- il y a 2 jours
How AI and Quantum will Reshuffle the World Order
Catégorie
🤖
TechnologieTranscription
00:30Hi, guys. Thank you so much for joining us here today.
00:33Good afternoon.
00:34Hi, Ingrid.
00:35I want to make sure I see you guys. You've got your Viva Juice.
00:37Yeah, this is the Viva Tech official juice.
00:40Yeah, awesome.
00:40If people have not tried it yet, tremendous energy. Tremendous.
00:45And it matches the stage.
00:47Right. So we've got so many things we can talk about with you guys because you're both, well, you're basically
00:52legends in the world of tech.
00:55But I'm going to try to focus it in on quantum and AI being the topic of the day.
01:00So let's just ask for quiet.
01:02Shh. Shh.
01:05Just have some quiet, please. Thank you.
01:09In honor of our moderator.
01:11Thank you.
01:11Thank you.
01:11Thank you, Jack.
01:12Jack. Okay. So, Jack, you guys are doing a quantum computing startup. Obviously, quantum computers are not a thing yet.
01:22They are talked about in a theoretical sense, but we don't have a wide market for them.
01:29So, I mean, the first question has to be, what is the opportunity and what are you doing in a
01:35quantum computing startup if we don't have quantum computers?
01:41No, great question, Ingrid.
01:42And that's the reason why we actually chose not to build a quantum computing startup, but an AI and quantum
01:48technology startup.
01:49It is true that most of the attention in the quantum world has been placed on quantum computers, but in
01:56fact, there's a whole range of quantum technologies well beyond computing that are actually here now.
02:02An example, just to give one example, is quantum sensing.
02:05We have a division building quantum sensors and the AI that gets the signal out of them, and they're right
02:12now in hospitals, on airplanes doing navigation, in hospitals doing cardiac diagnostics.
02:17You don't need millions of qubits. You don't need error correction. They are here today.
02:22So, quantum sensing is an example, Ingrid, of a quantum tech that is here today, doesn't rely at all on
02:28quantum computers.
02:29Okay. Is your idea, though, that you're building it for a day when there will be quantum computers, or is
02:35it sort of disaggregated?
02:36Totally, yeah, totally decoupled from that. Quantum sensing is its own business model.
02:41We have customers right now for these quantum sensors, and that is going to be something that I think people
02:47here are more about.
02:47But the only reason why quantum sensing can finally come to light is because of the combination of these wonderful
02:54sensors made of diamonds, or cesium vapor, and AI.
02:58They're so sensitive. They're so sensitive that only machine learning can actually pull the signal out.
03:04So these GPUs that we have, that you hear about from NVIDIA, that Jim and I have talked about,
03:08they are critical to actually making the quantum sensors active in today's noisy environment.
03:15So it's a combination of A and Q, AI and quantum, that make that.
03:19Not dependent at all on quantum computers. Don't need quantum computers at all to make that happen.
03:24Do you think we're ever going to see a quantum computer, or is this going to be what quantum is?
03:28I think quantum computers are being built now by about, let's call it four dozen very serious organizations.
03:34Some big tech, some small startups, some academics. They're making steady progress.
03:39I think we'll see progress over the next five to ten years in that area.
03:43They're hurtling a number of engineering challenges.
03:46But what we find is that there's a lot of different quantum technology that we can look at well beyond
03:52the computers themselves.
03:54Okay. So you think there is a future for an actual commercial deployment the way that we have today?
04:01People buying servers, or buying space and servers, or buying computers for the home.
04:06Absolutely. Quantum computers.
04:07There will be a quantum computer in every, a chicken for every pot, as they say.
04:12I don't think it's going to be the PC revolution in terms of people buying it.
04:15My recommendation, my strong recommendation to people is, don't buy a quantum computer, but use it on the cloud.
04:22Amazon, Microsoft, Google, IBM, all of them now have quantum computers on the cloud.
04:28This is one of the first ever, in the history of the hardware revolutions in the last 50 years,
04:34the first ever cloud-native hardware innovation where you can access it on the cloud.
04:39I absolutely believe, and we believe, in a CPU, GPU, QPU future, where the quantum processing unit will sit alongside
04:49these other processors for hybridized computing.
04:52That's, I think, where we're going.
04:54Is it tomorrow? No.
04:55But that's, it's going to take a number of years to get there.
04:58But it's an exciting one because you can do more things in that hybridized world than we can do today.
05:02Okay, so Jim, we're going to, on that point, it's very interesting because you guys are obviously an investor in
05:09Jack's company.
05:11You have been doing actually a ton of investing in, I would say, more enterprise-y, business-focused areas.
05:19You, of course, initially made your name in the world of venture with consumer, spotting great opportunities.
05:28If you don't know who this guy is, he is huge.
05:31He is basically, he gave Facebook its start.
05:35He's a big deal.
05:36And you've been at the forefront and been able to sense out all of these great consumer startups.
05:43Are you no longer looking really at consumer so much?
05:46Is this really the opportunity, what Jack is doing now?
05:50One, thank you.
05:52I appreciate it.
05:54No, I think consumer is still extraordinarily exciting.
05:59And in 2016, a number of things happened.
06:05I was on the board of Harvard and our president asked me to be point person on the dean of
06:11the medicine search.
06:12I interviewed 30 exceptional candidates from Johns Hopkins and Harvard and Stanford, a number of the European leaders.
06:24Only one of the 30, when I asked what influence will AI have on medicine, George Daly, the current dean,
06:33was the only one who could articulate that.
06:35And I received blank stares.
06:37And mind you, this is 2016.
06:40So Stan Druckenmiller, one of the world's great investors, called me and said, you have to go over to Memorial
06:48Sloan Kettering.
06:49And they are phenomenal in many ways, but on empathy, perhaps not.
06:55And on technology, even worse than that.
06:59So I spent nine months negotiating a spin-out, Page AI.
07:05And I've done now over a dozen of spin-outs where start with great medical data, very defensible in terms
07:18of the precise data, whether it's prostate cancer.
07:22My goal is to eradicate breast and prostate cancer in 10 years.
07:28And I think the only way we're going to do that is with AI.
07:31So that's the model, but I'm still very optimistic about how AI will influence consumer.
07:40Again, it's sometimes the most interesting places are the ones that are out of favor.
07:48And when I threw myself into AI, the number of people who said, that's crazy.
07:56It'll never happen.
07:57The world's a little different today, but I'm five years plus into a 15-year plan.
08:05Right.
08:06And you think AI and quantum are part of that plan fundamentally?
08:09Or is there something else that you're not mentioning?
08:12No, I think then you start to drill down within AI.
08:16Okay.
08:17But an AI talented team.
08:23And Jack Hittery and I spoke for three years before successfully spinning out Sandbox.
08:29I have a number of those currently underway.
08:33They take time.
08:34But AI as a linchpin have to have great machine learning talent.
08:41And it has to be a really right brain, left brain type of culture.
08:48And that's what I think is wonderful and different about AI.
08:53Back in the day, I could back the four A-plus students out of the Stanford Computer Science
09:00Department or MIT or others, and we'd build a good company, and we'd be able to sell it
09:06for a lot of money or go public.
09:11Today, it's interdisciplinary.
09:13And we absolutely need the biologists to work with the machine learning.
09:20Linear algebra, as Jack has articulated, if there's one skill set out there that is essential
09:31in the next 10 years in AI and quantum, it's mathematics with a very specific understanding
09:38of linear algebra.
09:40That's my advice to students, postdocs, study more math, because that is at the heart of
09:49AI and quantum.
09:51That's super interesting.
09:52I mean, I was going to ask you about this later, but as you just mentioned spin outs, I kind
09:56of want to get in on this, because I'm assuming there are a lot of startups walking around Viva
10:00Tech now, and maybe people listening in on this who are trying to start businesses and so on.
10:04The thing about spin outs is that they used to all sit within the universities for a very
10:09long time, and it was really the institutions that would fund them and nurture them.
10:15These days, I go to MIT, I went to go visit the engine, which is all about this idea of
10:20taking these PhD projects and commercializing them, you know, and it's super, super focused
10:25on it.
10:27At the same time, I'm looking at the venture industry, and you guys are not really backing.
10:31marketing.
10:32Everybody wants kind of verifiable commercial opportunities.
10:37France has got an incredible, you know, as we're in France and Paris, incredible kind
10:43of academic landscape.
10:46But I do wonder, are we still in an area where investors like VCs have the patience that's
10:53needed to fund a lot of these companies, which are a lot of times still quite moonshotty?
10:59They're not that immediately commercializable.
11:01And we're going to go back to you with this question in a second, too.
11:04But do you guys have the patience?
11:06Do you think your peers in VC have the patience that you have to back these companies for 5,
11:1410, 15 years before they can potentially either release a product or definitely before they
11:20can make a profit?
11:22I believe in many cases the answer is yes.
11:26I think for me...
11:28I mean, you're saying this is a very, very, very successful man who has some money to spend
11:33doing this sort of thing, you know.
11:35Your fund is, you know, so, yeah.
11:38To your point, this has been a great week.
11:41Maurice Levy, unbelievable, VivaTech.
11:45It's been a great several days.
11:48And through Maurice, I met the president of École Polytechnique.
11:52And I, as you can all gather, spend a lot of time with universities.
11:57And when he took me through the students, graduate students, postdocs, qualifications,
12:05I was blown away.
12:06And so, whether it's École Polytechnique, whether it's Paris, France, other university
12:14centers, the Briar Capital thesis is there has to be a great university, there has to be
12:23great technical talent, and the university's silos need to be broken like crazy.
12:29And, again, I've been part of university leadership where the chemist, Nobel Prize winning chemist
12:39at the medical school had never met, literally, the Nobel to be prize winning chemist in the
12:50college.
12:51And so, that is part of the model.
12:55I try to bring this together, knock down some silos.
12:58Yeah.
12:58And I have a 10-year view.
13:00I'm not sure I would say 15 because there are some research projects that should remain
13:05in universities who have a 100-year horizon in some cases.
13:09But I constantly think through Spotify, Facebook, all these companies were 10-year journeys.
13:16Yeah.
13:17Right.
13:17So, it never happens as quickly as people think.
13:20Yeah.
13:21So, Jack, it's true.
13:22So, Jack, this is the thing, and I wanted to go back to you, and I want to make sure
13:26we
13:26get onto this.
13:27So, you guys spun out of Google.
13:29You spun out.
13:29You're not a normal spin-out.
13:31You spun out with $500 million in funding at a $4 billion valuation.
13:38That is not a normal Google moonshot spin-out.
13:42That's actually far from it.
13:43So, can you just tell me a little about what happened there?
13:46Why did you guys spin out?
13:47If it was such a valuable business, why didn't Google hold onto it?
13:50No.
13:51Great question.
13:51I just want to make sure also I just build on what Jim was saying on investing in deep
13:56tech because we are finding, for example, here in France, Xavier Lazarus, the founder
14:02of Alaya Capital, which is one of the leading deep tech early stage investors here.
14:08So, complementary to what a Briar Capital would do, who would put in probably more money,
14:12Alaya Capital is well positioned here in France to invest the earliest seed money in something
14:18that might be considered a moonshot, but they really helped the company.
14:22In fact, we at Sandbox just bought a company that Alaya Capital invested in.
14:28And that company is based here in Paris.
14:30What company is that?
14:31It's called CryptoSense.
14:32It's a spin-out itself, Ingrid, of Enria, of the research institution here in France.
14:38And the technology spun out a number of years ago.
14:42Alaya Capital and Amadeus Capital, out of England, invested in it, and then we acquired that company.
14:49So, there's not a lot, but I think there's a growing number of investors like Jim, like Alaya, like Amadeus,
14:56and now also Quantonation, a dedicated early stage fund based in Paris just for quantum companies, based right here.
15:06And so, I think there's a growing recognition that deep tech, Ingrid, is actually not as risky as people thought.
15:12Actually, it might be less risky because it gives you a deeper moat around the castle.
15:18It's harder to get commoditized if you're into deep tech than if you have a more consumer-oriented kind of
15:27thing,
15:27which could be successful, but also could be easily commoditized.
15:31So, I want to kind of note that I think there's a rising trend of the Briar Capitals in some
15:37of the larger check sizes,
15:38and then Alaya Capital and others in the smaller check sizes to really start building a deep tech foundation.
15:45And you mentioned the universities. I think that's a very key point.
15:50Universities starting incubators has a long history, but a spotted history, right?
15:55I think you'd agree that not exactly always successful.
15:58I think what they're doing now, what I see, for example, I just came back from 10 days in Israel
16:03with the Shiva Medical Center hosting 35 not only startups, but also big companies coming in to help the startups.
16:11So, I think that after many years of spotty success and failure, the universities are starting to understand
16:19how to actually get this done in partnership with capitalists and with others as well.
16:25So, I think things are moving now towards a deeper trend of deep tech.
16:29Yeah, I'm seeing that a lot at Oxford and Cambridge, but let's go back to what I said.
16:32Oh, yeah. How did you spin out?
16:34So, why did you spin out if it was so valuable? Why did Google let you go?
16:37I think what's critical to any division of any organization, what's unique about Alphabet is that
16:43Alphabet was born with this concept of creating entities and very often they would spin out.
16:51So, Waymo was the first spin out for self-driving cars. Verily is another spin out.
16:55And what we realized for our particular division, Sandbox AQ, is that it really needed to be neutral
17:02in terms of clouds out there. We really couldn't be centered just on one cloud.
17:07Right.
17:08As much as we had a deep relationship with Google, and we still do, we needed to be on Amazon,
17:12AWS.
17:13We needed to be on Microsoft as well.
17:15So, you asked to go.
17:16So, it was a mutual conversation around, hey, we need to spin out now. The team, you know, we, the
17:23team,
17:24put that forward. And we then said, what's the structure in which that would make most sense?
17:29And we really needed that independence. And to Alphabet's credit, I want to give massive credit
17:34to the board of Alphabet, to Sundar, to Ruth, that they really saw the big picture here.
17:39That for this technology that is impacting life sciences, we're impacting drug discovery
17:45with our simulation software, we're impacting cybersecurity with our encryption software,
17:51we're impacting diagnostics of hearts with our diagnostic quantum sensor.
17:55This is world-altering technology. We can't just keep it on one platform. We have to get it out
18:02in a platform-neutral way on all the clouds. And I really want to give kudos and, you know,
18:07and thanks to Alphabet for looking at that big picture. Board members and the C-suite really backed
18:13us in that way.
18:13So, you are, your long-term vision, since Quantum is all about the long-term vision,
18:18is that you guys would remain independent. Like, if Amazon came along and said, or Microsoft,
18:23which seems to be the company making all the deals in AI right now, if they came along and said,
18:28we want to buy you guys for $8 billion, would you go for it?
18:33We have so much growth ahead of us at Sandbox AQ. I mean, really, we're just beginning. And the fact
18:39that we're able to raise a large round, as you mentioned, why did we raise such a large round?
18:43It was really because of the opportunity ahead of us is something that we need to scale
18:48now. We need to be global. We're in 12 countries now, including a great presence here in France.
18:55A third of our people are here in continental Europe. Why would a startup starting in Palo Alto,
19:00suddenly within one year have people all over the world, including here in Europe? Because the
19:05opportunity is global and because the customers are global. If we want to help a pharmaceutical company
19:11that has branches all over the world, take their molecules and make them into medicines,
19:16we have to be in France. We have to be in the UK. We have to be in the US.
19:21We have to work with them
19:22in all their R&D centers. They're global, so we have to be global. So this scale, and I think
19:28that
19:28might have surprised some people, Ingrid, when we came out with such a large round, but I think we're
19:32going to see more of that, not less of that. Does that mean you're raising money?
19:35No, we're done. We're done. We finished raising in December 22, but I'm saying we'll see more of
19:40that from other companies. Okay. That is, we'll see other companies. Have you got enough runway
19:44now for the long term? We have runway for years to come, thank God. So no one has to test
19:48that 4
19:49billion valuation, I guess, right? Well, we did test it because we bought several companies now and
19:53investing companies using that stuff. So we have tested that and people have accepted it. Okay. I'm so
20:00sorry. There are so many things I wanted to get to, which we have not gotten to. It is the
20:06first
20:06time in Viva Tech history where there are three of us on the afternoon panels or sessions, including
20:13this fellow Elon Musk, the genius from Austin, Texas. So go Austin. Go Austin. And I think people can
20:23sit here and watch the live stream. I believe that's the case. Yeah, I think you can stay here. People
20:27can
20:27stay here and watch the live stream. No mad dash. No mad dash necessary. Yeah. You can stay here.
20:32There's going to be some amazing things on this stage too. I think you've got. Thank you all.
20:36Yeah. Thank you so much. Bye everybody.
Commentaires