- il y a 4 semaines
From Ads to Armor
Catégorie
🤖
TechnologieTranscription
00:01I will have a moment of real pleasure, fun, because I will have a conversation with a very old friend.
00:13We know each other since more than 20 years.
00:17He is an incredible, incredible operator.
00:22And please welcome very warmly Nikesh Arora, Chairman and CEO of Palo Alto Networks.
00:32Nikesh, my friend, so good to have you.
00:50I think that the first time we met, it must be 2005 or 2006, something like this.
00:572005.
00:582005.
00:59I was learning how to sell ads from you and you told me this online stuff, it's okay.
01:07And we made an agreement, the first of the kind.
01:10That's right.
01:11In your office right by the Champs-Élysées.
01:15Yes, it's still there.
01:16We have no chance, which is not your case.
01:20You have done an incredible job with Google by transforming the ad business.
01:30And if Google has been able to capture such a large share of business, you have a very important responsibility.
01:44Fantastic.
01:46So, congratulations.
01:47It's really one of the most important building piece of Google Alphabet.
01:58And after one or two tries, you are…
02:05I used to work with Josiah.
02:07He was at Alibaba.
02:08I was at SoftBank.
02:09Yeah.
02:10He did better than me there.
02:12You were with SoftBank.
02:15Yes.
02:15And you have had a formidable, unique impact on shaping the very future of Palo Alto networks.
02:33So, I'm not sure that people know everything, but to start, I would say that your career at Google was
02:45nothing short of legendary.
02:47Honestly.
02:48Thank you.
02:49You have not only transformed and built the business, but you have fundamentally transformed the advertising landscape.
03:00Because if agencies today work as they do, and if clients are buying space as they do, you have your
03:09share of responsibility.
03:11So, what is interesting to understand is the catalyst.
03:17What is driving you?
03:19How you are inspired for the transformation?
03:25So, maybe you can take us a little bit in your journey and explain to us how you have been
03:32able to do such a fantastic job.
03:35Well, thank you, Maurice, and a big round of applause for the legendary Mr. Maurice Levy.
03:40Thank you.
03:44So, I think Maurice, I put it down to natural curiosity.
03:49I am forever curious.
03:52I was, and I work in reverse order.
03:54For the last few weeks, I am paranoid that I'm getting old.
04:01And I've been spending my time with, effectively kids, 25 to 27 years old, who are running these new AI
04:09companies.
04:10I went last night for dinner for two hours with the CEO of Perplexity.
04:14I was with the CEO of Codium two days before that.
04:17He's 26 years old.
04:19And he's explained to me how the whole world is going to fundamentally change.
04:22And I'm sitting there paranoid saying, yes, he's right.
04:24This world is going to fundamentally change.
04:26And so, if you track backwards, you know, seven years ago, I looked at cybersecurity and said, wow, everything that
04:32is connected will stay connected.
04:36And everything that is not connected will get connected.
04:38And if everything is connected, you know, and there are 10 billion artifacts connected to the Internet, in our industry
04:43now, in cybersecurity, we call it, it expands the attack surface.
04:47So, now you can attack things in so many different ways.
04:50So, I've always sort of gravitated towards growth industries.
04:54I don't know how to do non-growth industries.
04:56And that's the same case at Google.
04:57When I started, we are a $2 billion revenue company.
05:00And Europe was about $800 million.
05:03And I remember the ad market was $700 billion.
05:07I think Google revenues are close to $300 plus billion in advertising now.
05:11So, you can see that from $200 to $300 is not a bad progression.
05:17I think similarly, we're likely to see that in cyber.
05:22And clearly, AI is going to change everything that we know anyway.
05:25What is interesting is that you are moving from something which was already quite sophisticated, interesting, and even fascinating, which
05:38was Google, to a world which is much more technical, much more scientific, much more complex.
05:46When you speak about cybersecurity, a lot of people don't understand what you are talking about.
05:52I'm still trying to figure it out.
05:53So, you don't either?
05:54I kind of do.
05:56I kind of do.
05:56I kind of do.
05:57So, how have you been able to make such a jump from something you grew up into it, you understand
06:07it inside out, to a new territory which is highly sophisticated?
06:14So, for the first one year, I used to meet about 15 cybersecurity startups a week.
06:22And every morning on my way to work, I would call our chief product officer and ask him all the
06:26stupid questions.
06:27And every evening on the way back home, I would call our founder.
06:30So, I would bookend it for an hour every morning, hour every evening, discussing the evolution of all the different
06:36things.
06:36And eventually, I developed a perspective.
06:39And I think that's kind of helped us out, which is the same thing I'm trying to do with AI.
06:43Because I think AI is going to change everything we know and make it very, very different.
06:50We can expect that AI will transform society.
06:55Yes.
06:56And AI will transform all the jobs, the way we operate.
07:00AI will have also an impact in everything we do.
07:04Yes.
07:05So, today, Palo Alto Networks is solving a lot of issues.
07:14Can you tell us what is your vision, coming from where you come and knowing almost everything
07:24in our industry of the future of cybersecurity with AI inside it?
07:33So, let's talk about AI for a few minutes, and then we'll switch it to how cybersecurity works.
07:38I remember when I met you in 2005, when I was at Google, you know, you're still not very well
07:44known.
07:44And I think the biggest impact things like Google or search or technologies had is called democratization of information.
07:50Everybody has access to roughly the same amount of information or same kind of information.
07:55You don't have to be rich.
07:56You don't have to live in a capital city to get the news.
07:59You get the news as it happens on your phones and how you want it.
08:02And that's amazing.
08:03That allows us to be much better in many things.
08:06I interrupt you for one second because you are prompting, like in AI, something.
08:12I'm saying, when we made the strategic agreement between Google and Publicis, our executive committee
08:23was divided.
08:24Yes.
08:24And there were other people who were telling me, Mois, you are wrong.
08:29We should not go with Google.
08:31We should go with Yahoo.
08:33Yes.
08:35So you made the right decision.
08:37No.
08:38This means that at that time, it was bad.
08:42Yes.
08:43Yes, it was.
08:44It was bad.
08:45And this can be the case for almost everything in the digital world.
08:51I agree.
08:51Because Yahoo was there.
08:53Google was growing.
08:55But a lot of people, including Microsoft, you remember, they wanted to buy Yahoo at that
09:01time.
09:02We are believing that they will be the winner.
09:05So it's incredible what you have done out of what was a nascent company.
09:15Sorry for the interruption.
09:17And I think the movie is going to play itself out in AI.
09:19So if you think about, we talked about democratization of information.
09:22I think AI has the ability to democratize intelligence.
09:27Because we will all walk around with roughly the same amount of intelligence in our pocket.
09:32Which means we will know how to solve known problems.
09:36AI is fast moving to a place where every known problem will be figured out.
09:43And your mobile phone or your computer will be able to actually tell you how this problem
09:47was solved.
09:48Now the biggest fear or the biggest opportunity is when AI becomes super intelligence where
09:52it starts to solve unknown problems.
09:55It will find a cure for cancer.
09:57It will find a cure for many malaysies.
09:59It will find better solutions than human beings can.
10:03But imagine the world when you are democratizing intelligence.
10:08Everybody is equally intelligent as a starting point.
10:12And then the question becomes, what do you do with that base level of intelligence?
10:15That's going to change the creative industries.
10:16Because now we are equally as creative.
10:18Now we have to find that base knowledge to be uber creative.
10:22No pun intended.
10:23And similarly, you know, in cyber security we have the same challenge.
10:26We have to figure out how to not just be as good as the attacker, but how to be better
10:33than any attacker in the world.
10:35So you mean that AI will have to be superior to any AI of the attacker?
10:43We will have, yes, and we will, as human beings.
10:45You will have to think further than any of the possible attacker.
10:52That's right.
10:53Similarly, you know, in creative worlds, if every one of us can have a device which Google's
10:59new VO3 and, you know, new creative capabilities, every person with a computer will have the ability
11:05to design creative advertising perhaps equally well.
11:09Now the question is, what do I do with that base level of knowledge?
11:11How do I beat everybody around me?
11:13And what does the superior skill look like?
11:16So, building on that, finally, AI is a little bit like Janus.
11:23Is it like jazz?
11:24Janus.
11:25Janus is the ancient Roman goddess which had two faces.
11:34Yes.
11:35One good and one bad.
11:37Yes.
11:37So that is AI.
11:39Yes.
11:40Because with AI, there is a lot of things which are positive which can happen.
11:45Yes.
11:45And there are a lot of things which are either uncertain or frightening or worrying for the
11:54people.
11:54When it comes to cybersecurity, it is also both.
11:59No, we only have one objective.
12:01To make the bad face go away.
12:04Okay.
12:05We're staying on the good side, yes.
12:07Okay.
12:08But I mean, just expanding on what you just said, from an AI perspective, you're right.
12:11Because on one hand, if we train something how to find the cure for cancer, the same model
12:20can be used to create a bioweapon.
12:22So our job becomes how do we make sure that AI is not used to build the bioweapon instead
12:26is used to cure cancer.
12:28Or the same AI which we are using today, we call it the brain.
12:32You know, I don't know.
12:33I haven't seen a driverless car in Paris.
12:35Is there a driverless car in Paris?
12:36Yeah.
12:36There is.
12:37There is.
12:37Okay.
12:38So I call that giving AI agency.
12:41Giving AI arms and legs.
12:43When you start letting a brain decide how to drive, when to stop, when to turn, you're
12:47just given AI arms and legs.
12:49And in the next five years, you will see a lot of things will be given arms and legs
12:53around us, which will be managed by AI.
12:56And again, as a bad actor, the question is, can I take over that model and make it do things?
13:03Imagine you sit in your driverless car and it doesn't stop.
13:06It just keeps driving away with you.
13:07It's a lot easier.
13:08When you have an e-commerce company like our friend Joe Tsai, Alibaba, if they are making
13:20a mistake or they are missing a new innovation, they do some harm to the business.
13:32But at the next edition, the next version, they can catch up and it's okay.
13:45When it comes to cybersecurity, you have not that luxury.
13:50No.
13:51You have to be right all the time and you cannot be late on any innovation.
13:59So, tell me how you are organized and what kind of investment people, et cetera, you are
14:08investing in in order to stay current.
14:11So, Maurice, when I started seven years ago, we were about an $18 billion company and we
14:16used to be in one product swim lane.
14:18And I remember Larry Page told me that in tech, if your product is bad, you will die.
14:25Which is the example you gave of Yahoo or the Sun Microsystems example.
14:29It is quite prevalent in the tech industry.
14:33Our tech was good.
14:34But I looked around me and every year, 2,000 cybersecurity companies get funded in the world.
14:39I said, one of these guys is going to figure out something we don't know and will be extinct.
14:44So, in the last seven and a half years, we have bought 20 companies.
14:50We have doubled our R&D spend in terms of the number of people.
14:56If you are not smart enough to build it, we are humble enough to buy it.
15:00So, we are constantly trying to stay at the bleeding edge of innovation.
15:04We always buy number one and number two in the market because we believe they have the best technology
15:08and therefore we must have it.
15:10So, we have turned what was an $18 billion company now into a $130 billion company by spending
15:15about $5 billion in the last seven years and making sure, as I was seeing this morning,
15:21we have hit the 97th Fortune 100 company that uses our products.
15:26So, we have made sure that our product portfolio is amazing.
15:28I've also made sure that most of the large customers in the world have Palo Alto inside.
15:34With AI, one of the biggest threats is fake.
15:38Yes.
15:39Fake voices, fake face, fake messages.
15:44So, you are a good CFO.
15:50You receive a fake message from your CEO, his face, his message, his voice, and it says,
15:58okay, wire to my friend Nikesh, $20 million.
16:03That's a good voice.
16:04A real voice.
16:04Yeah, exactly.
16:05I thought that you will improve it.
16:07Please send it.
16:07And that's the reason.
16:08So, how do you deal with that?
16:12I think it's going to be fascinating to return to the office.
16:16We'll be back sitting across from each other to make sure we can do it in real time.
16:21But I think on a more serious note, I think that is the biggest risk with AI, with the ability
16:26to create this ability where we can have fake phone calls or fake videos and fake messages.
16:33There are a series of techniques being utilized right now around watermarking and making sure that when you create a
16:39tool that creates these fake videos, that the tool will actually advertise that I am a tool that can be
16:45used to make fake videos so you can be more aware.
16:47But just as is the case with every technology, and you call it Janus, which has a good side and
16:52a bad side, I think that is a bad side we'll have to protect against and make sure that we
16:56have enough safeguards in place to avoid that from happening.
16:59So, this means that one of the risks for society that we see tomorrow, a fake message from Zelensky, or
17:12a fake message from a scientist inviting people to use something, etc.
17:19Is Palo Alto Network a solution to de-risk the society, or is it something which is a lost cause
17:28and we have to educate the people and live with that kind of issues for the future?
17:36So, we have tools that we can make available. I think the opportunity there is going to belong to more
17:41like the Googles and the Facebooks and the Apples of the world,
17:44because they are the mass distribution players in the market.
17:48Exactly.
17:48They will have to produce capability or introduce capability with our tools.
17:52But you can sell them.
17:53We're trying.
17:54Yeah.
17:55Because, huh?
17:56We're trying.
17:56You are quite a good salesman.
17:57Do you know anybody?
17:58If I remember well, you are a pretty good salesman.
18:03Learning from the best.
18:05Yeah.
18:05So, you can sell.
18:07You can develop and sell products.
18:09Because that is one of the most frightening issue of our society.
18:16The development of those fake things which can really kill the trust in society.
18:27Yeah.
18:28And I think, as I said, we are going to have to work with the mass providers of social media
18:32and other distribution platforms to make sure we can do that.
18:36So, Nikesh, you have worked in Europe.
18:41You know Europe inside out.
18:43You have even, I remember, created an event in Sicilia.
18:49Yes.
18:51Google Cash.
18:52Thanks to your help.
18:52Which is still existing.
18:54Yes.
18:55And when you look at the old continent, unfortunately, till now, we have not been able to create one large
19:05company.
19:07In the field of internet.
19:09The largest we have in Europe is SAP and…
19:15You have ESML, you have SAP, you have Philips, but they are all very…
19:19Yeah.
19:19…somewhat old world perhaps.
19:21Yeah.
19:22And what would be the recommendation that you would put to President Macron or to Ursula von der Leyen,
19:34in order that we, if not catch up, at least we have some very strong capabilities and a kind of
19:47sovereign AI ecosystem in Europe.
19:52Otherwise, we'll be just subservient to the US and to China.
19:57Yeah.
19:58I think we had an opportunity with cloud, which was a missed opportunity because I know that a lot of
20:03work has gone in trying to create sovereign, you know, cloud data centers with partnership with many of the US
20:10companies.
20:10I think it's a fallacy to think about these large tech companies as US companies only because they actually operate
20:16in almost every part of the world and deliver services from every part of the world.
20:21I think…
20:21I know Jensen is going to be here anytime soon.
20:23I think we should encourage some version of sort of re-ignition of AI capability in every part of Europe
20:32in combination with people like Nvidia or people like Microsoft or people like Google and say, listen, make us a
20:38first-party citizen.
20:39Start your innovation here, make sure that European innovation is headquartered here.
20:44Because right now, I think the biggest moat is capital.
20:48If you think about Microsoft is going to spend $80 billion this year building AI data centers.
20:52I was in the UK this morning, they have announced a billion pounds.
20:57I don't think a billion pounds quite competes with $80 billion with one company.
21:01If you add the total amount of money being spent by tech companies, it's $300 billion this year in the
21:07next 12 months in building AI capability.
21:12More likely, more than likely that is going to be…
21:15If that is true, this means that the battle is lost.
21:19I don't think so.
21:20I don't think that battle is lost to the extent where if you can encourage those companies to divert some
21:25of that spending to building European AI data centers and European AI capability, the battle is not lost.
21:30Because that's not money that's been fully spent.
21:32I think part of that requires the right regulation, the right access to energy, the right sort of acceptance of
21:40risk taking.
21:41And that's not easy to do overnight.
21:44Risk taking.
21:45This is something which is not in the European dictionary.
21:48Maybe we should include it.
21:52Nikesh, I would like to thank you very warmly.
21:55We will have, in a few minutes, President of the Republic will be arriving and I will have to go
22:01to welcome him.
22:02And later on, I will be back with Jensen Hong and Arthur Mensch.
22:11We will discuss about the agreement that they concluded recently and they will be announcing the detail of this agreement.
22:19And then we will be joined by President Macron and we will have a further conversation with some interesting start
22:32-ups, French start-ups.
22:34Amazing.
22:35Which means that we still believe that start-ups can scale up.
22:40And we still believe that Europe can fight at least in a segment.
22:46I would like to thank you very, very warmly.
22:49Thank you for making the trip.
22:51I knew it was not easy.
22:53A rose of applause for my friend Nikesh.
22:57Thank you.
22:58Thank you so much.
22:59Thank you very much.
22:59Thank you.
23:00Thank you.
23:01Thank you very much.
23:01Thank you very much.
23:02Thank you very much.
23:03Thank you very much.
Commentaires