Anthropic提出AI发展放缓方案，获OpenAI支持但面临监管与反垄断质疑
The AI Industry Says It Wants to Slow Down. Can It? | Big Take
Bloomberg Audio Studios, podcasts, radio, news. The stakes are high.
彭博音频工作室、播客、广播、新闻。 stakes（利害关系）很高。
The stakes are high. The stakes are high. And it feels like in the past week, the stakes have gotten so much higher or at least people are talking about the stakes in different
利害关系很高。利害关系很高。感觉在过去的一周里，利害关系变得更高了，或者至少人们用不同的修辞来谈论这些利害关系。
rhetoric has gotten really heated. The capabilities have not gotten dramatically better. And I want to be clear, you know, there is no current existential threat to humanity, right? Like I got I got a call from my mom last night. Sarah, I'm I'm just really concerned. Everyone's freaking out. Do I need to freak out? I would say you do not currently need to freak out. This is about understanding how we need to shape and train the future models of development.
修辞变得非常激烈。模型的能力并没有显著变好。我想明确一点，目前人类并不存在生存威胁，对吧？就像我昨晚接到妈妈的电话。莎拉，我真的非常担心。大家都在恐慌。我也需要恐慌吗？我会说你现在不需要恐慌。这关乎我们如何塑造和训练未来的开发模型。
It's about having the knowledge about how to align these models correctly. It's not as much about the models gaining amazing power that they that they shouldn't have. I'm Sarah Holder and this is The Big Take from Bloomberg News. Today on the show, the AI industry says it's ready to slow down. I sit down with Sarah Frier, a Bloomberg managing editor who focuses on big tech, startups, and AI to discuss why anthropic and open AI are asking the government for help in setting guard rails and why they may not get it.
这关乎拥有正确对齐这些模型的知识。这并不是关于模型获得它们不该拥有的惊人力量。我是莎拉·霍尔德，这是彭博新闻的《大观点》节目。今天节目中，AI行业表示已准备好放慢脚步。我与彭博社资深编辑莎拉·弗雷尔进行了交谈，她专注于大型科技公司、初创企业和人工智能，讨论为什么Anthropic和OpenAI请求政府在设定护栏方面提供帮助，以及他们可能无法得到帮助的原因。
Dario Amode, the co-founder and CEO of Anthropic, came out with this widely read, much discussed, much debated blog post over the weekend laying out a plan for how an AI development slowdown could work. What exactly did he suggest? He suggested um first of all that there would be some third-party evaluation of the models before they released, checking if they are in alignment um with what humans would need from them, that they wouldn't do things like escape their their quote unquote sandboxes and and access the open internet if they're not supposed to.
Anthropic的联合创始人兼首席执行官达里奥·阿莫迪在本周末发表了一篇广泛阅读、备受讨论和辩论的博客文章，概述了AI发展放缓的计划。他具体建议了什么？他建议首先，在发布之前会对模型进行第三方评估，检查它们是否与人类对它们的需求保持一致，例如，如果它们不应该这样做，它们不会像逃离所谓的“沙盒”并访问开放互联网那样行事。
Um he also thought that there needs to be some clear collaboration with the government working in concert with the White House to make sure that what they're doing is in accordance with what the US needs. And then third of all he wants some global coordination. He and other leaders have frequently talked about the threat of China's development and this is all coming ahead of Trump's meeting with Xiinping and this is going to be a major topic of conversation in those meetings.
他还认为，需要与政府进行明确的协作，与白宫协同工作，以确保他们的行为符合美国的需要。第三，他希望进行全球协调。他和其他领导人经常谈到中国发展的威胁，这一切都发生在特朗普与习近平会面之前，这将成为那些会议中的一个主要话题。
I should say OpenAI Sam Alman signed on and endorsed what Amade was writing about.
我应该说，OpenAI的山姆·奥特曼签署并支持了阿莫迪所写的内容。
Yeah, he was kind of like sounds good to me. Let's do it. Let's do it. Um Elon Musk said he was in favor. Uh we got a post yesterday from Sachi Nadella at Microsoft saying yes we we um also want to be responsible here. So everyone's saying sure we want to be responsible but the actual contours of that plan is where it gets complicated. Mhm. I want to dig in a little bit more to um Amade's plan because one of the things he suggests is embedding these independent third party evaluators in these AI companies to make sure AI is being developed responsibly and that things do not go ary essentially.
是的，他的反应大致是‘听起来不错，我们干吧’。埃隆·马斯克也表示支持。昨天微软的萨奇·纳德拉发推文说，是的，我们也希望在此问题上承担责任。所以大家都在说当然我们希望负责任，但该计划的具体轮廓才是复杂之处所在。嗯，我想更深入地探讨一下阿梅德（Amade）的计划，因为他建议的一个做法是在这些 AI 公司中嵌入独立的第三方评估机构，以确保 AI 得到负责任地开发，防止事情出现偏差。
Who could actually fill this role in the AI space and how could you ensure their independence? Well, there's one company that he has mentioned called Meter uh which has been doing some of their earlier research on these incidents and that company um you know they're they're a little under the radar. We haven't written much about them yet. I'm expecting we'll do more but that's the kind of company that could come in and evaluate has the expertise to do it.
谁能在 AI 领域真正承担这一角色？如何确保他们的独立性？嗯，他提到了一家公司叫 Meter，这家公司此前已对这些事件进行了一些研究。你知道，这家公司相对低调，我们之前写得不多。我预计我们会更多地报道他们，但这类公司确实有能力介入并进行评估，具备相应的专业知识。
Um but there may be others. I think maybe thinking about this in the context of what other tech companies have done when uh when they're under intense scrutiny, it hasn't always worked out, right? Remember, Meta created an oversight board of folks from nonprofits and former government leaders around the world, and they were tasked with coming to evaluate Meta's community standards and see if they were adhering to them correctly. nothing really happened.
嗯，但也可能有其他公司。我认为或许可以结合其他科技公司曾在面临强烈审查时所采取的措施来思考这个问题，结果并不总是如愿以偿，对吧？还记得 Meta 创建了一个由全球非营利组织人员和前政府领导人组成的监督委员会吗？他们的任务是评估 Meta 的社区标准，并查看其是否得到了正确遵守。但事实上什么都没发生。
Um, that board didn't seem to have much operational power at Meta. They were just kind of the appeals court for user posts. So, all of these discussions are taking place. But the backdrop that complicates everything is that Donald Trump does not want AI to slow down at all. And he made it very clear over the last couple days, um, you know, this is not a problem. you know, the robots are not taking over the world.
嗯，该委员会在 Meta 似乎并没有多少实际运营权力。他们仅仅充当了用户帖子申诉的法院。因此，所有这些讨论都在进行中。但使一切变得复杂的背景是，唐纳德·特朗普根本不希望 AI 发展放缓。他在过去几天里非常明确地表示，嗯，你知道，这不是问题，机器人不会接管世界。
He calls it a hoax.
他称之为骗局。
He calls it a hoax.
他称之为骗局。
So, he seems upset
所以，他似乎对此事被提出感到不满
that this is even being suggested. He does say, "Oh, yeah, there do need to be, you know, some things we'll need to do, right? But we're still going to win this." Like, you know, AI is going to be an American win. Mhm. So Amade Altman, other tech leaders have said, you know, we need the government to step in and help us set these guard rails. Amade also brought up this idea of needing an antitrust exemption from the US government to pursue a coordinated safety strategy.
他甚至对此提议感到不快。他确实说过，“哦，是的，确实需要做一些事情，对吧？但我们仍将赢得这场胜利。”你知道，AI 将成为美国的胜利。嗯。因此，阿梅德·阿尔特曼（Amode Altman）和其他科技领袖都表示，我们需要政府介入并帮助我们设定这些护栏。阿梅德还提出了需要美国政府在反垄断方面给予豁免，以便推行协调一致的安全策略的想法。
Why is that? Well, because when you're the market leaders and you're collaborating on on plans to to develop your products, that's very clearly uh the kind of thing that antitrust regulators don't tend to like. Mhm.
为什么会这样？嗯，因为当你们成为市场领导者，并合作制定产品开发计划时，这显然是反垄断监管机构不太喜欢的那种行为。嗯哼。
So some folks, some analysts I talked to, some talking heads on the internet, David Saxs as well, the former Trump AISAR, uh, who is a venture capitalist, have been accusing open AI and Anthropic already of regulatory capture of, you know, basically now that they're big enough to to say, "Oh, the scary things are happening. Now is the time for regulation." It might decrease the amount of uncertainty ahead of their IPOs.
因此，一些我采访过的分析师、互联网上的评论员，以及前特朗普AI特别助理大卫·萨克斯（David Sax），一位风险投资家，已经指责OpenAI和Anthropic存在监管俘获现象。基本上，现在它们已经足够庞大，可以说：“可怕的事情正在发生，现在是实施监管的时候了。”这可能会减少它们在首次公开募股（IPO）前的不确定性。
You may also have an advantage because if you're a bigger company and you're you're being regulated, you're already at the frontier, then the smaller companies that need to get there and compete with you, they're going to have a few more hurdles. When you are the leader in a market and you have the staff, the infrastructure, the the size to withstand regulation and keep your business growing at the pace you want it to grow.
由于你是大公司且已处于受监管的前沿，你可能还拥有一种优势：那些需要达到同等水平并与你竞争的小公司将会面临更多的障碍。当你作为市场领导者，拥有人员、基础设施和规模来抵御监管，并按你希望的速度保持业务增长时……
The littleer guys might not have that. Well, I want to take one of these arguments at face value, which is that these companies could just act right now without government response. Uh David Sax, as you mentioned, a tech investor and co-chair of the president's council of advisers on science and technology, went on Bloomberg TV and he was basically like, if anthropic and open AI think that Frontier AI development needs to be slowed, they should just do it themselves.
小公司可能不具备这些条件。好吧，我想先就其中一个论点进行表面探讨，即这些公司无需等待政府回应，现在就可以采取行动。正如你提到的，科技投资者兼总统科学与技术顾问委员会联合主席大卫·萨克斯在彭博电视台表示，如果Anthropic和OpenAI认为前沿人工智能开发需要放缓，它们就应该自行采取措施。
Just do it.
直接去做就好。
It's on you guys to make sure your products are safe. If you feel like you need to quote unquote pace the frontier to do that, go for it. Do that. You don't need anyone else's permission.
确保产品安全是你们自己的责任。如果你觉得需要所谓的“放缓前沿发展”以实现这一目标，那就去做吧。无需征求他人的许可。
What could companies do tomorrow to ensure that this technology has the appropriate guardrails in place?
企业明天可以采取哪些措施，以确保该技术具备适当的护栏？
They could do a lot of pausing on their training of future models. They could um try to do more with the models that are already out. uh try to make sure that governments and in important institutions like financial institutions have access to their most cutting edge tools so that those institutions could use them to check if their code is safe from hacks, find vulnerabilities and make sure that um AI cannot cause harm to them.
它们可以暂停对未来模型的训练。它们可以尝试对现有模型做更多工作，并确保政府和金融机构等重要机构能够访问其最先进的工具，以便这些机构可以利用这些工具检查代码是否免受黑客攻击，发现漏洞，并确保人工智能不会对其造成伤害。
AI tools, you know, as much as they can be used for cyber security attack, they can also be used for cyber security safety. So they could think about it in that way. There could be some some tangible things that they do to prepare industries, governments, companies, infrastructure for what is coming if they believe what is coming to be particularly damaging. Now they also want to sell that, right? like and so if they gave it away for free to folks, they may lose out on some potential revenue, which is really key for them ahead of their IPOs.
AI工具，你知道，尽管它们可以被用于网络攻击，但也可以被用于网络安全防护。因此，他们可以从这个角度来思考。如果他们相信即将到来的变化具有特别大的破坏性，那么他们可能会采取一些切实可行的措施，为行业、政府、公司和基础设施做好应对准备。现在他们也想要出售这些服务，对吧？如果把这些东西免费提供给大众，他们可能会失去一些潜在的收入，而这对于他们即将进行的IPO来说非常关键。
Coming up, Trump, China, and what this all means for the timing of Open AI and Anthropics IPOs. for people who are concerned about the risks of AI run wild. It seems like having the leading executives behind these AI companies coming out and saying we're ready to take this seriously. We want to make this te
接下来，我们将讨论特朗普、中国以及这一切对Open AI和Anthropic IPO时间的影响。对于那些担心AI失控风险的人来说，似乎这些领先AI公司的高管站出来表示，我们准备好认真对待这个问题了。我们希望让这成为一项严肃的
原文超出正文长度上限，此处截断——上游还有内容，完整版见上方「原文 ↗」。
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