El CEO de Flagship Pioneering, Noubar Afeyan, habla sobre health tech en Semafor World Economy

Good to be here.

Well, I'm going to be joining a panel with Gary Cohn from IBM and we're going to talk about how the current moment presents some opportunities and challenges technologies, particularly in my space of healthcare and the changes that are going on and how we can actually compete more effectively.

Well, there's a way that it's affecting every sector and it's also affecting healthcare, which is productivity, simplifying work, and making things more efficient. But then there's a unique way in which AI is affecting science, which is really completely different territory. And that is the whole scientific method is increasingly being leveraged by AI and then leveraging of AI in the sense that we're using generative AI to come up with new hypotheses, not one at a time, but hundreds of them at a time. We're using AI to design experimental protocols that otherwise hundreds of scientists would be involved in doing. And then we're actually running experiments increasingly in automated labs and then interpreting the data. That whole wheel of science that's historically been very important as to its output, but very inefficient and slow, is becoming AI-ified, if I can call it that. And it's really a marvel to see because we don't really know kind of where it goes. It's a little like Waymo for scientific discovery, to put it in a kind of a trivial way, but it's really potent and we're just beginning to see the effects of that.

Well, so many things I think have contributed to our innovation edge. I'd say first and foremost, our migration policy has been a huge part of that in the sense that we've attracted the best and brightest from the world and we've been able to incent them to stay here and to contribute and that's been a very big driver. I think that the policies that we use in terms of funding basic research that leads to some of the major breakthroughs. There's a whole continuum. People don't realize when they see a breakthrough, when some satellite gets shot up in the sky and does wonderful things, that where does that come from? And the answer is a consistent investment in basic research that follows by applied research, follows by entrepreneurial research, all the way to companies that can create these products. There is a real battle going on as to whether that research is worth it and whether we should just get out of the business of funding scientific research by and large and that's really scary in my view. Then there's competitiveness issues, you know, we're competing with China in certain sectors, particularly in biotechnology and we definitely are feeling the effects of that because they have a centrally planned economy that's focusing on some areas of relative advantage for them, particularly new drugs that can be developed very efficiently there and the US biotech industry is feeling that. Now, competition is good in general cuz it makes you even better, but competition based on policies that are very different as to what it takes to take something into the clinic and test it in humans or the ability to aggregate data from lots of places so you can make much better decisions. These are things that are essentially not possible in the US. We're not organized that way. So, we've been advocating that there needs to be attention paid to some of the conditions for us to maintain our competitive edge. So, I think that the challenges bring opportunities. The startup community that's very healthy in the US, we need to make sure we're protecting that advantage. Obviously, large companies are also competing. So, how do we create the dynamics right to make sure that happens? All of that goes into competitive advantage and innovation.

Well, AI, if you think about artificial intelligence, maybe I'll start with this. You know, we've had human intelligence as the only form of intelligence since humans invented language. So, one of the interesting things when you grow up you realize that animals are supposed to be not intelligent because humans said so. We have no idea if they're intelligent or not, but that's what we say. Now all of a sudden in the last few years we've made room in our lexicon for another form of intelligence. We call that artificial, it's machine intelligence. I'd rather call it machine intelligence. And it's a different kind of intelligence. But now that we've done that, I would say the interesting new emerging frontier is that we're going back and looking at nature and realizing that it too is all sorts of forms of intelligence. A cell in your body is a form of intelligence. Otherwise, you couldn't use immune cells to go after cancers. A virus is an intelligent entity in that it otherwise can't take down hundreds of millions of people around the world with infection. So, that form of intelligence we're beginning to realize can be better understood by machine intelligence than necessarily human intelligence. That's a really profound thing. I call this poly-intelligence, this multiple forms of intelligence. So, AI in my view would be a way for us to understand more about nature and do more good with it. Of course, we have to be careful of the bad that can be done with it, but the good can be better medicines, better foods, better climate mitigation, you name it. I think we can improve a lot if we understand how nature does what it does. And that's what AI is a key to.

You know, there's so many technologies including the ones that Flagship's been working on that are making biology a programmable field. You know, it's interesting when I went to school some 40 some years ago, you know, there was computer science and then there was engineering and basic sciences. And computer science since then has really taken off in information technology. That's a big deal. In the biological chemical sciences, we kind of haven't done things that differently. It's been incremental. What excites me is that we're beginning to now realize that biology and medicines can be programmable. We can start using the same advantages that you see in IT in terms of modular, reusable ideas that can be used to scale things in the biotechnology field, in medicine, in healthcare, in agriculture. So, what excites me is this programmability making a field that was quite artisanal much more industrial and scalable. And that from an economic impact standpoint and an impact to the globe is just we're just seeing little signs. And interestingly, the pandemic gave us a glimpse of that because we took a technology, our technology in mRNA at Moderna, and applied it to a global threat that created massive economic loss, trillions of dollars of economic loss. And just with that one technology, we scaled it from having only ever made hundreds of doses of anything to a billion doses in a year. There isn't IT examples of that where you can go from 100 to a billion units, iPhones, you know, the Fitbits and all. I think that's going to begin to usher in a new era where the scalability of biotechnology and its impact will be felt. And so now and one of the things that I ironically expect is that people's expectations of our field will go up as a result. And as expectations go up, capital follows. And as those expectations are met, hopefully increasingly more effectively, there's going to be even more possibilities. So, it's really an interesting flywheel effect that I look forward to.

We generally don't do anything in the short term. I mean, it's just something like we have an extreme version of innovation. We call this pioneering. We only ever work on things that are disconnected with the current reality. So, when we innovate, we base things on science that's not yet at all proven. And so, what we do is we try to figure out what if only it could be made is worth making and could be impactful. And then we come back from that vision to how do we show that it can be done? And a handful of the times that back-horse traveling journey is borne out in a sense that we envision something and then we come back and say how do we connect that with today's reality? And if we're successful, suddenly we've built something that should only have existed 5 years from now, except we do it today. What does that mean? We have patents all over the place. We can protect the idea. We're the only game in town. We can define the language around the idea. All of those things, what we call platform technologies, give us a huge advantage. So, it's all long-term. In fact, it's so long-term that our short-term is in service of the long-term, not in furtherance of the present, which is how most people think about short-term.

I think that in science, there's long-term uncertainty. And in the economic sector, there's short-term and long-term uncertainty, but more short-term today. We're in the business of actually resolving uncertainty. In fact, what you just used the word uncertainty is a big word for us. And just for your audience, I'll just explain. Uncertainty is a good word to use to describe things for which we can't assign risk. Because if it's never been done before, you have no idea if it can be done. Just saying it's risky is actually not true. You just don't know. The minute it can be done, and that's why we call it uncertainty resolution, now you can say how risky is it? How hard is it? How many can... So, we are in the business officially at Flagship at scale to take uncertainty and convert it into risk. So, short-term economic uncertainty effectively kind of affects our cost of capital. And so, if it's more expensive for us to access capital to be able to pursue these things, we're going to be able to do less of them. If we get more efficient with AI, we can do more with less. So, you can imagine how this dynamic is shaping is that kind of the more people have optimism around what can be done, the more they're going to put money in, the lower the cost of capital based on supply and demand, the further we can do with capital. But the less capital we have, we can't just give up. We just basically have to say, 'Okay, then how can technology get us to do more with that limited capital?' And that's exactly what we've been doing across Flagship. We have about 16 different companies that are all AI-based, all applied to science of all kinds, material science, life science. We have a company called Lila Sciences, which is a very, very transformative kind of next-generation scientific company, 3 years old, 300 people using AI to drive the entire scientific method. This is not something that presumes short-term economic certainty, but it is based on confidence that the long-term opportunity entirely lies in long-term uncertainty. In other words, if you only aim at short-term opportunities, it's a commodity. Everybody's there. If you start aiming at long-term opportunities, you will be the only game in town, and that's kind of what we're trying to advocate.

Thank you so much. Thank you.