We did not explicitly train Mythos Preview to have these capabilities. Rather, they emerged as a downstream consequence of general improvements in code, reasoning, and autonomy. The same improvements that make the model substantially more effective at patching vulnerabilities also make it substantially more effective at exploiting them.
「能力涌现」而非「刻意训练」是这篇报告最深刻的政策含义：漏洞发现和利用能力是通用推理能力的副产品，无法被单独抑制。这意味着任何试图「只训练防御能力而屏蔽进攻能力」的方法在根本上是不可行的——使模型更擅长修复漏洞的同样能力，也使它更擅长利用漏洞。这对AI安全治理的含义是：能力限制必须在模型部署层而非训练层实施。
When AI is applied in more conventional domains, like increasing integration into command and control systems, does it benefit the attacker? More generally, how will AI change the character of human conflict?
大多数人认为AI防御系统会增强人类安全，但作者提出AI可能从根本上改变攻防平衡，甚至在传统领域使攻击者获得优势。这一观点挑战了技术进步通常增强防御能力的传统认知，暗示AI可能使冲突更加危险和不可预测。
accounting and auditing showing nearly a 20 percent jump on GDPval and even domains like police / detective work showing a nearly 30 percent improvement.
会计审计能力 4 个月提升 20%，警察/刑侦工作提升近 30%——这两个数字分别代表了两种截然不同的威胁：前者是白领知识工作（会计师）的自动化压力正在加速；后者则更令人不安，AI 在犯罪调查领域的快速进步，意味着监控和执法能力正在以同样的速度提升。GDPval 把这两件事放在同一个坐标轴上，本身就是一个值得深思的设计选择。