TE
科技回声
首页24小时热榜最新最佳问答展示工作
GitHubTwitter
首页

科技回声

基于 Next.js 构建的科技新闻平台,提供全球科技新闻和讨论内容。

GitHubTwitter

首页

首页最新最佳问答展示工作

资源链接

HackerNews API原版 HackerNewsNext.js

© 2025 科技回声. 版权所有。

Eliezer Yudkowsky's six-minute TED talk on AI

6 点作者 johnaspden大约 2 年前

2 条评论

johnaspden大约 2 年前
He gets a standing ovation at the end, which is nice to see.<p>Apparently this was released by accident and the youtube copy has been set to &#x27;private&#x27;, but luckily some quick-witted person downloaded it and has put a copy where it can be found.<p>I made a copy of that with:<p>$ wget <a href="https:&#x2F;&#x2F;files.catbox.moe&#x2F;qdwops.mp4" rel="nofollow">https:&#x2F;&#x2F;files.catbox.moe&#x2F;qdwops.mp4</a><p>and I&#x27;d be happy to stick that somewhere if this one disappears.<p>Just in case anyone needs to be reassured that this is a thing that can be believed by at least two people, I endorse what he&#x27;s saying here wholeheartedly, and I don&#x27;t think we&#x27;ve got very long left at all.<p>Timescales are the dodgiest bit of this whole argument, and all we have is intuition, but I&#x27;d be absolutely amazed if humanity still existed in ten years time, and pleasantly surprised if we make it to the end of this year.
评论 #35863993 未加载
transfire大约 2 年前
I’m sorry. But he is all sorts of wrong and is dangerously fear mongering. Whether he is doing this for personal gain or if in fact he actually believes all this doesn’t really matter.<p>I’m not going to go into it all but I’ll touch on one of the first mistakes me makes… that a super intelligence won’t follow orders. Well you better hope a super intelligence won’t just blindly follow orders because THAT would be dangerous. A super intelligence won’t just be able to do things better than us, it will be able to understand things better than us and in so doing it will also have a superior mortality than humans do. Notice how his immediate train of thought on AI is conflict — a battle between us and them. Why? At worst an AI will be indifferent. So the only conflict is one we create out of fear and foolishness.<p>Bombing AI data centers? Yeah, We will surely be the reason for our own destruction. Not AI.
评论 #35864496 未加载
评论 #35865910 未加载
评论 #35866481 未加载