Anthropic’s Claude Code and Opus 5.5 are reportedly refusing requests to recreate existing games or perform some reverse-engineering tasks. Users say the restrictions have interrupted legitimate development, including work on original games, familiar mechanics, tabletop-style projects, anti-cheat systems, and security reviews. A rule circulated by a user online says Claude may help create original games but must not reproduce a specific existing title, even under a different name; Anthropic has not confirmed that wording. Other complaints describe refusals involving the Reverse Engineer Anything tool, crash-dump analysis, and removing unwanted software from a television. Reports have increased in recent weeks, though the timing and relationship between the restrictions remain unclear.
The restrictions may reflect concern that AI assistance could enable copyright infringement, game cloning, or other software misuse, especially as AI-generated copies of commercial products and games proliferate. Anthropic has also limited access to advanced cybersecurity capabilities through its Cyber Verification Program, raising questions about whether security safeguards contribute to the refusals. The impact is especially troubling for small studios, which face risks of having unfinished demos copied, while developers say the safeguards can obstruct legitimate work. Some users recommend switching to local, open-weight, or Chinese AI models; OpenAI has reportedly taken similar action in at least one case. The debate highlights a tension between AI companies’ use of copyrighted material for training and their efforts to prevent users from applying AI to reproduce protected work.
Summary generated by AI from the linked article. hn.today is not affiliated with Hacker News or Y Combinator.
Anthropic has restricted Claude from helping with video game modding, decompiling, and reverse-engineering, amid growing use of AI to alter games. The limits are broad enough to disrupt legitimate development and basic troubleshooting, and Anthropic has not publicly explained the restrictions. (mashable.com)
OpenAI’s GPT-6.1 Terra is a midrange model focused on visual reasoning, computer use, and technical communication, with faster computer-use performance than Astra and Sol. It costs $1 per million input tokens and $5 per million output tokens, positioning it between the cheaper Luna and pricier Sol. (simonwillison.net)
Pope Leo XIV argued that AI’s future depends on human choices and that systems should serve people, not be treated as persons. He called for human oversight of high-stakes decisions, steps to ban lethal autonomous weapons, and fair access to AI’s benefits. (garymarcus.substack.com)
Researcher Jiahao Shao says Anthropic blocked or downgraded his access after flagging his work as training a competitor. He says Claude only helped run jobs and analyze results, not generate training data. (twitter.com)
GPT-5.6 Sol completed a peaceful Factorio run in 13.5 hours of game time using a Python tool to control the game. It launched a rocket despite late-game factory problems, improving on an earlier GPT-5.5 attempt that failed to use conveyor belts. (ryanmadden.net)
A language model was post-trained on short Dwitter demos to generate plausible continuations of unfinished code and explore alternative visual effects. The project treats demo remixes as a kind of programming grammar, where tiny code changes can produce strikingly different results. (datavorous.com)
Today's best Hacker News stories, summarized and screenshotted, one email a day.