- 1 hour ago
Category
🤖
TechTranscript
00:16Wow, what a big room.
00:20Bonjour.
00:23Are you enjoying today?
00:25Yeah?
00:28A bit slippy.
00:34Okay, now you will have a great, great, great session.
00:39Joe Tsai from Alibaba with Christophe Jakubizine from Les Echos.
00:46As you know, Vivatech is a worldwide platform and we are pleased to welcome more than 170
00:53countries.
00:54And when you walk through the alleys, you will see a lot of pavilions from countries and a
01:02lot of nationalities.
01:03We hear a lot about our friends from the US, from Europe, of course, but too often we tend
01:12to forget the other half of the world and that's why our next guest is here.
01:22He co-founded Alibaba.
01:26Alibaba.
01:27People do not know very well his name, but they will remember forever because he has helped
01:36transforming Alibaba.
01:39Alibaba has created Quen and I'm sure that you don't know, but Quen is the most widely
01:49used open source model in the world.
01:53The most widely open source AI model.
02:01It's not ChatGPT.
02:02It's not Claude.
02:05It is Quen.
02:08And Alibaba released its latest version of Quen a few months ago and we are pleased to welcome
02:16someone who has been at the helm of the company since its inception and he was the man behind
02:26the open source of AI.
02:28So please welcome very warmly Christophe Jakubizil from Les Echos and a big round of applause for
02:38Joe Tsai, chairman and co-founder of Alibaba.
02:53Joe, welcome.
03:03Hello everyone.
03:04Thank you to be here.
03:05We are very, very pleased to, as Maurice said, to welcome Joe Tsai, the co-founder and chairman
03:11of Alibaba.
03:13Alibaba went through many revolutions and of course we'll speak about the last one, about
03:18the open source LLM model.
03:20But first, can we talk about your history?
03:23Because many people in France still think you are a marketplace, B2B, B2C, we know AI
03:30Express here.
03:31I know you are the largest marketplace in China.
03:35So could you tell us more about the evolution of your company, of your group?
03:41Thanks, Christophe.
03:42Maurice, it's great to be here.
03:44So Alibaba started in 1999 as a B2B marketplace and at the time the business idea was very simple.
03:52We put these small Chinese manufacturers and trading companies online to sell to the whole
03:59world wholesale.
04:00So that kind of was on the cusp of China entering the WTO.
04:05So the export business was about to boom.
04:07And then we got into the consumer B2C business and launched a website called Taobao and which
04:16today is the largest consumer e-commerce business in China.
04:20How many customers?
04:20We serve 820 million Chinese consumers.
04:25And by the way, this platform helps European companies and brands to sell 30 billion euros
04:33to Chinese consumers every year.
04:35So you would say that Alibaba is quite important for European companies doing business around the world.
04:41But we went big into AI and cloud.
04:49And this is actually quite a long history.
04:53We've been investing in cloud technology since 17 years ago.
04:57And that's because of necessity.
05:01We were seeing that we were managing so much data from our e-commerce business and transactions
05:07on a daily basis that if we continue to be dependent on other people's technology like database
05:14and storage technology from others, we would end up giving all of our profits to the technology providers.
05:23And hence, there was an effort to develop our own proprietary technology to manage all the data.
05:29All right.
05:30And that's when our cloud business was born.
05:33Basically, we ate our own dog food before we launched the business to offer that service to third-party customers.
05:42But I think that's sort of from the micro perspective.
05:45From a macro perspective, we're all in on AI.
05:49Actually, the logic is very simple.
05:52AI is today, if you ask me, how big is the market?
05:57What is the demand?
05:58What is the TAM, total addressable market?
06:00I would tell you that it's much, much bigger than anybody's IT budgets.
06:07It's much, much bigger than the software market.
06:09Because what is AI?
06:11AI is producing units of human intelligence and human productivity.
06:16And if you look at global GDP, over $100 trillion of GDP, at least half of that,
06:21$50 trillion, $50 trillion is about human productivity and human intelligence.
06:27And that is the TAM of AI.
06:29And that's why we're all in on AI.
06:32So you think AI will really make us have a higher productivity?
06:36Because we just had, like, one hour ago, a round table where we were questioning that.
06:42People are investing a lot, but they don't see the results yet.
06:47They don't see the results yet.
06:49I think a lot of corporate CEOs will tell you that their engineers are burning a lot of tokens.
06:54Cost goes up.
06:56But I think I would say we're at the cusp of real productivity gains.
07:01Right now, a lot of people are experimenting.
07:03And there are definitely, like we see it in our company, our own engineers, there are super users that use
07:12the coding tool, for example.
07:13And they are not just doing their work that's within their domain, but they're also experimenting with other stuff.
07:20Because when you give engineers toys, they will always play with it more than, and they didn't realize that the
07:28company is actually paying for it.
07:30So this is what's happening right now, but I think what I truly believe, this is really a belief of
07:35whether, you know, production of artificial units of intelligence will be able to add value to human intelligence.
07:45And it's like religion.
07:47I don't want to convince all of you that that's going to happen, but we believe that that's going to
07:54happen.
07:54So if we go back to the Alibaba abolition, in which layer of AI do you invest most? Is it
08:01infrastructure, models, cloud, services?
08:04So we are in at least four layers of AI.
08:09At the bottom, we're not involved in energy, because in the China context, energy is actually very efficient, less expensive.
08:21And the Chinese government has invested over the last 15 years in the national grid that makes energy delivery, production
08:29and delivery very, very efficient to all the users.
08:32But we are at the front and bottom, we're involved in chips, infrastructure in terms of our cloud business, and
08:39also the model.
08:39We have, as Maurice very, you know, kindly pointed out that our Quint model is actually one of the most
08:49popular open source models in the world.
08:51And then at the application layer, we have an e-commerce ecosystem.
08:55We have an ecosystem for online shopping, grocery delivery, travel, maps, and all of that, where we can infuse and
09:06deploy AI for our users.
09:08So we're involved in all those stacks of technology.
09:12And we think that there's huge benefits to the full stack tech approach, as opposed to just taking one layer
09:21of the stack.
09:23Because I think over time, nobody can tell you right now where the value will accrue.
09:29Which part of the stack? Is it in chips? Is it in cloud? Or is it in the model?
09:33Right now, the model companies, the pure model companies are very hot.
09:36They seem to accrue a lot of the value.
09:40But over time, that may not be the case.
09:43So having an integrated approach of being involved in the full stack makes a lot of sense to us.
09:50And that's our core strategy.
09:51That's very interesting. You're right.
09:53We don't really know where the value will be.
09:54But when you see the investment in infrastructure for AI, do you think there is a kind of bubble there?
10:01Do we have all those needs to make the model work?
10:04Because some models are more efficient than others and they need less capacity than others?
10:09So what about that?
10:11I don't think so.
10:13I mean, the numbers really are quite astounding because if you just look at the American hyperscalers, the four or
10:21five companies combined will invest over $800 billion of CapEx.
10:26And next year, it's going to go to over a trillion dollars.
10:30It's eye-popping type of CapEx investment.
10:34And I think it's natural that people will ask whether there's going to be over capacity.
10:38But again, like I said, we're going against, we're trying to tackle a total addressable market of 50 trillion US
10:49dollars.
10:50And that's why we're optimistic.
10:53And in the China context, we're actually very under invested in infrastructure and in the supply chain of AI.
11:02And so it actually behooves all the companies in China to step up their investment.
11:09Obviously, we're not investing at the same level as some of the American hyperscalers, but it is still very, very
11:15substantial.
11:16Why don't you?
11:18Oh, well, you know, sometimes you're limited by your capital.
11:23You're limited by the free cash flow that you're generating.
11:26And the lucky thing is Alibaba is one of the very few companies that actually has a core business, which
11:32is e-commerce.
11:33We generate $25 billion of free cash flow every year from our e-commerce business that can fuel our investments
11:40in AI.
11:40So we're actually one of the better positioned companies out there.
11:45I think today your business in marketplace is still 80, 85% of your turnover?
11:52Yeah, it is still 80 plus percent of our revenues in the e-commerce marketplace.
11:58We're very lucky that that produces the cash flow that allows us to make future investments.
12:04To invest in AI and cloud business.
12:05Yes, absolutely.
12:06What about your, Maurice mentioned, the Quen offer and the Quen model?
12:11Because it's open source now, so what kind of customer you can have and you can help with that?
12:18Well, I thought Maurice was too charitable in crediting me to, as the open, behind the open source movement.
12:27That's totally untrue.
12:28I would say that our colleagues at Alibaba have worked last several years in pushing the open source movement in
12:37these frontier models and to a high degree of success.
12:42And we're continuing to do that.
12:46What's the significance of open source?
12:48So I've been spending the last couple of weeks in Europe.
12:51And as I talk to European companies and CEOs and scientists, people in Europe, one of the biggest key words
13:00here is sovereignty.
13:02But what is sovereignty?
13:04If you ask 10 Europeans about what sovereignty means, you'll get 12 different answers.
13:13But to me, they mean basically two things.
13:16One is technology independence.
13:19They are all worried about the kill switch, that being reliant on some other country's technology, they could shut off
13:27the kill switch.
13:27We just saw a very live example of that in the recent days.
13:32With Fable, right?
13:34And the second is data privacy.
13:37People wanted to be able to use AI, use technology with data that's proprietary to them within their own environment
13:45and hence build up the firewall to protect their own data.
13:50I think open source solves both problems because it is basically a free piece of software that you can download
13:57into your data center.
13:59You can download the Quint model in your own notebook computer, for example.
14:04And that's completely independent from the original maker.
14:10So if you use our Quint open source model, it will have nothing to do with Alibaba.
14:15I mean, we have to figure out how we charge for it, but we can't.
14:19So that's number one.
14:20That's independence.
14:22What's important is you can take the open source model and then you could take your own data and train
14:27it, further train it, fine tune it, do post training of the model, and keep your entire data private within
14:37your own firewalls.
14:38And I think that's a really, really important point for companies in Europe to recognize that open source is actually
14:46one, I'm not saying it's the panacea, it's not the only approach, it's one of the approaches to achieve some
14:52degree of sovereignty.
14:53And today, the open source movement is actually driven by Chinese companies.
15:00All the American players have closed source their model, so they want you to use their models through an API,
15:08and you have no idea where your data is going.
15:10When you're chatting with a chat bot, all of your most private questions and your confessions go into some pool
15:23where they use that to further train their model.
15:25So you have no idea where that's going.
15:28To be honest, we, you know, Europe, sovereignty is a very great concern today, and I think we just realized
15:34how dependent we can be from the US or from China, from the industry, for instance.
15:38So how, I mean, I agree with the open market, I mean, the open model, etc. But, but still, we
15:47can be worried, there is, you know, bypass for the Chinese company to maybe to one day, if they want
15:52to cut the access or to, to be able to stop the model to function.
15:56How can, how can we trust China as we, because we were disappointed by the US in the last days,
16:02but, but we could be also by China. I mean, it's, it's a big risk for Europe.
16:05Yeah, you can't. That's the short answer. Because you cannot be reliant on a third party government to say that
16:17they won't do something that's detrimental to you.
16:19But here's the thing. Right now, all of your eggs are in one basket. Why not get a second basket
16:26to put your eggs in two baskets, even though Europe in the long term may develop their own basket.
16:33But at least now you have two eggs, sorry, two baskets to put your eggs in.
16:38Yeah, it's true.
16:48Maybe we can, you, you can give example of how you work with the German company like BMW, you, you,
16:53you, you have them and or Siemens in Germany.
16:55What, what was the base of the collaboration and cooperation you did with them?
16:59Yeah. So these German manufacturing businesses are absolutely fascinating.
17:04They're all clients of ours, Alibaba Cloud in China, in their Chinese operations.
17:10Uh, and we work with them in, in the manufacturing contest, uh, in terms of design, uh, testing, quality control.
17:20And, uh, uh, we think that, uh, this is in the future is going to be a very interesting segment
17:28because today, most of the application in AI, you know, you have chat GPT that's consumer.
17:34You have like clock code that's, uh, uh, focused on, uh, coding and the knowledge worker space.
17:40Uh, but in the future, uh, these manufacturing businesses are very valuable because they will have very valuable data that's
17:49proprietary to them in the manufacturing process.
17:52And that data is, is very, very good, high quality data that you can use to train your own model,
17:58uh, to improve your manufacturing process.
18:01And this is the approach that we're taking, uh, with these companies.
18:05Uh, you, you also, uh, you mentioned BNW, uh, Siemens.
18:09Uh, we also have worked with Bosch.
18:11Last week I was at the Bosch Connect World Conference.
18:14Uh, we work with them because they are, uh, using AI, uh, to develop their, uh, assisted driving, autonomous driving.
18:23And, uh, that requires a lot of compute.
18:26Uh, uh, so, you know, uh, there, there's a lot of very interesting going on in the manufacturing sector.
18:33So, if I understand you, you consider the ban on, um, Entropy class model by the U.S. an opportunity
18:39for you and your model to be, uh, to be adopted by, uh, other European customers?
18:44Our model?
18:45Yeah, your model, yeah.
18:47Uh, yeah, well, there's two approaches to it.
18:50Uh, one is, uh, through, you know, they could just take our open source model and then, uh, they will
18:55put that, deploy that into their own infrastructure if they have their own data center, whatever.
19:00And, uh, uh, but our infrastructure is developed hand, uh, sort of, you know, hand in glove with the model.
19:11So, we actually have one of the most efficient infra to help people train their models.
19:16And, uh, so, if they use our open source model, they could also buy compute from us, right?
19:22Okay.
19:22So, that's a symbiotic, very symbiotic relationship between the model and infra.
19:29Um, so, uh, that's one approach.
19:31Another approach is there are now emerging, uh, uh, a number of, uh, companies that, uh, develop, uh, these, uh,
19:40inferencing platforms, uh, that offer people a choice of models.
19:46Uh, different kind.
19:47You don't have to use Quinn.
19:48You could use somebody else's.
19:49Uh, they also can access closed source models as long as there's an agreement between the model maker and the
19:56inferencing platform to open up the weights, uh, the gates to the weights.
20:00Uh, in a private context.
20:03Uh, and then customers can go on the, go on to these inferencing platforms to use the models.
20:08I have a more philosophical question about, um, uh, your vision of future of AI, LLM, uh, balance between humans
20:16and, uh, agents, uh, and even in humanity.
20:20How do you see the, the, the next 10 years of the humanity?
20:24Uh, so, today I was, um, I had a, a talk with, uh, uh, the audience.
20:30Alibaba Paris office colleagues, uh, it's located.
20:35We just moved to a new office.
20:36It's located in, uh, I think the second or third floor of a beautiful building.
20:41And I looked out the window.
20:43There's a cafe and people are sitting in the cafe, uh, outdoors because the weather is pretty nice having a
20:53great time.
20:54And I point to them and I tell my colleagues, this is the future of AI.
20:59They may be, you think they may be drinking coffee and, you know, having a good time, not doing anything,
21:05not being productive.
21:06But the reality is, uh, they have deployed agents that are doing the work.
21:12So when you're going to sleep, you still have agents working for you.
21:16Yeah.
21:17So think about the productivity gains.
21:20But, uh, 24 seven, you can actually have somebody else work for you.
21:24So you are, you have, you have the same philosophy that the, the, the people from the Silicon Valley, you
21:28think that many people won't work and we will make the, the agent and robots work for them.
21:34Well, I, I think it'll free, definitely free up, uh, people's time, uh, to do, to enjoy life, to enjoy
21:41their family, to do more entertainment.
21:43This is why I'm very big on live entertainment.
21:46I mean, when people, uh, spend less time in the office, where do they want to go?
21:50They don't want to just sit at home.
21:51They want to go to concerts.
21:53They want to go to, uh, football matches.
21:55They want to go to basketball games.
21:57Uh, that's why.
21:58Don't put that to French people because they will, they will want to work less.
22:04Well, a lot of them are focused on the world cup right now.
22:06I think you Chinese work a lot.
22:08When you, when you, when you look at the Chinese, uh, engineer, they, they have lots of working hours.
22:13This is even with agent, even with AI.
22:16Uh, there will always be people that will work a lot more than others.
22:20Uh, but I think most of us, uh, want to be able to enjoy life a little bit.
22:24Want, want to be able to spend time with our family more.
22:27I think you are, you are, you are a very big, a big fan of the backs basket, Brooklyn nets,
22:32and even the Liberty New York supporter.
22:35I think you are even supporting a, a, a, a play.
22:38I even don't know lacrosse.
22:39Uh, um, what about soccer world cup?
22:42Are you, are you enjoying the, the, the, the soccer?
22:46Well, first of all, about basketball.
22:48I went to the French league championship game last night between Paris and Monaco.
22:54Uh, this is a second game of the five game series.
22:57Uh, and I thought it was very interesting.
22:59And I, and then I realized that, uh, our women's basketball team, the New York Liberty have three French players.
23:05Three players that, uh, play for the French national team.
23:09And then we have on the men's, uh, Brooklyn nets have one French player.
23:12So there's a lot of very big French connection.
23:15Uh, I think, uh, France is probably, uh, the most important non American influence in the NBA and also in
23:24the WNBA, uh, today.
23:25So it's very interesting.
23:27Uh, so about the world cup, um, I, uh, I don't have a horse in this race.
23:34That's the problem.
23:35Uh, I think I just saw some social media, um, that says there's a Chinese referee in the world cup
23:43and he is getting sponsorships.
23:45And I hope at some point China will have a strong team, but, uh, I think that development will take
23:52some time.
23:52Yeah.
23:53Thank you very much.
Comments