Anthropics' Claude cracks a 25-year-old NSA cryptography mystery in under 30 minutes

Anthropic's Claude model successfully identified the seeds behind NIST’s standardized digital signature curves, a mystery that has persisted for 25 years since the NSA first introduced them. By reverse-engineering these parameters, the AI addressed long-standing suspicions regarding potential hidden backdoors in government-approved encryption standards. This result demonstrates a significant leap in the ability of large language models to assist with complex cryptanalysis and security auditing tasks.