.@lumpenspace says OpenAI and Anthropic may have a 5–6 month window to use their cyber lead defensively before open-source models catch up: "Right now, the big labs are at an advantage in cyber. If I were OpenAI or Anthropic right now, I would distribute my cyber-enabled AIs as far and wide as possible to all tech companies that ask for them in order to do some hardening." "In the future, there can be open source models with similar capabilities. Unlike the OpenAI or Anthropic models, people…