WEBVTT - BONUS EPISODE: Interview with Microsoft CEO Satya Nadella

0:00:00.080 --> 0:00:03.800
<v Speaker 1>Microsoft has been working on AI for decades and chatnots

0:00:03.800 --> 0:00:07.840
<v Speaker 1>actually aren't anything new, but all of a sudden everyone

0:00:08.000 --> 0:00:11.320
<v Speaker 1>is salivating. Why do you think the moment for AI

0:00:11.560 --> 0:00:15.600
<v Speaker 1>is now? Yeah, I mean it's actually you're absolutely right,

0:00:15.600 --> 0:00:19.840
<v Speaker 1>which is AI has been here in fact its mainstream, right,

0:00:19.920 --> 0:00:22.800
<v Speaker 1>I mean, search is an AI product, even the current

0:00:22.840 --> 0:00:28.880
<v Speaker 1>generation of search, every news aggregation, recommendation, and you know

0:00:29.160 --> 0:00:33.760
<v Speaker 1>YouTube or e commerce or TikTok or all AI products,

0:00:34.240 --> 0:00:37.479
<v Speaker 1>except they're all I would say, today's generation of AI

0:00:37.880 --> 0:00:41.240
<v Speaker 1>is all autopilot. In fact, it's a black box that

0:00:41.360 --> 0:00:46.000
<v Speaker 1>beaches sort of use that is dictating in fact, how

0:00:46.000 --> 0:00:49.240
<v Speaker 1>our attention is focused. Whereas going forward, the thing that's

0:00:49.280 --> 0:00:52.600
<v Speaker 1>most exciting about this generation of AI is perhaps we

0:00:52.680 --> 0:00:56.920
<v Speaker 1>move from autopilot to copilot, where we actually prompted. I mean,

0:00:56.960 --> 0:01:00.640
<v Speaker 1>think about it, right, what we are learning to program

0:01:00.680 --> 0:01:04.040
<v Speaker 1>AI as with just natural language, right, and it gets

0:01:04.040 --> 0:01:06.760
<v Speaker 1>smarter every time you use it. Yeah yeah, and also

0:01:07.000 --> 0:01:10.280
<v Speaker 1>you are making it. It's just it's ultimately a stochastic

0:01:10.319 --> 0:01:14.480
<v Speaker 1>machine that you're using as a tool to help reason

0:01:14.560 --> 0:01:18.640
<v Speaker 1>about what you're learning, what you're creating, what you're doing.

0:01:19.040 --> 0:01:22.520
<v Speaker 1>And so yes, I think this shift from autopilot to

0:01:22.560 --> 0:01:27.080
<v Speaker 1>copilot is actually, yes, the next phase of AI, which

0:01:27.240 --> 0:01:30.360
<v Speaker 1>in fact is perhaps going to put us as humans,

0:01:31.120 --> 0:01:34.600
<v Speaker 1>you know, more in the center of using AI to

0:01:34.680 --> 0:01:39.920
<v Speaker 1>our benefit. With Copilot, you are deeply weaving lams across

0:01:40.080 --> 0:01:45.880
<v Speaker 1>all of Microsoft's products Word, Excel, power Point, Outlook. You're

0:01:45.920 --> 0:01:50.720
<v Speaker 1>also basically giving folks their own personal business chatbot. How

0:01:51.080 --> 0:01:54.120
<v Speaker 1>transformative a change do you think this will be in

0:01:54.680 --> 0:01:58.080
<v Speaker 1>how we work? Yeah? To me, that is it right.

0:01:58.200 --> 0:02:02.520
<v Speaker 1>Having built now up Copilot, having built the web co

0:02:02.680 --> 0:02:06.680
<v Speaker 1>pilot with being, and even what we did with Dynamic,

0:02:06.720 --> 0:02:09.680
<v Speaker 1>this is the big next step for us to put

0:02:09.720 --> 0:02:13.200
<v Speaker 1>it in the tools everybody uses every day for their work.

0:02:13.919 --> 0:02:16.840
<v Speaker 1>I think it does three things emily for me. You know,

0:02:17.440 --> 0:02:19.399
<v Speaker 1>one of the things that I've always said is, God,

0:02:19.400 --> 0:02:23.440
<v Speaker 1>there's so much functionality in Word or Excel and PowerPoint.

0:02:24.280 --> 0:02:26.799
<v Speaker 1>How do we make it such that people use this

0:02:27.360 --> 0:02:31.160
<v Speaker 1>in powerful ways to create great content, great documents, great

0:02:31.160 --> 0:02:35.400
<v Speaker 1>PowerPoint art, learn how to do analysis that's pretty sophisticated

0:02:35.400 --> 0:02:38.239
<v Speaker 1>in an interesting way. Now without having to say, let

0:02:38.240 --> 0:02:42.480
<v Speaker 1>me learn all the commanding of office, I just literally

0:02:42.560 --> 0:02:46.519
<v Speaker 1>can use natural language. So the power of thirty plus

0:02:46.600 --> 0:02:51.880
<v Speaker 1>years of office creation, the sophistication of these tools is

0:02:51.880 --> 0:02:55.200
<v Speaker 1>just available to every user. Same thing with even teams

0:02:55.240 --> 0:02:58.600
<v Speaker 1>and teams co pilot. I think about how meetings can

0:02:58.680 --> 0:03:01.440
<v Speaker 1>be more effective team's co pilot, But I think the

0:03:01.560 --> 0:03:04.560
<v Speaker 1>probably the biggest difference maker will be business chat because

0:03:04.960 --> 0:03:07.639
<v Speaker 1>if you think about the most important database in any

0:03:07.680 --> 0:03:12.480
<v Speaker 1>company is the database underneath all of your productivity software.

0:03:12.800 --> 0:03:15.880
<v Speaker 1>Accept that data is all siloed today, But now I

0:03:15.960 --> 0:03:18.480
<v Speaker 1>can queriate with natural I can say, oh, I'm going

0:03:18.520 --> 0:03:20.440
<v Speaker 1>to meet this customer. Can you tell me the last

0:03:20.480 --> 0:03:22.280
<v Speaker 1>time I met them? Can you bring up all the

0:03:22.360 --> 0:03:25.440
<v Speaker 1>documents that are written up about this customer and summarize

0:03:25.440 --> 0:03:27.480
<v Speaker 1>it so that I'm current on what I need to

0:03:27.520 --> 0:03:32.240
<v Speaker 1>be prepped for that ability to interrogate that database, queriate

0:03:32.480 --> 0:03:36.120
<v Speaker 1>and do it without learning some new syntax of querying language.

0:03:36.120 --> 0:03:38.560
<v Speaker 1>But just natural language is super powerful. How do you

0:03:38.600 --> 0:03:41.320
<v Speaker 1>make sure it's not clippy two? Point out that it

0:03:41.400 --> 0:03:44.720
<v Speaker 1>is helpful, delightful, doesn't want to make me click out

0:03:44.720 --> 0:03:47.840
<v Speaker 1>a sap. There are two sets of things. One is

0:03:48.240 --> 0:03:52.440
<v Speaker 1>you know if you're laughing, because because look like our

0:03:52.520 --> 0:03:55.640
<v Speaker 1>industry is full of lots of you know, examples from

0:03:55.680 --> 0:03:58.960
<v Speaker 1>clippy to even let's a current generation of these assistants

0:03:59.000 --> 0:04:01.200
<v Speaker 1>and so on, they all griddle. I think there are

0:04:01.240 --> 0:04:04.600
<v Speaker 1>two things that this generation of way I do. One

0:04:04.840 --> 0:04:08.320
<v Speaker 1>is when we say they understand natural language, they truly

0:04:08.400 --> 0:04:11.560
<v Speaker 1>understand natural language. So it's not like, Okay, we'll understand

0:04:11.560 --> 0:04:13.600
<v Speaker 1>you only if you stay in sort of these very

0:04:13.720 --> 0:04:18.040
<v Speaker 1>narrow rails that we've defined for you. Then the comprehension

0:04:18.760 --> 0:04:21.479
<v Speaker 1>and what they can do for you right is also

0:04:21.560 --> 0:04:25.159
<v Speaker 1>not narrow, and so I think that's why. But that said,

0:04:25.240 --> 0:04:27.400
<v Speaker 1>I think we are also going to have to learn

0:04:27.560 --> 0:04:32.360
<v Speaker 1>that ultimately these are tools. They're stochastic in nature. Just

0:04:32.440 --> 0:04:35.520
<v Speaker 1>like anytime somebody sends me a draft, I review the draft,

0:04:35.560 --> 0:04:38.160
<v Speaker 1>I just don't accept the draft. We will do that

0:04:38.279 --> 0:04:40.200
<v Speaker 1>like interesting enough, we learn a lot and get up

0:04:40.200 --> 0:04:42.159
<v Speaker 1>co pilot. In fact, that first time we get up

0:04:42.160 --> 0:04:44.800
<v Speaker 1>copiot came even softa abouper saying oh yeah, this does

0:04:44.920 --> 0:04:48.960
<v Speaker 1>make mistakes, except in even a few months people say,

0:04:49.000 --> 0:04:50.880
<v Speaker 1>oh yeah, but I know how to correct those mistakes.

0:04:51.640 --> 0:04:55.359
<v Speaker 1>And so that ability to work with this co pilot,

0:04:56.240 --> 0:04:59.320
<v Speaker 1>give it feedback, know how to verify it. Even this

0:04:59.400 --> 0:05:01.680
<v Speaker 1>chain of thought, reasoning and excel. Like the one feature

0:05:01.680 --> 0:05:03.000
<v Speaker 1>I don't know if you saw this, but that is

0:05:03.080 --> 0:05:05.479
<v Speaker 1>very cool, which is we said, okay, what's the design

0:05:05.640 --> 0:05:07.880
<v Speaker 1>choice we can make so the users get in the

0:05:07.960 --> 0:05:10.880
<v Speaker 1>habit of not just accepting whatever AI is saying, but

0:05:11.080 --> 0:05:14.640
<v Speaker 1>even ask it to show you it's scratch bad work. Right.

0:05:15.480 --> 0:05:18.360
<v Speaker 1>It's literally like inspecting somebody's homework, right, which is, hey,

0:05:18.400 --> 0:05:20.279
<v Speaker 1>tell me exactly how you did this and so that

0:05:20.320 --> 0:05:22.680
<v Speaker 1>I can verify. Those are the kinds of things that

0:05:22.720 --> 0:05:26.039
<v Speaker 1>we learn. You're trying to reinvent search with this AI

0:05:26.120 --> 0:05:29.680
<v Speaker 1>powered thing, and I believe it's been using GPT for

0:05:29.680 --> 0:05:33.880
<v Speaker 1>for a while. Now what's worked? What has it? I

0:05:33.920 --> 0:05:38.000
<v Speaker 1>think what they're too. One thing that we're learning is

0:05:38.040 --> 0:05:43.800
<v Speaker 1>the search context, right, So conversational search is a thing.

0:05:44.520 --> 0:05:48.800
<v Speaker 1>So this grounding of your conversation with search data, I

0:05:48.839 --> 0:05:52.640
<v Speaker 1>think is one mode. And then there is a completely

0:05:52.640 --> 0:05:54.960
<v Speaker 1>different mode that we're also learned, which is people just

0:05:55.000 --> 0:05:59.480
<v Speaker 1>want to chat. And so we are now getting good

0:05:59.560 --> 0:06:01.919
<v Speaker 1>at even the product design so that we make that

0:06:02.120 --> 0:06:05.080
<v Speaker 1>an explicit choice. So for example, when we launched Being,

0:06:05.480 --> 0:06:09.640
<v Speaker 1>we didn't have these three modes. We now have how

0:06:09.640 --> 0:06:11.760
<v Speaker 1>precise do you want it to be? Or how creative

0:06:11.800 --> 0:06:13.400
<v Speaker 1>you want to be or you want to be? Balanced

0:06:13.800 --> 0:06:15.359
<v Speaker 1>that I think is one of the biggest learnings we

0:06:15.440 --> 0:06:17.600
<v Speaker 1>launched it at Oh Wow. People do, in fact want

0:06:17.600 --> 0:06:21.200
<v Speaker 1>to engage even in what is chat inside of search

0:06:21.520 --> 0:06:23.279
<v Speaker 1>in different ways, and we've got to put the user

0:06:23.320 --> 0:06:25.320
<v Speaker 1>control back. How much market share do you think you

0:06:25.360 --> 0:06:30.039
<v Speaker 1>can really take from Google? Like prediction? Give me. We

0:06:30.600 --> 0:06:35.320
<v Speaker 1>are a real I'm thrilled to be in search. We're

0:06:35.320 --> 0:06:39.640
<v Speaker 1>a very small player in search, and I look forward

0:06:39.680 --> 0:06:43.800
<v Speaker 1>to every inch we gain is a big game. You're

0:06:43.800 --> 0:06:46.559
<v Speaker 1>coming for Search, They're coming for Office. They're now putting

0:06:46.600 --> 0:06:50.320
<v Speaker 1>AI in there. You know, Google Docs and cheets and Gmail.

0:06:50.920 --> 0:06:53.040
<v Speaker 1>Are we just going to see you and Sundar are

0:06:53.040 --> 0:06:54.920
<v Speaker 1>trying to one up each other every week in this

0:06:55.160 --> 0:07:00.919
<v Speaker 1>race to greatness. You know, I just warnt bardon being

0:07:01.000 --> 0:07:04.359
<v Speaker 1>both to thrive. I just want Google work Space and

0:07:04.520 --> 0:07:06.760
<v Speaker 1>Microsoft three six five both to thrive. I mean, look,

0:07:06.800 --> 0:07:09.279
<v Speaker 1>at the end of the day, the fun part of

0:07:09.360 --> 0:07:12.680
<v Speaker 1>being in this industry and competing is but you know,

0:07:12.760 --> 0:07:16.920
<v Speaker 1>it's the innovation and competition. Is the last time I

0:07:17.000 --> 0:07:20.800
<v Speaker 1>checked off a fantastic thing for users and the industry,

0:07:20.840 --> 0:07:23.280
<v Speaker 1>and so yeah, so let there be good sort of

0:07:23.600 --> 0:07:26.080
<v Speaker 1>you know, good innovation coming. And I think you know,

0:07:26.120 --> 0:07:28.160
<v Speaker 1>Google's going to do you know, is a very innovative

0:07:28.160 --> 0:07:30.320
<v Speaker 1>company and uh, and we have a lot of respect

0:07:30.360 --> 0:07:33.720
<v Speaker 1>for them, and I expect us to compete in multiple categories.

0:07:33.840 --> 0:07:37.960
<v Speaker 1>In my decade plus covering Microsoft, I can't remember you

0:07:38.080 --> 0:07:41.640
<v Speaker 1>releasing this much in quick succession. Why is it all

0:07:41.640 --> 0:07:44.840
<v Speaker 1>happening so fast? Yeah, it's you know, it's sort of

0:07:44.880 --> 0:07:48.520
<v Speaker 1>sometimes it feels it's all happening fast. It's we started

0:07:48.560 --> 0:07:51.200
<v Speaker 1>working on this, you know, good four years ago, right,

0:07:51.240 --> 0:07:52.880
<v Speaker 1>I mean in some sense, if you think about when

0:07:53.000 --> 0:07:56.040
<v Speaker 1>open ai and Microsoft came together and said, hey, this

0:07:56.200 --> 0:08:02.040
<v Speaker 1>next generation of large language models need infrastructure. Let's build

0:08:02.040 --> 0:08:07.880
<v Speaker 1>the infrastructure, tune the infrastructure, Let's understand even what AI

0:08:07.960 --> 0:08:10.760
<v Speaker 1>safety and alignment looks like for these what are the

0:08:10.880 --> 0:08:14.000
<v Speaker 1>use cases? And this has been four years plus in making.

0:08:14.320 --> 0:08:19.040
<v Speaker 1>So once we started seeing the scaling effects, the promise

0:08:19.280 --> 0:08:23.560
<v Speaker 1>of the emergent capabilities, even that started showing up in

0:08:23.640 --> 0:08:26.240
<v Speaker 1>these large language models. That's why last years in fact

0:08:26.280 --> 0:08:30.360
<v Speaker 1>that perhaps for me, the application of these large language

0:08:30.400 --> 0:08:33.319
<v Speaker 1>models inside or get up Copilot was the big it's

0:08:33.360 --> 0:08:37.120
<v Speaker 1>the biggest LM product out there today, and so that

0:08:37.200 --> 0:08:39.640
<v Speaker 1>gave us confidence that hey, we now can apply it

0:08:39.720 --> 0:08:43.800
<v Speaker 1>in more context. So yes, it feels that we've launched

0:08:43.800 --> 0:08:45.760
<v Speaker 1>a lot of things just in a hurry this year,

0:08:45.880 --> 0:08:48.000
<v Speaker 1>but it's been four years in the making, and obviously

0:08:48.040 --> 0:08:52.040
<v Speaker 1>it's a great partnership with open AI. Microsoft just reportedly

0:08:52.080 --> 0:08:56.360
<v Speaker 1>laid off a team focused on ethical and responsible AI. Meantime,

0:08:56.360 --> 0:09:00.360
<v Speaker 1>you've got the Center for Humane Technology calling the too

0:09:00.440 --> 0:09:03.959
<v Speaker 1>AI a RaSE to recklessness. How do you respond to that?

0:09:04.320 --> 0:09:07.679
<v Speaker 1>I mean, first of all, in terms of impact on

0:09:07.800 --> 0:09:10.800
<v Speaker 1>anybody at Microsoft, this is just probably the thing that

0:09:10.880 --> 0:09:14.240
<v Speaker 1>weighs on me heavily, because after all, any restructuring is hard,

0:09:14.280 --> 0:09:17.360
<v Speaker 1>hard on the people who are mostly impacted. That said,

0:09:18.040 --> 0:09:21.280
<v Speaker 1>two things. One is this is no longer a side

0:09:21.360 --> 0:09:24.400
<v Speaker 1>thing for Microsoft, right because in some sense, whether it's design,

0:09:24.480 --> 0:09:28.440
<v Speaker 1>whether it's alignment, safety, ethics, it's kind of like saying quality,

0:09:28.600 --> 0:09:32.840
<v Speaker 1>performance and design, core design. So I can't have now

0:09:32.920 --> 0:09:35.960
<v Speaker 1>an AI team on the side. It's all the mainstream.

0:09:36.120 --> 0:09:38.800
<v Speaker 1>It's sort of like there's no get up. There's no

0:09:38.840 --> 0:09:42.280
<v Speaker 1>get up without copilot. There is no Microsoft three six

0:09:42.320 --> 0:09:45.280
<v Speaker 1>five without co pilot. So in some sense, the hard

0:09:45.360 --> 0:09:48.400
<v Speaker 1>process of companies like ours are going to constantly go

0:09:48.520 --> 0:09:51.400
<v Speaker 1>through a lot of change, and what you as something

0:09:51.440 --> 0:09:54.600
<v Speaker 1>that was done on the side today is now mainstream.

0:09:54.640 --> 0:09:58.880
<v Speaker 1>So that's what's happening. And then I think, if anything, debate,

0:09:59.320 --> 0:10:04.760
<v Speaker 1>dialogue and scrutiny on what is the space of innovation?

0:10:06.000 --> 0:10:11.160
<v Speaker 1>Is it really creating benefits for society? I think are absolutely,

0:10:11.200 --> 0:10:13.400
<v Speaker 1>and I'll welcome it. Right, I look at them and say,

0:10:13.480 --> 0:10:18.040
<v Speaker 1>no one can run faster than the benefits to the

0:10:18.160 --> 0:10:21.960
<v Speaker 1>broader society. And then the norms that we enforce as

0:10:21.960 --> 0:10:26.400
<v Speaker 1>a democratic society on any technology. And so I feel

0:10:26.400 --> 0:10:28.240
<v Speaker 1>like we are at the very early stages of it.

0:10:28.280 --> 0:10:30.440
<v Speaker 1>So I will ask us to be open to it,

0:10:30.880 --> 0:10:33.640
<v Speaker 1>but at the same time scrutinize it and let's have

0:10:33.679 --> 0:10:36.720
<v Speaker 1>a dialogue on what the benefits are. And in that context,

0:10:36.800 --> 0:10:41.720
<v Speaker 1>let's also recognize, especially with this AI, well why were

0:10:41.800 --> 0:10:44.880
<v Speaker 1>we not asking ourselves, like the AI that's already in

0:10:44.920 --> 0:10:48.600
<v Speaker 1>our lives, and how what is it doing? It's right

0:10:48.679 --> 0:10:51.760
<v Speaker 1>we've gone straight to say, oh, wow, these lns have

0:10:51.840 --> 0:10:55.360
<v Speaker 1>some hallucination. Guess what. Right, there's a lot of AI

0:10:55.400 --> 0:10:57.640
<v Speaker 1>that I don't even know what it's doing. And except

0:10:57.679 --> 0:11:02.240
<v Speaker 1>I'm happily clicking away and accepting the recommendations. So why

0:11:02.240 --> 0:11:06.040
<v Speaker 1>don't we in fact educate ourselves to ask all of

0:11:06.040 --> 0:11:08.199
<v Speaker 1>what AI is doing in our lives and say how

0:11:08.240 --> 0:11:10.800
<v Speaker 1>to do it safely in our line way. Elon Musk,

0:11:10.800 --> 0:11:12.920
<v Speaker 1>who co found it open ai and then left, has

0:11:12.960 --> 0:11:16.000
<v Speaker 1>said it's not what he intended. It is closed, sourced

0:11:16.000 --> 0:11:19.760
<v Speaker 1>and effectively controlled by Microsoft. How would you respond? Look,

0:11:19.800 --> 0:11:21.800
<v Speaker 1>I mean, I think, first of all, open ai cares

0:11:22.080 --> 0:11:26.959
<v Speaker 1>deeply about their mission and doing it in the most

0:11:27.000 --> 0:11:29.800
<v Speaker 1>safe way and in the most open way. And in

0:11:29.840 --> 0:11:35.600
<v Speaker 1>some sense there's an interesting trade between openness and safety.

0:11:36.240 --> 0:11:38.040
<v Speaker 1>So that is sort of one of the reasons why

0:11:38.080 --> 0:11:40.600
<v Speaker 1>they have what they have in terms of their governance,

0:11:41.559 --> 0:11:45.200
<v Speaker 1>architecture and you know, and so therefore at some level

0:11:45.240 --> 0:11:49.199
<v Speaker 1>they have been very very clear on what principles drive them. Similarly,

0:11:49.320 --> 0:11:51.560
<v Speaker 1>we have been very very clear on the principles that

0:11:51.679 --> 0:11:55.600
<v Speaker 1>drive us around AI safety and responsibility, and we'll stick

0:11:55.600 --> 0:11:59.240
<v Speaker 1>to them. I have to ask you a question about

0:11:59.320 --> 0:12:04.600
<v Speaker 1>the economy and whether you're concerned about a prolonged tech bust.

0:12:04.720 --> 0:12:07.160
<v Speaker 1>I mean, we've seen the collapse of three banks over

0:12:07.160 --> 0:12:09.920
<v Speaker 1>the course of the last week, including Silicon Valley Bank,

0:12:10.480 --> 0:12:13.800
<v Speaker 1>tighter money, more uncertainty. How are you thinking about this?

0:12:14.520 --> 0:12:17.920
<v Speaker 1>I mean, I think at the highest or levels. I

0:12:17.960 --> 0:12:21.959
<v Speaker 1>think there was an aberration of maybe a ten year

0:12:22.080 --> 0:12:27.560
<v Speaker 1>period of low interest rates and everything that came with it,

0:12:27.880 --> 0:12:30.760
<v Speaker 1>not just in tech but in the broader economy. And

0:12:30.840 --> 0:12:33.080
<v Speaker 1>I just think that we're just getting back to normal,

0:12:33.200 --> 0:12:36.480
<v Speaker 1>like at least the thing that perhaps we have to

0:12:36.480 --> 0:12:39.320
<v Speaker 1>remind ourselves. Mostly the world looked like this, which is

0:12:39.520 --> 0:12:45.880
<v Speaker 1>interest rates were higher than zero. Inflation was perhaps maybe

0:12:45.960 --> 0:12:49.920
<v Speaker 1>structurally going to be higher, just given everything that's happening

0:12:49.920 --> 0:12:53.760
<v Speaker 1>with supply chains in the geopolitics, and we all as

0:12:53.800 --> 0:12:56.560
<v Speaker 1>businesses have to be accountable to how to manage in

0:12:56.600 --> 0:13:02.240
<v Speaker 1>that environment, and tech is one sector. And so I

0:13:02.320 --> 0:13:03.760
<v Speaker 1>kind of look at this and say, hey, it's a

0:13:03.800 --> 0:13:06.679
<v Speaker 1>return to normal as opposed to anything sort of that

0:13:06.720 --> 0:13:09.120
<v Speaker 1>to be we need to be worried about as being prolonged.

0:13:09.160 --> 0:13:12.000
<v Speaker 1>In fact, this is the long run. The economies have

0:13:12.080 --> 0:13:14.240
<v Speaker 1>to sort of be, you know, more real. All right,

0:13:14.280 --> 0:13:17.520
<v Speaker 1>So this is normal to you, I think. So, I mean,

0:13:17.559 --> 0:13:19.920
<v Speaker 1>I think that sometimes we sort of say, you know,

0:13:20.760 --> 0:13:24.840
<v Speaker 1>the last ten years can never be sort of the

0:13:24.920 --> 0:13:28.800
<v Speaker 1>way way forward on and it's good. I think it's

0:13:28.880 --> 0:13:32.680
<v Speaker 1>better to have businesses that are run efficiently, that are

0:13:32.800 --> 0:13:36.720
<v Speaker 1>actually measured on the way both whether it's on the

0:13:36.760 --> 0:13:41.719
<v Speaker 1>societal impact or on real economic impact. In nineteen ninety five,

0:13:41.760 --> 0:13:44.959
<v Speaker 1>Bill Gates Centemmo calling the Internet a title way if

0:13:45.000 --> 0:13:47.160
<v Speaker 1>that would change all the rules and was going to

0:13:47.160 --> 0:13:50.800
<v Speaker 1>be crucial to every part of the business, is AI

0:13:50.840 --> 0:13:53.839
<v Speaker 1>and epic. Yeah. I mean, in fact, I sort of

0:13:53.880 --> 0:13:57.000
<v Speaker 1>say the chat GPT when it first came out was

0:13:57.080 --> 0:14:00.360
<v Speaker 1>like when Mosaic first came out I think in ninety

0:14:00.440 --> 0:14:04.320
<v Speaker 1>three as the first browser. And so yes, it does

0:14:04.480 --> 0:14:07.240
<v Speaker 1>feel like you know, to the Bill Memo in nineteen

0:14:07.360 --> 0:14:10.160
<v Speaker 1>ninety five, It does feel like that to me. So

0:14:10.200 --> 0:14:12.240
<v Speaker 1>it's as big as the enter. I think it's as big.

0:14:12.320 --> 0:14:14.079
<v Speaker 1>It's just like in all of these things, right we

0:14:14.200 --> 0:14:16.880
<v Speaker 1>in the tech industry or you know, classic experts and

0:14:17.040 --> 0:14:21.440
<v Speaker 1>over hyping everything. So the question is can we really

0:14:21.520 --> 0:14:24.680
<v Speaker 1>like I hope at least that What motivates me is

0:14:24.720 --> 0:14:28.160
<v Speaker 1>I want to use this technology to truly do what

0:14:28.280 --> 0:14:30.440
<v Speaker 1>I think at least all of us are in tech

0:14:30.480 --> 0:14:33.720
<v Speaker 1>for which is democratizing access to it. So when someone

0:14:33.760 --> 0:14:36.320
<v Speaker 1>says to me, hey, here is how a farmer in

0:14:36.440 --> 0:14:41.640
<v Speaker 1>rural India, you know, can use this technology to express

0:14:41.680 --> 0:14:44.560
<v Speaker 1>a complex thought on how to get a subsidy from

0:14:44.600 --> 0:14:48.160
<v Speaker 1>a government program and can do that successfully. That gives

0:14:48.200 --> 0:14:50.280
<v Speaker 1>me a lot of sort of you know, hope. How

0:14:50.320 --> 0:14:53.600
<v Speaker 1>confident are you that Microsoft is going to catch the wave? Look?

0:14:53.840 --> 0:14:55.840
<v Speaker 1>We are you know, at the end of the day,

0:14:56.320 --> 0:15:02.440
<v Speaker 1>my confidence is high only if I can do useful

0:15:02.520 --> 0:15:05.520
<v Speaker 1>things every day for users. So in other words, what

0:15:05.560 --> 0:15:09.520
<v Speaker 1>we did with being, what we did with getub copilot,

0:15:09.560 --> 0:15:11.480
<v Speaker 1>and what we're about to now do with Microsoft three

0:15:11.520 --> 0:15:14.040
<v Speaker 1>six five copilot is what gives me confidence. Ultimately, it's

0:15:14.080 --> 0:15:16.200
<v Speaker 1>not about my confidence, but it's about the products people

0:15:16.280 --> 0:15:19.080
<v Speaker 1>use and find useful. I want to ask about jobs

0:15:19.080 --> 0:15:22.480
<v Speaker 1>because obviously Microsoft makes software that helps people do their jobs,

0:15:22.480 --> 0:15:25.920
<v Speaker 1>and I wonder if AI laid in software, we'll put

0:15:25.960 --> 0:15:29.480
<v Speaker 1>some people out of jobs. Sam Altman has this idea

0:15:29.800 --> 0:15:33.440
<v Speaker 1>that AI is going to create this kind of utopia

0:15:33.520 --> 0:15:35.720
<v Speaker 1>and generate wealth that's going to be enough to cut

0:15:35.760 --> 0:15:39.200
<v Speaker 1>everyone a decent size check, but eliminate some jobs. Do

0:15:39.240 --> 0:15:41.880
<v Speaker 1>you agree with that? You know? Look, I mean you

0:15:41.880 --> 0:15:44.880
<v Speaker 1>know from Kenes, do I guess? Altman? They've all talked

0:15:44.880 --> 0:15:47.920
<v Speaker 1>about the two day work week, and I'm looking forward

0:15:47.920 --> 0:15:51.240
<v Speaker 1>to it. But the point is, Look, the lump of

0:15:51.360 --> 0:15:56.440
<v Speaker 1>labor fallacy has never proven out right, which is in

0:15:56.480 --> 0:15:59.320
<v Speaker 1>some sense there is displacement, and in fact of anything,

0:15:59.360 --> 0:16:02.680
<v Speaker 1>what we have to do is really do a fantastic

0:16:02.760 --> 0:16:07.200
<v Speaker 1>job as a society to deal with any displacement. Because

0:16:07.400 --> 0:16:10.200
<v Speaker 1>one job turns into another job, you have to then

0:16:10.320 --> 0:16:13.240
<v Speaker 1>skill people on an other job. And in fact, in

0:16:13.280 --> 0:16:16.640
<v Speaker 1>an interesting way, here's one thing even in this Microsoft

0:16:16.640 --> 0:16:19.760
<v Speaker 1>three six five tool, like, there's this power automate tool.

0:16:19.880 --> 0:16:21.840
<v Speaker 1>Up to now we've called it the low code no

0:16:22.000 --> 0:16:26.160
<v Speaker 1>code tool for doing workflo automation. Interestingly enough, you now

0:16:26.240 --> 0:16:30.000
<v Speaker 1>can automate workflows just using natural language. Guess what that means.

0:16:30.120 --> 0:16:33.320
<v Speaker 1>Anybody who's in the front lines in healthcare and retail

0:16:33.920 --> 0:16:37.640
<v Speaker 1>can automate or be part of the IT journey. That

0:16:37.720 --> 0:16:42.000
<v Speaker 1>to me gives means their new jobs and better wage support.

0:16:42.280 --> 0:16:45.720
<v Speaker 1>So I feel, yes, there's going to be some changes

0:16:45.760 --> 0:16:48.040
<v Speaker 1>in jobs. There's going to be some places where there

0:16:48.040 --> 0:16:51.480
<v Speaker 1>will be wage pressure, there will be opportunities for increased

0:16:51.480 --> 0:16:55.200
<v Speaker 1>wages because of increased productivity. We should look at it

0:16:55.240 --> 0:16:57.960
<v Speaker 1>all and at the same time being very clear eyed

0:16:58.080 --> 0:17:00.920
<v Speaker 1>about any displacement risk because that one thing that we've

0:17:01.000 --> 0:17:03.920
<v Speaker 1>also learned in the last twenty years is that any

0:17:04.080 --> 0:17:08.439
<v Speaker 1>society that doesn't really pay attention to who are the

0:17:08.480 --> 0:17:12.080
<v Speaker 1>winners and the losers, and to make sure that as

0:17:12.160 --> 0:17:16.679
<v Speaker 1>a society we are not really you know, imbalanced in

0:17:16.760 --> 0:17:19.760
<v Speaker 1>it in terms of economic opportunity, will be better off.

0:17:19.920 --> 0:17:22.760
<v Speaker 1>I think a lot about my kids and how AI

0:17:22.880 --> 0:17:25.600
<v Speaker 1>will have something that I don't, which is an infinite

0:17:25.600 --> 0:17:28.760
<v Speaker 1>amount of time to spend with them, and how these

0:17:28.840 --> 0:17:32.280
<v Speaker 1>chatpots are so friendly, and how quickly that could turn

0:17:32.320 --> 0:17:35.280
<v Speaker 1>into an unhealthy relationship or you know, maybe it's nudging

0:17:35.280 --> 0:17:38.880
<v Speaker 1>them to make a bad decision. As a parent, does

0:17:38.920 --> 0:17:42.159
<v Speaker 1>any part of that scare you? So that's kind of

0:17:42.160 --> 0:17:45.800
<v Speaker 1>one of the reasons why I think this moving from

0:17:45.880 --> 0:17:50.680
<v Speaker 1>autopilot to this copilot hopefully gives us more control, whether

0:17:50.720 --> 0:17:53.199
<v Speaker 1>it's as parents are more importantly, even as children. Like

0:17:53.280 --> 0:17:56.600
<v Speaker 1>one of the things that was very cool yesterday to

0:17:56.840 --> 0:18:01.440
<v Speaker 1>see in the launch of GPT four was the demo

0:18:02.040 --> 0:18:05.800
<v Speaker 1>or the launch of Khan Academy stuff. And sal sent

0:18:05.880 --> 0:18:07.600
<v Speaker 1>me this last night and I was looking at his

0:18:07.720 --> 0:18:11.119
<v Speaker 1>algebra class. It was so engaging, right, I mean, think

0:18:11.160 --> 0:18:13.439
<v Speaker 1>about it, like one of the dreams we've always had

0:18:13.960 --> 0:18:18.800
<v Speaker 1>is can I have a personalized tutor that is engaging,

0:18:18.840 --> 0:18:21.639
<v Speaker 1>that is actually trying to teach me so to yours point,

0:18:21.760 --> 0:18:25.520
<v Speaker 1>I think maybe we should, of course be very very

0:18:25.560 --> 0:18:28.639
<v Speaker 1>watchful of what happens. But at the same time, I

0:18:28.680 --> 0:18:32.480
<v Speaker 1>think this generation of bots, in this generation of AI,

0:18:32.720 --> 0:18:37.760
<v Speaker 1>probably just go from engagement to more giving us more

0:18:37.840 --> 0:18:41.760
<v Speaker 1>agency to learn. It's fascinating such a NATA. Thank you,

0:18:42.000 --> 0:18:43.080
<v Speaker 1>thank you for watching us.