我们的使命是确保 AGI(通用人工智能)造福全人类。 Our mission is to ensure that AGI (Artificial General Intelligence) benefits all of humanity. ¶
开始指向 AGI* 的系统正在进入人们的视野,因此我们认为,理解我们所处的时刻很重要。AGI 是一个定义模糊的词,但一般来说,我们用它指代一种能够在许多领域以人类水平应对日益复杂问题的系统。 Systems that start to point to AGI* are coming into view, and so we think it’s important to understand the moment we are in. AGI is a weakly defined term, but generally speaking we mean it to be a system that can tackle increasingly complex problems, at human level, in many fields. ¶
人类是工具的建造者,天生就有理解和创造的驱动力,而这让我们所有人的世界变得越来越好。每一代人都站在前人发现的基础上,造出更强大的工具——电、晶体管、计算机、互联网,以及很快将至的 AGI。 People are tool-builders with an inherent drive to understand and create, which leads to the world getting better for all of us. Each new generation builds upon the discoveries of the generations before to create even more capable tools—electricity, the transistor, the computer, the internet, and soon AGI. ¶
随着时间推移,人类创新的步伐虽有起伏,却始终向前,给人们的生活几乎每一个方面都带来了此前难以想象的繁荣与改善。 Over time, in fits and starts, the steady march of human innovation has brought previously unimaginable levels of prosperity and improvements to almost every aspect of people’s lives. ¶
从某种意义上说,AGI 不过是我们共同搭建的这座越来越高的人类进步脚手架上的又一件工具。但从另一种意义上说,它又是某种全新事物的开端,让人很难不说一句「这次不一样」;摆在我们面前的经济增长看起来令人咋舌,而且我们现在已经可以想象这样一个世界:治愈所有疾病,有更多时间陪伴家人,充分释放我们的创造潜能。 In some sense, AGI is just another tool in this ever-taller scaffolding of human progress we are building together. In another sense, it is the beginning of something for which it’s hard not to say “this time it’s different”; the economic growth in front of us looks astonishing, and we can now imagine a world where we cure all diseases, have much more time to enjoy with our families, and can fully realize our creative potential. ¶
十年之后,也许地球上的每一个人所能成就的,都会超过今天最有影响力的人。 In a decade, perhaps everyone on earth will be capable of accomplishing more than the most impactful person can today. ¶
我们持续看到 AI 发展的飞速进步。以下是关于 AI 经济学的三个观察: We continue to see rapid progress with AI development. Here are three observations about the economics of AI: ¶
1. 一个 AI 模型的智能水平,大致等于用于训练和运行它的资源的对数。这些资源主要是训练算力、数据和推理算力。看起来,你投入任意多的钱,都能得到连续且可预测的提升;预言这一点的缩放定律(scaling laws)在横跨许多个数量级上都是准确的。 1. The intelligence of an AI model roughly equals the log of the resources used to train and run it. These resources are chiefly training compute, data, and inference compute. It appears that you can spend arbitrary amounts of money and get continuous and predictable gains; the scaling laws that predict this are accurate over many orders of magnitude. ¶
2. 使用某一给定水平 AI 的成本,大约每 12 个月下降约 10 倍,而更低的价格会带来大得多的使用量。你可以从 2023 年初的 GPT-4 到 2024 年中的 GPT-4o 的 token 成本变化中看到这一点:那段时间里,每 token 的价格下降了约 150 倍。摩尔定律以每 18 个月翻 2 倍的速度改变了世界;而这个速度之惊人,还要强得多。 2. The cost to use a given level of AI falls about 10x every 12 months, and lower prices lead to much more use. You can see this in the token cost from GPT-4 in early 2023 to GPT-4o in mid-2024, where the price per token dropped about 150x in that time period. Moore’s law changed the world at 2x every 18 months; this is unbelievably stronger. ¶
3. 线性增长的智能所产生的社会经济价值,本质上是超指数级的。由此而来的一个推论是:我们看不到指数级增长的投资在近期有任何停下来的理由。 3. The socioeconomic value of linearly increasing intelligence is super-exponential in nature. A consequence of this is that we see no reason for exponentially increasing investment to stop in the near future. ¶
如果这三个观察持续成立,对社会的影响将是巨大的。 If these three observations continue to hold true, the impacts on society will be significant. ¶
我们现在开始推出 AI agent,它们最终会让人觉得就像虚拟同事。 We are now starting to roll out AI agents, which will eventually feel like virtual co-workers. ¶
让我们想象一个软件工程 agent 的情形——我们预计这一类 agent 会尤其重要。想象一下,这个 agent 最终能完成一家顶尖公司里一位有几年经验的软件工程师能做的大部分事情,可以处理最长几天时间的任务。它不会有最了不起的新想法,它需要大量的人类监督和指导,它会在某些事情上非常出色,又在另一些事上差得出人意料。 Let’s imagine the case of a software engineering agent, which is an agent that we expect to be particularly important. Imagine that this agent will eventually be capable of doing most things a software engineer at a top company with a few years of experience could do, for tasks up to a couple of days long. It will not have the biggest new ideas, it will require lots of human supervision and direction, and it will be great at some things but surprisingly bad at others. ¶
尽管如此,还是把它想象成一个真实但相对资浅的虚拟同事。现在想象 1000 个这样的同事。或者 100 万个。再想象一下,这样的 agent 遍布知识工作的每一个领域。 Still, imagine it as a real-but-relatively-junior virtual coworker. Now imagine 1,000 of them. Or 1 million of them. Now imagine such agents in every field of knowledge work. ¶
从某些方面看,AI 在经济上最终可能会像晶体管一样——一项重大的科学发现,可以很好地扩展,并渗透进经济的几乎每一个角落。我们不会多想晶体管,也不会多想晶体管公司,而它带来的收益分布得非常广泛。但我们的确期待我们的电脑、电视、汽车、玩具等等能创造奇迹。 In some ways, AI may turn out to be like the transistor economically—a big scientific discovery that scales well and that seeps into almost every corner of the economy. We don’t think much about transistors, or transistor companies, and the gains are very widely distributed. But we do expect our computers, TVs, cars, toys, and more to perform miracles. ¶
世界不会一下子改变;它从来不会。短期内,生活大体上还是老样子,2025 年的人们基本上还是会用 2024 年的方式度过时间。我们依然会坠入爱河、组建家庭、在网上吵架、去野外徒步,等等。 The world will not change all at once; it never does. Life will go on mostly the same in the short run, and people in 2025 will mostly spend their time in the same way they did in 2024. We will still fall in love, create families, get in fights online, hike in nature, etc. ¶
但未来将以一种无法忽视的方式向我们袭来,对我们社会和经济的长期改变将是巨大的。我们会找到新的事情去做,找到新的方式彼此成就、彼此竞争,但它们可能和今天的工作看起来大不相同。 But the future will be coming at us in a way that is impossible to ignore, and the long-term changes to our society and economy will be huge. We will find new things to do, new ways to be useful to each other, and new ways to compete, but they may not look very much like the jobs of today. ¶
能动性(agency)、意志力和决断力可能会变得极其宝贵。正确地决定要做什么、弄清楚如何在一个不断变化的世界里航行,将有巨大的价值;韧性和适应力会是值得培养的有用技能。AGI 将是有史以来加在人类意志力之上的最大的杠杆,它会让单个个体拥有比以往更大的影响力,而不是更小。 Agency, willfulness, and determination will likely be extremely valuable. Correctly deciding what to do and figuring out how to navigate an ever-changing world will have huge value; resilience and adaptability will be helpful skills to cultivate. AGI will be the biggest lever ever on human willfulness, and enable individual people to have more impact than ever before, not less. ¶
我们预计 AGI 的影响会是不均衡的。虽然有些行业几乎不会改变,但科学进步很可能会比今天快得多;AGI 的这一影响可能会超过其他一切。 We expect the impact of AGI to be uneven. Although some industries will change very little, scientific progress will likely be much faster than it is today; this impact of AGI may surpass everything else. ¶
许多商品的价格最终会大幅下降(目前,智能的成本和能源的成本制约着许多东西),而奢侈品和少数本质上有限的资源(比如土地)的价格可能会涨得更加惊人。 The price of many goods will eventually fall dramatically (right now, the cost of intelligence and the cost of energy constrain a lot of things), and the price of luxury goods and a few inherently limited resources like land may rise even more dramatically. ¶
从技术上讲,我们面前的道路看起来相当清晰。但公共政策以及社会对于应如何把 AGI 融入社会的集体看法,关系重大;我们之所以选择尽早且频繁地发布产品,原因之一就是让社会与技术有时间共同演化。 Technically speaking, the road in front of us looks fairly clear. But public policy and collective opinion on how we should integrate AGI into society matter a lot; one of our reasons for launching products early and often is to give society and the technology time to co-evolve. ¶
AI 将渗入经济和社会的所有领域;我们会期待万物皆智能。我们中的许多人预计,我们需要比过去给人们更多对技术的控制权,包括更多地开源,同时承认安全与个人赋能之间存在一个需要权衡取舍的平衡。 AI will seep into all areas of the economy and society; we will expect everything to be smart. Many of us expect to need to give people more control over the technology than we have historically, including open-sourcing more, and accept that there is a balance between safety and individual empowerment that will require trade-offs. ¶
我们从不希望鲁莽行事,而且在 AGI 安全方面很可能会有一些不受欢迎的重大决定和限制;但从方向上讲,随着我们越来越接近实现 AGI,我们相信朝着更多个人赋能的方向走很重要——我们能看到的另一条可能的路径,是 AI 被威权政府用来通过大规模监控和剥夺自主权来控制其人民。 While we never want to be reckless and there will likely be some major decisions and limitations related to AGI safety that will be unpopular, directionally, as we get closer to achieving AGI, we believe that trending more towards individual empowerment is important; the other likely path we can see is AI being used by authoritarian governments to control their population through mass surveillance and loss of autonomy. ¶
确保 AGI 的收益被广泛分享至关重要。技术进步的历史影响表明,我们关心的大多数指标(健康状况、经济繁荣等等)平均而言、长期而言都会变好,但平等程度的提高似乎并不是由技术自动决定的,要做好这件事可能需要新的思路。 Ensuring that the benefits of AGI are broadly distributed is critical. The historical impact of technological progress suggests that most of the metrics we care about (health outcomes, economic prosperity, etc.) get better on average and over the long-term, but increasing equality does not seem technologically determined and getting this right may require new ideas. ¶
尤其是,资本与劳动之间的力量平衡似乎很容易被打破,而这可能需要及早干预。我们对一些听起来奇怪的想法持开放态度,比如给每个人一份「算力预算」,让地球上的每个人都能大量使用 AI;但我们也看到,有许多路径只要坚持不懈地把智能的成本压到尽可能低,就能达到同样的效果。 In particular, it does seem like the balance of power between capital and labor could easily get messed up, and this may require early intervention. We are open to strange-sounding ideas like giving some “compute budget” to enable everyone on Earth to use a lot of AI, but we can also see a lot of ways where just relentlessly driving the cost of intelligence as low as possible has the desired effect. ¶
到 2035 年,任何人都应该能够调动相当于 2025 年全人类总和的智力资源;每个人都应该能够调用无穷无尽的天才,按自己的想象去指挥它。当下有大量人才因为没有资源而无法充分施展自己,如果我们改变这一点,全世界由此产生的创造产出将给我们所有人带来巨大的收益。 Anyone in 2035 should be able to marshall the intellectual capacity equivalent to everyone in 2025; everyone should have access to unlimited genius to direct however they can imagine. There is a great deal of talent right now without the resources to fully express itself, and if we change that, the resulting creative output of the world will lead to tremendous benefits for us all. ¶
特别感谢 Josh Achiam、Boaz Barak 和 Aleksander Madry 审阅本文草稿。 Thanks especially to Josh Achiam, Boaz Barak and Aleksander Madry for reviewing drafts of this. ¶
*我们在这里使用 AGI 一词,目的是清晰地沟通,我们无意改变或解释那些界定我们与微软关系的定义与流程。我们完全预期将与微软长期合作下去。这条脚注看起来很蠢,但另一方面,我们知道有些记者会为了流量写一些很蠢的东西,所以我们就先在这里把蠢话堵回去了…… *By using the term AGI here, we aim to communicate clearly, and we do not intend to alter or interpret the definitions and processes that define our relationship with Microsoft. We fully expect to be partnered with Microsoft for the long term. This footnote seems silly, but on the other hand we know some journalists will try to get clicks by writing something silly so here we are pre-empting the silliness… ¶