Anthropic’s actions this time are really a bit baffling—treating defensive research as something to be repelled “by attack” is effectively giving hackers a boost; it puts security researchers in shackles, and in the long run, it’s like moving a stone to drop on your own foot.

Wu Shuo learned that Uniswap founder Hayden Adams posted a statement opposing the new model Claude Fable 5 introduced by AI company Anthropic's safety diversion mechanism. He pointed out: "Fable 5's safety protection mechanism detects and intercepts requests related to cybersecurity. This is an extremely bad decision because legitimate requests aimed at strengthening and defending systems (white hats) are very likely indistinguishable at the underlying logic level from malicious hacker attack requests."