← Sat, Sep 12 digest

Google stole open source code without crediting the authors (Artemis/Minitap)

read article ↗ hn thread ↗

summary

Google's Artemis project allegedly used code and prompts from the open-source mobile-use project without proper attribution, even removing the original authors' names from the repository history. This lack of credit undermines the open-source ecosystem by discouraging developers from sharing their work if they believe it will be appropriated by large corporations. Proper attribution is not just a legal requirement under licenses like Apache 2.0, but a fundamental ethical practice that sustains collaborative innovation.

key points

  1. Evidence of plagiarism includes identical code for Android device connections, word-for-word prompt instructions, and the reproduction of specific, arbitrary naming conventions and bugs.
  2. The authors discovered that their names were present in an early version of the Artemis package file but were subsequently removed via a force push to the repository.
  3. The issue extends to public benchmarks, where the authors' superior performance results were ignored by leaderboard maintainers and omitted from Google's own comparison charts.

who should read it

Software engineers and open-source contributors who care about intellectual property, licensing, and the ethical treatment of community-driven projects.

discussion skeptical

consensus

Commenters generally view Google's actions as a deceptive attempt to hide license violations and attribution removal. The force push to overwrite author history is widely interpreted as a deliberate cover-up rather than a technical mistake. Most users believe that adding credits after being caught is a minimal effort to regain compliance without genuine transparency.

top insight

The force push of the pyproject.toml file is particularly damning because it didn't just change the author, but effectively erased the record of the original authors from the public commit history. This suggests an intentional effort to obscure the origin of the code rather than a simple mistake in project maintainer lists. This distinction shifts the conversation from accidental non-compliance to systemic dishonesty.

Generated by an LLM from the article and thread. Check the sources before relying on it.