互联网大规模被AI攻击成近忧，而非超级智能
Wake up, people! What we should actually fear, near term, is not so much rogue superintelligence as...
Gary Marcus说，现在用AI攻击服务器很方便，花1万亿美金就能获得管理员权限，花0美金就什么也得不到，这很可怕。
Gary Marcus指出，目前互联网正面临AI驱动的黑客攻击威胁。他提到，只要有足够资金，就能指挥无限数量的AI机器人尝试通过任何公开方式攻击任何服务器，且不会停止。这一现象已发生，且不需要超级智能。目前只有硅谷少数人尝试，因为他们拥有海量计算资源。
Wake up, people! What we should actually fear, near term, is not so much rogue superintelligence as...
Wake up, people! What we should actually fear, near term, is not so much rogue superintelligence as unleashed agentic AI hacking the internet at scale: “[If you have enough money], you can now instruct an infinite number of bots to randomly attempt to hack any server through any possible published mechanism anywhere on the internet, and never stop until it is accomplished” a. That is basically already happening. b. It doesn’t require superintelligence or RSI or “civilizations” or any other speculative bullshit. it’s here now. c. We are doing literally nothing to stop it. d. It will be bad. Dr_Gingerballs @Dr_Gingerballs Hear me out... You can now instruct an infinite number of bots to randomly attempt to hack any server through any possible published mechanism anywhere on the internet, and never stop until it is accomplished. One button could conceivably lead to access to infinite money. Yet the only people who have attempted it are a handful of people in Silicon Valley. What do they have that everybody else does not? Unlimited access to $1T of computing resources. So $1T and AI gets you admin access to a few secure databases. And $0 and AI gets you admin access to nothing. I think I see the problem here and it's not AI. 🔗 View Quoted Tweet 💬 1 🔄 0 ❤️ 7 👀 1061 📊 2 ⚡