① 他说过的话

9 条

但现实是，至少就目前而言，AI 模型吐出的很多代码在根本上并不安全。
But the reality is, at least as it sits today, a lot of code that AI models spit out is not fundamentally secure.
—— Vineet Edupuganti · [05:48]

指向原始笔记的链接

防守者面临的动态是，要做的工作远多于做这些工作的人。
And the dynamics for defenders are there’s a lot more work to do than people to do it.
—— Vineet Edupuganti · [06:09]

指向原始笔记的链接

但我认为人们正在意识到，从五年的视角来看，当你被机器攻击时，你不能坐在那里等变更控制流程。
But I think people are realizing that as you take a five-year view, when you’re being attacked by a machine, you can’t sit around waiting for change control.
—— Vineet Edupuganti · [12:33]

指向原始笔记的链接

而对于我最关键任务的生产服务器，我仍然想要人在回路中。
And for my most mission critical production servers, I still want a human in the loop.
—— Vineet Edupuganti · [14:27]

指向原始笔记的链接

但你需要达到 99.99% 的准确率。95% 是不够好的。
But you need to be 99.99% accurate. 95% is not good enough.
—— Vineet Edupuganti · [22:52]

指向原始笔记的链接

所以预测二，也就是 Cogent 正试图引入的这个世界的本质，是我们基本上就是那个自主蓝队，持续地发现你的暴露点，并在攻击者真正攻进来之前缓解它们。
And so prediction number two, and really the world that Kojin is trying to usher in here, is we are basically the autonomous blue team that is continuously finding your exposures and mitigating them before the attacker can actually get in.
—— Vineet Edupuganti · [25:04]

指向原始笔记的链接

我们在过去几十年里建立了一整套保护人类的原则和控制措施，但没有任何人在思考我们如何保护这些 AI 智能体。
There’s this whole set of principles and controls that we’ve built up over the past few decades to go and protect the human, and nobody’s thinking at all about how do we secure these AI agents.
—— Vineet Edupuganti · [25:31]

指向原始笔记的链接

我见过市场销售团队的成员，按各种标准衡量他们都不该会写代码，但这些人正在构建应用程序和工具，去解决他们自己遇到的问题。
I’ve had members of the go-to-market team where by all measures they should not be able to write code and these guys are building applications and tools to go and solve the problems that they have.
—— Vineet Edupuganti · [28:40]

指向原始笔记的链接

我能够把他们的 30 年经验，在大约八个月的时间里装进我的脑子，对吧?
I was able to take their 30 years of experience and get it into my brain in like eight months, right?
—— Vineet Edupuganti · [36:32]

指向原始笔记的链接

② 出现在这些集

1 集

• 《当攻击用 AI、防守靠智能体：Abnormal 与 Cogent 聊智能体网络安全》 — 作为联合主持

③ 他谈到的

点进去有真内容 —— 本页主要出口

Corinne Riley · Evan Reiser · Abnormal · Cogent Security · 智能体 · 漏洞管理 · 自主修复 · 邮件安全 · 身份威胁检测 · 人在回路

④ 也在聊「智能体」的人