A shocking mishap at Anthropic resulted in the accidental leak of Claude's entire source code, raising serious questions about AI safety and security. Discovered by a security researcher within ...