Manage episode 191699281 series 1539200
Companies across the board are investing in AI, and certainly automotive companies are ahead of the pack. The progress towards autonomous vehicles, more intelligent driving assistance, and even things like automated parking and crash avoidance are all evidence of AI impacting the automotive industry in deep ways. In addition, there’s a wide range of non-AI technologies that are also providing competitive advantage to automotive companies. In this podcast, we interview Samantha (Sam) Huang of BMW iVentures, a strategic corporate VC fund backed by BMW. She shares with us insights into investment trends in AI in automotive, and some thoughts on the challenges and opportunities for emerging companies in the AI landscape
- BMW iVentures
- Venture Capital Perspective: The Three Rules of AI Investing [Sam Huang’s Medium Blog]
[00:00:01] Welcome to The AI Today podcast, produced by Cognilytica, cuts through the hype and noise to identify what is really happening now in the world of artificial intelligence. Learn about emerging AI trends, technologies, and use cases from Cognilytica analysts and guest experts.
[00:00:22] This podcast is sponsored by QS. For over 25 years, QS has been helping prospective MBA candidates make informed decisions about choosing the right business school. At our upcoming Dallas event, you can meet face-to-face with admissions directors from top-ranked U.S. and international business schools – including UT Austin, SMU, Rice, IE, Hult, and many more. Find out more at topmba.com.
[00:00:50] Kathleen: Hello and welcome to the AI today podcast. I’m your host Kathleen Walch.
[00:00:55] Ronald: And I’m your host Ronald Schmelzer. Our guest today is Sam Huang, senior associate at BMW iVentures. Hello Sam!
[00:01:01] Sam: Hi Ron, hi Kathleen. How’s it going?
[00:01:05] Kathleen: Hello and welcome to our podcast. We’d like to start by having you introduce yourself to our listeners. Tell them a little bit about your role at BMW iVentures and your experience in A.I.
[00:01:16] Sam: Right. So I’m a senior associate at BMW iVentures. We are a 500 million euro fund, that invests in companies and interesting technologies that are related to the automotive ecosystem. What that means is that we invest in all of the technologies and start-ups and services that it takes to really make a car. That could include things like industry 4.0, autonomous driving, enabling technologies for the car – even intelligent systems and so forth. And then we also invest in all of the technologies, services and start-ups that it takes to retain the customer, after the car is in his or her hands and continue to build interesting experiences on top of automotive ownership, or alternative mobility experience as well. That can mean things like automotive insurance marketplaces or used cars, and just generally different modes of mobility. And then in terms of my experiences that I’ve had: previously I was at SK Telecom ventures and then I was also at Bosch ventures. Incidentally I’m also a lawyer, which is kind of funny and most people don’t really see the connection. But interestingly enough, law kind of focuses a lot on Boolean theory and that kind of underpins a lot of the logic behind AI too. But I spend a lot of my time thinking about AI problems and AI technologies, and I guess you could say I’m the dedicated AI guy on the team. Though I think everyone focuses on AI to some extent as well.
[00:02:47] Ronald: Great. So let’s get into the artificial intelligence side a little bit more specifically. Talk about some ways that you see some of the investments in AI, and some of the recent or more intriguing investments you’ve made. Maybe you could tell us about what BMW iVentures is doing specifically with their AI investments.
[00:03:01] Sam: Yeah. When you think of AI investments, you can really categorize it into two ways. One is the general purpose AI companies. These are what I call great companies with really smart people, and they have AI technology, but they don’t really have a specific business problem that they’re trying to solve. And then we have the second camp of AI companies, that are really targeting a specific business application using AI technology. I call these guys the ‘applied AI’ companies. In terms of where I think the great investment opportunity lies, it’s really in the applied AI camp. And the reason why, is because even though we might have some really good exits with some of the general purpose AI companies, that was really when a lot of bigger corporations that didn’t have really formulated AI strategies and didn’t have the talent. So a lot of the former, or even some more current acquisitions, are focused on acquihires and just building out expertise. Once you have that, you’re going to see a lot more acquisitions around applied AI companies. For me, when I think of what’s a good company to invest in, I would say: always a company that thinks really hard about a specific business problem and uses AI technology as a means to get there, but really does not view AI technology as the end. Let me give you an example of an applied AI company, just for illustrative purposes. An applied AI company would be something like a robotics company that uses AI technology to create a robot that also serves a specific business function. Whether that’s transporting goods throughout the warehouse or bringing meals to people in a senior folks home.
Other than these two camps of applied AI and general purpose AI companies, the third way that I really look at AI companies or the AI market, is looking at it in terms of big data. Companies that can clean the data, provide high quality data and at scale. I view these also as within the AI landscape. Mainly because any sort of machine learning project or AI project or deep learning project relies on the quality the of data. Without investing in these companies that clean the data and make sure the data is good, then you really don’t have an effective or sufficient AI technology or product.
And so in terms of where we’ve invested in AI technology – or I’ve invested in AI technology – I kind of operate under the assumption that AI is everywhere and even when you’re not investing specifically in AI technology, it’s somehow going to be related. One of the interesting investments that I did recently, was in DSP Concepts, and what they do is provide this great audio processing software that would enable the quicker iteration of intelligent assistants and voice interactivity across products. So this really capitalizes on, what I would call, the move towards having more engaging and highly interactive experiences with products using voice technology. Which is, of course, AI technology. Another interesting company that we’ve invested in is Xometry, which is a marketplace for machining services that allows corporations trying to do prototyping and production of a part, to get these parts really more cheaply and quickly. And while people are like ‘is this an AI company?’, for me it is. Because the way that they are so unique and different and effective, is because they use AI technology to find the best machining service for a big corporation. So I think in terms of interesting investments, those are kind of the top two that come to mind. We recently also invested in an autonomous driving start-up. It’s not disclosed yet, but this company is operationalizing autonomous driving technology. So of course, you know, all automotive OEM has to have some sort of an autonomous driving investment and this is the one for us.
[00:06:55] Ronald: Very cool.
[00:06:55] Kathleen: Yes. So what are some of the challenges to adopting AI in the auto industry that you see?
[00:07:02] Sam: The automotive industry is really like any other industry, in how it’s trying to tackle AI. And it really comes down to two things. One is: how do you incorporate AI technology across the enterprise, so that you can streamline workflow and optimize cost reduction and things like that. And then the other way that enterprises are trying to integrate AI technology is, of course, in creating sticky and more addictive products. Those are two general things that any sort of AI company, or any sort of company really, is trying to solve in terms of the big problems. Going specific to automotive though…
One of the key challenges I see is that with a car, the R&D time is most likely around four to six years from thinking of a concept to actually going into production. So what this means is that your feedback loop, into the type of data that you are able to get from developing a car, is limited. You’re not going to be able to have as many iterations on the car, because the development times are just so long. But then… I think a lot of people think about the automotive industry and autonomous driving, ’cause that’s really the hot thing. And the key AI technology problem with cars and AD cars right now is having bad data. There’s a lack of data to train your AD parts sufficiently. Right now, to train an autonomous driving car, it takes – give or take – 6 billion miles to train an autonomous driving car. So some ways that OEMs are trying to get around that problem, is to create simulations, so that you don’t have to really run those 6 billion miles in real life. So those are some of the big problems we’re trying to solve. It’s really about the data and then leveraging your AI technology on top of that data, to streamline workflows and create really sticky and addictive products across enterprise.
[00:08:57] Ronald: We actually briefly talked about this, about the role that big data plays with AI in general. How do you think data issues are impacting artificial intelligence, both good and bad, in terms of adoption?
[00:09:08] Sam: So data is a huge problem within the AI field, but it’s also one of the drivers of the proliferation of AI technology today. So what I mean by that is… With any sort of AI problem or machine learning problem, you have your algorithms and you have the data. In order to train your algorithms effectively, you need this data. What this means for automotive OEMs like us is: when we don’t have enough data, or we don’t have enough data about the miles of a car that has driven on the road, we’re not able to train our system effectively. The second problem with data is that if you have data that is dirty, you’re going to train your system completely wrong. Just as an example: if you tell a child growing up, that a picture of a cat is in fact a dog – and the child learns or thinks that that cat is a dog – you’re feeding that child the wrong data. So that child is going to have the wrong output. That’s a very basic machine learning example. That’s the type of problem that really resonates across data and AI. Without, you know, solving these key issues of bad quality data and lack of data, we’re not gonna be able to solve the big AI problems down the road.
[00:10:22] Ronald: So some of what you’re looking at may be big data specific, that may not solely even be AI specific? Just sort of more underlying issues of helping organizations get better data, clean that data better, maybe try to eliminate some biases in the data? That may be applicable to other domains that are not just AI specific? Have you been looking sort of more broadly beyond AI in terms of big data?
[00:10:42] Sam: Yeah, that’s exactly my investment focus. Applied AI and companies that can simulate AI data or actually clean parts of the data, before you run your machine learning algorithms on top of them.
[00:10:55] Kathleen. Okay! And then, as a last note, we’d like to end with: what do you believe the future of AI is in general? And then its application to corporations and beyond.
[00:11:04] Sam: I believe that all corporations are going to have to quickly design and implement an AI strategy, or they will fade fast or be cannibalized in the market. And the way that they have to define their strategy is really in two ways. They have to answer these two questions. Number one is: how are they going to integrate AI technology and leverage the data that is existing across the enterprise, to streamline collaboration across the business space? To implement cost reduction, to improve the enterprise as a whole. The second question that they’ll have to ask is: how are they going to use the data that they have, that is specific to their industry, to basically get a competitive advantage over their competitors. So what this means is, for example, Google has so much data on its search engine, and people have been using Google’s search engine for nearly the last two decades. I don’t think anyone can ever go and beat Google in developing a new search engine. Google has acquired such a large dataset for search engines, that it’s going to stay the winner in this field. So for AI companies like Facebook and Google, that have really amassed a market lead in either social media platforms, creating sticky newsfeeds and feedback loops based on interactions with the sites, or whether it’s a search engine like Google: you’re not going to see entrance into that market. So what does that mean for start-ups? Start-ups who want to win and want to leverage AI technology, which I believe is necessary for them do to, are going to have to find a very specific market niche and be very good at it. The best example I have is from Andrew Ng — he is one of the foremost experts in AI technology. And he has this example of – Blue River systems was recently acquired by John Deere technologies, and the founders of the river were his students at Stanford. The acquisition of that was $300 million dollars. What Blue River did was – the students at the time were just taking pictures of heads of lettuce and they just have millions of millions of millions of pictures of heads of lettuce. So they had the biggest collection of heads of lettuce. Now: Google doesn’t have that, Facebook doesn’t have it, IBM doesn’t have a lot of heads of lettuce pictures, but these students did. What they did was build a company on top of that dataset, so that they could create a technology – basically it’s a robot that allows you to see which heads of lettuce need to be sprayed with pesticides, or which leaves need to be picked out. And what this meant is, at the end, farmers could have a 90 percent reduction in the amount of pesticides they use, so it was a general optimization in the processes in the farm. So that company got bought out by John Deere for $300 million. As you can see, finding a specific niche when you have players like Google and Facebook which are so far in what they’ve done in terms of the datasets and just their experience – you’re going to have to find a specific niche, or it’s going to be very hard to win.
[00:14:32] Ronald: Good to know. Well, I think that’s one of the things we’re definitely focusing on, as part of both our research and our podcast; we’re trying to find all those niches. ‘Cause artificial intelligence is definitely going across that big hype cycle right now, everybody is talking about it. We’ve gone to many events and there’s certainly a lot of great things happening in artificial intelligence, but also a lot of things that are just spreading a veneer of AI technologies on their products.
[00:14:53] Sam: Yeah, ‘AI light’ as we call it.
[00:14:54] Ronald: Exactly, ‘AI light’. Well we really want to thank you. You did a great job, we really appreciated you joining us on this podcast.
[00:15:02] Sam: Thank you. I had a great time.
[00:15:04] Kathleen: And listeners, as always, we’ll post articles and concepts discussed in the shownotes. And I know that Sam’s starting a blog as well, so we’ll link to that as well. Sam, we’d like to thank you once again for joining us.
[00:15:14] Sam: Thank you so much, both of you.
[00:15:16] Kathleen: And listeners, we’ll catch you at the next podcast.
[00:15:19] And that’s a wrap for today. To download this episode, find additional episodes and transcripts, subscribe to our newsletter and more, please visit our website at cognilytica.com. Join the discussion in between podcasts on the AI Today Facebook group, and make sure to join the Cognilytica Facebook page for updates on this and future podcasts. Also subscribe to our podcast on iTunes, Google Play and elsewhere to get notified of future episodes.
Want to support this podcast and get your message out to our listeners? Then become a sponsor. We offer significant benefits for AI Today sponsors, including promotion in the podcast and landing page, and the opportunity to be a guest on AI Today Show. For more information on sponsorship, visit the Cognilytica website and click on the podcast link.
The post AI Today Podcast #011: BMW investing in AI – Interview with Sam Huang of BMW iVentures appeared first on Cognilytica.
25 episodes available. A new episode about every 6 days averaging 18 mins duration .