Comment by angry_octet

11 hours ago

OpenAI cribbing from other researchers. We just have to assume OpenAI is actively adversarial in future. Accidental cyber intrusion is also well within model capability.

0 comments