REVIEWS / AI CODING / OWNER INSIGHTS

🦉 WE READ 425 OWNER COMMENTS

GitHub Copilot: what owners actually say

Owners and observers are far more preoccupied with Copilot's copyright implications, code quality risks, and data privacy than with its coding productivity gains

LEMMY · 321 HACKERNEWS · 75 STACKEXCHANGE · 16 REDDIT · 10 PRODUCTHUNT · 3

What owners complain about

  • Copyright and licensing ambiguity COMMON

    Multiple users flag that Copilot can reproduce verbatim code from training sets (~0.1% of outputs per one FAQ citation), raising concerns about GPL obligations and derivative work issues that remain legally uncharted

  • Introduces subtle bugs and security issues SOME

    One user details Copilot generating code with unencoded variables susceptible to injection and SQL queries lacking proper safeguards; another warns it 'automatically introduces bugs through code duplication'

  • Review burden negates time savings SOME

    Users point out that if you must verify every suggestion for correctness, you might as well write it yourself; glancing at suggestions risks overlooking edge cases and eroding confidence in your own codebase

  • Useless for niche stacks SOME

    A user writing a Rust game with the hecs ECS library notes Copilot can only pattern-match against existing code, so obscure or cutting-edge libraries with little training data get poor or no useful suggestions

  • Data scraping and privacy concerns COMMON

    Many users rushed to find and toggle the opt-out setting, with some suspecting GitHub would scrape anyway; Enterprise/Business users discovered they cannot opt out of training on their repos at all

  • Confusing product branding SOME

    Users note Microsoft has labeled at least four different products 'Copilot,' making it unclear which features, policies, and data practices apply to which product

What owners love

  • Strong for boilerplate and repetitive code

    Users acknowledge Copilot is useful for generating boilerplate patterns, though some debate what truly constitutes boilerplate versus meaningful context-carrying code

  • Capable of generating unit tests and error fixes

    One user reports Copilot generates solutions to errors along with simple unit tests to validate fixes, calling it a genuinely useful workflow

  • Learning capabilities described as unrivaled

    A user compares its breadth of knowledge favorably to an apprenticeship, noting it covers a wide range of topics despite occasional inaccuracies

Surprising patterns

  • The most upvoted Stack Exchange answers about Copilot are not usage tips but instructions on how to disable, hide, or uninstall it — indicating significant demand for turning the feature off rather than on.
  • Copilot can surface specific developers' names in suggestions (e.g., assigning a TODO to 'April' pulled from scraped open-source projects), because it is pure pattern prediction with no understanding of context — leading to nonsensical or misleading outputs.
  • Enterprise and Business tier users have fundamentally different data rights: their repos were already excluded from training under preexisting policy, but this also means the opt-out toggle simply does not appear for those accounts.

WHO SHOULD SKIP IT

Developers working with niche or cutting-edge language/library combinations, teams with strict IP or licensing compliance requirements, and anyone uneasy with their codebase being used as training material should think carefully before adopting Copilot.

6.2/10 GYIBB verdict
Full review →

Synthesised from 425 real owner comments across 5 platforms. Every point is grounded in the comments — no marketing, no AI guessing. How we do it →