Full Keynote: Satya Nadella at Microsoft Inspire 2023 - YouTube
发布时间 2023-07-17 16:00:00 来源
How do we empower every person and every organization on the planet to achieve more? It starts with one. From one employee to a team collaborating with another team. From one partner to their customers to their customers customers. Your passion. Your ingenuity. Your hard work. Supporting inclusive opportunity. Protecting fundamental rights. Committee to a sustainable future. Earning trust. For our society. Our economy. Our world. That's the impact. That's the impact. That's the impact each of us can have. That's our collective opportunity.
如何使地球上的每个个体和组织都能取得更多的成就？这始于一个人。从一个员工到一个与另一个团队合作的团队。从一个合作伙伴到他们的客户，再到他们的客户的客户。你的热情，你的独创性，你的努力工作。支持包容性机会，保护基本权利。致力于可持续发展的未来。赢得信任。为了我们的社会，我们的经济，我们的世界。这就是影响力。这就是每个人都能产生的影响力。这是我们共同的机遇。
Empowering 8 billion people. We're entering a new era of AI. Where we will be able to build as fast as our imagination. Where creativity will have no constraints. And we can focus on the big picture. Together we can ensure these benefits reach. Every country. Every industry. And every individual. When we empower every person and every organization on the planet to achieve more. We empower the world.
为80亿人赋能。我们正在进入人工智能的新时代，在这个时代里，我们会有能力按照我们的想象力来快速建设。在这个时代里，创造力将不受任何限制。我们可以专注于大局。只有当我们赋予地球上每个人和每个组织实现更多的力量时，我们才能赋能这个世界。这样我们就能确保这些益处能够惠及每个国家、每个行业和每个个体。
Hello everyone it's fantastic. It's fantastic to be back here and inspire. And to me as a platform company. The most important thing we get to do is to come together as an ecosystem. With all of our partners. And reaffirm our commitment to our customers. To innovation on behalf of our customers. And to serve our customers across every community. Every country. Across every sector of the economy. So it's a very special time for us to be at a partner conference. To really celebrate not only the technology but most importantly. The excellence in our partner ecosystem.
大家好，真是太棒了。能够回到这里并激发灵感真是太棒了。作为一个平台公司，对我们来说最重要的事情就是与整个生态系统的合作伙伴们团结一致。再次确认我们对客户的承诺。为了客户的创新，为了服务于我们的客户，涵盖每一个社区、每一个国家和经济的各个领域。所以，这对于我们参加合作伙伴大会来说是非常特殊的时刻。我们不仅仅是在庆祝技术，更重要的是庆祝我们合作伙伴生态系统的卓越。
And talking about excellence I want to first start by congratulating. All of the partner of the year award winners that we have sitting here. So thank you on behalf of our customers for leading this ecosystem and the innovation around this ecosystem. Congratulations. We have a ton of exciting things to talk about throughout this keynote and beyond throughout this conference. And especially inspired to me at a time when there is massive changes happening around the platform. Makes it even more special.
谈到卓越，首先我想要恭喜所有在座的年度合作伙伴奖获奖者。所以代表我们的客户，感谢你们在引领生态系统和生态系统创新方面所做的努力。恭喜！在这次主题演讲和全程会议中，我们有很多令人兴奋的事情要讨论。特别是在这个平台发生大规模变革的时候，这让它变得更加特殊。
There's no question we're in the midst of a massive platform shift with the new generation of AI that's going to transform pretty much every sector. Every category of computing. Perhaps a good place to start is to sort of ground ourselves in what is that shift. What is the characteristics of the shift that we all as an ecosystem need to both understand and deeply innovate on top of. There are two dimensions to it.
毫无疑问，我们正处在一个巨大的平台转变中，新一代人工智能将彻底改变几乎所有行业，所有计算领域。也许最好的起点是先了解这种转变的本质是什么，我们作为一个生态系统需要深刻理解并在此基础上进行创新。这个转变具有两个方面的特征。
The first dimension is around the user interface. If you think about the arc of computing the last 70 plus years has been one constant pursuit to create human computer interfaces that are more intuitive, that are more natural. Right. Everything that we did I grew up through the graphical user interface. The mouse and the keyboard were very revolutionary when they came out changed how we sort of remediated computer human interaction. Obviously multi touch on the phone was another massive shift. But we now have arrived I would say at that next logical but big leap. Which is natural language. To be able to have natural language as the interface we have with computers is being the pursuit. And we get to exercise that across every application we build every interface we build.
第一个维度围绕用户界面展开。如果你考虑计算机科学的发展轨迹，过去70多年一直在不断追求创造更直观、更自然的人机界面。没错，我们成长的过程中一直在使用图形用户界面。当鼠标和键盘问世时，它们是非常革命性的，改变了我们对计算机人机交互的理解。显然，手机上的多点触控是另一个巨大的进步。但是现在我们可以说已经迈向了下一个逻辑性的大跃进，也就是自然语言。能够用自然语言作为我们与计算机交互的界面一直是我们追求的目标。而且我们可以在构建的每个应用程序、每个界面中实现这一目标。
The second piece is around information management. The entire history of computing has been about digitizing people places things and making sense of them. That's really essentially what we've done. Information management as Bill used to always say is the only category that actually exists. Which is you really digitize the world and try to reason and understand the world. In that context we now have a new superpower. We have a reasoning engine that allows us to start with a draft for any activity we start. You know we want to complete any task we want to complete. We have a reasoning engine that gives us predictive capability. We have a reasoning engine that will give us more insights. That's the two ships. When we talk about AI you can distill it into a more natural interface using natural language. A reasoning engine that works on top of all your data giving you more power. Those are the two things that we should keep in mind and ground ourselves.
第二个方面涉及信息管理。整个计算机发展历史都是关于将人、地方、物品数字化并加以理解。这实际上就是我们一直在做的。正如比尔经常说的那样，信息管理是唯一真实存在的类别。也就是说，我们真的将世界数字化并尝试推理和理解这个世界。在这个背景下，我们现在拥有了一种全新的超能力。我们拥有了一种推理引擎，它使我们能够从任何活动的起草开始。你知道，我们想要完成任何任务，我们都可以利用这个推理引擎。我们拥有一个能够提供预测能力的推理引擎。我们拥有一个能够提供更多洞察的推理引擎。这就是这两艘船。当我们谈论人工智能时，你可以用自然语言来简化它，而这个推理引擎则在你的所有数据之上工作，为你提供更大的能力。这是我们应该牢记和立足的两个要点。
And speaking of that every category right in our parlance we talk about solution areas. We've got to talk about every room of the house being remodeled by AI and that's what is going to happen. Whether it's work, whether it's business process and business applications, whether it's security, whether it's software development, both low code, no code and professional development. Every software category as we know of it is fundamentally going to be transformed by these two changes in the platform.
说到这一点，我们在技术语言中提到了解决方案领域的每个类别。我们必须讨论的是AI将重建房屋的每个房间，这将会发生。无论是工作、业务流程和业务应用，还是安全性、软件开发，包括低代码、无代码和专业开发。我们所知的每个软件类别都将根本性地被这两个平台变革所改变。
Now of course it's going to create a massive platform opportunity. And we've seen this before. If you think about it in my 30 plus years at Microsoft I've lived through these three transformations. And each one of them, whether it's a PC server, the web internet, mobile cloud, have been massive in terms of creating partner innovation, partner opportunity.
现在当然会创造一个巨大的平台机会。我们以前已经见过这种情况。如果你想一想，在我在微软的30多年里，我经历了这三个转变。每一个转变，无论是个人电脑服务器、网络互联网、移动云，都在创造合作伙伴的创新和机会方面是巨大的。
Reaching in fact one of the things that has given me the greatest pleasure is to see the composition of our partner ecosystem and how representative it is of the geographies of the world and also all the sectors of the economy. Now the question is what happens next? With this AI comes more opportunity. In fact, if you look at the data that's coming out, this is probably going to be the most profound shift as a percentage impact on GDP of technology. Something like 10% there will be additional GDP growth. So if you have an economy that's around 100 trillion, we may have 7 to 10 trillion more of GDP growth driven by this next generation of AI technology. And that translates as a percentage of GDP growth into a massive partner opportunity.
事实上，其中一件使我最愉悦的事情是看到我们合作伙伴生态系统的组成，以及它如何代表世界各地的地理位置和经济的各个领域。现在的问题是接下来会发生什么？AI的出现为我们带来了更多机会。实际上，如果你看看正在涌现的数据，这可能会成为对技术对GDP影响最深远的转变。大约有10%的额外GDP增长。所以，如果一个经济体大致为100万亿，我们可能会通过下一代AI技术带来7到10万亿的GDP增长。从GDP增长的百分比来看，这将转化为巨大的伙伴机遇。
Look, we are already as an ecosystem participating in a $4 trillion opportunity that could in fact grow by another 50%. So that means approximately $2 to $2.5 trillion of additional time is going to be addressable by all the innovation across our partner ecosystem. So there's no question the opportunity is tremendous and our partner ecosystem, the uniqueness of what you all represent really is what gets me most excited. It's the map of the world. We can reach every community in every country in every sector of the economy, both public and private. That's what's exciting.
看，我们作为一个生态系统，已经参与了一个价值4万亿美元的机会，实际上它还可能增长50%。这意味着，我们合作伙伴生态系统中的创新大约有2到2.5万亿美元的额外市场空间。因此，毫无疑问，机会是巨大的，而我们的合作伙伴生态系统，你们所代表的独特性，正是让我最激动的地方。它是地图，我们可以触达每个国家的每个社区，涵盖经济的每个领域，无论是公共还是私营。这就是令人兴奋的地方。
The rate of diffusion, in fact one of the most fascinating things in the last six months as I've been going to different parts of the world, talking to lots of you, what's unbelievable is the rate of diffusion of this technology. And thanks to sort of all your capability is what's most exciting about this particular revolution and platform shift. And that's what really grounds us in our mission to empower every person and every organization on the planet to achieve more.
在过去的六个月里，我去了世界上不同的地方，与许多人交谈，发现扩散速度实际上是最令人着迷的事情之一。让人难以置信的是，这项技术的扩散速度。多亏了你们所有人的能力，这次革命和平台转型才如此令人兴奋。这正是我们的使命所依托的东西，我们的使命是赋予全球每个个人和组织实现更多的能力。
In some sense this is a collective mission, right? We all work with us, partner with us, co-created, innovate with us in order to drive that equitable economic growth in every part of the world, to ensure that we protect those fundamental rights, to ensure that we protect those fundamental rights, to ensure that we protect those fundamental rights, to ensure that the innovation that we do is protecting our finite resource of the planet. We build trust across everything we do. These are the core pillars of our mission.
从某种意义上说，这是一项集体任务，对吧？为了推动全球各地的公平经济增长，确保保护那些基本权利，确保我们对地球上有限资源的创新能够得到保护，我们都与我们合作、合作创造、合作创新。我们在所做的每一件事中都建立信任。这些是我们使命的核心支柱。
And we've been hard at work. In fact, one of the things that people tell me is like Satya, you must have decided to do a bunch of press releases starting January because we've been really doing a lot of innovation. Believe me, we have been hard at work over the last multiple years and for sure the pace of innovation is picked up in this, the last six months. And we've sort of started in fact with get up co-pilot, we've followed up with everything that we're doing with Microsoft 365 co-pilot, the Dynamics 365 co-pilot, the infrastructure in Azure for both inference and training, our coding tools for low code, no code. And so there's just tons and tons of innovation.
我们一直在努力工作。实际上，有人告诉我，萨蒂亚，你肯定决定从一月开始发布一系列新闻稿，因为我们确实进行了很多创新。请相信，我们在过去的几年里一直在努力工作，而在最近六个月里，创新的速度确实有所加快。实际上，我们一开始就使用了合作驾驶功能，在这之后，我们还推出了所有与Microsoft 365合作驾驶、Dynamics 365合作驾驶以及用于推理和训练的Azure基础设施、低代码及无代码的编码工具相关的创新项目。因此，创新的内容非常多，多不胜数。
And it's just not the innovation in our products, but we also able to distill all of this into a first class platform construct. That's what we do as a platform company. We've built out this co-pilot stack and everything from the infrastructure layer all the way to the foundation, to how we manage data. In fact, something like Fabric, I think is going to be very, very transformative going forward in how we think about the AI age, how we think about the foundation models themselves, these reasoning engines that then reason over that data, the orchestration layer, which is going to be one of the most important middle tier pieces. And then at the application level, not only do we have what it takes to create the user interface, right? Co-pilot is even a design pattern. It's not about auto-pilot, remember, it's about putting humans in the loop and designing it in such a way that the human is always in control with a co-pilot that's powerful and helping them with every task.
不止我们的产品创新，我们还能将这一切提炼成一流的平台构建。这就是我们作为一个平台公司所做的。我们已经构建了这个共同驾驶（co-pilot）堆栈，从基础设施层一直到基础层，再到数据管理方式。事实上，像Fabric这样的东西，在我们思考人工智能时，以及思考基础模型本身时，将会带来非常深远的变革，它们是推理引擎，能够在数据上进行推理，并且编排层将成为最重要的中间层之一。而在应用层面，我们不仅具备创建用户界面的能力，是的，共同驾驶甚至是一种设计模式。这与自动驾驶无关，要记住，它是让人类参与其中并始终保持控制的一种方式，共同驾驶是强大的，并在每个任务中帮助人类。
So we've really brought everything from user experience design patterns to technology into a first class construct that then in your hands can be translated into unbelievable innovation everywhere. And that's what the co-pilot ecosystem represents, right, whether it's ISVs, SIs, startups, all coming together. And the combination of the co-pilot stack, the co-pilot ecosystem is what drives the AI advantage for us collectively, and really the AI advantage for our customers that we can deliver.
所以，我们真正将用户体验设计模式和技术融入了一流的构架，然后在你们手中可以转化为令人难以置信的创新。这就是联驾生态系统所代表的，无论是独立软件开发商、系统集成商还是初创企业，都汇聚在一起。联驾技术栈和联驾生态系统的结合，共同推动着我们的AI优势，也真正为我们的客户提供了AI优势。
So that's what's really exciting, this particular event, this particular conference, again, we're going to have 40 plus announcements, 40 plus innovations, and I wanted to highlight five of them. The first is I am really thrilled to be announcing Bing Chat Enterprise. Like, this is the feature, this is the product, this is the moment in some sense that every CIO, CISO, business user has been waiting for. Give me chat GPT that I can use by logging in through AAD with all the enterprise data privacy and security guarantees, and you have it now with Bing Chat Enterprise, you log into Bing with AAD, and you really have all the enterprise promises and in combination with the Microsoft 365 co-pilot between Bing Chat Enterprise and the Microsoft 365 co-pilot, knowledge work and knowledge workflow is going to be completely transformed. So to give you a little bit of a flavor for this, let me invite up on stage, Yousa Medi, to show you a quick demo of all of this innovation. Yousa.
这就是真正令人兴奋的原因，这个特别的事件，这个特别的会议，我们将会有40多项宣布，40多项创新，我想突出其中的五项。首先，我非常高兴地宣布必应企业聊天。就像这个功能，这个产品，从某种意义上说，对于每个首席信息官、首席信息安全官和商业用户来说，这是他们一直在等待的时刻。给我一个可以通过AAD登录并带有全面企业数据隐私和安全保证的聊天GPT，现在你拥有它了，通过AAD登录必应，你真正拥有了所有的企业承诺，并且与Microsoft 365 co-pilot相结合，必应企业聊天和Microsoft 365 co-pilot之间，知识工作和知识工作流程将会完全转变。为了让您对此有所了解，让我邀请Yousa Medi上台，给大家展示一下所有这些创新的快速演示。Yousa。
Good morning everybody, how are you? So great to be with all of our great partners today. In February, we introduced the new Bing, your AI powered co-pilot for the web, and the response has been nothing short of incredible. We've had over 1 billion chats to date, and since that launch, we've had literally over 100 new improvements and new features we've rolled out, and today I'm excited to introduce you to the next panel. I'm excited to introduce to you two of them. Multimodal visual search and Bing Chat Enterprise that Sachi just announced. First, let's talk about the new visual search. Leveraging the power of GPT-4, Bing Chat now accepts images as well as text. It'll actually analyze an image that you upload and give you insights related and answers related to that image. Let's take a look.
大家早上好，你们好吗？能够与我们伟大的合作伙伴们在一起真的太棒了。在2月份，我们推出了新的Bing，它是你的AI辅助网络导航员，反应非常惊人。到目前为止我们已经有了超过10亿次的对话，并且自推出以来，我们有了超过100个新的改进和新功能，并今天我很高兴向你们介绍下一个面板。我很激动地向你们介绍其中的两个功能：多模态视觉搜索和刚刚宣布的Bing聊天企业。首先，让我们谈谈新的视觉搜索。借助GPT-4的强大能力，Bing聊天现在不仅支持文本，还支持图片。它实际上会分析你上传的图片，并提供与该图片相关的见解和答案。我们来看一下。
I love architecture, and I want to learn a little bit more about the interesting details of a building that I've seen when I'm walking around. I'll take this image and upload it, and it'll tell me, hey, this is actually a cartouche, and it can actually detect on the image elements, like the presence of a map, which signifies the representation of the age of exploration. And there's a wreath of leaves that talk about the victory and glory age of Roman and Greek inspiration from architecture. It's incredible. Imagine being able to take pictures of anything out there and then have Bing come back and tell you. For example, we're an international adaptor work in a certain country. It's rolling out today, and we can't wait for you to try it.
我热爱建筑，当我四处走动时，我想多了解一些建筑中有趣的细节。我将使用这张照片并上传它，它会告诉我，嘿，这是一个镶嵌装饰，它实际上可以检测图像元素，比如地图的存在，这代表了探险时代的象征。还有一个叶环，它谈论着罗马和希腊建筑启发的胜利和荣耀时代。真是太不可思议了。想象一下，能够拍摄任何东西然后让必应告诉你。例如，我们是一个国际适配器，正在某个国家推出。它将在今天上线，我们迫不及待地想让你试用它。
Next, let's show you Bing Chat Enterprise, your AI-powered chat for work. I think this might be one of the biggest things that happens in generative AI this year. Since we've launched Bing, we've heard many corporate customers who are clamoring for these powerful AI tools, but they're worried about their company data that it won't be protected. In fact, some of these concerns from companies have actually led them to ban AI at work. Well, today, this morning, all that's going to change. Bing Chat Enterprise now delivers commercial data protection for your AI-powered chats. This means your confidential business data won't leak outside your organization. Your employee's data is not commingled with web data. Data's not saved. No one at Microsoft can view your data. And the chat conversations in Bing Chat are not used to train the underlying AI model. All right, you ready to take a look at how this works? Let's take a look.
接下来，让我们向您展示Bing Chat Enterprise，这是一款为工作场景提供的AI聊天工具。我认为这可能是今年生成AI领域发生的最重大的事情之一。自从我们推出了Bing以来，我们听到了很多企业客户迫切需要这些功能强大的AI工具，但他们担心企业数据无法得到保护。事实上，一些公司对此表示担忧，甚至禁止在工作中使用AI。但是，今天上午，这一切将发生改变。Bing Chat Enterprise现在为您的AI聊天提供商业数据保护。这意味着您的机密业务数据不会泄露给组织外部。员工数据与网络数据分开存储，数据不会被保存。微软的任何人都无法查看您的数据。Bing Chat中的聊天对话不用于训练底层AI模型。好了，准备好看看它是如何工作的了吗？让我们来看一下。
Here's Bing Chat Enterprise. And I can take full advantage of the power of AI to research and create. All I need to do is log in with my work credentials, and now you can visibly see that the experience has changed and that on every search and every chat your data is protected.
这是Bing企业聊天。我可以充分利用人工智能的能力进行研究和创造。我只需要使用我的工作凭证登录，您就可以明显地看到体验有所改变了，在每次搜索和聊天中，您的数据都得到了保护。
So let's stay on this architecture theme. I'm working on a confidential proposal to design a new building headquarters, let's say, in downtown Cleveland. And with Bing Chat Enterprise, I can prompt Bing now with my confidential bid information, my pricing, my guarantees, my unique certification capabilities. And Bing will automatically create a proposal, an introductory proposal paragraph. And I can do this now with the confidence that my confidential information data is protected. Look how great that is. I type in the confidential data and I get the paragraph.
所以让我们继续谈论这个建筑主题吧。我正在为设计一个新的总部大楼的保密提案而工作，就说在克利夫兰市中心吧。有了必应企业聊天，我现在可以用我的保密竞标信息、定价、保证和独特的认证能力来向必应提问。而必应会自动创建一个提案，包括一个引导性的段落。我可以放心地做这个动作，因为我的保密信息数据得到了保护。看起来多么好啊，我输入保密数据就能得到段落。
But let's go further. I want to have Bing Chat Enterprise now compare my confidential specifications to one of the buildings that are out there in Cleveland. And in no time, I get this beautifully formatted table that pulls information from the web about key tower in Cleveland with my internal confidential data. And you can see that Bing Chat Enterprise combines the power of GPT-4 with Bing's search data. You get the latest and most up-to-date information. We also provide you the links. So if you want to go learn more about that information with one click, you can get there. And finally, I could even have Bing Chat Enterprise create a SWOT analysis for my strategy planning by comparing confidential building data with public data. So I can simply ask it to create a SWOT analysis. And you'll see it comes back now with a whole bunch of great bullets. Again, all of this is analysis happening. The ability to get sort of a next level of productivity and creativity is just unmatched. You can now do all this with full confidence that your data is protected.
但我们来谈谈更进一步。我现在想用Bing Chat Enterprise将我的保密规格和克利夫兰的某座建筑进行对比。很快，我就能得到一个漂亮的格式化表格，该表格从网络上提取了有关克利夫兰主要塔楼的信息，并结合了我内部的保密数据。你可以看到，Bing Chat Enterprise结合了GPT-4的强大功能和Bing的搜索数据，使你能够获取最新和最实时的信息。我们还提供链接，所以如果你想通过点击一个链接了解更多信息，你可以轻松地实现。最后，我甚至可以让Bing Chat Enterprise通过将保密建筑数据与公开数据进行对比来创建一份SWOT分析报告，以供我进行战略规划。我只需简单地要求它创建一份SWOT分析报告，你会看到它将返回一系列精彩的要点。所有这些都是分析自动生成的。获得更高生产力和创造力的能力是无与伦比的。现在，你可以放心地进行所有这些操作，因为你的数据受到保护。
All right, I think you're going to love it. Yeah. And if you love what you can do with Bing Chat Enterprise, wait until you see what you can do with Microsoft 365 co-pilot. It has all of the capabilities of Bing Chat Enterprise plus so much more. It lets you reason over all of your business data in the Microsoft Graph. And it's built on Microsoft's comprehensive approach to security, so you know it's enterprise grade. Let's take a quick look here. I'm going to go in Teams now, and I can ask co-pilot to access the information from a PowerPoint presentation to create a starting point for my written proposal. Notice how nice it is just to use these powerful AI capabilities within the flow of my work. I'm just in Teams. I just ask, I pointed at the document and up comes this fantastic analysis. Microsoft 365 co-pilot is integrated into the world's most popular productivity app tools like Word, PowerPoint, Excel, Outlook, and Teams. And as we shared at our initial announcement, there are an incredible array of new capabilities like automated team summaries, which I use every day now, which offer me a whole new way to work.
好的，我认为你会喜欢它的。是的，如果你喜欢Bing Chat Enterprise的功能，等你用上Microsoft 365 co-pilot就会更惊喜了。它拥有Bing Chat Enterprise的所有功能，还有更多。它可以让你在Microsoft Graph中分析所有的业务数据。而且它是建立在Microsoft全面安全的基础上的，所以你知道它是企业级的。让我们快速看一下这里。我现在要进入Teams，然后可以让co-pilot访问一个PowerPoint演示文稿的信息，为我的书面提案创建一个起点。注意在我的工作流中使用这些强大的人工智能能力是多么方便。我只需要在Teams中提问，指向文档，就会呈现出这个精彩的分析。Microsoft 365 co-pilot已经集成到世界上最受欢迎的生产力应用工具中，如Word、PowerPoint、Excel、Outlook和Teams。正如我们在最初的发布中所述，它还有许多令人难以置信的新功能，比如自动团队摘要，我现在每天都在使用，它为我提供了全新的工作方式。
So to summarize, three things today. First, Bing Chat is your free AI powered co-pilot for the web, and I'm delighted to announce that as of this morning, you can go out and try the multi-modal visual search. Take a look and let me know what you think of it. Next, Bing Chat Enterprise. It's your AI powered Chat for work, first of its kind. It's available today at no additional cost for the over 160 million people who use Microsoft 365 business offerings, and soon we'll have a standalone option. And finally, Microsoft 365 unlocks a whole new way of working. Based on the positive feedback from our pilot testing, we're pleased to share our final pricing today. Microsoft 365 co-pilot will be available at $30 per user per month. So as our great partners, this is an incredible opportunity to accelerate your customers' move to Microsoft 365. Today, we unlock a new era of creativity and productivity as we bring the power of AI to work.
综上所述，今天的三个重点。首先，必应聊天是您在网络上的免费AI辅助工具，我很高兴地宣布，从今天早上开始，您可以尝试多模态视觉搜索。请看一下，并告诉我您对此的想法。接下来，必应聊天企业版。这是您的AI辅助工作聊天，前所未有。它今天可以免费为超过1.6亿使用Microsoft 365商业套件的人提供，很快我们也会推出独立选项。最后，Microsoft 365开启了全新的工作方式。根据我们在试点测试中获得的积极反馈，我们很高兴地在今天分享我们的最终定价。Microsoft 365 co-pilot每个用户每月30美元。因此作为我们伟大的合作伙伴，这是一个难得的机会，加快您的客户转向Microsoft 365。今天，我们开启了创造力和生产力的新时代，将AI的力量引入工作中。
Thanks very much. Thank you very much, Yousif.
非常感谢。非常感谢你，Yousif。
A multi-modal chat, work identity, and the Microsoft 365 co-pilot is fundamentally going to transform knowledge work as we know of it, and really exciting to see. And now talking about knowledge work, the next thing that I want to announce is the Microsoft sales co-pilot. This is a super exciting sort of next step in the evolution of the scope of co-pilots. If you think about it, we've talked about the horizontal knowledge work, we've talked about individuals, and individuals collaborating. But now we uplevel this to the level of the organization and business processes that really span the organization. And this allows anyone of us, who's in sales, to stay in the flow of our activity with our customers, right? So whether you're responding in email or you're in a team's meeting, you get to stay in the flow with the customer and ground that activity of customers.
一种多模式聊天、工作身份和Microsoft 365合作伙伴，将从根本上改变我们所认识的知识工作，并且非常令人兴奋。现在谈论知识工作，我要宣布的下一件事是Microsoft销售合作伙伴。这是合作伙伴范围演进的非常激动人心的下一步。如果你仔细思考一下，我们已经谈论了水平知识工作，谈论了个人和个人之间的合作。但现在我们将这个提升到组织和贯穿整个组织的业务流程的层面上。这使得我们销售人员中的任何一员都可以与客户保持活动的连续性，对吧？无论您是在回复电子邮件还是参加团队会议，您都可以与客户保持活动的连续性，并将这些客户活动联系在一起。
And the activity of customer interaction with CRM data, whether it's from Salesforce or whether it's from Dynamics 365, and to be able to complete your task without having to go out of band to your system of record. That's the dream that we've been waiting for for a long, long time. I think every SAS application category is going to be fundamentally rewired because of this capability. We are very, very excited to both announce and launch the Microsoft Sales Co-pilot.
客户与CRM数据的互动活动，无论是来自Salesforce还是Dynamics 365，都能够在不离开工作系统的情况下完成任务。这是我们期待已久的梦想。我认为每个SAS应用类别都将因为这种能力而进行根本性的重组。我们非常非常激动地宣布和推出Microsoft Sales Co-pilot。
Now, another third announcement I have is also related to business process. In fact, the most important tool that perhaps all of us in our partner ecosystem are going to use is to identify the opportunity for how we drive that AI advantage to every customer. It's all about process mining. So I'm very excited to announce the general availability of the process mining tools in Power Automate. This is the tool chain that everyone of us can use to be able to first map out the processes, sales processes, finance, human resources, everything that you can think of, which is a business process. And then optimize that business process, optimize that business process with the power of AI, things that can be automated, where you can deploy co-pilots to give the human in the loop the advantage to advance the business process. So process mining is perhaps the tip of the spear for you to be able to then drive additional value to every customer interaction, every customer deployment and every customer project that you are engaged in.
现在，另一个重要的公告也与业务流程相关。实际上，我们合作伙伴生态系统中可能所有人都将使用的最重要的工具是用于确定如何将人工智能优势运用到每个客户身上的机会。这一切都与流程挖掘有关。因此，我非常高兴地宣布Power Automate中流程挖掘工具的一般可用性。这是我们每个人都可以使用的工具链，用于首先绘制出销售流程、财务流程、人力资源流程等一切您能想到的业务流程。然后利用人工智能的力量优化这些业务流程，自动化那些可以自动化的部分，并使用协作机器人为循环中的人类提供优势，推动业务流程的进一步发展。因此，流程挖掘或许是您能够为每次与客户的互动、每次客户部署和每个客户项目带来附加值的关键。
Now, talking about the fourth big area for announcement is Azure AI updates. The first thing I want to start with is OpenAI. Simply put, Microsoft loves OpenAI. You know, every one of these platform shifts that we've been engaged in has also come with an iconic industry shaping partnership that we've been involved in. And in this age and in this era, it's the partnership between Microsoft and OpenAI. We love everything about OpenAI. What they stand for, think about it, right? It's a nonprofit entity that then has a per-profit capped entity, right? It's a profit capped entity that we are deeply partnered with innovating in. We build the infrastructure to train their large models, they're innovating on the algorithms and the training of these frontier models. We innovate on applications on top of it. You all get to innovate on top of it. It's a very unique way for us to do things that are going to really fundamentally shape the world going forward. And they care deeply about AI safety. They care about their mission. And all of that shows through in all the innovation that we've sort of been able to bring forth to date and what's going, you know, is going to come going forward.
现在，我们来谈谈第四个重要的领域公告，即Azure AI更新。我想首先谈谈的是OpenAI。简单来说，微软非常喜欢OpenAI。你知道，我们参与的每一个平台转型都伴随着一个具有标志性行业塑造合作伙伴关系。而在这个时代，微软与OpenAI之间的合作伙伴关系就是其中之一。我们对OpenAI的一切都非常热衷。想想看，他们的理念是什么？这是一个非盈利组织，但却有一个有限利润的子公司。我们与他们有着深度的创新合作。我们建立基础设施来训练他们的大型模型，他们则在算法和模型训练方面进行创新。我们在此基础上创新应用程序，而你们也可以在此基础上进行创新。这是我们以非常独特的方式塑造未来的重要途径。他们非常关注人工智能安全，关心他们的使命。所有这些都在我们迄今为止所带来的创新中得以展现，并将继续展现。
Now, today I'm also thrilled to take what we've been doing with OpenAI and expand it to the next phase. The regional expansion of OpenAI APIs, we've now got more regional expansion in North America, in Europe, and for the first time we're going to be able to deploy this in Asia. So we're very, very excited about all that's coming with OpenAI and what you all can do on top of OpenAI.
现在，今天我也非常激动地将我们与OpenAI合作所做的事情扩展到下一个阶段。OpenAI API的区域扩展，我们现在在北美、欧洲有更多的区域扩展，并且首次我们将能够在亚洲部署这一技术。所以我们对OpenAI即将带来的一切以及你们可以在OpenAI之上做的事情非常非常兴奋。
Now, speaking about all of the AI innovation, the other element is Microsoft loves open source. And this is something we are one of the largest contributors to open source, and when it comes to AI, it's no different. And so one of the things that we are very excited today is the announcement of Meta's, Lama 2, coming to both Azure and Windows. Meta has been doing phenomenal work in innovating in the open models, and Lama is sort of really captured the imagination of what open source and AI foundation models can do. And now with Lama 2, it takes it to the next level, and we are going to bring all of that to both Azure. And, but frankly, one of the most exciting things is to see these models even being optimized for Windows, so that application developers now can take advantage of both the frontier models of OpenAI, as well as the models that are available in open source to be able to transform every software category there is. And to show you all of this in action, I want to invite up on stage Jessica Hart to show you Azure AI in action.
现在，谈到所有的人工智能创新，微软热爱开源是另一个方面。这是我们在开源方面最大的贡献之一，而在人工智能方面也不例外。所以我们今天非常激动的一件事就是Meta的Lama 2将在Azure和Windows上发布的消息。Meta在创新开放模型方面做了非凡的工作，Lama已经真正激发了人们对开源和人工智能基础模型的想象力。现在，Lama 2将把它提升到一个新的水平，我们将把这一切都带到Azure上。但说实话，最令人兴奋的事情之一就是看到这些模型甚至为Windows进行了优化，这样应用开发人员现在可以利用OpenAI的前沿模型以及开源中可用的模型来改变每一个软件类别。为了向大家展示所有这些的实际效果，我想邀请Jessica Hart上台向大家展示Azure人工智能的实际应用情况。
Jessica? Azure AI is the cloud platform offering the best tools to ground, test, and deploy both open source and frontier models and always do so safely. As a former partner myself, and now as the leader of the data, AI, and digital applications business here at Microsoft, I know just how important it is to stay current with everything that Microsoft is bringing to market.
Jessica，Azure AI是最佳的云平台，提供了最好的工具，以确保我们能够安全地进行模型的搭建、测试和部署，无论是开源还是最前沿的模型。作为一位前合作伙伴，现在作为微软数据、人工智能和数字应用业务的领导者，我深知与微软市场推出的一切保持同步对于我们的重要性。
And so today, I'm thrilled to show you just how easy we're making this for you with the Azure AI studio, the place where a developer has everything they need to bring AI-powered experiences to their applications. So you can see I'm starting with a selection of models, including OpenAI's groundbreaking GPT-4, which is available in the Azure OpenAI service, along with all the other OpenAI frontier models. There are also great open source models and thousands of models from Hugging Face. And you'll notice it now also includes the llama2 models from Meta.
今天，我非常激动地向您展示Azure AI工作室，这是一个让开发人员轻松将人工智能体验引入他们的应用程序的地方。您可以看到，我从一系列模型开始，包括OpenAI的创新性GPT-4模型，在Azure OpenAI服务中可以使用，以及其他的OpenAI前沿模型。此外，还有来自Hugging Face的许多优秀开源模型和成千上万个模型。您还会注意到，现在它也包括Meta的llama2模型。
After choosing a model, I can see sample inputs and outputs, try models out, and fine tune and deploy models into their own private sandbox running in Azure. And I like to call this the AI version of the old file Save As move. And I want to pause here because this is where magic happens. This means organizations can quickly and easily benefit from the power of these foundational models, yet they have their own private instance of the model.
在选择模型之后，我可以看到示例的输入和输出，尝试模型，调整并部署模型到在Azure中运行的私有沙盒中。我喜欢把这个称为旧的"另存为"功能的人工智能版本。我想在这里暂停一下，因为这就是魔法发生的地方。这意味着组织可以快速、轻松地从这些基础模型的强大功能中获益，同时拥有自己的私有模型实例。
And why does this matter? Because application usage and fine tuning of the model makes the model more effective. And model effectiveness is the currency of AI. And most importantly, as I deploy the model, I can ensure it follows responsible AI best practices because Azure AI Studio has content safety built right in. I'll come back to that in a minute.
为什么这很重要呢？因为应用程序的使用和对模型的微调使得模型更加有效。而模型的有效性是人工智能的货币。而且最重要的是，当我部署模型时，我可以确保它遵循负责任的人工智能最佳实践，因为Azure AI Studio内建了内容安全功能。我一会儿会再回到这个问题上。
And so with just a few clicks, I can deploy a brand new llama2 endpoint and start integrating it into my application straight away, and it can be that easy. But I'd like to take it a step further, because we know, in this new world of generative AI, prompt engineering, which if you don't know, is the process of choosing the right words, phrases, symbols, and formats to guide the model, is critical to getting that great result from the large language models.
所以只需点击几下，我就可以部署一个全新的llama2端点，并立即开始将其整合到我的应用程序中，而且这也可以如此简单。但我想进一步，因为我们知道，在这个生成式AI的新世界中，提示工程至关重要，提示工程是选择正确的词语、短语、符号和格式来引导模型的过程，这对于从大型语言模型中获得优秀结果非常关键。
And so to do this, I'm using prompt flow, which supports commonly used frameworks like Lang Chang and semantic kernel, or I can just use native Python. And it works with multiple data sources, both structured and unstructured, so I stay in the flow of my development work, never needing to move to other tools.
为了完成这个任务，我使用了 prompt flow，它支持常用的框架如 Lang Chang 和语义核心，当然也可以直接使用本地的 Python。它同时适用于多种数据源，包括结构化和非结构化的数据，这样我可以保持开发的连续性，无需切换到其他工具。
And we know that safety is paramount, and it is the most important priority for our Azure AI platform. Using Azure AI content safety, again built right into the Azure AI Studio, I can directly add safety controls to the deployed model and choose from a range of high, medium, and low sensitivity labels and other controls for an even more finely tuned safety experience.
我们知道安全至关重要，并且对于我们的Azure AI平台来说，安全是最重要的优先事项。再次使用集成在Azure AI Studio中的Azure AI内容安全功能，我可以直接为已部署的模型添加安全控制，并选择一系列高、中、低灵敏度标签和其他控制，以实现更加精细调整的安全体验。
So now that I've got the safety controls where I want them to be, prompt flow has tooling for bulk testing to evaluate the results from the prompts. And you can see how the answers are being evaluated for how grounded the answer is to the original question on a scale of one to five. I can also use built-in evaluations for things like accuracy, relevancy, or coherence, or I can even just create my own parameters.
现在，我已经将安全控制设置在我希望的位置上，提示流已经具备批量测试工具，用于评估提示结果。你可以看到答案在原始问题上的可靠程度被评估为从一到五的等级。我还可以使用内置评估功能来评估准确性、相关性或连贯性，或者我甚至可以创建自定义参数。
And once testing is complete, deployment is easy, creating the final API, which includes all of the power that the foundational model offers with the additional value from the prompt engineering and safety work that I just added. I can even use the built-in testing chat window to see the model being queried via the new prompt flow. And once deployed, I can always come back here to monitor how the model is performing as the application is actually getting used.
一旦测试完成，部署就变得很容易了，可以创建最终的API。该API不仅包括基础模型提供的全部功能，还包括我刚刚添加的提示工程和安全性工作所提供的附加价值。我甚至可以使用内置的测试聊天窗口，通过新的提示流程来查询模型。一旦部署完成，我可以随时回到这里监控模型的性能，因为应用程序实际上正在被使用。
Now Satya also announced that we've optimized the Lama2 models to run on Windows. There's been amazing advances in Windows hardware in the last few years, with better GPUs and now in PUs, unlocking the power to run models locally. When you couple that computing power with the utility of the Windows terminal and the Windows subsystem for Linux, things like local support for container-based development and great tools like VS Code and advanced AI tools like the Onyx runtime, you have a fantastic place to do AI development locally.
现在萨蒂亚宣布我们已经将Lama2模型优化，使其能在Windows上运行。过去几年中，Windows硬件取得了令人惊人的进步，有更好的GPU和现在的PU，释放本地运行模型的能力。当您将这种计算能力与Windows终端和Windows子系统用于Linux的实用性结合在一起时，诸如基于容器的开发的本地支持以及像VS Code和高级AI工具（如Onyx runtime）等出色工具，您就拥有了一个非常出色的环境来进行AI开发。
So here you'll see that I have the 7 billion parameter metal on the two model running locally on my Windows PC via Onyx and direct to MEL being called by C-sharp when UI application. As you can see, the model easily generates a very detailed reply.
在这里，你会看到我在我的Windows PC上通过Onyx和直接调用MEL的C-sharp来运行的两个模型，其参数量为70亿。正如你所看到的，这个模型轻松生成了一个非常详细的回复。
At Microsoft, we are committed to offering the most comprehensive selection of open and frontier models like Lama2. We're bringing together the best services and tools available to empower developers and organizations to bring AI-powered applications to life easily, safely, and at the pace our customers in this era of AI demand.
在微软，我们致力于提供最全面的开放和前沿模型选择，如Lama2。我们正在集结现有的最佳服务和工具，以便赋予开发者和组织轻松、安全地将基于人工智能的应用程序落地，同时满足我们这个AI时代客户的要求。
Back to you, Satya. Thank you so much. Thank you so much, Jessica. That really, I think, perhaps best captures the AI advantage that everyone in the partner ecosystem has. The ability to use Azure AI with the frontier models and the best open source models to ultimately drive value to our customers. It's very, very exciting.
回到你的话题，Satya。非常感谢你。非常感谢你，Jessica。我认为这真的可能是最好的方式来强调伙伴生态系统中每个人所拥有的人工智能优势。能够利用Azure AI与前沿模型和最好的开源模型，最终为我们的客户提供价值。这非常非常令人兴奋。
So my last announcement is around the partner program. In fact, Nicole was going to come later and talk about the Microsoft AI cloud partner program in great detail, but it's really exciting to see us completely revamp the investments, the incentives so that you all can really take advantage of everything that this AI age has to offer and drive both business growth for yourselves and most importantly, value for our customers.
所以，我最后的公告是关于合作伙伴计划。实际上，Nicole本打算稍后详细介绍微软人工智能云合作伙伴计划，但让我们兴奋的是，我们彻底改革了投资和激励政策，让你们都能充分利用人工智能时代所提供的一切，并为我们的客户提供业务增长和最重要的价值。
And this is going to span everything, right? It's going to be about partners who are startups, ISVs, and system integrators of all sizes and across all industry segments. And so we're very excited about the partner program.
这个将囊括一切，对吧？它将涵盖创业公司、独立软件供应商以及各规模和行业领域的系统集成商等合作伙伴。因此，我们对这个合作伙伴计划非常兴奋。
But one of the things that I always go back to is how any given partner innovates with new technology, and I want to just share one great example with you. Epic. In the United States, as a percentage of GDP, health care is 18, 19% of our economy. And Epic is the leader when it comes to clinical management systems. And what they are doing is pretty phenomenal, right? They're not only bringing all of their software to the cloud, they're now taking advantage of some of the advances in AI with nuance and diarization of the patient physician interaction and incorporating it right into their application.
但是，我经常回头看的一件事是任何合作伙伴如何借助新技术进行创新。我想和你分享一个很好的例子，那就是Epic。在美国，卫生保健在国内生产总值中所占的比例为18%到19%。在临床管理系统方面，Epic是领军者。他们所做的事情相当了不起，对吧？他们不仅将所有软件引入云端，现在还利用人工智能的一些进展，如理解与患者和医生的交互内容以及口述内容，并将其直接纳入他们的应用中。
But perhaps most importantly, they're fundamentally changing all clinical workflow through their innovation on top of Azure OpenAI. And to me, to be able to sort of really see an ISV like Epic, who has been transformative in health care already, to take this next step forward with this platform shift and have an impact in perhaps one of the most critical sectors of our economy, where you can have better outcomes for patients, better outcomes for clinicians, and better outcomes for every provider out there is fantastic. That's what it sort of is all about at the end of the day.
然而，可能最重要的是，通过在Azure OpenAI上的创新，他们从根本上改变了临床工作流程。对我来说，能够真正看到像 Epic 这样在医疗领域已经引起了革命性变革的 ISV，向着这个平台转变迈出下一步，并且对我们经济中可能是最关键的部门产生影响，从而为患者、临床医生和每个提供者获得更好的结果，这是非常棒的。这才是最终的目标。