Vitalik Buterin posted in support of Kimi 3.1 proactively removing advanced cyberattack capabilities from its training data, arguing that restricting AI use cases is more reasonable than restricting specific users. He warned that, in the face of cybersecurity risks posed by AI, the more important priority right now is to buy time, harden systems, review old code, and formally verify new code. Vitalik is also concerned that if the initial AI