我 采 访 了 苏 妈 ！| 专访AMD CEO Lisa Su【大咖谈芯09】

Summary

在这次访谈中，Lisa分享了她对人工智能（AI）的积极看法，强调AI是过去50年来最具变革性的技术。她认为AI已经能够真正地在今天为我们的生活带来便利，通过云AI或新的AI专用PCs提高生产力和能力。她还提到AMD致力于开发跨应用领域的AI技术，包括边缘计算和客户端应用，并强调了构建强大生态系统的重要性。此外，Lisa讨论了AI在芯片设计和软件开发中的作用，以及AMD如何利用AI技术来提高芯片设计的速度、性能和质量。

Takeaways

• 🌟 AI是近50年来最具变革性的技术，对生活和工作方式产生了重大影响。
• 🚀 当前AI技术已经能够在日常生活中带来实际益处，提高生产力和效率。
• 💡 未来AI的发展将更加快速，AI个人电脑（AI PC）将变得越来越普及。
• 🔧 AI技术的应用将不仅限于云计算，还将扩展到边缘计算和客户端应用程序。
• 🛠️ AI技术的发展需要强大的生态系统支持，包括硬件和软件的协同发展。
• 🔄 AI的集成方式多样，包括CPU、GPU和专门的AI处理器（NPU）。
• 🔍 AI在芯片设计和验证领域的应用前景广阔，可以提高设计速度和质量。
• 💻 软件在AI时代的重要性日益增加，易用性是推动AI发展的关键因素。
• 🌐 开放的生态系统对于AI技术的发展至关重要，鼓励软件开发者参与共建。
• 🤖 AI作为个人助理，能够节省时间并帮助完成日常任务，提升个人生产力。
• 🔑 硬件公司如AMD致力于使硬件易于使用，以便软件开发者能够快速开发应用。

Q & A

AI如何真正地改变我们的日常生活？

AI能够通过云服务或新的AI专用硬件，提升我们的生产力和能力，使日常任务更加高效。

传统PC是否仍能满足我们的日常工作需求？

虽然传统PC仍然可以满足日常工作，但AI的引入可以使我们的生活更加便捷和高效。

AI在个人生产力方面的作用是什么？

AI可以作为个人助理，帮助我们完成日常任务，节省时间，提高生产力。

AMD在AI技术发展中的角色是什么？

AMD致力于开发适用于所有应用领域的AI技术，包括云服务、企业和边缘应用。

AI在芯片设计中的作用是什么？

AI可以帮助设计更快速、更强大的芯片，提高芯片设计的质量。

AI芯片通常包含哪些组件？

AI芯片可能包含CPU、GPU和特定的AI处理器（NPU）。

如何使AI技术更易于使用？

通过建立强大的生态系统和软件开发环境，使硬件易于使用，让软件开发者能够快速开发应用程序。

AMD如何支持软件开发者？

AMD通过提供易于使用的软件能力和创建开放的生态系统，鼓励软件开发者与他们合作，共同推动技术发展。

AMD的软件战略是什么？

AMD的软件战略是使软件尽可能易于使用，并扩展其软件能力，以支持整个产品组合。

AI如何帮助AMD提高产品质量？

AI可以在硬件和软件开发方面提供帮助，使产品更快、更强大、质量更高。

AMD对于AI未来的展望是什么？

AMD预计AI将继续快速发展，AI专用硬件将变得更加普及，每个人都会希望拥有AI专用PC。

Outlines

00:00

🤖 人工智能在日常生活中的实用性

本段落讨论了人工智能（AI）对日常生活的影响和实用性。讲者认为AI即使在今天也能真正地为我们带来好处，并且通过使用AI，我们可以构建更好的体验。讲者鼓励团队成为AI的首批用户，并强调构建开放生态系统的重要性。讲者还提到，AI技术的发展将会非常快速，并预计在2024年及以后，AI个人电脑（AIPC）将变得越来越流行。此外，讲者强调了云计算和AI个人电脑在提高生产力和能力方面的作用。

05:00

🔧 AI在个人生产力和技术发展中的作用

在这一段中，讲者探讨了AI在个人生产力提升和技术发展中的作用。讲者比喻AI为个人助理，能够帮助节省时间并完成日常任务。从AMD的角度来看，他们的目标是开发适用于所有应用领域的AI技术，包括云计算、企业和边缘计算。讲者强调了生态系统的重要性，并提到在北京举行的活动重点关注AI个人电脑。此外，讲者讨论了将AI引擎集成到CPU中以及其他技术如何帮助提升整体计算机系统的AI能力和计算效率。

Keywords

💡人工智能

人工智能（AI）是指由人造系统所表现出来的智能行为，能够通过学习和经验来优化执行特定任务的能力。在视频中，人工智能被视为一种变革性技术，可以极大提高生产力和个人能力。例如，通过云AI或新的AI专用PC，人工智能可以帮助我们更高效地完成日常工作。

💡生产力

生产力是指在生产过程中创造产品或服务的效率和效果。在视频中，通过将人工智能技术融入到我们的日常工作和生活中，可以显著提高生产力，让我们能够更高效地完成任务。

💡云计算

云计算是一种通过互联网提供计算资源（如服务器、存储、数据库、网络、软件等）的服务模式。在视频中，云计算与人工智能结合，为用户提供强大的计算能力和存储能力，使得AI服务更加普及和便捷。

💡生态系统

生态系统在这里指的是一个由硬件、软件、服务和用户等组成的互相依赖和互动的复杂网络，它们共同促进技术的发展和应用。在视频中，强调了建立开放的生态系统的重要性，以便于软件开发者能够快速开发出应用程序，推动新技术的快速发展。

💡边缘计算

边缘计算是一种分布式计算架构，其核心思想是将数据处理任务从中心数据中心转移到网络边缘的设备上，从而减少延迟，提高响应速度和数据处理效率。在视频中，边缘计算被视为AI技术发展的一个重要方向，特别适用于需要快速响应的应用场景。

💡AI处理器

AI处理器是专门为人工智能应用设计的处理器，它们通常具有优化的架构和指令集，能够更高效地执行AI相关的计算任务。在视频中，AI处理器（如NPU）与CPU和GPU一起被提及，作为构成AI个人电脑的关键技术组件。

💡芯片设计

芯片设计是指对集成电路或半导体芯片的功能、结构和制造工艺进行规划和优化的过程。在视频中，芯片设计是讨论的重点之一，强调了AI在提高芯片设计速度、性能和质量方面的潜力。

💡软件开发

软件开发是指编写、测试和维护计算机程序的过程，以满足特定的业务需求或创建新的应用程序。在视频中，软件开发的易用性被强调，说明了为了使硬件易于使用，需要大量工程师努力使软件平台如Rockham更加用户友好。

💡开放生态系统

开放生态系统是指一个允许多方参与、共享资源和促进创新的平台或环境。在视频中，开放生态系统的理念被提倡，目的是为了让软件开发者能够与公司合作，共同构建强大的技术平台，快速推出新的应用程序。

💡AI时代

AI时代指的是人工智能技术得到广泛应用并对社会各个领域产生深远影响的时期。在视频中，AI时代的到来意味着软件的重要性日益增加，高性能和高效的硬件需要与易用的软件相结合，以支持AI计算的发展。

Highlights

AI是过去50年最具变革性的技术

AI今天已经真正能够为我们带来好处

通过AI，我们可以构建更好的旅程

我们应该首先使用AI，以开放的生态系统哲学

AI可以提高我们的生产力和能力

AI和PCs的结合是未来的趋势

AMD致力于开发跨所有应用空间的AI技术

边缘计算和客户端应用程序对AI至关重要

AMD专注于创建强大的生态系统以推动新技术发展

AI引擎集成到CPUs是提升AI能力的一种方式

AI可以帮助芯片设计，提高设计速度和质量

AI在硬件和软件开发方面都能发挥作用

软件在AI时代的重要性不容忽视

AMD的软件能力称为Rockham，旨在易于使用

开放生态系统对于软件发展至关重要

AMD致力于使硬件尽可能易于使用

Transcripts

I actually believe that AI is can truly

benefit us uh even

today and um we will build better trips

uh with the use of AI I tell my team we

must be the first users of

AI our philosophy is to have a very open

ecosystem actually uh this is something

I'm very passionate

[Music]

about

here

[Music]

[Music]

hell

hi Lisa uh it's a great pleasure to talk

to you and share with you because I'm

big fan some my my teammates a big fan

of you very inspirational um and uh bit

of introduction for myself I'm Ki um and

I'm I've been working on my social media

channel Sheil can talk for the past

three or four years as you can see my

channel is all about semiconductor chip

umch technology as well as AI so today

I've got a couple of questions uh about

the future of AI and your perspective of

the future development AI particularly

so let's jump into it yeah that's great

thank you for having me okay perfect

thank you and my first question is you

have stated a couple of times and in

this Summit as well AI is the most

transformative technology uh for the

past 50 years right well AI is

definitely helpful and it does make many

changes but a lot of people thinking

like um so many of our daily Works can

still be properly managed by

conventional PC if you like so in your

opinion how long do you think AI will be

truly or genuinely practical in our

daily lives well um I actually believe

that AI is can truly benefit us uh even

today your comment that uh we can of

course we can live with the traditional

PC of that's definitely true uh but I

think the power of uh bringing AI into

our lives whether you're using Cloud um

you're using AI in the cloud or you're

using AI uh with these new AI PCS is uh

is really to make all of us uh more

productive and more capable so um I do

think it's just the beginning for AIP

PCS but it will go very fast and you

know I'm very excited about some of the

the products that are in the market

today um and in uh uh later in 2024 and

I expect over the next few years aips

will become more and more popular and

everybody will want to have an aipc okay

thank you and a followup question on

that so in terms of the Practical

utility of AI um what types of

Technologies do you think will you know

uh enable or support this uh this

transformation and what what role will

MD play in this uh transformation well

the way I like to describe uh the role

that AI can uh bring for personal

productivity is you know we're all so

busy every day um if you have a um a

personal assistant which is your aipc

and it can save you um a few hours a

week or five hours a week or even more

than that uh because it helps you um

accomplish uh many of your daily tasks

then that's a very positive thing and

you from AMD standpoint our goal is to

develop AI Technologies for um across

all of the application space so uh we

developed uh you know the largest uh

foundational infrastructure for uh the

cloud and Enterprise uh but we also are

very very excited about uh Edge

applications and client applications so

you know today here in Beijing we were

really focusing on

aips and what is so important about this

new area is that you need to bring all

of the partners together and that's

really what we have um uh been trying to

do do is to create a very very strong

ecosystem uh to enable the new

technology to go as fast as possible I

see so um in terms of aipc um I see like

there's way to integrate AI engines into

the CPUs but in addition to that um is

there any other ways or Technologies can

help to improve the computer as a whole

system like in terms of their AI

capabilities or the computing power or

efficiency of AI computation yeah you

see um today

uh you know aips may have uh CPU GPU and

then uh what we call npu or specific AI

processor I think all of these

Technologies are important and depending

on what your uh use case is whether

you're uh you know very focused on uh

personal productivity or notebook form

factor uh you will use one set of

Technologies or if you're very focused

in the desktop area or in the

workstation area uh you'll have a

different combination of these

Technologies but in the end um the key

is to make the technology very very easy

to use and that's why the ecosystem is

so important and having uh you know the

software development ecosystem our goal

is to make the uh Hardware as easy to

use as possible so the software

developers can come up with their

applications as fast as possible uh so I

was very excited today to see uh so many

of the AI startups and um uh local isvs

uh really making uh so much progress

with rise in AI thank you so much so um

as of myself I'm currently working on

scientific research of Chip design

verification and recently I see many of

the researchers and the research Works

being focused on AI AED chip development

um so in your perspective what's your

take on uh AI role or if AI can help

with chip development and also I'm very

curious about if

has uh used AI to help with the chip

development at all oh definitely

actually uh this is something I'm very

passionate about um I think the uh so

you know as a technology company and

since we're building so much AI I tell

my team we must be the first users of AI

in um um in our uh in our world so

definitely as you said um AI can help

chip design uh our goal is to design

chips you know uh faster more powerful

and um with a higher quality and AI can

help us do that in many ways on both the

hardware standpoint um as well as um you

know from the software development so I

think you're absolutely right it's a

very uh it's a great area for uh you

know research and uh development and um

you we will build better chips uh with

the use of AI okay thank you so much um

and my last question is about software

and I think if you talk about software

uh already and AMD is uh provides very

high performance and efficient hardware

for AI Computing um but we both know

that software cannot be neglected so in

your perspective uh what's the

importance of software especially in the

ai ai era yeah it's uh extremely

important um you're right you know it's

not easy to program uh these um you know

AI devices and so you know our job as a

um AI company is to make sure that our

software is very easy to use and from

that standpoint um our philosophy is to

have a very open ecosystem so um our

software uh uh capabilities are are

called Rockham and um our our Rockham um

you know software capabilities have uh

really um uh improved a lot over the

last uh you know 12 to 18 months uh we

have thousands of Engineers working on

making Rockham as easy to use as

possible and we're also extending um

across um our you know entire uh

portfolio so uh it's a a very important

part um I think the open ecosystem is

also very important and so uh that's why

we really want all of the software

developers to work together with us and

um you know together we will build uh a

very uh strong ecosystem and get

applications out there as fast as

possible okay thank you so much Lisa uh

it's it's a great um answer I I learned

a lot and thanks for your time for today

that's all for me excellent thank you so

much for having me thank you so much so

shall we take um take pictures and and

stuff yes ABS so much yeah

great