Subscribe
Share
Share
Embed
Born & Kepler is named after the mathematician and scientists Max Born and Johannes Kepler. This bilingual podcast, offered in both German and English, dives into the expansive world of Artificial Intelligence (AI), exploring its foundations, evolving technology trends, academic search, and its impact on businesses and society.
Born & Kepler will feature a diverse lineup of experts from academia, venture capital, private equity, journalism, entrepreneurship, CTOs, and policymakers. Each guest offers unique insights into how AI is reshaping their sectors and what we might expect in the future.
Our goal is to provide a deep understanding of the core principles and breakthroughs in AI, enabling you to stay updated with the latest advancements in AI technologies and how they are transforming industries. During our episodes, we will explore how AI is influencing business strategies, optimizing operations, and driving innovation. We will also explore the ethical, social, and regulatory aspects of AI in everyday life.
Andreas Deptolla (00:01.519)
Good morning Francis, how are you? Very, very good. Good morning to Berlin. Let me maybe start there. You're originally from France, you spent a lot of time internationally. How is Berlin treating you as a city for you and your family?
Francis De Vericourt (00:02.956)
I'm good, you?
Francis De Vericourt (00:19.214)
Well, it's true. moved around a lot. And if you ask a French man, I'm French indeed, why he's moving so much, you need to follow his wife. And my wife is German, so that took us back to Europe. Berlin is a fantastic city to grow kids and to have also a nice, rich personal life. So I'm enjoying it.
Andreas Deptolla (00:44.624)
Maybe tell us a little bit about your academic journey.
Francis De Vericourt (00:50.744)
Well, most of my studies I did in France in the very, very French system of teaching style. And I did my PhD in Paris and I thought I will never go to academia because I was not very inspired with the French academic system. I don't want to hurt the feelings of my German friends, but I don't think the German academic system is that inspiring either. But I got a chance to do a postdoc at MIT.
Andreas Deptolla (01:14.063)
Yeah
Francis De Vericourt (01:20.492)
And I thought before working for some consulting company, would enjoy that. And then I fell in love again with academia in the US, mostly because people, it was a very flat hierarchy and you had the most well-known people around the world who were coming to my office and asking me to explain my research to them. And also it was very multidisciplinary. I could talk to anybody, whatever their field. And so it was Disneyland for me quite.
And then that convinced me to stay in the US. Did a big part of my career there and as I just mentioned, after falling in love, followed my wife who is a pediatrician who wanted to save babies and so we moved back to Europe.
Andreas Deptolla (02:03.162)
Mm-hmm.
Andreas Deptolla (02:06.762)
So you mentioned at MIT the, yeah, attracts the best talent, right, flag hierarchies. How would you describe kind of the differences between the US system and maybe the European system in terms of the intersection of academia and business? Is there a closer relationship? And if so, how does it manifest?
Francis De Vericourt (02:09.294)
Mm-hmm.
Francis De Vericourt (02:13.656)
Mm-hmm. Mm-hmm. Mm-hmm.
Francis De Vericourt (02:29.896)
You know, I think, so first I think Europe, France and Germany also are doing very outstanding research. I think this is still a field where we as a continent can still compete and the number of fields where we can still compete is shrinking, so let us celebrate that. The German system at least is a Kaiser system and you know even so now if I were to the classical German system I would be a
probably a Kaiser and I have people working for me and I don't find that inspiring. The main difference in the US is that whether you are a young scientist or a very well established professor, at least this is what I experienced, your interaction is more or less at the same level. I remember, for instance, when I was at Duke University, I was a young assistant professor and they were very well established professor across
across the corridor was not in my field. And I had just published a paper in a major outlet and he came to my office. He had the journal in his head. I mean, there was still paper at the time and he put the papers on my desk. I did not know him and he said, congratulations. We're so happy you're here. You're doing amazing work. And then he left. You know, I never experienced that in Europe. I never saw a European professor being French or Germans to come to the people at the
Andreas Deptolla (03:50.163)
Hmm
Francis De Vericourt (03:57.644)
bottom of the food chain and just say amazing work, I love what you're doing. And so that's, you know, that's, it's just a mindset. This is positivity and recognitions, which when you're young, need. I don't see it in Germany. I think it's more suppressing talent and everything is scattered to support the Kaiser basically. And I'm going to make a lot of enemies, but the whole system in Europe is scattered to that, like provide resources to one person at the top.
Andreas Deptolla (04:22.288)
Hehe.
Francis De Vericourt (04:28.012)
which is not, I think, the best way to make breakthrough in science.
Andreas Deptolla (04:32.053)
When we talked the other day, also talked about how in the US oftentimes, like in Germany, said, like, you know, there's a lot of great foundational research, right? But then startups don't quite pick it up, right? There are no companies that truly scale right here in Europe. Obviously, a couple of exceptions to it, right? Why is that?
Francis De Vericourt (04:46.675)
Francis De Vericourt (05:00.248)
No, you're right. And this is indeed a fundamental difference between what's going on in Europe and in the US. I think, I mean, more in France than in Germany, but still Germany, they believe that because you throw a lot of money at fundamental research, it's going down the road to have spillover effect on technology that is going to help the economy and grow the market.
But it's research for money for fundamental research. It's just a necessary conditions In the US this is mentality that Applications business and scientific research of fundamental research. There's not a big disconnected. It's all mixed up and This is ability to or you know, I'm I'm doing fundamental research, but yeah, I can
see the applications and I can engage and take the risk to do this amazing venture, make money of course, but also have implications for people and save life quite frankly. So I think this is dichotomy. Germany is better than France because it has a long history of having a close relationship between industry and research. What is missing for Germany is risk taking because
Andreas Deptolla (06:24.555)
Mm-hmm.
Francis De Vericourt (06:25.186)
The country is good to work with established manufacturer, established big corporate, and let us try to improve these processes and innovate, which is good. But when you have really disruptive innovations, you need to take risk. And it's really not the culture of the German culture, which is extremely risk averse. Investors also, VCs, and very risk averse. And that's killing innovation.
Andreas Deptolla (06:52.46)
Yeah, it's interesting that you say that. assume, correct me if I'm wrong, that like your top students, probably want to work for McKinsey, Bain, or maybe they work for one of the big German automotive companies or whatever that might be, right? Or maybe they continue academia, but like creating their own startup is probably like a less troubled field, so to speak.
Francis De Vericourt (07:21.306)
I think you're right. mean, so if I look at the young students that are studying business and doing their master, I think I find some of them that have a change of mentality here and wants to engage a bit more, but they want to work for McKinsey and the like. The issue is with scientists and when you want to have to break through, the founder and the innovation is pretty much always, not always, but most of the time coming from a scientist.
Andreas Deptolla (07:50.264)
Mm-hmm.
Francis De Vericourt (07:50.478)
And you the appeal for the academic career is so strong. It's also a status story in Germany. You know, when you have the title professors, your semi-god is the only country I know in the world where you put that your professor on your passport. It's insane. It's like, if you're in the US, if you're a professor, you're a loser because if you're so smart, why aren't you rich? But here it's like, okay, look, I'm a professor. I want to be a professor. So this is appeal. so...
Andreas Deptolla (08:02.438)
Mm-hmm
Francis De Vericourt (08:18.414)
That's really on the scientists that I'm starting trying to work hard to show them that this is path where the innovations you can write papers and it's good and it's beautiful, but they will always stay papers if you let them. But if you want to change really the life of people, that's the path you need to take. it's risky, it's time consuming, but it can really be rewarding. And so that's, yeah, that's you're right. mean, this is mindset.
risk appetite, that is a big issue in students.
Andreas Deptolla (08:51.574)
What maybe in more generic terms would you recommend to your students in the age of AI, right? I mean the world is going to change in next five, ten years, right? What kind of skills, framework, regardless of what career they're choosing, what will be important? And maybe on the flip side, is there some kind of advice that you often hear people are giving to students where you would say like, I don't agree with, right? I'd love to hear your two cents here.
Francis De Vericourt (09:03.02)
Yes, it's solid.
Andreas Deptolla (09:21.307)
video.
Francis De Vericourt (09:21.934)
So I think it's a fantastic time to be a student right now. mean a student more in terms of if you are at the master level or because I feel we are reaching the point where you do not need to have a master or PhD in artificial intelligence to have a huge impact with AI. You start to have really off the shelves libraries already but even now
you can use Chatipity and all that, where it's a bit late 90s, early 2000s with the advance of the internet, and where the technology was there, and then you start to see the breakthrough in terms of business. Let's say that we're not coming from engineers. mean, Google is an exception, and not an exception, a counterexample, but Facebook and all those Amazon...
they came because the technology was mature enough. And still today in the field of AI, we do not quite know which one, which venture, which new applications will really bring the next big, big, big company. So I think this is an exciting time from that respect because it's an open field. Now, to go back to maybe more working in the
AI age, if you want the advice or what I hear and I hear that less and less, so it's good news. But a couple of years, were these talks about what makes us humans different, which is, you know, we have emotions, emotional intelligence, and the machines will maybe never be able to do that, or we have intuitions. And I think, no, in fact, if the machine is good at is to make
better decision when it comes to intuitive decision that we have is you see that with with AlphaGo for instance, I mean it's an old story now but the way Go grandmasters play they really use their intuitions and that's why it was shocking for everyone that machines can beat the human at its own game to emulate intuitions. So what I'm a very strong advocate is to what really makes us different than the machine is that we can
Francis De Vericourt (11:49.784)
build our own representation of the world. So we can build mental models that are extremely powerful and that to this day, the machine cannot do those. AI, when it makes a prediction, a recommendation, it's really stuck to the data on which it is trained. And we do not think like that. And we can do amazing thing that the machine does not do. So I would tell the students, train yourself to
think in frames, in mental models to reframe problems because machines are not reframing anything, but you can. And so that's where the complementary is going to come from.
Andreas Deptolla (12:28.572)
Maybe tell us a little bit more about the idea of framing, right? And I think in the field you publish a book, right? I think it would be interesting for us to hear more about the core idea behind it, how it can be applied, and then maybe also your motivation to even publish a book, right? What kind of led to that?
Francis De Vericourt (12:50.602)
Okay, I mean you want me to talk for a long time now. Let me start with the motivation and then the trigger. So the motivation is I think one thing that has inspired me in my studies right from the start is the realizations of how thinking in the abstract has in fact huge implications and real
Andreas Deptolla (12:53.318)
Yeah
Francis De Vericourt (13:19.574)
real life applications. Any technology that is changing our life today started in the mind of some people as a very controlled way to to harness the imagination. So it's not just flight of fancy, like they use again, I'm going to use this notion of mental models to think about those problems in the abstract and that had ripple through effect and to have those implications.
application. So now the trigger is that this is basically that's that was my work in decision making and and I have a very good friend Victor who is a professor at Oxford who wrote a beautiful book on big data in the mid 2010 I think it was like more than 10 years ago and he came to me and we had I think a coffee I mean just just to catch up and he asked me this question so what do you think is going to be the role of human
human decision making more precisely in the age of artificial intelligence. And my answer was the one that I gave to you and I give to my students is we have an ability to represent the world and the machine won't be able to do that. So that's what really triggers everything came together with that questions. And then the answer, well, let's write a book about it. And the rest is history, I guess.
Andreas Deptolla (14:40.813)
If you look at decision making, and maybe specifically if you look at this now from a business perspective, right, the C-suite of a company, where do you see applications right now where AI can already support decision making? And how do you think will this develop over the next, let's say, five to 10 years?
Francis De Vericourt (14:46.944)
Mm-hmm
Francis De Vericourt (15:03.758)
So, you know, to be totally honest, I think that the biggest impact right now that AI can have is one of mostly efficiency. And so it's repeated task that is time consuming and where there's a lot of data, stable environment and where AI can do the work better, faster. So, GPT is the perfect example. It's extremely versatile.
Andreas Deptolla (15:30.581)
Mm-hmm.
Francis De Vericourt (15:34.094)
Now if you ask what excites me most, and I think that's what you're going to see. So it's more the mundane decisions, but can really improve efficiencies. And quite frankly, this is also an advice I would give to students, we do not quite know yet the true application of those tools, such as GPT, that already exists. So it's really early on to play with those tools, explore and
individually people will start to find all sorts of applications that fits your needs and that others may not have thought about. So it's like really practicing finding the gain of efficiency, how you can make the best of those tools and this is act of creativity like to find that match. But what excites me the most I guess and in the future is what I see more in deep tech is it seems that, it's not that it seems, I'm kind of convinced that AI
help us innovate. mean, it's really where I would see in the future where you're going to have the biggest breakthrough that it help not only help the innovation process to be more efficient. That's one thing, but it's really given us new ideas that we may not have thought about without AI basically. And there is an amazing potential and I think it can really transform humanity quite frankly. I mean, I'm a bit dramatic here, but it's...
Andreas Deptolla (16:34.37)
Hmm.
Francis De Vericourt (17:01.947)
I have this strong belief that this is what might come.
Andreas Deptolla (17:06.177)
I'm sure like there's a fear, a justified fear, right, in society that AI will also...
cut jobs, right? You mentioned like, you know, these kind of repetitive tasks, right? We're seeing now the Klana case study that went around, right, where AI already has a pretty big impact on customer support, right, kind of automating some of these tickets with pretty high customer satisfaction. Where would you see in the short and midterm kind of like the major impact to society and the job market as AI is evolving?
you
Francis De Vericourt (17:46.05)
I mean, you know, it's not the first time that you have a new technology that is disrupting markets. this story of having new technology taking over jobs has been the story of humanities in the industrial revolutions. And so I think, yes, there will be a job that has been cut. mean, think about, for instance, translations. just, so unless you want to translate a book,
Andreas Deptolla (18:12.569)
Mm-hmm.
Francis De Vericourt (18:16.248)
then there is a future for translators that will use AI, but then really add maybe one layer, which is important. But now you don't need any translator, unless it has to be certified, but you can do that automatically. generating pictures, there's a lot of jobs that may be at risk or tasks.
Andreas Deptolla (18:22.053)
polish it.
Francis De Vericourt (18:42.786)
That being said, if you take at the whole overall market, I strongly believe that you're going to create more jobs. The key questions, and it's a very important one, is how do you help people who are going to lose their job to find their new ways? And that's always been a challenge and that creates resistance. I I understand rightly so because I'm losing my job and what do I do now? So it's the task of societies to have this shift and it's important to
Andreas Deptolla (18:55.407)
Hmm.
Francis De Vericourt (19:12.83)
aware of it and manage it because otherwise now we are going to go to the political sphere but then you end up with populisms and you end up with crazy things and so being aware of it helping the people who may lose their job transit is key and it will open new jobs for sure.
Andreas Deptolla (19:33.66)
What are your recommendations to European politicians to best handle this? So undoubtedly that there will be a change, right? People are talking about re-skilling. That's not easy to do, right? How can you take society on this journey successfully?
Francis De Vericourt (19:55.448)
Well, I mean, unfortunately, I think if I had the recommendation for the policymakers of Germany and Europe is more wake up to the rise of AI and even not AI, mean, science based technology in general, because the change we are talking about is, you know, it's more for the US quite frankly, because we are not there yet where we are embracing those new technology the way we should do.
Andreas Deptolla (20:02.535)
Hmm.
Francis De Vericourt (20:25.002)
It's more like, you know, wake up and get ready not to be left behind because we already... I mean, Europe missed the AI revolution. It means the internet revolution, it missed the AI revolution. So if we do not wake up, that's it. I mean, it's our last chance now. Then we are losing a century. I mean, more than a decade. It's a century of drama that may waiting for us.
Again, it may be a bit dramatic, but I'm very frustrated by, I mean, I do not want to blame politicians because it's also the populations that is asking for that. it's not part of the political discourse today, really, that Europe is missing out the coming deep technological revolutions. And I'm going to say one thing on that.
Sorry to speak too much about it, but I'm passionate about it. It's like, you look at the industrial revolution, which is really the first scientific-based revolution, this is where deep tech innovation was invented by humanity. Before that, science did not really exist. So innovations were trials and errors and very slow. And that industrial revolution brought a lot of wealth, mean, drama as well, but wealth and innovations to Europe.
And China, who just before was in fact more developed than Europe, missed that opportunity. And it took a century for China to recover with a lot of drama. And now it's the other way around. mean, China is overtaking Europe. And if not, we do not react and we do the way China reacted at the time of industrial revolutions.
European are tossed. I mean, I'm sorry to say it. And so that frustrates me that I don't hear that in the public discourse. I know there is drama going on in the world, but if you think long term, that's big, big, big, big issue that we are facing.
Andreas Deptolla (22:29.599)
Yeah, China certainly not without its challenges, right? If you look at like, you know, economy, right? If you look at like, you know, demographics, you know, some, some other things. you know, I think to your point, what's, what's interesting to see is like, you know, the automotive industry, right? Where like, where are we going these days here in Germany, right? Versus the Chinese and it's challenging.
Francis De Vericourt (22:55.34)
I mean, let me defend, mean, China is not that I'm defending the Chinese government, but China is still competing with AI with the US. China had its own Amazon in its own Google. And they didn't miss the AI revolution. They didn't miss the internet revolution. They have a big market. We also have a big market. No, yes, I mean, they have political issues and it's not that.
Andreas Deptolla (23:03.625)
Mm-hmm.
Francis De Vericourt (23:19.394)
because of the industrial revolutions, Europe didn't have its political issues. I there were two world wars. so, indeed, I mean, but if you look at the long-term and the trajectories, it's very hard to recover when you miss that type of revolutions because what is fundamental is, I said in the beginning that Europe, Germany is competing in terms of fundamental scientific research with the rest of the world.
and missing out in this translation of technology. But the problem in that to stay at the top, even in the academic research, you need the technological innovations. it's both where there's a feedback loop. So if we miss that, if we do not innovate in terms of technology, our ability to have scientific breakthrough is also going to be suppressed. It's just a question of time. And then we are really losing out.
Andreas Deptolla (24:18.892)
Yeah, I heard the other day like common like, you know, the United States is a great place to, you know, innovate, build companies, make money and Europe is nice to spend a, you know, a nice lifestyle, right? So, you know, hopefully that doesn't mean I think that's a trend we want to, yeah, totally.
Francis De Vericourt (24:40.408)
We want to protect and if you want to protect that lifestyle, I mean, I'm a European to the core. mean, I'm a European with very strong connections and link with the US, but that system, that lifestyle, we need to defend it. And the way to defend it is to play that game. Otherwise, yes, mean, Americans and Chinese will spend nice vacations among us and they are welcome to come, but we are not going to benefit from.
Andreas Deptolla (24:46.465)
Ahem.
Francis De Vericourt (25:10.168)
from those advancements.
Andreas Deptolla (25:12.318)
Let's talk about some concrete things that you're working on that can help you. The Deep Tech Institute in Berlin. What's the mission? What inspired you and the team to start it?
Francis De Vericourt (25:27.787)
And so, you know, we, in fact, in many ways, we always talk about what inspires to start it, which is the realization that Europe is losing its way in deep tech innovations. we wanted to help. mean, we're not going to say we are going to change Europe there, but we are in a very good position to play the role of a catalyst. And we are business school, school of management. So it's not that we are developing the technology.
Andreas Deptolla (25:35.658)
you
Francis De Vericourt (25:56.524)
But we are in fact building bridges between, and funnily enough, between different scientific institution of Germany who do not like to talk to each other, typically, I there's competitions, especially between Munich and Berlin, I discovered that. So the vision is to try to help enhance the ecosystem around it. And to in particular,
Andreas Deptolla (26:12.738)
Mm-hmm.
Francis De Vericourt (26:25.07)
the scientists talk to the business people, so train the scientists to become more business mindset. Also convince our students and entrepreneurs who are not necessarily into deep tech but to use their amazing skills to help translate scientific breakthrough in new technologies. So that was one of the main motivations. The other one is more of an academic if you want but
very well established frameworks today, which you're all familiar with, to help scale up startups, Ling startup in particular, but Airbnb Dropbox, all of them manage to leverage. So we understand what is the best practice, what works, even though it's still very, very risky, most of the time you are not going to succeed, but we know what is the best approach. The problem in that those approach do not work in Deep Tech.
because those approaches are about de-risking the market. Deep tech is also about de-risking a technology and that's a different game. And we do not know, I mean, there are best practices, but there's not a full understanding about what really works and why. So that's the second mission of the Institute is to try to uncover that.
Andreas Deptolla (27:43.762)
That's great. the US there are incubators and what nots, maybe somewhat of a similar model that were very successful, right? If you look at like Y Combinator or Techstars, they have created many, many multi-billion dollar companies, right? It might be a little bit earlier on the...
Francis De Vericourt (27:51.832)
Yeah, yeah, true, true.
Francis De Vericourt (27:59.288)
True, true.
Andreas Deptolla (28:05.269)
German site here, but are there certain outcomes already of your work there? Any companies that either launch products, launch companies, venture capital?
Francis De Vericourt (28:17.096)
So yes, the institute is two years old, so we are brand new, but so we have some, so we have a mentorship program which is very close to what US are doing. So it's not quite incubator because they are not on site, but we create a bench of mentors from Europe, which also one of the benefits of that is that
Andreas Deptolla (28:21.201)
Mm-hmm.
Francis De Vericourt (28:43.762)
mentors themselves, so the VCs and the entrepreneurs who are maybe in different ecosystems they meet. that's also and so we have some companies that managed to raise rounds thanks to us after going through the programs. I cannot give you a unicorn yet. I hope they will make it. Some of them I'm quite confident but I you know I should not be over confident. So we start to have some some attractions here.
Andreas Deptolla (29:00.658)
Mm-hmm.
Francis De Vericourt (29:13.646)
One thing that in terms of outcome, is not necessarily on the venture quite yet, we are now partnering with friends with HACC, which is in France, where we have created a European stream trying to bridge both ecosystems together around quantum computing, new generations computing.
And there, really the benefit is to try to have the two ecosystems, the venture from Paris coming here, discovering the German ecosystem and vice versa. And I have a lot of hope there. Yes, and that's a very positive development for me because that's what needs to happen in Europe if we want to be successful.
Andreas Deptolla (30:05.523)
Let me maybe ask you a personal question here as well, whether this is now like from the Institute or anything else that you see in the market, if you would have to invest $1,000 into a startup, like is there something where you say, this is, whether it's a specific company or maybe just a sector, where it's like, this is really hard, where would you deploy the capital?
Francis De Vericourt (30:07.587)
Yes.
Francis De Vericourt (30:20.808)
yes, yes, yeah, yeah. Well, if I had your money. So first, I think the general trend that I really believe in is how AI can, especially in biotechnology, how AI is used to
enable the discovery of new drugs or new pathways. And I think here there's a huge potential and with big impact. so, I mean, you probably maybe you know in silico medicine, so probably it's a bit late to be an investor in that, but they are doing amazing work to do that. And we are just seeing the beginning. So there's today only
Andreas Deptolla (30:51.528)
Mm-hmm.
Francis De Vericourt (31:18.53)
I do not know 10 drugs that are AI based that are in clinical trials. So we'll know, know, soon if these approach works, but if they work, you know, then there's a whole world that can be done. So I would target those. Now, if you want some that went through our programs, there's one that I really like and I have no stake in it, no money in it. In fact, I,
It's called bio memory. is based in Paris, so not Germany, but what they do is incredible. One of the big issue with data center today and is going to be a big problems in the future that they are consuming tons of energy, huge CO2 impact and with the rise of AI and the rise of data is going to be worse and worse and worse and worse. And they came up with a technology which is
based on DNA. So instead of storing information on hard drives or on physical support, they use the beauty of evolution that found a very efficient way to store information through the DNA. So it's not binary. It's maybe a code with three. And it's much smaller. The condensation, the space is much smaller. But more importantly, once it's encoded, it's just a molecule.
Andreas Deptolla (32:30.859)
Mm-hmm.
Francis De Vericourt (32:44.308)
it uses no energy whatsoever. So you spend energy to read it, to write into it, but that's it. And I think they are really starting to really, so now they are trying to scale the technology so they can store limited information now, but they can still store a couple of books on DNA. The goal is to try to scale it up and I...
I'm not a specialist in that, but I see no reason why they would not be successful on the technology side. there are plenty of reasons, but it's really looking good. So that, yeah, if you want to invest, talk to them.
Andreas Deptolla (33:20.281)
Well, it's certainly a massive leverage, If you now look at the AI landscapes to your point, so much money is spent on energies and data centers. As a matter of fact, I read that Microsoft now is the first company is buying an all nuclear plant in the United States. whether this novel goes through or not, it shows how important that topic is.
Francis De Vericourt (33:47.988)
That's true and the fundamental reason for that is the architecture for computers were not designed for AI. mean they were designed for other tasks but not for AI and so yes we need more energy if we use the old infrastructure or come up as my examples illustrate with new architecture that are radically different and I believe in that so you can buy nuclear plants and that's certainly be helpful.
But there might be innovations coming to say, maybe you do not need this old type of computer. You can do different architecture where less energy, more speed, et cetera.
Andreas Deptolla (34:30.162)
Thanks for your prediction here. And of course, with all of these startup or venture investing, these are very risky investments and whatnot.
Francis De Vericourt (34:32.776)
Sure.
Francis De Vericourt (34:39.106)
Yes, yes, yes, it's risky. Yes, yes.
Andreas Deptolla (34:42.386)
But it's always fun to speculate a little bit if you you now you know look more at the at the you know current leaders of European you know companies here like You know you look at return investment, right? We talked earlier about like, you call centers, you know, there's a Klana case study, right? Where else like where would you say like what are what are the the case studies right that maybe you have worked with companies or you?
Francis De Vericourt (34:59.534)
Mm-hmm.
Andreas Deptolla (35:12.059)
have heard about where people had an impact today.
Francis De Vericourt (35:17.976)
Yeah, I'm sorry, I'm going to bring you back to what I just said. think the, I'm looking really to big impact. I mean, it's like, and I really think that is going to be about how you manage to use AI to help you innovate. Now, you know, of course today, I do not want to say low hanging fruit because it can have massive impact, but
Andreas Deptolla (35:27.566)
Mm-hmm.
Francis De Vericourt (35:47.202)
The approach you want to have is to look to an industry where there's a lot of data, unused data, messy data, which has some stability because if the industry changed too fast, AI cannot follow. I mean, you need to retrain. So, know, immediate application is retailing, e-commerce when you place your ad, even medicine. And all that is today.
grab. mean we have all the tools we can make that super efficient. I mean there is a business model to have and all that but that's going to work for sure. What I really see is how AI can help you look at things differently and which is is paramount for innovation. I mentioned
Andreas Deptolla (36:37.733)
Mm-hmm.
Francis De Vericourt (36:45.27)
in silicon medicine. But I think, you know, the new technology that I will want to see more, and I know some people are working on that, is how do you train an AI machine that surprises us in a meaningful way? So there is a bit of a paradox here because I'm talking about surprising, right? So it's something we do not know, but it has to be meaningful. So it has to be something that we know.
Andreas Deptolla (37:01.948)
Mm-hmm.
Francis De Vericourt (37:14.734)
So that's where I see the big, big, big, big return in the near future. If you manage to do that, I think you can have a huge impact. But of course I can mention the current application that you can have in retailing, e-commerce, et cetera, where there's a lot of data and there's a lot of automatic tasks. In healthcare, still, we are way, way, way beyond what we can do with AI. Yeah.
Andreas Deptolla (37:43.988)
On the innovation front, you think about the intersection of the human and the machine, what are maybe, or do you see any limitations with the service machine? Because there's a lot of fear, so to speak, that the AI and the machines will take over.
Francis De Vericourt (38:07.606)
Yeah, so first I strongly believe that the machine, at least the way we know them and the way the technology is developed, will never replace humanity. I don't believe in the terminator story. No, know, why do I know? It's like, you know, maybe someone will develop a
Andreas Deptolla (38:30.162)
Mm-hmm.
Francis De Vericourt (38:36.184)
develop a new machine. But the technology as it is developed now, I don't see a path, quite frankly. And in the end, the innovation is still coming from us. The way a machine decide, I mean, we talk a little bit about that is, or make predictions or generate answers is basically, I'm going to simplify a lot, but it's...
Andreas Deptolla (38:42.334)
Mm-hmm.
Francis De Vericourt (39:02.772)
in the end using the data on which it is trained. And so the type of innovation it can produce has to be embedded somewhere already in the data. So it's really us humans who miss it because it's too complex, it's hidden, it's a blind spot for us, but it's there. The innovation is somewhere there, the idea is somewhere there, and we just do not see it. The way the human...
Andreas Deptolla (39:18.867)
Mm-hmm.
Francis De Vericourt (39:29.918)
human brain works, the way we innovate is radically different. mean, to go back in time, maybe it's not as inspiring, but if you go to Einstein, I mean, think about the guy, he was working in his patent office and he could make predictions about how the world works at the speed of light, right? Himself, he never went, you know, faster than 100 kilometers an hour, maybe two, but
You know, he could still imagine how the physics at the speed of light, he could imagine black holes and all that. And no amount of data could have done that because this is data that did not exist yet. It's like the black holes didn't exist. I mean, we never observed one and plus it's impossible to directly observe. And so our ability is to really innovate far away from the data, to really imagine the world in a very coherent way.
Andreas Deptolla (40:11.584)
Hmm.
Francis De Vericourt (40:28.716)
with not much data and we can do that because of a mental model as I briefly mentioned. And so there is a huge complementarity here where machines can switch through a massive amount of data, show pattern we overlooked, but we can take that and go way beyond what the machine can do. So when I say AI can help us innovate, the thing that AI does is to say, look at that. In this data, I see that. I cannot tell you why it's going to work, but you better explore that.
And then we take over and we explore and we reframe and we rethink and then innovation is coming from us. So I think now, I think the human will still be in common there.
Andreas Deptolla (41:08.636)
So maybe putting this in maybe naive sentence, right? What I'm hearing is maybe AI can inspire us, show us some blind spots, right? Maybe in our thinking, but then the quantum leap still has to come from the humans.
Francis De Vericourt (41:26.54)
Yes, and I can give you a concrete example of that. A couple of years ago, there's a team at MIT, one of the first of those type of approach where they train an AI system to find new molecules that could kill bacteria. And they find one that they call halicin. And the machine just said, look at that, right? Or look at those couples. mean, it output maybe 10 of them.
Andreas Deptolla (41:45.567)
Mm-hmm.
Francis De Vericourt (41:55.566)
machine says, I believe it's going to work. Why? I have no clue, but look at it. And then they looked at it and in the lab they could find that it was working. But the really surprising and important fact is that the way halicin killed the bacteria is a totally new way, which is instead of cutting the cell membrane or the cell wall, it would preempt, inhibit the ability of bacteria to process energy basically. So it would slowly die out.
This understanding is not coming from the machine, it's coming from the scientists and nobody thought about killing bacteria this way. And so now it's opening up, maybe it's not going to be highly seen itself, but let us explore new ways of killing bacteria that we didn't think about. So here you have machine to say, hey, look at that. I do not know why it might work. have no idea. I'm just telling you my data tells you that. And then now the human brain takes over and start to rethink the whole approach in a way that the machine can't.
basically. So that's what I see as the biggest potential for AI to help us and the role of humans in that equation.
Andreas Deptolla (43:06.128)
I assume something similar what you're describing about innovation applies for decision making as well, right? So if we look at and,
know, top management, think a lot of the decisions, I mean, of course, it's all quantifiable, right? Like I think that there's this huge element of like intuition, so to speak, right? And experience, right, of management. How do you think about that? like, how can maybe management for decision making interact with AI to come to better outcomes?
Francis De Vericourt (43:40.702)
So, you know, I would say that, so first management and executive should be careful with their intuition. There's not a lot of research and case study that show that, sorry for my French, but we suck at it. And intuition really work where we have been in very familiar context before, and we had an ambiguous feedback about
Andreas Deptolla (43:56.951)
Mm-hmm.
Francis De Vericourt (44:09.004)
or decision. So you take a medical doctor who's working in the ER emergency room for three years in a row, that you should trust her intuition. And the intuition, if you go deeper into it, is always coming up for a representation of the world. It's like, it's your brain pick up two, three informations and have an embedded mental model that connect the dots and say, you need this surgery, or I need to do, hire or buy this company.
But if you think deeper into it, you will be able to have a reason of the why. So I think in the end, good decision makers at that level who have an intuition is grounded in their mental model, in the framework they develop. And my recommendation to them, I I have many, interactions with C-level executives and I'm training them also is,
Andreas Deptolla (44:45.253)
Mm-hmm.
Francis De Vericourt (45:10.018)
The strength is to leverage diversity of representation. So don't get anchored on you only, open yourself to others, wait to understand the world and out of this diversity, something beautiful is going to come up with. What is new is AI, which I find really, really, again, inspiring and exciting is in the end, and I'm going to be bit nerdy, apologize for that, but.
why is our mental models working so well is because it constrains the way we think. And so really the innovation in the end is coming from the constraints we put on ourselves. I mean, I'm going to take you a little bit away, but why is Bach music so beautiful? It's because he pick up his own constraints and within those constraints, it creates beauty. Well, it's the same thing when we innovate. It's how we
Andreas Deptolla (45:50.331)
Mm-hmm.
Francis De Vericourt (46:03.822)
force ourselves to think that we'll create something new. And when we use AI, we do something similar because we constrain the machine. We constrain it by our choice of data and we constrain it by reinforcing the machine to look at this data in a certain way. If you use neural networks, it's going to be a hierarchy, structure of this data as different layers. If you use ChatDPT, it's about sequence. And we choose two types of different type of constraints. Then what is new in that?
It's not the human brain who is doing the job. You let the machine, boom, come up with something. And I think at the executive levels, it's like trying to learn to use these machines and these AI in this way, whatever their problem is, and then complement it with their own way of thinking. Maybe that AI will tell you, go to this market.
and why I do not know. So let me know, explore, and maybe I'll discover something that will help me think differently about my business model, my strategy.
Andreas Deptolla (47:05.45)
Yeah, it's super interesting that the idea of these constraints in our own heads, right? And I think going back to the recommendation that you gave to students, Probably the same thing applies, right? are probably like, companies the same thing, right? Like the bigger that we think, right? The more we will accomplish.
Francis De Vericourt (47:15.214)
completely
Francis De Vericourt (47:22.766)
I agree and there's something that I also saw since going back to the student and also I tell executive because it's a nice catch word but it's dangerous to try to think outside a box. And they study that show that if you try to think outside a box, you're less innovative, you're happier, no constraints, but you're less innovative. The beauty is choose the right box and learn to choose the right box and have many box to choose from.
Andreas Deptolla (47:36.509)
No.
Andreas Deptolla (47:46.909)
Mm-hmm.
Francis De Vericourt (47:52.492)
Yeah, that would be the advice I would give to students. AI can help you through the right box.
Andreas Deptolla (47:58.505)
That's great advice. Maybe my last question for today is what AI tools are you personally using today? Whether this is now in your personal life, right, in the academic world, when teaching, what has created value for you?
Francis De Vericourt (48:18.19)
I mean, I'm going to give you a very mundane answer, but ChadGPT has really embedded my life big time, both on a daily basis to write better, to translate, but also in research. So now I'm doing research using a script of VCs discussions with Venture. And then I'm using ChadGPT to say, what are the main themes has been discussed there? And you see how the
Andreas Deptolla (48:23.209)
course.
Andreas Deptolla (48:30.728)
Mm-hmm.
Andreas Deptolla (48:40.796)
Mmm.
Francis De Vericourt (48:47.63)
themes changes over time during the discussions. I mean, it's a bit of a programming and then I have amazing collaborators who also do that very well. So it's still a bit of effort, but it's this is the purpose. I'm sure there are better AI that do that better, but it's good enough for what I'm doing.
Andreas Deptolla (49:08.565)
That's great. Well, Francis, thank you so much for the time today. Would love to continue the conversation next time in Berlin.
Francis De Vericourt (49:11.479)
My pleasure.
Francis De Vericourt (49:16.706)
Same here, we should meet in person.
Andreas Deptolla (49:18.687)
Let's do that. Awesome. Thank you so much.
Francis De Vericourt (49:21.08)
Take care, bye.