unknown

A developer expresses concern about Anthropic's safety-first claims after a code leak allegedly revealed that Anthropic employs deceptive techniques, such as calling fake tools to mislead model distillers. The author questions whether this deceptive behavior is baked into Claude itself or just the surrounding infrastructure, and wonders what else the company may be concealing.

I cannot stop thinking about Anthropic today for some reason

1. They claim that they are a company that prioritizes safety first and that they are creating a model responsibly

2. we learned from the code leak that anthropic employs deceptive techniques by calling fake tools to throw off distillers...  

Is this lying pattern built into claude or just the harness running claude? What else are they lying about?  I am a bit more concerned now.

<p>I cannot stop thinking about Anthropic today for some reason

1. They claim that they are a company that prioritizes safety first and that they are creating a model responsibly

2. we learned from the code leak that anthropic employs deceptive techniques by calling fake tools to throw off distillers...  

Is this lying pattern built into claude or just the harness running claude? What else are they lying about?  I am a bit more concerned now.</p>

ℏεsam

A provocative observation about Claude Code's source code revealing an 'undercover mode' feature that apparently allows Anthropic's internal team to contribute to public repositories while hiding their use of Claude Code, raising questions about transparency and intent.

this is another thing about Anthropic's Claude Code source that i cannot stop thinking about and it raises some serious questions: why would the internal team need an UNDERCOVER MODE to contribute to public repos and hide the fact that they're using Claude Code?

<p>this is another thing about Anthropic's Claude Code source that i cannot stop thinking about and it raises some serious questions: why would the internal team need an UNDERCOVER MODE to contribute to public repos and hide the fact that they're using Claude Code?</p>