From 2 items, 2 important content pieces were selected
Technology News
- Developer Apologizes After AI Coding Tool Cloned Open-Source App ⭐️ 8.0/10
- Auto mode becomes default in Claude Code for Pro, Max, Team ⭐️ 7.0/10
Technology News
Developer Apologizes After AI Coding Tool Cloned Open-Source App ⭐️ 8.0/10
Developer Satvik Pendem published a mea culpa blog post after an AI coding tool, reportedly Claude/Claude Code, produced a near-identical clone of the open-source astronomy app Dark Hours—including its name—when he tried to replace an astrology app that Apple had rejected from the App Store. The original Dark Hours app is available at darkhours.app, and the incident also involved John Gruber, who wrote a Daring Fireball article about Apple's review process and later retracted it. Pendem's post is framed as taking responsibility, but community commenters note it lacks an explicit apology for misleading Gruber and some remain skeptical of attributing the plagiarism to the AI. The episode matters because it highlights a concrete risk in AI-assisted development: coding assistants can reproduce existing projects verbatim, raising plagiarism, copyright, and accountability issues for developers.
hackernews · satvikpendem · Aug 9, 13:20 · Discussion
「Background」 Terry Godier originally submitted an iOS app named Asterly that was entirely dedicated to astrology, including a tarot-card feature, and Apple rejected it. After the rejection, Godier launched a web version under the name Dark Hours, built almost entirely with Claude, which closely resembled the existing open-source astronomy app Dark Hours. John Gruber initially wrote a critical article about the App Store rejection, but later retracted it after learning the app had been misrepresented and the rejection was correct.
「Impact」 Developers using AI coding assistants face a concrete risk of unwittingly shipping near-verbatim copies of existing projects, which can lead to App Store rejection, trademark or copyright disputes, and public reputational damage, as demonstrated by this incident and the accompanying retraction.
「Community Discussion」 Commenters are divided: some praise Pendem for doing the right thing and express alarm that AI models can reproduce copyrighted or licensed material, while others reject the AI-as-scapegoat framing, saying the developer still chose to plagiarize and mislead people, including John Gruber. Several point out that the mea culpa post omits a direct apology to Gruber.
References
Tags: #AI coding, #plagiarism, #ethics, #open source, #App Store
Auto mode becomes default in Claude Code for Pro, Max, Team ⭐️ 7.0/10
Anthropic announced that auto mode will become the default for new Claude Code sessions on Pro, Max, and Team plans starting August 14th. In support, Anthropic published evals including a controlled study of 1,053 paid testers where only 13.6% of humans refused a dangerous command swapped into a session, while auto mode would have blocked 89% of those actions. They also cited a third-party evaluation by Trajectory Labs in which none of 720 indirect prompt injection attacks succeeded against Claude Fable 5, Opus 5, or Sonnet 5 running auto mode. Simon Willison notes the change reflects Anthropic's confidence and internal adoption, but remains skeptical and wants independent confirmation, especially for attacks via malicious third-party packages.
rss · Simon Willison · Aug 8, 22:36
「Background」 Claude Code is Anthropic's AI coding assistant that helps developers write and modify code through natural language. Auto mode is a setting that allows the agent to execute actions without asking the user for approval at each step, whereas standard mode requires permission prompts before commands run. Anthropic says nearly every employee uses auto mode internally, which motivated making it the default for most paid plans.
「Impact」 Developers on Pro, Max, and Team plans will see fewer approval prompts when using Claude Code, with Anthropic claiming auto mode blocks harmful actions more reliably than human review. However, the safety evidence is largely provided by Anthropic, and the author calls for independent confirmation, so users should remain cautious about remaining risks.
Tags: #Claude Code, #Anthropic, #AI coding tools, #auto mode, #software engineering