Future of Business: Baidu’s Robin Li on the Technology Trends that Will Transform Business
21:34 Share

Welcome to the H. B. Idea cost from harvard business review. I'm alson beard.

I do launched in two thousand as a search engine platform, fast forward two decades, and it's now one of the few companies in the world that offers a full AI stack. Its core businesses spend mobile, cloud, intelligent driving in other growth initiatives, and its products and services have attracted hundreds of millions of users and hundreds of thousands of enterprise customers. Today's guest is leading all of that.

For the third episode in our special series on the future of business will hear from Robin lee, the cofounder, chief executive and chairman of video. He explains how his company has built general AI into its business, the technology trends he's keeping an eye and how he anticipates these tools will transform our lives. Robin spoke to H B.

R. Editor in chief, audionav es, and took questions from the audience during our recent virtual future of business conference. Here's their conversation.

So Robin ly.

I know IT is very late your time in beijing, so we appreciate your joining this life for this welfare.

Hi, I thank you for having me. It's great to be here.

Well, it's great to have you. Before we start, let me just remind everyone in the audience to put any questions you have for Robin in the asked speaker chat. And I will try to get to as many I can later.

But Robin, let's get to IT. So why do your company introduced a ChatGPT like product on boat last year that last I saw has more than three hundred million users? I assume it's been a learning experience. You can you talk a little bit about what you've learned since the first version came out and how IT has evolved? And just tell a little .

bit bit of that. Yeah sure. We launched the early but um I think march sixteen uh of last year, um I think that that was the first ChatGPT like chatbot for all the public companies around the world.

Uh because we we been investing in A I especially uh nature language related A I for quite a few years, we were able to quit launch A A uh a chatbot based on are a large language models over the past a year and half A A lot has happened. Uh the the technology has evolved very quickly and and dramatically. Um this week we learn there. There are a lot of things I should mention. Uh the first is that a lot of people, uh users, developers, uh, customers, they not only care about the efficacy of of the model, they also care about the the response speed.

They also care uh, about the cost of the influence, uh, cost so after march of last year, we have rolled out a series of language models or foundation models to all kinds of different needs in in difference scenario, uh uh meaning that the the model size could vary greatly and the inference cost could very different to and in certain cases, uh users don't mind to wait you know ten seconds to get the best answer and in other narrow, you will have to do IT very quickly subsequent a time. And also in terms of cost, we've uh uh being able to reduce the cost by about nineteen nine percent, meaning the current inference cost is about one percent of the original cost. When we first launched that, uh, I think said all of that, I would say that uh, probably the most significant change 啊 we we think， uh, over the past, you know eighteen to twenty months is the accuracy y of those some answers from from the uh, large language models.

And I I think over the past eighteen months, that problem has pretty much being solved. Meaning that when you talk to a chat pot, A A frontier model based a chatbot, you can basically trust the answer. That's a huge difference.

Now from my perspective, maybe this is a us perspective. There was a huge way of the excitement about AI, particularly with the release of general AI products. Know you've talked about search.

Um there don't seem to be a lot of or or as many interesting use cases as as maybe some of us had expected by now. So i'm interested in your view, are we are we in an AI bubble this point? What's the trajectory of of the technology?

I think like many other h technology, waves bubble is kind of inevitable. When you past the stage of initial excitement, people would be disappointed that the technology doesn't meet the h high expected generated the through the the initial excitement. We've saying this, uh, many times when the internet took off in the mid to late nineties and and uh there was a huge bubble for mobile internet.

Similar things happened and this time for generative A I I think we will also go through that kind of um uh period two, but I think it's it's also healthy. IT will wash out a lot of those fake innovation or or uh products that doesn't have a market fit. After that, probably one percent of the companies will stand out and become huge and will create A A lot of value, will create uh uh value for for the people, for the society. And I think we are just going through this this kind of process this year. The sector is probably cooler than the last year, but I I think is also healthier than than last year.

What's what's the right business model? I mean, there are some large models. You know, mas lama, for example, are open source.

Others are close source, like open a eyes. You know, GPT by do I think is abdicated for a close source approach. What's the thinking behind that? And you know how does that set by due up to capitalize on on AI technology .

yeah you mission close source but I would prefer called a commercial grade uh model foundation model. I I think uh you when you look at the most advanced language model, function model, most of them uh are closed. And when people talk about open source is kind of misleading to me, is different from the open source of linux or or or pitch.

Because for an open source model, you what you basically get is, is a bunch of parameters. You don't know how those parents were derived and and you you have no way of changing of those are so IT doesn't have the effect of many, many people from different part of the world and contribute back to the main branch and make IT Better and Better. Language models are very different and uh you you can use an open so called open source model to do things by the it's very hard for you to contribute back.

And another perception people have on the open source and stuff is that it's pray or IT. It's actually cheaper, at least the cheaper than the the uh commercial once. But uh in the a foundation model area, that's also not true for urban model. I I think h we try to support all kinds of applications of all kinds of customers, both external are uh you know class computing customers and also our internal customers by do search, by do math and know 百度 文库， there are a lots of applications that leverage the power of uh only about and uh basically we we charge uh the the inference cost for those A A P I calls.

So I want to switch now to um robot taxi. A few days ago, tests announce its robot taxi plan. wyo. Has expanded its service.

What do you think are we at a point where the large scale development of robot taxi is including your own product? Has IT arrived this the moment? Yeah.

we ve been investing in the south driving technology for over a decade. IT took us a very long time to get to this point a meaning that uh, we now have more than four hundred cars in the city of 武汉 that covers about nine million people and uh， a lot of people in in that city already used to taking a robot taxi. Our brand is called Apollo goal.

They pay a fair that's typically you know cheaper than uh regular taxi. And I I think the technology is ready in certain restricted areas. It's it's not ready for uh anywhere, anytime uh in um are uh person terms is called the level. Level five is that you can drive anytime, anywhere.

But for level four, I think we are at this stage the we we are at level four, which means that when you know where which area uh you are in and you you you can get rid of the human driver and provide a right healing service. Uh we we cannot do that in the most crowded, most the complicated traffic area yet, but we can do that in most of the the areas in in in, uh most cities. I think U S.

Is probably at the similar uh, stage right now. The the bottom neck is is more about regulation in I think in most cities around the world, taxi service without A A driver is not allowed yet. So we basically try to go to those select few cities, uh, that regulations allow us to Operate such a service. And IT is a gradual process, is is probably gonna take party another another decade for the robot taxi service to be really everywhere to to become mainstream. But a IT will become uh service that people uh prefer.

So you talk about the inevitability of robot taxi putting human drivers out of work. I'm curious your perspective. You can talk about AI without at least discussing the issue of job displacement more broadly. What do you think do you think AI general AI will replace humans on on a large scale? And if so, how do we prepare for that?

Yeah uh, a lot of people compare the general V I revolution to the industrial revolution. If you look at the look back the industrial revolution, similar things, uh, happened. Uh, a lot of old jobs will uh uh you know get rid of but the more new jobs were created every time when, uh, innovation, when when, uh, in a technology revolution happens.

The jobs that got lost are those hardest, toughest job, those jobs that's not so pleasant to human being and the job the new jobs that got created or the jobs that that more comfortable, more decent and less, uh, stressful. Uh, i'm optimistic that this way of innovation or generate A, I will do the same thing. Another point I would like to make is that this kind of process is not something that's gonna en overnight.

IT will take, you know ten years, maybe twenty years or maybe thirty years to happen. So I think human being have time to prepare for that. And uh, uh, we need to be proactive. And companies, organizations, governments and ordinary people all need to prepare for that kind of for.

So I love to talk about more specifically and you know china's approach to A I um how you know do you see a difference between china's path to to generate I development and the approach taken by the rest world?

Yeah I I do see some differences. The most obvious difference is that china is more application driven. We hear more about what kind of applications can benefit from this kind of frontier models， and a lot of starts trying to find a way to leverage the the power of foundation model companies like I do.

Our strategy is to reconstruct and rebuild almost every one of our existing product based on the early about a foundation model. We have already seen uh very big changes um in our existing product to search being the first and foremost thing uh right now over eighteen percent of the by two surgical lots are are they generated by uh early bot. Um we also have this kind of phenomenon in china, life dreaming shopping.

I know it's not that popular in the U S, but it's it's a big business in china. But live streaming requires A A real human to devote to that kind of um um service um for time。 But a we now can create digital human to do life streaming shopping。

Uh the the scripts uh can be generated by early bot IT look like a real person, is is is not a avatar, pacy, uh, can be, can look very real. Sometimes the consumers that the shoppers can not just tell whether it's a digital human or or it's a real human. And during the lifestream, the dial human can also interactive with the audience and answer questions and the react to um their shopping activities, things like that.

Um I I think we just got started. There are there are a lots of lots of uh use cases. We saying that by leveraging the power for generated A I B is the can um uh get much Better r ee, they they can get more revenue and they can save a more cost.

So just to follow up and that I don't want to get into politics at all, but you know there's there's a concern certainly in the us. About dealing with with chinese tech companies.

Tiktok, uh, is face in the situation now that the the that the data you know could be misused, that our data could be just given the difference between the chinese system in the american system, in the western system, that data is not safe, that the government could have access to IT IT said. That how do how do you respond that in terms of your your company, your products and about those concerns? Uh.

well, first of all I do is masked listed, the public company and we would follow, we would comply with any so for whatever uh practice we have, we would fully disclose that we do respect uh user a privacy. We have uh uh uh data uh compliance committee within A R company with a very uh high ranking executive um uh being the the chair and a we we uh we take this very seriously.

And the chinese consumers are not different from any consumers for the rest of the world. They they care about their data. They don't want other people to to look look at their data.

And in order to uh gain the trust of our users, uh, we need to do the right thing and in order to uh stay on as a public listed company in the U. S. And so we we also need to comply with all the affected loss in in the U. S. To okay.

I want to get to a couple of audience questions and there a couple, one from cj, one from Vicky. Not sure where they're from, but they're both asking about a sustainability and environmental cost. So here's one, you know, how do you think about the environmental cost of AI when you consider a return on investment? When is the use case you too taxing on the environment too much for the environment to be valuable as a product or to remain valuable to society?

Yeah, that's a very good question. Uh, but I I I would argue that the efficiency gain, the value created through this kind of innovation, well more than uh you know offset the the power conception. We will find ways uh to generate Green energy faster than without generating A I uh we will be able to complete as other kinds of packs Better and and faster and at at the lower uh cost so that that I I think generate A I well have a positive impact on the moment.

So we probably have time for just one more question. This is an audience question, is from brian and do what you want with this. When brian asked, what do you think the world will look like in ten to fifteen years? How what will our interactions with technology be like then?

Yeah, it's very exciting to think about that. I think a generate A A I is a is a really uh, disruptive A I think IT will give people a, give a lot of the ordinary people the power to be a programmer. 我， what does that mean nowadays? The the the engineers, the programmes get paid like, let's say, you know two hundred thousand per year.

And that's because they they are quite powerful. They they can program things, they can create software that's very valuable. And I say ten probably doesn't take more than doesn't take ten years.

We should see that around five to ten years. And uh uh and a person who can speak natural language, be at english or or chinese can have the power of stock age. Here you can imagine that how uh how much um you know productivity we we can have just by having that kind of capability.

And when I was in college, we we learn program language with them. We learn a second language. Nobody use assembly language to do program anymore.

Nowadays, people use pye song or c plus plus five, ten years form. Now, nobody use pies on or or, or c plus plus anymore. They are just to use english or chinese to do what is the word they want.

And everyone can do that. So think about that. The world will be completely different ten years from now.

That was Robin lee, the cofounder, chief executive and charmante by do. Speaking to h. Br editor and chief audiencia at our recent virtual future of business conference, I hope you listen to all of our future of business series and all the episodes we have on the H B.

Dear cast about leadership strategy and the future of work. Find us at h br adult org flash podcasts or search h br and apple podcast or whatever you listen. And if you don't already submit to H B R, please do is the best way to support our show.

H br dot org flash. Subscribe to learn more. Thanks to our team, senior producers and sanny mary dee, associate producer annabeth at your product manager em fox and senior production specialist road ecard. And thanks to you for listening to the h very idea I am, albeit.