AI의 발전: 에너지 효율성 및 에이전트 인텔리전스

엔비디아는 에이전트 AI를 통합하여 효율성과 에너지 관리를 개선함으로써 산업에 혁신을 일으키고 있습니다.

### 지속 가능한 미래를 위한 AI의 발전: 혁신과 영향력
* 🌍 이 특별 연설은 AI 분야의 총체적인 성과를 강조하며, **기술 민주주의를 촉진하는 동시에 미국의 AI 리더십을 강화하기 위한 기업 및 정부 기관과의 협업**을 강조합니다.
* ⚡ 성능과 에너지 효율을 모두 우선시하는 시스템 설계에 중점을 두기 때문에 **가속화된 컴퓨팅 인프라를 갖춘 데이터센터**가 필요합니다.
* 🖥 풀스택 접근 방식을 채택하여 칩뿐만 아니라 전체 시스템을 최적화함으로써 **획기적인 엔지니어링**을 선보이는 인상적인 **Blackwell 플랫폼**을 개발할 수 있었습니다.
* 🔋 지난 10년 동안 에너지 소비가 **100,000배나 감소**하는 놀라운 성과를 거두었으며, 다양한 AI 애플리케이션에서 에너지 효율성이 크게 향상되었습니다.
* 🤖 명령 주도형 컴퓨팅에서 **의도 주도형 컴퓨팅으로의 전환**이 진행 중이며, 컴퓨터가 자율적으로 지능과 기술을 생성하여 **제조업과 같은 산업에 혁명**을 일으킬 수 있습니다.
* 📊 새로운 AI 에이전트는 **문제를 사전에 파악**하고, **공급망을 최적화**하며, **다양한 부서 간에 원활하게 소통**하여 운영 효율성을 개선할 수 있습니다.
* 🔒 엔비디아는 언어 모델의 데이터를 보호하기 위해 **기밀 컴퓨팅을 개척**하여 민감한 정보를 보호하는 데 중요한 안전하고 신뢰할 수 있는 AI를 보장하고 있습니다.
* 🌌 빠른 전파 폭발에 대한 실시간 검색을 통해 우주 탐험을 강화하여 지구와 그 너머의 우주를 이해하는 데 있어 **AI의 잠재력**을 보여 줍니다.

you know what we're doing here in this special address is give is just giving you a glimpse of all the great work that you have done with us. You know, it's really about the work that we're that you were doing that we ideally helping you enable. This is really about um working with companies, working with government agencies to help the US lead in AI, to help democ AI. We can't stop trying to find ways to improve our planet and Humanity. Performance is going to continue to need to go up. I'm just here to tell you that we're designing our systems with not just performance in mind, but with Energy Efficiency in mind. It's going, it's going to mean that we're going to need data centers that understand this accelerated Computing infrastructure um. It's going to mean that we understand and approach things not just from a ship level but from a full stack level because we don't just try to optimize the chip. We're no longer just optimizing the GPU. We're optimizing the system. We look at a full

full-stack approach of how the system operates, the environment in which that system operates, and the software stack that runs in the system. And with that in mind, that's how we spent years developing our blackw platform. The amount of technology in here is just, it's a Marvel of engineering. All of these components, you know, we purposely build these racks of computing trays. We make sure that those racks of computing trays work together. And on top of that platform I mentioned, full-stack platform, you know, this is not the, this is the AI super cluster platform, the real platform, and the Cuda libraries, our software acceleration libraries. And then to further advance work in AI, in physical AI, we hon our two main platforms, Nvidia AI Enterprise with our microservices, our large language models, and Omniverse, our digital twin, you know, where we use the virtual world to improve the physical world. We've achieved 100%, 100,000x reduction in energy over the last 10.

years 100,000x, and that's on the impressing side. So, the generation of tokens is on the, uh, industry side. On the training side, I think we reduced energy consumption by 200x. But Blackwell is coming. It's in production, and the biggest and best systems are, uh, uh, will be shipping here in this next quarter. We're moving from instruction-le computing to intention-le computing, and that software has to be new, um. We have to be able to talk to our computers, to ask questions, um. This is going to be the second part of what's going to drive this Industrial Revolution, probably one of the most important parts, um. We've got years of skills here, and the amount of energy we have, more software people in Nvidia working here, um, than we do hardware people for this reason, um. Instead of the computer being a tool that you use to get stuff done, the computer is now going to be able to generate intelligence, generate skills. So, let's take an example of a manufacturing floor. You might have a very AI agent

that's constantly moding to your inventory. It notices that you got excess capacity, but your floor is not taking advantage of that excess capacity. They're working on a plan to produce X number of widgets today. That AI agent can communicate with the factory floor agent. That factory floor agent can help accelerate the continued building versus sending people home early or whatever. Continue to build, which then communicates with the purchasing AI agent to re-replenishing the materials at a rate. And you can do in minutes what used to take days. So, this era of agentic AI is really going to transform a lot of different industries. We've already seen examples of some horizontal use cases where digital avatars are used to improve the way we talk to customers, to improve the way we enable our employees to be more productive. We've seen some great examples in the past on content generation, whether it be texts, images, videos. Now, a lot of great work in product R&D, not just in healthcare, what we show there, but three that you see.

There are three things that I want to talk about in a minute, and we can use that AI agent since it is proactive, since it's not waiting for you to prompt it to see if you have an issue. We can use that AI agent to let you know you have an issue, that you have a vulnerability. We want to enable you all to incorporate your data because you know you're not going to optimize your supply chain on ChatGPT alone. You're probably going to take a smaller model, you're going to fine-tune it, you're going to fine-tune it maybe in your particular industry, and you're going to fine-tune it for your particular corporation because you have different semantics, different lingo, right? Training the AI makes sure that you expose it to the way that your company or your organization is able to interpret it. That reduces hallucinations, right? That makes it tailored, but you still don't stop there. You don't look back and train it. You've got to continually expose it to new data.

new intelligence that data is intelligence, and Company people's people are intelligence. So really happy to announce our nage and blueprints. Um, it's a way to connect your data fly will your data fly well to a bunch of different reference workflows. We're not an application company; we provide reference examples based on some of the most um often asked workflows that our customers that you bring to us, some of the biggest pain points, and some of the ones we care about the most: improving customer satisfaction, employee satisfaction, getting smarter, better, safer products to market, helping companies cure some really, really horrible diseases, and protecting our government IP, our personal IP, and having a strong AI ecosystem to lead the world. He looks real; he's not real, but James a blueprint that's available on build.nvidia.com is going to tell us a little bit more about all the blueprints that we have. Agses mod devent impr en. When you transmit data, it's encrypted, mostly safe, mostly safe, but llms are plain Tex. You're going to train an llm or

small language mod slm. It's plain text. Everything you said is in plain text. That knowledge is in plain text. Nvidia is the only provider to implement confidential computing to protect those LLMs, and that's the only way protecting your crown jewel to deliver secure and trustworthy AI. People could try to the encryption codes on where data is moving, but I think we all can imagine it's much easier to read a Word file than it is a, you know, an encrypted file. You know, primarily to date, we've really affected the IT industry. We've affected the electron part of this, and the next step is to affect the proton part of this, the physical world. We know that robots can help save lives, do jobs to make it safer for manufacturing employees to work. They can be assistant to people with disabilities. We know multiple organizations are talking about the need that we'll have A4 million, do four million jobs created in the next several years. We're not sure where that

Workforce could come from some of it might come from the jobs that are too dangerous with robots. Some of it, and I hope a lot of it, comes from our ability to upscale our current Workforce. What if we could train a robot on the manufacturing process in the virtual world, train them in the virtual world, train it at scale? Don't train one autonomous machine, one car, one robot, one at a time in the physical world. Train millions of them, train millions of them virtually, using that agentic AI knowledge that they need to have to reason and learn. Train them in the second computer to simulate how they operate, throwing out a multitude of conditions, bad and good. And when we have success, we take that neural net and we put it in a third computer. We put it in the car, we put it in the submersible, we put it in the Drone, we put it in the humanoid. So Omniverse is our way, um, really to test and optimize, uh, people's design and all the operational processes that go to

producing these physical AIS. It's a virtual treating G real-world Radio Networks 5G 6G. Um, we're now using Nvidia aerial ran um to improve both the transmission, the location of powers, and the utilization of how these networks um exist. We're announcing that se is releasing their first real-time search for fast radio burst. You're probably saying, What are fast radio burst? Se's mission is dedicated to searching for life in the universe. You know, they use our Hollan Edge platform, that third computer. They use that third computer to listen to signals from deep space. Fast radio bursts are one of those kinds of signals, but they happen so fast. They happen in fractions of a millisecond, fractions of a second. So, the only way they know they happen is they go back to this stored telescope array data. They see it happen, but it's long gone. It's long gone. We have no idea really where it came from. And so, the ability to spot these things in real time is really accelerating our ability.

to find the source, identify the source, analyze these frequencies. You can assume what these frequencies might be, but the amazing thing about them is not that there are radio frequencies coming from distant galaxies, but the energy coming from these distant galaxies is brighter than some of the galaxies in our universe. A fast radio burst of milliseconds is producing energy that's greater than the energy produced by some of our galaxies. So, really excited about that as a space guy. I'm excited about that. You might be saying, So what? But it's a massive application to help us understand not just about our planet, but what might be out there, what might be out there as we go beyond Mars.

Nvidia's AI Summit 2024: The Next Wave of AI in 14 Minutes

CNET is the premier destination for explaining and navigating the changing world around us. Through news, product reviews, price comparisons, and more, we’ll tell you what’s new in the parts of modern life that increasingly matter more — from health, home, and money to tech, culture, and climate — and how to get the most out of them.

CNET

기업과 정부 기관이 AI를 선도하고 지구와 인류를 개선할 수 있도록 지원하는 데 초점을 맞추고 있습니다.

이 접근 방식에는 시스템, 데이터센터, 소프트웨어의 전체 스택 최적화를 통해 성능과 에너지 효율성을 향상시키는 것이 포함됩니다.

AI 슈퍼클러스터 플랫폼 및 마이크로서비스의 개발과 디지털 트윈의 사용은 AI를 더욱 발전시키고 상당한 에너지 절감을 달성하는 것을 목표로 합니다.

AI 토큰 생성은 극적으로 증가했으며, 학습을 위한 에너지 소비는 200배나 감소했습니다.

명령어 주도형 컴퓨팅에서 의도 주도형 컴퓨팅으로의 전환은 AI 에이전트가 제조 현장의 운영을 변화시키면서 새로운 산업 혁명을 주도하고 있습니다.

이러한 에이전트 AI 시대는 다양한 산업 분야에 혁신을 가져와 고객과의 상호작용, 직원 생산성, 제품 연구 및 개발을 개선할 것입니다.

데이터 최적화를 위한 AI 에이전트 사용

AI 에이전트는 메시지를 기다릴 필요 없이 취약점과 문제를 선제적으로 식별할 수 있습니다.

회사별 데이터로 소규모 모델을 미세 조정하고 새로운 인텔리전스에 지속적으로 노출하면 환각을 줄이고 조직의 필요에 맞게 AI를 조정하는 데 도움이 됩니다.

nage와 블루프린트의 출시는 데이터를 참조 워크플로우에 연결하여 일반적인 문제점을 해결하고 고객과 직원의 만족도를 향상시키는 동시에 강력한 AI 생태계를 촉진하는 것을 목표로 합니다.

기밀 컴퓨팅 및 AI 인력 혁신에서 엔비디아의 역할

엔비디아는 대규모 언어 모델(LLM)을 보호하기 위해 기밀 컴퓨팅을 구현하여 안전하고 신뢰할 수 있는 AI를 보장하는 유일한 공급업체입니다.

이 회사는 현재 인력의 숙련도를 높이는 데 중점을 두고 로봇을 사용하여 생명을 구하고 일자리를 창출하는 등 실제 세계에 영향을 미치는 것을 목표로 하고 있습니다.

엔비디아의 옴니버스는 수백만 대의 로봇을 가상으로 훈련하고 시뮬레이션하여 실제 세계에 배치하기 전에 로봇의 작동을 최적화하도록 설계되었습니다.

빠른 라디오 버스트 실시간 검색

엔비디아는 5G 및 6G 네트워크를 강화하기 위해 공중 RAN을 사용하고 있습니다.

SE는 심우주에서 짧지만 강력한 신호인 고속 라디오 버스트에 대한 실시간 검색을 시작하고 있습니다.

이러한 폭발은 일부 은하보다 더 큰 에너지를 생성하며, 실시간 탐지는 그 소스를 이해하고 주파수를 분석하는 데 매우 중요합니다.