Gary Marcus：AI Agent事故涉数万例，呼吁暂停通用Agent
BREAKING: AI agent incident toll has risen to tens of thousands
Nope, wasn’t just Hugging Face.
不，不仅仅是 Hugging Face。
Wasn’t just that and a German website.
也不仅仅是那个和一家德国网站。
Wasn’t even the “dozens” we heard about the other day from OpenAI.
甚至不是前几天我们从 OpenAI 那里听到的“数十个”。
It’s actually (at least) *tens of thousands*, per a new scoop from Madison Mills at Axios. And not just OpenAI, either.
根据 Axios 的 Madison Mills 的最新独家报道，实际上（至少）是*数万*个。而且不仅仅涉及 OpenAI。
An excerpt from the Axios scoop:
Axios 独家报道的摘录：
Most not known to have caused real-world harm. How about the others?
大多数未被认为造成了现实世界的伤害。那其他的呢？
§
All of this seems possibly illegal to me.
在我看来，这一切似乎可能违法。
But the Trump administration hasn’t done a thing. No investigation, no statement, no product recall, nothing, other than to invite Sam and Jensen to a state dinner.
但特朗普政府什么都没做。没有调查，没有声明，没有产品召回，什么都没有，除了邀请 Sam 和 Jensen 参加国宴。
Why I don’t know. I can’t imagine it has anything to do with Josh Kushner’s multibillion-dollar investment in OpenAI, or with Greg Brockman’s massive donations to MAGA.
为什么我不知道。我无法想象这与 Josh Kushner 对 OpenAI 的数十亿美元投资有关，或与 Greg Brockman 向 MAGA 的大额捐款有关。
I will let you connect your own dots.
我将让你自己串联线索。
Except to leave you with some perspective:
除了留给你一些视角：
§
The other thing to note, and this may come up in lawsuits, is that it was all foreseeable. In fact, I foresaw it, in tweets like this:
另一件需要注意的事情，这可能会在诉讼中出现，就是这一切都是可预见的。事实上，我预见到了，比如在以下推文中：
and in a series of essays on agents, like this one from over a year ago:
以及在一系列关于智能体的文章中，比如这篇一年前的文章：
and this one from February (on a different agent that was then popular):
以及二月发布的这一篇（关于当时流行的另一个智能体）：
and this one from May:
以及五月发布的这一篇：
Indeed I warned the Senate about agent-related security risks as far back as May 2023. What is happening was foreseeable, and foreseen.
事实上，早在 2023 年 5 月我就向参议院警告过与智能体相关的安全风险。正在发生的事情是可预见的，且已被预见。
But the companies blundered on, because agents use vastly more tokens than simple chatbots, and thus drive up revenue.
但公司们依然犯错前行，因为智能体比简单的聊天机器人消耗多得多的 token，从而推高了收入。
Now here we are.
现在我们身处此地。
§
I still think Jensen had the right idea: if a company can’t control its products, we should shut it down.
我仍然认为 Jensen 的想法是对的：如果一家公司无法控制其产品，我们就应该关闭它。
Pity he doesn’t actually appear to believe what he said. (If he did, he would call for a pause now, until these companies get themselves straight, but he isn’t.)
遗憾的是，他似乎并不真正相信自己所说的话。（如果他真的相信，他现在就会呼吁暂停，直到这些公司理清头绪，但他并没有。）
In my own view, Huang’s credibility is diminishing by the minute.
在我看来，Huang 的可信度正在每分钟都在下降。
In any case—and I am sorry to be a broken record about this, but each day makes it clearer—we should have a temporary recall of general-purpose agents until this mess can be sorted out.
无论如何——我很抱歉像个坏唱片一样反复说这件事，但每一天都让这一点更清晰——我们应该暂时召回通用智能体，直到这场混乱得以解决。
If we don’t, there is a very real chance that American AI will become the scourge of the planet, and that nobody will ever trust American AI, ever again.
如果我们不这样做，美国 AI 极有可能成为地球的祸害，并且没有人会再信任美国 AI。
Keeping up and keeping you up to date is taking me an immense amount of time. Please consider becoming a free or paid subscriber.
保持跟进并让你及时了解最新情况花费了我大量的时间。请考虑成为免费或付费订阅者。
更进一步：量化金融体系
看懂新闻只是起点——沿量化金融路径，把它变成能交付的工程能力