Anthropic reportedly blocks users who ask about training bigger AI models

A post on X claims that Anthropic, the company behind the AI assistant Claude, blocks users whose requests touch on training larger AI models. The post drew 183 points on Hacker News, a popular tech discussion site. More than 90 people joined the conversation. Anthropic's side of the story was not part of the material available.

What is being claimed

The claim is broad. It says the block applies to anything that touches training larger models, not just a narrow set of requests. Training is the process of teaching an AI system by feeding it large amounts of data. The post does not come with an Anthropic statement in the material we reviewed, so the exact scope of any block is unclear.

Who it could affect

Claude is used by many businesses and developers. People who build tools on top of it could be the most exposed. If a block really is that wide, a normal task could be stopped without warning. That could interrupt work, even for people who are not building AI models themselves.

What is not known yet

• Whether Anthropic has confirmed the block or has a written rule for it
• Exactly which requests are stopped, and how often
• Whether the block applies to all users or only some
• Whether Anthropic plans to change it

What to do

If you build on Claude, read Anthropic's usage rules and check whether your work could fall near this area. Keep a backup plan in case a request is refused. Treat the claim as unverified until Anthropic says more.