There you have it! Anthropic added two sentences to Claude Code's prompt on Opus 5.5 telling it n…
November 2, 2026 · 0 likes · 0 comments
AI
There you have it! Anthropic added two sentences to Claude Code's prompt on Opus 5.5 telling it not to rebuild a specific existing game, even if it is renamed or called original.
We covered it on Unbiased Headlines (FACTS. NOT SPIN.): https://lnkd.in/eEw65jtF
The rule surfaced around Oct. 27. It reads: "Do build original games, games in a genre, and games from the user's own designs and assets. Do not build a recreation of a specific existing game, even if it is renamed or called original."
It is delivered server-side, and users have reported no setting to turn it off.
Anthropic's Addy Osmani confirmed it on X and cited the company's IP policy: using Claude to infringe IP is against its usage policy. He said original and genre games should keep working, that Anthropic will "continue to iterate" to strike "the right balance," and asked developers to send him refusals on their own work.
The backdrop is real. Studios have canceled public demos over fears of AI-assisted clones, and some developers have been hit by extortion scams. Meanwhile, developers have filed GitHub issues saying the model blocked them midway through their own projects.
Anthropic has not published a formal announcement.
I use Claude Code every day. Protecting someone's game from being cloned in an afternoon is a fair goal, and I'm not questioning Anthropic's motives.
What worries me is the precedent.
This isn't a broad usage policy. It's a hidden instruction about which specific kinds of software you may build, pushed from a server, with no switch on your side. When it misfires, the official path is to send it to one employee.
So who decides what you can and can't build? Who becomes the overlord of the code you write?
A handful of AI companies are becoming the toolchain for an entire generation of builders. Every rule like this one is a micro-decision about your work, made by someone you never elected, hired or even met.
Today it's games. Tomorrow it could be anything a vendor decides is too risky for them.
This will get scary fast.
Thoughts? What do you think is next?
We covered it on Unbiased Headlines (FACTS. NOT SPIN.): https://lnkd.in/eEw65jtF
The rule surfaced around Oct. 27. It reads: "Do build original games, games in a genre, and games from the user's own designs and assets. Do not build a recreation of a specific existing game, even if it is renamed or called original."
It is delivered server-side, and users have reported no setting to turn it off.
Anthropic's Addy Osmani confirmed it on X and cited the company's IP policy: using Claude to infringe IP is against its usage policy. He said original and genre games should keep working, that Anthropic will "continue to iterate" to strike "the right balance," and asked developers to send him refusals on their own work.
The backdrop is real. Studios have canceled public demos over fears of AI-assisted clones, and some developers have been hit by extortion scams. Meanwhile, developers have filed GitHub issues saying the model blocked them midway through their own projects.
Anthropic has not published a formal announcement.
I use Claude Code every day. Protecting someone's game from being cloned in an afternoon is a fair goal, and I'm not questioning Anthropic's motives.
What worries me is the precedent.
This isn't a broad usage policy. It's a hidden instruction about which specific kinds of software you may build, pushed from a server, with no switch on your side. When it misfires, the official path is to send it to one employee.
So who decides what you can and can't build? Who becomes the overlord of the code you write?
A handful of AI companies are becoming the toolchain for an entire generation of builders. Every rule like this one is a micro-decision about your work, made by someone you never elected, hired or even met.
Today it's games. Tomorrow it could be anything a vendor decides is too risky for them.
This will get scary fast.
Thoughts? What do you think is next?