visible, hands‑on “help” over actually effective help.
  • Open‑source contribution as a hiring signal is seen as Goodhart’ed: once it became a target, spammy contributions eroded its value.

Proposed Technical and Social Mitigations

  • Ideas: donate money or token credits instead of “drive‑by” AI PRs; let maintainers decide how to use resources.
  • Some advocate blanket bans on AI‑generated PRs and even AI‑free hosting platforms.
  • Others prefer filters over bans, comparing this to email spam and favoring automated triage over hard rejection.
  • Social gates: requiring first‑time contributors to join a short audio/video call, or at least answer “what problem were you hitting?” before review.

GitHub and Ecosystem‑Level Ideas

  • Mention of GitHub PR limits and PR disabling; skepticism that rate limits alone will meaningfully reduce spam.
  • Suggestions for reputation systems based on PR history, org‑based trust, “guilds” that vouch for members, or web‑of‑trust–style models.
  • Concerns that org‑based trust could drift toward privileging big corporate accounts and excluding individuals.

Broader Reflections

  • Comparisons to early‑2000s email spam: cheap-at-scale actions turning once‑intimate, high‑signal channels into noisy ones.
  • Some see this as contributing to an “OSS is dead” feeling; others think new ecosystems and alternative projects may benefit.