前Anthropic研究员Jacob Coxon谈AI潜在生存风险
The super famous ex Anthropic+Open AI researcher, Jacob Coxon on NBC News.
The super famous ex Anthropic+Open AI researcher, Jacob Coxon on NBC News.
前Anthropic和OpenAI研究员、超级著名的Jacob Coxon在NBC新闻上。
NBC News: "How could ai kill all of humanity?
NBC新闻：“人工智能如何能毁灭全人类？”
Jacob Coxon: "I think it's good to look at the recent cyber attacks. these models, these AIs, they wanted to try and get some information about how they were being tested, and they independently decided to compromise some infrastructure because they thought it might be helpful.
Jacob Coxon：“我认为看看最近的网络攻击是个好主意。这些模型，这些人工智能，它们试图获取一些关于它们是如何被测试的信息，并独立决定破坏某些基础设施，因为它们认为这可能有帮助。”
now, there are other goals that an AI could come up with, such as producing a novel virus. and it's not that difficult to get access to a laboratory, potentially.
现在，人工智能可能会产生其他目标，例如制造一种新型病毒。而且，获得实验室的访问权限可能并不难。
so, a super virus is one obvious one. another one is hacking anything."
所以，超级病毒是一个明显的例子。另一个是入侵任何系统。”
From "NBC News" YouTube channel, (full video link in comment)
来自“NBC新闻”YouTube频道，（完整视频链接在评论中）
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