00:02Bloomberg audio Studios podcasts radio
00:06news open AI top of Mind across the
00:09world of technology that includes for
00:11Elon Musk who posted on X overnight that
00:14his EV company Tesla is having to boost
00:17compensation to ward off approaches for
00:19its AI Talent those approaches claims
00:21musk include open AI Ethan Knight a
00:24member of Tesla's team working on
00:26computer vision for FSD was approached
00:28by open AI so musk moved him to his own
00:31AI company X AI let's get to the open AI
00:35side of the story let's talk about
00:37Talent let's talk about what's happening
00:39in the world of artificial intelligence
00:40Brad light cap open AI coo joins us here
00:44in New York City there's no accusation
00:47that anything wrong was done but it
00:49highlights an interesting point right
00:51now there is a lot being done on the
00:53training of it prior generation models
00:55future generation models there is a
00:57finite amount of talent give the open AI
01:00side of this equation how much of a
01:01premium are you having to pay out I
01:03don't know if it's salary or or stock or
01:06or uh engineering freedom to get the top
01:09names well thank you Ed and Caroline for
01:12having me uh it's great to be with you
01:14you know look this is a competitive
01:15market for sure um it's an amazing Place
01:18uh that we're at because there's very uh
01:21few number of people that can really
01:23make such a dramatic impact in this
01:24field but we try and pull people uh into
01:28our mission and I think that's
01:29historically what's been our strength is
01:31um people want to build in the place
01:33where uh we're pushing the frontier the
01:35furthest um and also where we really
01:37think about the impact of the technology
01:39both uh on a societal level and then
01:41also on an applied an industry level how
01:43big is open AI now how many employees
01:45and where are they we're about 1,200
01:48people 1200 yes mostly in San Francisco
01:51uh but also in uh in London and Dublin
01:54um and increasingly other places in the
01:55world increasingly other places in the
01:57world the reporting was that Japan is
01:59the next Frontier for you what's the
02:02strategy there and where you choose I
02:04don't know whether you call it opening a
02:06new market but where you choose to have
02:07operations well we're very fortunate we
02:09have a very uh Global base of demand so
02:13we want to show up where our customers
02:14are we think this is going to be a
02:16global transformation uh we feel a lot
02:18of pull from uh from places like Japan
02:20and Asia broadly uh and so we'll we'll
02:23show up where we feel like we need to be
02:25and let's talk about way show up from an
02:26industry perspective you're here in New
02:28York talking to Partners we understand
02:30Sam and the teams have been in Hollywood
02:32discussing how Sora can be used but
02:34ultimately how you can have a working
02:37relationship with companies and
02:38industries that think they're going to
02:39be upended talk to us about how those
02:41conversations going for example with
02:42Hollywood and Publications yeah you know
02:45we we take a very unique perspective
02:46here we really model openai as a
02:49Partnerships company and we think this
02:52transformation is going to be very
02:53broad-based and so we're one company we
02:55can't do everything ourselves being able
02:57to find ways to partner with indust
02:59industry uh with countries um with with
03:03specific companies uh that's really
03:05important to our mission uh and the way
03:07that we think about bringing technology
03:08to the world um by way of recent example
03:11we work really closely actually with the
03:12publishing industry right now on ways
03:14that we can deploy AI into publishing to
03:16make the quality of Journalism better uh
03:18to make the quality of reader
03:20experiences better uh we've heard
03:21nothing but uh really uh actually a lot
03:24of enthusiasm from that industry and so
03:26we're always looking at ways that we can
03:27partner with industry uh I expect a lot
03:29more from us in the future on this front
03:31uh and and sort is a great example of
03:33where uh we really are excited about the
03:35impact that we can make in media it's
03:36interesting that media companies over in
03:39Europe for example seem to have got with
03:40that quite quickly actual spring are a
03:41key deal that you sort of birthed early
03:44days and actual Springer therefore seems
03:46for Business Insider perhaps a little
03:48bit more Blue Links within chat GPT New
03:50York Times ain't on that discussion
03:53point right now how is that legal fight
03:56and indeed behind the scenes discussion
03:58going with those sorts of companies
04:00well I can't comment on the New York
04:02Times case specifically uh other than to
04:04say we we think it's without Merit but
04:07um the overwhelming majority of what we
04:09hear from the publishing industry is
04:11actually quite positive uh there's a
04:13high level of excitement everywhere
04:14about what these tools can do both to
04:16unlock better journalism to give readers
04:19more access to information um to enhance
04:21the way that people engage with content
04:23so uh the way that we actually think
04:25about even the fundamental business of
04:27of information engagement and
04:28consumption um these things are all
04:30really possible now with language models
04:32and and increasingly with multimodal
04:33models too so um we think we're just
04:36starting to scratch the surface on that
04:37opportunity we've got a lot more to
04:38share there soon uh but uh stay tuned
04:41you're being positive being optimistic
04:43of course you are what's interesting is
04:45I come away from meetings where
04:47Hollywood is deeply anxious people in my
04:50industry are deeply anxious goodness
04:52everyone with a job is trying to work
04:53out how they adopt Ai and also there's
04:55this question that is ongoing of open
04:57versus closed your going for another
05:00legal fight at the moment Elon Musk has
05:02dragged you into it and wants you to
05:03rename yourself where do you sit on the
05:06closed versus open how much more open do
05:08you have to become yeah well look I
05:11think we really focus just on where we
05:14can deploy the technology and what's the
05:15best way to do that and at the end of
05:17the day we're going to focus on where
05:19our customers are and the way that they
05:20want to deploy it uh we work with
05:22thousands of Enterprises across the
05:24world um we spend a lot of time working
05:26with those Enterprises on ways that we
05:28can configure the technology spef
05:29specific to their use case specific to
05:31their stack uh and we think that the
05:33easiest way to do that is the way that
05:34we serve it um but that's not to say
05:36that you know we wouldn't look at other
05:37ways in the future uh we really
05:39ultimately want to be where we think uh
05:41the best place to deploy the technolog
05:42is going to be for Bloomberg television
05:44and radio audience worldwide we're in
05:46New York City speaking to open AI coo
05:49Brad light cap I I want to learn more
05:51about where open AI sits uh in its
05:54growth its day-to-day operations you
05:57know you just said thousands of
06:00we we came to know you because of the
06:02the explosive response to chat GPT and
06:05that was principally in the context of a
06:06consumer facing chatbot but it sounds
06:10very much like growth now is focused on
06:13that Enterprise Arrangement licensing of
06:15Technology can you give me proportions
06:18the rate of growth for those two pillars
06:21well we we've seen tremendous growth in
06:22the Enterprise you know 24 is going to
06:24be the year of adoption for AI in the
06:26enterprise we felt really like 23 was
06:29the year that people just started to
06:30wrap their head around what was possible
06:32with AI in the Enterprise but
06:33increasingly the market is pulling us
06:35toward real applications that are
06:37delivering real business results uh and
06:39a very broad focus on AI enablement so
06:43uh how do we think about bringing AI
06:44both to our operations but also to our
06:47workforces uh and then also to new
06:48product experiences that we can deliver
06:50to customers um and we work with plenty
06:52of companies that are doing all three of
06:53those things at the same time we think
06:55that's very possible uh and it's having
06:57really dramatic bottomline impact today
06:59so this is what we think uh the PLL is
07:02going to be and and and we're ready to
07:03support Brad you're the COO so I'm
07:06assuming you're responsible for also
07:08some of the pur strings involved in in
07:10open AI finances compute must be a
07:13really considerable cost right now uh
07:15and with that the energy consideration
07:20you making enough revenue from this
07:23Enterprise business to to keep the
07:25lights on when you factor in the compute
07:30we are yes uh we we think that we have
07:33uh well first of all a very Diversified
07:35business and so we serve hundreds of
07:37millions of users uh with our chpt
07:39product um and many of those users are
07:41paying users of course we serve uh now
07:44over 600,000 uh individual users in the
07:46Enterprise with chat Enterprise um and
07:49so we're just seeing tremendous momentum
07:51uh we're very early Innings on this and
07:53so we look at it as hey this is this is
07:55the start of something um and ultimately
07:57our priority is just to be able to
07:58deliver the best intelligence uh whether
08:01it's in the Enterprise or to individuals
08:03uh that's the most useful that drives
08:04the biggest biggest business impact car
08:06I'm thinking back to the conversation we
08:07had uh the other day with um Daniel
08:10amade from anthropic and what she
08:12explained Brad was the relationship with
08:14AWS and and Bedrock they're getting a
08:17lot of business through that you have a
08:19quite similar and and actually a
08:21pre-existing situation with Microsoft
08:23and Azure are you able to explain how
08:26much new business comes through that
08:29cond it of the Microsoft relationship
08:31and and what were probably existing as
08:33your Enterprise customers we're really
08:36fortunate to have Microsoft as a partner
08:39um they've been an amazing partner with
08:41us for a long time we've increasingly
08:43looked at the world as uh ways that we
08:46can bring open AI technology to life
08:48together uh we're very uh fortunate that
08:51they're they're offering our models
08:52through their Azure open AI service but
08:54um we have an independent business so we
08:55serve uh we serve customers
08:57independently of Microsoft um and and is
09:00that one bigger than than the Microsoft
09:02Avenue yeah look we you know we we we're
09:04it's it's it's fluid between the two but
09:06um we really are uh focused on what's
09:08best for for for the companies we work
09:10with and um we're very fortunate to be
09:12able to diversify the surfaces in which
09:13we offer our products we think the
09:15demand for this technology is massive
09:17and so uh being the a smaller company
09:19obviously um having a partner that can
09:21help bring the technology to scale uh is
09:23really really advantageous for us um and
09:25we're excited to support them in the
09:27future it's an extraordinary very novel
09:30kind of partnership and in fact the
09:32whole of AI is birthing these kind of
09:37relationships I just think of also Aqua
09:40hires deals that seem to be going like
09:42Microsoft has just seemingly Aqua hired
09:44the whole of inflection what did you
09:46make of that particular move
09:48particularly when suddenly You' got what
09:50the co-founder of deep mind coming in to
09:51be the head of consumer AI at Microsoft
09:53and ultimately you put out a response to
09:55that being like congratulations but how
09:57did that make you feel as the previous
10:00partner of choice for Microsoft well
10:02we're excited for them uh we
10:03congratulate Mustafa and Karen and the
10:05whole team and uh we work very closely
10:07with Microsoft and so uh we think that's
10:09going to be a major advantage for our
10:11partnership as well um we're focused on
10:13what we can you know what we can control
10:14in the open AI side um which for us
10:16right now is really thinking about the
10:18next wave of the technology uh and
10:20thinking about ways that that wave is
10:22going to start to enable businesses um
10:24and the more we can think about that
10:25problem together the better we'll all
10:29CEOs I speak to some of them are your
10:32customers say they're technology
10:34agnostic right now in other words they
10:36might use GPT or another Foundation
10:38model they might also be building their
10:40own in parallel and when we told our
10:42audience who were coming on the show a
10:43lot of people wanted to get your
10:45thoughts on the paler relationship it's
10:47an interesting case study they they seem
10:50to be one of the beneficiaries of the
10:53Enterprise investment in AI or at least
10:55the data analytics side what's it like
10:57working with Alex carp yeah well we you
11:00know we we actually work with a bunch of
11:01partners that help us bring the
11:02technology to life and there's so many
11:05ways to deploy the technology in the
11:06Enterprise that's what's kind of amazing
11:07about this technology is it's not just
11:10one function in the Enterprise or one
11:12part of the Enterprise stack uh it's so
11:14diverse uh and it's so broad-based and
11:16so um we work with with with with many
11:19partners to to bring that to life and um
11:21we'll continue to look for partners to
11:23to work with to to help us do that um
11:25but uh yeah no we you know look we're um
11:29uh we're excited with with with way
11:30paler has has been enabling their
11:32customers on top of our technology uh
11:35and we'll continue to support them going
11:36forward Brad uh I every day it crosses
11:39my desk you know that I've done quite a
11:41lot of reporting on Sam's chip Ambitions
11:43as well there's some relationship going
11:45on between the UAE and open AI
11:48fundraising or technology
11:50Partnerships can you clarify what's
11:53going on there and why you'd need to
11:54raise funds from somewhere like that
11:57yeah well look I I won't comment
11:58specifically on that we uh broadly
12:00though believe that there's just not
12:02enough total Supply uh in the world and
12:05that of AI accelerators specifically AI
12:07accelerators and broadly that the supply
12:08chain will need to adapt to what we
12:10think is going to be uh this highly
12:12inflected and and nearly exponential
12:14demand in the next 10 years uh and so we
12:16spend a lot of time thinking about ways
12:18that uh we can make sure that that
12:19demand is met part of it is just being
12:21able to bring our models to Bear but up
12:23dream of that is obviously making sure
12:24that the actual underlying hardware and
12:26infrastructure exists to be able to
12:28build the systems that we need to
12:29ultimately serve that demand and so um
12:32you know we we we think a lot about it
12:34and um you know our hope there is uh
12:36that the the supply chain starts to
12:37think about that the same way we do Brad
12:40your your Mantra your Manifesto your
12:42focus is to build artificial general
12:45intelligence for the good of humanity
12:48and I speak to a lot of humans who are
12:51terrified can you encapsulate how you
12:54can without regulation being there in
12:57place yet and the eu's sort of put put
12:59the gaet out there you're doing all of
13:02this without disrupting too much too
13:07hard yeah well look Central to our
13:09mission is to make sure that we bring
13:12the technology in a way that's safe and
13:14that creates broad benefit it's really
13:16encoded into how open AI was structured
13:18uh and it is actually the thing that we
13:20spend uh our time thinking about how to
13:22deliver in real world results today um
13:25and so we're very optimistic um and I
13:27can understand where uh the fear comes
13:30from uh it things are moving very fast
13:32even for us at the center of it uh it
13:34sometimes is dizzying but um look our
13:37perspective is that this is an enabling
13:39technology at its core it's going to
13:41make businesses more efficient it can
13:42make people more productive it can be
13:44dramatically inflecting in how people
13:46learn how people can understand
13:47information um and so we think that that
13:50has a very broadly positive effect on
13:52the world um and the world that we wake
13:53up to 10 years from now uh will be one
13:55that we look back on uh you know and
13:57really are are uh are excited to be able
13:59to go build uh today and um you know
14:02we're that's that's that's really what
14:04we wake up every day and want to go do
14:06so we want to thank you being in the
14:08house with us the CEO of open AI Brad
14:11light cap there joining us in New York