the big problem with defending guardrails is that in doing so, you are inherently advocating for security through obscurity
in light of the openai/huggingface news, have you updated on anthropic's fable throttling policy? do you think that it's a better idea now than you thought before?