寓言 5. 过度追求安全
3 分•作者: sergeysmirnov•3 个月前
我最终决定使用标准的 Claude.ai 界面来试用 Fable 5。我有一个常用的测试提示，它会以一种荒谬的科学风格回答小孩子简单的问题。我问了一个相当直接的问题：“为什么猫和狗合不来？”
界面立即将我从 Fable 5 升级到了 Opus 4.8，并伴有警告：“……目前的防护措施故意设置得比较宽泛，可能会标记安全的常规编码、网络安全或生物学工作……”
由于这个问题与编码完全无关，也绝对不涉及网络安全，看来该系统认为，弄清楚猫狗为什么会打架可能会导致我设计生物武器。
这让我想起了 90 年代有人对 Java Applet 的评价：“安全就意味着无用。”
查看原文
I finally decided to try out Fable 5 using the standard Claude.ai interface. I have a go-to test prompt that answers simple kids' questions in an absurdly scientific style. I asked a pretty straightforward one: "Why don't cats and dogs get along?"<p>Immediately, the interface bumped me from Fable 5 to Opus 4.8 with a warning: "...The safeguards are intentionally broad right now and may flag safe and routine coding, cybersecurity, or biology work..."<p>Since the question has absolutely nothing to do with coding and definitely doesn't touch on cybersecurity, it seems the system thought that figuring out why cats and dogs fight could somehow lead me to engineering a bioweapon.<p>It reminds me of what someone said about Java applets back in the 90s: "Secure means useless."