the company that markets itself as the safety-first AI lab just shipped its own source code to the public making it Anthropic's second major security blunder in a week.... You can't make this shit up
