PickVidsNicholas Carlini - Black-hat LLMs | [un]prompted 2026
推荐理由
This talk changed my view on where we're actually at
说明
这是2026年由unprompted发布的一部演讲，主题归在Security & hacking。它讲的是llm vulnerability research这一支里的具体问题，而不是把整个领域扫一遍。适合已经能看懂基本术语、想继续往下走的人。内容上可以概括成：Nicholas Carlini of Anthropic talks at [un]prompted 2026 about how adversaries could misuse current LLMs for attacks. He shows recent models finding zero-day vulnerabilities in heavily tested software, including a SQL injection in Ghost.这些读者给出的理由，核心是：HN cites it as a concrete demonstration that LLMs have suddenly become strong at finding and exploiting vulnerabilities.它更像一份当时的现场记录，适合了解那个节点的争论。从2026年到2026年，有9位不同的读者在2个不同月份把它推荐出来，其中明确叫别人去看的有1次。适合一次性看完。讲者把一个判断讲透，演示是在给这个判断配例子。如果已经会了基本定义，就把注意力放在它怎么把llm vulnerability research变成能跟着复述的例子。
视频信息
- 频道
- unprompted
- 时长
- 26 分 28 秒
- 发布年份
- 2026
同主题更多
- Behind the Scenes of iOS Security
- Why Electronic Voting Is Still A Bad Idea
- USENIX Security '18-Q: Why Do Keynote Speakers Keep Suggesting That Improving Security Is Possible?
- Defcon 18 Changing threats to privacy Moxie Marlinspike Part
- 28c3: The Science of Insecurity
- DEF CON 25 - John Sotos - Genetic Diseases to Guide Digital Hacks of the Human Genome