Latest / What's Uptin? / The man who bet on OpenAI before anyone else has a warning for you

Transcript

The man who bet on OpenAI before anyone else has a warning for you · What's Uptin?

• 0:00
  Should an 18 year old today still go to college, and if So,
• 0:03
  what should they study? Steve Jobs once said that he
• 0:06
  worked every day at the intersection of engineering and
• 0:10
  the liberal arts. Microsoft is the fourth most
• 0:12
  valuable company in the world. It slipped from the top three
• 0:15
  because of Nvidia's surge. I'm sitting down with its
• 0:18
  president, Brad Smith, to figure out how we should be thinking
• 0:22
  about education, jobs and artificial intelligence
• 0:26
  disrupting everything we ever knew.
• 0:28
  How? Do we use AI to think more
• 0:30
  trustworthy? AI is what is fundamental.
• 0:34
  It's almost existential to the industry's long term success.
• 0:39
  You're not going to use AI if you don't trust it.
• 0:41
  This is What's Up tent with Microsoft president Brad Smith.
• 0:44
  Brad Smith, thanks so much for sitting down with me here at Web
• 0:46
  Summit. Thanks.
• 0:48
  Great to be. Here, Yeah.
• 0:49
  So first question here. Microsoft invested in Open AI
• 0:52
  back in 2019. What did you see in Open AI at a
• 0:56
  time where most people hadn't even heard of it?
• 0:59
  Satya Nadella, our CEO, had this intuition and he asked Kevin
• 1:03
  Scott, one of our most senior people, to go spend time with
• 1:06
  them. But the intuition that Satya had
• 1:09
  was that the large language model approach that Open AI was
• 1:13
  taking and it's scaling approach might just be the secret to
• 1:18
  bringing AI to life in a much bigger way.
• 1:21
  And the intuition was based on, I think, a very interesting
• 1:24
  insight. Satya said in a way, language is
• 1:28
  software for the brain. The brain is the hardware.
• 1:32
  Language became the software. If you could replicate
• 1:35
  everything that one can do with language, you might unlock the
• 1:39
  potential of artificial intelligence.
• 1:41
  Now obviously there's a lot more complexity that has emerged, but
• 1:45
  I think as we look back six years later, that insight is at
• 1:49
  the heart of everything. And Open AI was the company to
• 1:51
  to lead that. Absolutely.
• 1:53
  I mean, we knew and Satya knew, Sam Altman and I, and when Kevin
• 1:59
  Scott, our senior engineer, went over and spent time, he looked
• 2:03
  at what they were doing. He came back to Satya and he
• 2:07
  said, I think these guys are on the right track.
• 2:10
  Now Fast forward to today, a lot of your ecosystem is powered by
• 2:14
  open AI. How do you draw the line between
• 2:16
  partnership versus dependency? Well, in a sense, it's been an
• 2:20
  evolving relationship for both of us and for everybody in the
• 2:23
  ecosystem. Open AI and Microsoft are
• 2:27
  critical partners. We enormously value that
• 2:30
  partnership with Open AI, but what started out as sort of more
• 2:34
  symbiotic and exclusive is much more varied today.
• 2:39
  Open AI works with other companies for infrastructure.
• 2:42
  Microsoft Works with other companies for models.
• 2:45
  We're developing our own as well.
• 2:47
  Open AI is going to build its own infrastructure, and we
• 2:50
  collaborate in a wider variety of areas and we compete as well.
• 2:54
  And that was always part of the original plan.
• 2:57
  Our agreements were written to reflect that, and I think you're
• 3:01
  just going to see this kind of collaboration and competition
• 3:04
  not just between Microsoft and Open AI, but between a number of
• 3:08
  different companies in the AI sector.
• 3:11
  Microsoft announced over 15,000 layoffs this year.
• 3:14
  How do you respond to people like Andrew Yang who say this is
• 3:17
  the beginning of the end and these jobs are not coming back?
• 3:20
  Well, I think it's way too early to jump to that conclusion.
• 3:25
  What Microsoft did and what some other tech companies have done,
• 3:29
  but let me just speak for Microsoft, was we eliminated
• 3:33
  15,000 jobs in one area and then we created a comparable number
• 3:38
  in other areas. So the total employment level
• 3:41
  didn't necessarily change. And I think Andrew Yang's
• 3:44
  comment goes more to that. Now.
• 3:48
  I do think that we've learned from this year, and I think the
• 3:51
  fundamental learning for us is something we need to apply not
• 3:54
  just in Microsoft, but it's probably a question that is more
• 3:58
  societally relevant. If some jobs go away and other
• 4:02
  jobs get created, how do we help people move from one thing to
• 4:07
  the next? And this is where we're now
• 4:10
  focused as a company. It's an easy problem to
• 4:14
  describe. It's a harder challenge to
• 4:16
  solve. I do think that it requires
• 4:19
  several things and its heart is just more skilling, more work to
• 4:24
  equip people so they can do new things.
• 4:27
  And that in part involves AI skills, but there's a lot of
• 4:30
  other skills as well. I also think it really requires
• 4:34
  more forecasting, longer term efforts.
• 4:37
  So we have a sense of what are the new areas that where we're
• 4:40
  going to need people so we can work with our employees and help
• 4:44
  them make that transition. And you can use LinkedIn right
• 4:46
  for for a lot of that. Absolutely.
• 4:49
  Because LinkedIn is part of Microsoft.
• 4:51
  We, we can use that, you know, in ways that we, we should.
• 4:54
  I mean, if anybody's using LinkedIn, it should be
• 4:56
  Microsoft. We are.
• 4:58
  And I think LinkedIn has become just this extraordinary resource
• 5:01
  for the world in matching people who are seeking jobs with the
• 5:05
  jobs that need to be filled, but not just the jobs, the skills
• 5:09
  and the ability to turn those skills into credentials so that
• 5:13
  they're obvious to future employers.
• 5:14
  You know, absolutely LinkedIn is important.
• 5:16
  Should an 18 year old today still go to college, and if So,
• 5:20
  what should they study? Well I don't think that there's
• 5:22
  a single answer that is right for every 18 year old.
• 5:27
  My kids are a little older than 18, but if they were 18 I would
• 5:30
  advise them go to college. I would say learn a combination
• 5:35
  of computer science and engineering and liberal arts and
• 5:41
  think about what combination is right for you.
• 5:44
  It's not the same combination for everyone, but everyone
• 5:47
  should have a good dose of both. And why liberal arts?
• 5:50
  Well, think about the role that AI is playing in the world.
• 5:54
  You know, it's so societally impactful in the future.
• 5:57
  It's not just about learning to code.
• 5:59
  It's not just learning about, say, AI systems, It's really
• 6:04
  becoming familiar with the ethics of AI, the societal
• 6:08
  impact of AI, the economic effect, the sociological effect.
• 6:12
  I think one way to learn about all of that is to have the
• 6:15
  benefit of studying history, to have some grounding even in the
• 6:19
  humanities and and philosophy. Steve Jobs once said that he
• 6:23
  worked every day at the intersection of engineering and
• 6:27
  the liberal arts. I think part of what has made
• 6:30
  Microsoft a success for so many decades is we bring together
• 6:34
  that variety of disciplines. There are days when I'm one of
• 6:38
  the Champions of the liberal arts side of this and I interact
• 6:41
  with great engineering leaders. I am so thankful that our
• 6:45
  engineering leaders, starting with Satya, but everybody
• 6:49
  actually has this broad set of interests that I think is at the
• 6:54
  essence of what is required to make sure that AI serves people.
• 7:00
  Should governments be thinking about universal basic income as
• 7:04
  a result of of AI? I'm skeptical myself that that
• 7:09
  kind of solution will outperform the market.
• 7:13
  If it does, it means because we're going to see a market that
• 7:16
  collapses. I would much rather see the same
• 7:20
  market forces that have created prosperity for now 250 years, in
• 7:25
  an admittedly imperfect way, continue to be harnessed.
• 7:30
  What I think the future requires is real public private
• 7:33
  partnerships. Where is the market working?
• 7:36
  How can government help make it work better?
• 7:39
  How does regulation make sure that we avoid excesses?
• 7:42
  How does government fill in the gaps for people who do risk
• 7:46
  being left behind? But universal basic income is
• 7:50
  something that may fit, but probably in my view, less so
• 7:55
  than these other more targeted efforts.
• 7:57
  Our government's our government's taking the
• 8:00
  appropriate action right now when it comes to regulation of
• 8:03
  of artificial intelligence. Well, there's a debate going on.
• 8:06
  Has Europe gone too far? Is the US going to pull back too
• 8:09
  much? Where is the balance that needs
• 8:12
  to be struck? I do think there is a balance
• 8:15
  that will be needed and it will be struck over time.
• 8:19
  But in the meantime, if you're in the AI sector, the question
• 8:23
  or thought I would leave with everyone is this.
• 8:27
  You don't need a law to give you permission to do what you think
• 8:31
  is right, even if there's less regulation.
• 8:34
  I don't believe that we should all abandoned the field of
• 8:38
  responsible AI, ethical AI, because I believe that
• 8:42
  trustworthy IAI is what is fundamental.
• 8:46
  It's almost existential to the industry's long term success.
• 8:51
  You're not going to use AI if you don't trust it, and you're
• 8:54
  not going to be able to trust it if the companies that build it
• 8:57
  don't think hard about how to do this in a disciplined way.
• 9:01
  And thank goodness we and many companies are I you know, that
• 9:05
  is moving forward even while regulatory debate, political
• 9:09
  debate, has become somewhat more expansive and intense over the
• 9:13
  last year. Absolutely.
• 9:15
  Your new AI Diffusion report shows that AI hit 1 billion
• 9:19
  users faster than anything in in history in terms of technology.
• 9:23
  What comes next? Well, the first thing I think we
• 9:26
  should recognize is that AI. AI is accelerating, but gaps are
• 9:31
  opening up. We're approaching the three-year
• 9:33
  anniversary of the launch of Chachi PT.
• 9:36
  So three years into what I think we should regard as a many
• 9:40
  decades process of expanding the use of AI around the world,
• 9:44
  there are two things that are sobering. 1 is the growing gap
• 9:48
  between the Global North and the Global South.
• 9:50
  So you know, you, you see the EU, you see the US, you know,
• 9:54
  hovering around 30% of the population using AI.
• 9:58
  You see the UAE and Singapore, almost double that.
• 10:01
  But you see most of Africa in single digits.
• 10:05
  And there's an understandable reason.
• 10:07
  They lack electricity. They lack access to the Internet
• 10:11
  and connectivity. They lack access to digital
• 10:13
  skills. If we don't close those gaps,
• 10:17
  and this is going to exacerbate the divide between the global
• 10:20
  North and the global South. And there's one other divide
• 10:23
  that's opening up as well between urban communities and
• 10:26
  rural communities. We see that here in Portugal.
• 10:29
  We see it in my home state of Washington.
• 10:31
  We see it as a pretty universal element.
• 10:34
  AI is taking off as other accelerants of economic activity
• 10:39
  have over the last two decades in some areas, but not in
• 10:43
  others. Even within a state, even within
• 10:45
  a country. These are the two issues, issues
• 10:48
  that I think we need to start to talk about, think about and
• 10:51
  develop plans to address. Final question, you just came
• 10:55
  back from the UAE, Microsoft invested 15 billion.
• 10:58
  Why the UAE? Well, in some ways it's
• 11:00
  reflected in that data point. Right now, when we look at the
• 11:03
  AI leaderboard for countries around the world, the UAE is at
• 11:08
  the top. And of course, people start with
• 11:11
  things like access to cheap energy, not just say, natural
• 11:13
  gas, but some of the world's largest solar plants as well.
• 11:18
  I think the number one thing that has driven this growth in
• 11:22
  the UAE is broad adoption. It was the first country to have
• 11:25
  an AI minister and the first thing the AI minister did was
• 11:30
  saying let's skill the government and let's get the
• 11:32
  government using AI to provide better public.
• 11:35
  Services. It's a mandate, right?
• 11:36
  It it comes, you know, from the top, but it's really, I think
• 11:40
  almost now part of the culture of the country.
• 11:43
  When I was in Abu Dhabi last week, the two things that
• 11:47
  excited me the most were probably this one.
• 11:51
  The Department of Health has a command center using AI built on
• 11:54
  Microsoft Azure. If you have to call an ambulance
• 11:58
  in Abu Dhabi, you're going to get routed and your ambulance
• 12:02
  drivers are going to take you to the hospital that has the
• 12:05
  fastest, fastest drive, the best availability of doctors to help
• 12:10
  you as soon as you go in the available hospital beds for your
• 12:15
  category of illness. And then there's another service
• 12:18
  called Tam Tamm. It's Arabic for Consider it
• 12:22
  done. Every city in the world has
• 12:25
  potholes. Abu Dhabi is the one city where
• 12:28
  anybody who sees a pothole can take out their Cam app, use
• 12:32
  their phone, take a photograph, upload it on the app.
• 12:36
  The geolocation goes with it. It goes to the part of the city
• 12:39
  government that then is responsible for filling the
• 12:42
  pothole. When they do that, you will get
• 12:44
  a message back, but they don't close out the issue until you as
• 12:49
  the person who reported it says. Now it's fixed.
• 12:54
  That to me is the kind of benchmark that we should set for
• 12:59
  every government around the world.
• 13:01
  Let's use AI to save peoples lives, to fill potholes and
• 13:06
  address every other citizen service in between.
• 13:09
  That is why we don't go to Abu Dhabi to raise money, which is
• 13:13
  what other tech companies do. We go to the UAE to spend it
• 13:17
  because we're seeing it put to such good use.
• 13:19
  Incredible, you said. What excites you?
• 13:21
  What keeps you up at night? Well, I think to me, the first
• 13:24
  thing that does keep you up at night is this set of divides,
• 13:28
  the emerging AI divide, because I think that is something that
• 13:32
  risk creating a more fractured world.
• 13:34
  And then we always have to worry about the reality that AI, like
• 13:39
  every technology ever invented, will be used as both a tool and
• 13:43
  a weapon. So we do see AI being weaponized
• 13:46
  to, say, unleash more powerful cyberattacks.
• 13:50
  You know, we see a lot of questions emerging about what's
• 13:52
  the right balance, what's the right age for kids to start to
• 13:56
  use generative AI? How do we make sure that it
• 14:00
  reinforces and advances mental health and doesn't become a risk
• 14:04
  factor for it? I would say more broadly,
• 14:07
  there's a fundamental question for every one of us.
• 14:10
  How do we use AI to think more, not think less?
• 14:14
  It can do a lot of work for us fast.
• 14:16
  I use it that way every day, whether it's, you know, taking a
• 14:20
  look at what I'm writing, the copilot researcher is my
• 14:24
  favorite agent because I can go get a briefing memo in 10 or 15
• 14:29
  minutes that, you know, in the past I would have to ask people
• 14:32
  to go work on for a week. I don't stop thinking when I get
• 14:36
  it. I then ask people more
• 14:37
  questions. We should all use it as a tool.
• 14:40
  Think more, think less. Don't just rely on it to do your
• 14:45
  work for you, rely on it to help you do your work better.
• 14:49
  That to me, I think is going to be the key to for all of us in
• 14:54
  terms of personal growth and success, the ability to pursue
• 14:57
  better jobs, more successful careers in the years ahead.
• 15:01
  Brad, thank you so much for your time.
• 15:02
  Really appreciate it. Thank you.
• 15:03
  Thanks for listening to this episode of What's up 10 where I
• 15:06
  bring you inside the mines driving tech, money and power.
• 15:09
  Follow along on YouTube, Spotify, Apple or wherever you
• 15:12
  get your podcasts. You can also follow my personal
• 15:15
  account on Instagram at up 10. Would love to connect with you.