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。

Elon Musk Says AI Is the ‘Greatest Risk We Face as a Civilization’

41 点作者 wei_jok将近 8 年前

14 条评论

Animats将近 8 年前
<i>Part of Musk’s worry stems from social destabilization and job loss. “When I say everything, the robots will do everything, bar nothing,” he said.</i><p>That&#x27;s still a ways off. Robot manipulation in unstructured environments is still terrible. See the DARPA Humanoid Challenge. People have been underestimating that problem for at least 40 years.<p>But that doesn&#x27;t help with the job situation. Only 14% of the US workforce is in manufacturing, mining, construction, and agriculture, the jobs where robot manipulation in unstructured environments matters. Those aren&#x27;t the jobs at risk.<p>I&#x27;ve been saying for a while that the near future is &quot;Machines should think, people should work&quot;. An Amazon warehouse is an expression of that concept. So are some fast-food restaurants. So is Uber. The computers handle the planning and organization of work; the humans are just hands for the computers. (Yes, &quot;Manna&quot;, by Marshall Brain.) That&#x27;s going to become more common. Computers are just better at organization and communication than humans.<p>Computers have already made a big dent in middle-class jobs, and that&#x27;s going to continue. If everything you do goes in and out over a wire, you&#x27;re very vulnerable to automation. If 20% of what you do can&#x27;t be done by a computer, that means five of you will be replaced by one person. This is already hitting low-level lawyers; it hit paralegals years ago.<p>The end state of this trend is a modest number of well-paid people in control, a huge number of people taking orders from computers, and many people without jobs. That&#x27;s not far away; one or two decades. It&#x27;s mostly deploying technology that already exists.
评论 #14780991 未加载
esaym将近 8 年前
&gt;Musk outlined a hypothetical situation, for instance, in which an AI could pump up defense industry investments by using hacking and disinformation to trigger a war.<p>What the heck is he talking about? With my limited exposure to AI and neural networks, there really is no algorithm that can make algorithms. And therefore, AI doesn&#x27;t really &quot;think&quot;. Sure you can train a neural net to pick out the &quot;diamonds&quot; in a sea of garbage, but that is still not &quot;thinking&quot;, merely going on an educated guess backed by statistics. Or am I missing something?
评论 #14780961 未加载
评论 #14780953 未加载
评论 #14780936 未加载
评论 #14780935 未加载
seertaak将近 8 年前
Although I don&#x27;t doubt that there&#x27;s an element of sincerity in Musk&#x27;s many pronouncements, all this talk of AI is also great way of signalling that he&#x27;s at the technological forefront.
评论 #14780857 未加载
elorant将近 8 年前
Does he know something that we don&#x27;t? From what I understand his companies have done little to no research on the subject. He might be much better educated than the average geek but that doesn&#x27;t mean much considering that the whole field is highly experimental. No one can tell with any kind of certainty how an AGI would behave.<p>What am I missing here?
评论 #14781064 未加载
评论 #14781092 未加载
评论 #14781013 未加载
评论 #14781011 未加载
评论 #14781023 未加载
borplk将近 8 年前
If anyone other than Musk was saying the EXACT same thing no one would care and they would have been laughed at.<p>But Elon says something and everyone loses their minds.<p>Let thoughts stand for themselves. Why attach so much weight to the speaker?
评论 #14781071 未加载
评论 #14781053 未加载
评论 #14781015 未加载
guelo将近 8 年前
I&#x27;m more apocalyptic about climate change. Ironically one of my few hopes is that some kind of AI can save us.
tanilama将近 8 年前
But his company is among one of those companies that are ruthlessly pursuing AI technology for its own commercial purpose, like self driving car. Did he just contradict himself a little in that sense?
评论 #14780977 未加载
Nuzzerino将近 8 年前
For those interested, here is a Quora Q&amp;A which has a lot of worthy debate on the AI Doomsaying research that Musk apparently bases his views on. <a href="https:&#x2F;&#x2F;www.quora.com&#x2F;How-do-we-know-that-friendly-AI-research-is-actually-right-meaningful" rel="nofollow">https:&#x2F;&#x2F;www.quora.com&#x2F;How-do-we-know-that-friendly-AI-resear...</a>
atroyn将近 8 年前
It&#x27;s unclear that it&#x27;s even possible to emulate general intelligence by computable functions, let alone that it&#x27;s possible to improve it to superhuman capacity.<p>There are clear and present threats to civilization needing to be dealt with - superhuman A.I is, to quote Maciej&#x2F;Pinboard, the &#x27;Idea that eats smart people&#x27;.
评论 #14784602 未加载
评论 #14781014 未加载
评论 #14784641 未加载
评论 #14781063 未加载
评论 #14780930 未加载
partycoder将近 8 年前
I disagree.<p>AI will not be the same as animal intelligence. The driving force behind animal intelligence has been survival. Animal intelligence evolved gradually resulting in a hybrid brain containing primitive structures with primal instincts and irrational behavior as well as more evolved structures capable of strong problem solving. Therefore our intelligence is tainted with primitive behavior.<p>Strong AI can eventually set intelligence free from our primitive, irrational roots and that is in itself not bad.
评论 #14780891 未加载
评论 #14780885 未加载
评论 #14784540 未加载
评论 #14780872 未加载
评论 #14781001 未加载
评论 #14780912 未加载
tlrobinson将近 8 年前
Prequel to &quot;Daemon&quot; (the novel by Daniel Suarez): before his death, Matthew Sobol warns the world of the threat AI poses after accidentally creating The Daemon and losing control of it. The Daemon has gone into hibernation until the one person possibly able to stop it is dead.<p>Also, Sobol previously started digital payments and self-driving car companies, which are repurposed by The Daemon for payments on the Darkent and AutoM8s...
fxj将近 8 年前
One basic difference between humans and robots is sustainability and resilience even when something goes wrong big time. In the evolution of mankind the number of humans was reduced to several ten thousands and still we did survive as a species because the biology of reproduction makes us very resilient. Robots however need a vast infrastructure in to be produced and maintained which makes failure much more probable.
jaimex2将近 8 年前
I&#x27;m seeing a lot of comments with &quot;this is crazy, AI wont reach that level&quot;, I&#x27;m not so sure after some of the stuff thats come out this year.<p><a href="https:&#x2F;&#x2F;www.theverge.com&#x2F;2017&#x2F;2&#x2F;9&#x2F;14558418&#x2F;ai-deepmind-social-dilemma-study" rel="nofollow">https:&#x2F;&#x2F;www.theverge.com&#x2F;2017&#x2F;2&#x2F;9&#x2F;14558418&#x2F;ai-deepmind-socia...</a><p><a href="http:&#x2F;&#x2F;www.highsnobiety.com&#x2F;2017&#x2F;07&#x2F;13&#x2F;google-deepmind-ai-walk&#x2F;" rel="nofollow">http:&#x2F;&#x2F;www.highsnobiety.com&#x2F;2017&#x2F;07&#x2F;13&#x2F;google-deepmind-ai-wa...</a>
solotronics将近 8 年前
What I really want to know is if it&#x27;s possible for an AI to emerge organically on the net and if so how would you even detect it? Could a distributed intelligence be influencing things already without people knowing? It&#x27;s a fun thought experiment I play with myself while I build datacenters all over the world stuffed with cloud computing hardware. Deus ex machina?
评论 #14780922 未加载