Kirkpatrick: One of the things that’s really sunk in on us in the past few years, and particularly this year as we’ve prepared for this first full-day Techonomy Bio Conference—having done a half day at the same venue last year—is that the idea of being able to program life is getting more and more real. And for those of us, like all of the people at Techonomy, who kind of come from an IT mindset, where Moore’s Law and the Internet have changed everything in our minds, the idea that we’re moving into a whole new phase of socioeconomic progress, based on a new kind of programming that effectively goes even deeper into everything that surrounds us, and then of course combines with all that other Moore’s Law driven stuff, and Metcalfe’s Law, and all the other laws of extraordinary import in modern economic-techno life—the opportunities are just mind-blowing. So that’s why we’re here. And one of the key things that we’re trying to do here, and this first session will really focus a lot on, is this idea of the intersection of IT and life sciences—although, we’re going to go a little beyond that, but all four of these people could really go deep on that exact thing. They’re also pretty knowledgeable about the big picture, so we’re going to sort of blend both here.
So let’s get started. Welcome guys.
Waller: I’m director of scientific modeling platforms and responsible for observational research IT.
Kirkpatrick: That’s a fun thing. Basically, you’re a technologist who’s had your whole career inside pharma companies, but you have a degree in biology?
Waller: No, I have a degree in medicinal chemistry.
Kirkpatrick: Medicinal chemistry. But you’ve basically been in the IT side most of your career, right?
Waller: I’ve bounced back and forth. I’ve recycled a few times going through large pharma companies, started three times in large pharma on the research track and that didn’t last very long, so I’ve gone onto the IT side all three times.
Kirkpatrick: You kind of embody the theme of this conference.
Waller: I sit right in the middle.
Kirkpatrick: And when you hear what he’s doing with the Pistoia Alliance, which he helped create, and is one of our sponsors, you’ll see that there’s a lot of relevance there too.
Next to him, David Glazer, who runs Google Genomics, and also the Google Cloud for Life Sciences, right? I don’t know if you run that, but you’re very involved with that. And Google has made a big step towards getting involved in this arena and I want to hear you talk about that.
Steve Jurvetson, of Draper, Fisher, Jurvetson, one of the great VCs, one of our great friends at Techonomy, and somebody who has helped us get to where we are today, as Simone pointed out, and actually gave us the title for one of the sessions later, which is “Sitting on a Can of Miracles,” which he said on this stage last year on our investing session. Maybe you’ll have to define that again, Steve.
Finally, Drew Endy, who—well, actually, how do you define what you do, exactly?
Endy: I’m a bioengineer.
Kirkpatrick: That’s nice. Okay. Well, he teaches at Stanford too. So let me, actually, I meant to warn you—let me start with you, Drew. When I was talking to Andrew Hessel, who’s here and will be onstage later, from Autodesk, he told me he recently was visiting you in your office at Stanford and you had picked up a cellphone—he had said something about, “What are you doing?” and you picked up a cellphone and you said, “I’m trying to grow one of these.” What did you mean by that?
Endy: I’m trying to grow one of these.
Kirkpatrick: But how would that work?
Endy: So how is an object like this manufactured today? And could you imagine a future, if you live in the future and work backwards, where we could take surplus manufacturing capacity, now invisible to us because we just choose to ignore it, and say we’re going to repurpose a type of plenty we have to grow objects in a different way. So in Menlo Park, for example, the gardens of Menlo Park collectively ship 500 pounds of state-of-the-art nanotechnology that’s compiled from the atmosphere every year—that’s 16 million pounds for a city of 32,000 people—
Kirkpatrick: What, you’re talking about like—
Endy: Garden clippings, pinecones, grasses, leaves. We pay people to take that and compost it, and then we’re done with it. As we’ll hear later today, I saw in your program, you have folks from Ecovative and others. You could take a wood fungus that eats plant material and have that, recompile that material over a period of time to differentiate into an object you wouldn’t think of as being manufacturable via biology. It requires doing biosynthesis of hybrid materials. It requires program patterning—but I mean it quite literally.
Kirkpatrick: All those things you believe are possible and will happen in the not distant future?
Endy: Within two to three decades I think I can grow this.
Kirkpatrick: Really? Okay. And of course later today we have people growing bricks, so we’re moving in the right direction there. That’s probably relatively crude compared to a cellphone.
Endy: You have to put it in context, right? Look backwards in time where the Computer History Museum is proud that it’s been around back to the 1940s of tech, but what are we going to get good at over the next period of time? And if we got really good at engineering systems that handle bits, now we’re going to get really good at engineering systems that handle atoms, and biology is right in the middle of that.
Kirkpatrick: Very cool.
Endy: So that’s a lot of work to do.
Kirkpatrick: I think that’s a good way to start this day-long conference, because it’s big picture. And Steve, let me just continue on the big picture with you. When you said, “Moving from an oil-based to a bio-based economy,” five years ago what did you mean, and how far have we come, and where would you say we are in that spectrum now?
Jurvetson: Sure. Thanks. That, in sort of the heyday of energy and clean tech investing, was a metaphor I think that—
Kirkpatrick: Okay, so you were really talking about that?
Jurvetson: Well, no, but it was just oil is not a renewable feedstock, right? If you want to go to Mars and build things, you’re not going to use oil. If you want to like think of a solution over the last 500 years, it’s not going to involve oil as an input, and if it did it wouldn’t be for burning it, right? You might think of reorganizing those long chain hydrocarbons into something more useful, like a specialty chemical. But I think one way to take that comment, which is we’re moving from a traditional, inefficient, manufacturing era, right, where you poison the environment that you’re working in, you use a feedstock or a source of energy that isn’t renewable, and you just harness a huge amount of stored reservoir capacity underground, to a future where in every sense it’s renewable. The organisms that are manufacturing a cellphone in the future probably won’t poison the environment in which they live. They will do things in a much more elegant way. So if I step back more generally today, a few years on, I think what he said is a metaphor for the future, which is moving from a manufacturing designed, controlled, sculpted, and architected world where things do what you want to a grown, organic, robust future where the physical artifacts are grown and the software—the metaphor might be closer to parenting than engineering.
Kirkpatrick: Yeah, I like that. Okay, we’re going to skip David just for a second because you started nodding. So you’re in the healthcare, no, drug discovery and creation and sales business. When you hear that kind of talk, given that you are a trained scientist who’s also doing all this IT stuff, does that resonate with you?
Waller: I mean, it does. It’s a fascinating future to think about, right? And I’m interested—well, we’re going to have to talk more about this, but it does. I mean I don’t know if I have much more to say. I was thinking about “Interstellar,” actually, as he was talking because I watched the movie last night, and the future, a future world where we have to think about how we take better stewardship of the environment that we live in, right, and think about how we’re going to manufacture things and deal with things. And I don’t know that necessarily pharma thinks a lot about that currently. There are green efforts within it, but it’s not first and foremost.
Kirkpatrick: Well, it’s an interesting point. I mean, in a sense, you work in one of the biggest bio-based companies today, in some fundamental way, because you’re building drugs out of the tools of biology—certainly you’re interfering with biology with your products. And what you’re doing there at Merck is building modeling platforms, and working with the Pistoia Alliance, which is a collaborative effort of a bunch of pharma companies that you were the main person I guess, or one of the main people to help set up initially, eight years ago, when you were at Pfizer. So in a way, the creation of those kinds of platforms and systems could be said to be essential to get to the world that Drew was referring to later. Is that a reasonable way to think of it?
Waller: I think that makes a lot of sense, right? I think if you look back at what I’m actually trying to do with the development of this modeling platform that spans early concept through manufacturing, through marketing, through healthcare, right—how are people actually using the compounds that we’re making, and the entities that we make? What I’m trying to do is shortcut the process, because ideally, when we’re making drugs, there are a lot of questions that we ask about should we make it, can we make it, how can we make it, yadda, yadda. And a lot of information comes from the big data stacks that live underneath all of the various areas that we work. But there are lots of holes in the data, and we have to do a lot of experimentation. So I think I told you this story about rat experimentation, animal experimentation. Using models in lieu of animal experimentation not only saves animals, but it actually decreases the cycle times.
Kirkpatrick: You told me about a case of a specific drug project where you were able to eliminate the traditional phase of testing on rats by putting in a software component that basically mimicked the entire process of that and got to the exact same place on the other side.
Waller: We get to the exact same quality of compounds at the end of the day. We don’t completely eliminate it, because you do have to go back and do some spot testing to make sure that the models are still working. But at the end of the day, I’ll ultimately have a platform that underpins that entire process, that relies on information models in the absence—or to supplement information and knowledge, to drive that process in a much more efficient way and a biologically friendly way.
Kirkpatrick: I love that. Now, David, so at Google, which is a company we do know thinks a lot about data, you’re building platforms for something very analogous. Talk about what you’re doing and how it might contribute to this long-term trajectory that Drew and Steve were both talking about before.
Glazer: I think, to the, you know, the atoms and bits world—and that atoms, bits, and cells maybe is—what we’re trying to do is grease the bits, and if I think about the state of what’s been happening in the biology world and why I got excited about it, when I moved from being a software guy forever into trying to apply software in this space a couple years ago, is we’re at a transition point in the world of much of life science, genomics in particular, from artisan-scale to factory-scale processing, and that’s a tremendous opportunity. The folks who work on the wet side of things have gotten good, and they’re now starting to generate more data than they’ve ever needed to work with before, which means they’re moving right into Google’s traditional sweet spot. And we’re saying what can we do with the things we’ve already gotten very good at, with working with large amounts of information in many different domains and finding the value in them, unlocking that value, and in building ecosystems to help communities unlock value from large amounts of information. That’s about to be important in this area. Let’s provide the tools, provide the platforms, to really enable the kind of software-driven acceleration that happens—that happened in supply chains, it happened in retail, it’s happened in, you know, pick an industry. Anytime it becomes a software-driven and enhanced industry, you get this immediate acceleration. It jumps onto a new curve. Grease the bits of biology, life science, and genomics by applying what Google’s already good at to this new domain.
Kirkpatrick: So the presumption is, you know, Google has learned an enormous amount about managing gigantic datasets. This is one of the most interesting new gigantic datasets to deal with. Is Google doing it in order to make money doing it? Is that—what’s the motive, at the macro level, for Google to be involved in this?
Glazer: I think yes/and, right? Google’s overall mission is use technology to help humanity, and that’s ridiculously broad, so now you pick the places where, where are the things that we at Google can do that are playing to our strengths, that are the things that we can help more at than other people? And in particular, anything around taking 15 years of tools, infrastructure, and expertise for working with data applies.
You know, very practically, pragmatically, how can I make money immediately? Very simple. Google Cloud Platform is how we take our infrastructure and make it available to the world. And if you want to use our infrastructure to do anything, you can. If you want to use our infrastructure to do things with genomics, I’ll make it easier.
Kirkpatrick: Yeah, great. Drew, what do you think of this analogy of moving from artisan-scale to factory-scale? Is that the way you look at it? It seems a little different than the way you talk about it.
Endy: Yeah, I mean artisan doesn’t have to be small, right? You can have massive artwork involved in making a human therapeutic that goes all the way through trials—it’s an ad hoc process, incredibly expensive, and vulcanized as a vertical. The factory metaphor, or the platform metaphor, I think reflects an exciting transition that the group here really can make true, where all of the sudden you see organizations creating business-to-business opportunities that allow many other people to do many more things, right? So companies like Ginko, companies like Emerald, companies like Transcriptic, companies like Gen9, Synthetic Genomics, companies like Zymergen, they’re all positioning as platforms, like the services at Google, and potentially even the modeling platforms over at Merck. They’re all potentially, if you put a business-to-business wrapper around it, clicking into a future where many other people can get access to what has previously be a very exclusive artisanal, expert-driven process, and that is something I think is a cultural recombination between IT and hardware and wetware. And when I think about those intersections, it’s really the cultural intersections that are incredibly powerful to me. So for example, one of the things that makes synthetic biology go is that we’ve taken on the problem from hardware and software of, we’ll work on not just solving the problem right now, but we’ll also solve the problem of solving problems, meaning we want to get exponentially better at how we solve problems. That’s a cultural transition.
Another one, which is linked into these earlier comments, has to do with cooperation and community. GitHub, amazing, amazing free software. Amazing. How do you get people to be able to coordinate labor, reuse things over time and across locations? And if you open up that possibility, then collectively humans do things which is individually impossible. There’s no way I grow a cellphone by myself, if I have to work on that by myself. But if the whole world begins to work together on that, and many other applications, then we open up huge possibilities.
Kirkpatrick: And you’ve spent a lot of time building one particular platform yourself. Could you talk a little bit about that?
Endy: Well, it’s associated with the launch of what became iGEM, and what Randy Rettberg and others have taken forward, and that’s now shipped to 20,000 people, right, and has taken over the Hynes Convention Center in Boston every year now. The BioBricks Foundation is now a decade old, as a public benefit charity to bootstrap free-to-use languages.
Kirkpatrick: BioBricks Foundation?
Endy: BioBricks Foundation of free-to-use languages for programming life. When we got started, that was really a political stake in the ground, because the early days of this community were all about—one organizations would show up, they’d go, “We want to own it all,” and that just seemed really, really evil. And so BBF got started, said, “No, you know, we’re going to represent something different.” We’ve been punching above our weight class and ahead of our time—and I can’t share the details right now, but for the first time in the history of the BioBricks Foundation, we now have qualitative resourcing to go build some stuff that will really I think be transformative, and we’ll let you know about that in a couple months.
Kirkpatrick: And a lot of your work’s been on genomics, which we haven’t even really mentioned, but the idea of this—talk about datasets, right?
Endy: Yeah, yeah.
Kirkpatrick: For doing all these genomes, and each one is massive, right? Of individuals and other stuff, and figuring out how they compare to one another, what the variances mean, and—
Endy: Yeah. Look, you’ve got to take an engineer’s view on this. We sequenced on an order 1015 bases of DNA. Earth has 1035 bases of DNA on it, total. Right, so we’ve got 1020-fold to go. We should finish by 2090, this century—
Kirkpatrick: But finish doing what?
Endy: Sequencing everything.
Kirkpatrick: You think by 2090 we will sequence everything on earth?
Endy: If we haven’t done that, we’ve failed, right?
Kirkpatrick: Well, I’d say given Moore’s Law and all that stuff, it’s not crazy to say that.
Endy: But just live in that future for a moment and work backwards and say how do we maximally benefit from having reached that goal and how do we discover all the functional genetic elements? How do we discover the things that we shouldn’t have to worry about, right? And how do we remix those in support of making new things?
Kirkpatrick: Good. Steve, take us further, whatever you want to say.
Jurvetson: I can build off what he just said, because it certainly resonated—this always happens, whenever Drew speaks, I get very excited. He’s one of my favorite professors at Stanford in this regard. So you mentioned something about these platforms and learning occurring about how something works and how profound that is, and I want to dwell on that for a moment. It’s almost like, if you think about the greatest inventions in history, the scientific method itself is probably one of them. And I think we’re in the middle of something not quite as profound as that, but it’s close, and that is a massive shift in how we do engineering. If you think again about this future you described and chain back to the present, that’s a future world where we are dealing with datasets and building complex systems—be that software or things—that exceed human understanding, that are, without a doubt, beyond the capacity of anyone to ever understand, even teams of people. Yet we will be building artifacts—and we are already today, but this future world, most of what we build will be in that domain. And I use the world ‘build,’ by the way, generically. It could be grown, it could be parented, it could be ushered into being—if we ever get to AI topics. But what it means is, there are different methodologies. It’s not the Germanic approach of reusable subcomponents that you fully understand and an artifact that you’ve built where every element of it is fully understood and nothing goes wrong. It is a bit much more like the biological world or the parenting metaphor. And I think today where you see that are experiments in high throughput experimentation, experiments with directed evolution, experiments, frankly, on the IT side, in machine learning and deep learning, that are, in my mind, almost synonymous with this whole methodology. How do you set up a series of iterative algorithms to grow software, perhaps mimicking the brain, perhaps mimicking something else from biology, that’s very powerful, that allows us to transcend human limitations, to do just like evolution itself, to build complexity that’s greater than that of its antecedents?
Now, that’s all very abstract, but we see it in the companies we invest in today. They can build products, be it a microbe that makes a chemical or a piece of software that recognizes faces in a crowd, that’s better than any engineered artifact, right? And that’s now creeping into medical imaging, where you can do better than radiologists, or better than the people reading mammograms or reading MRIs, using these software techniques that mimic biology. It’s got to be the case in the not too distant future, when you’re designing multicellular organisms within which there are regulatory feedback loops—this complexity compounds so dramatically that you’re going to need approaches like the Google approach to everything, if you will. That data science approach to everything, writ large, in the sciences world. So it’s a different way of doing things.
Kirkpatrick: Great. Anything you want to say, but I have a question, if you want to go.
Waller: No, I was just going to say, I love the vision, and I think the challenge that we’re going to run into, and I’m struggling with every day—Drew and I talked a little bit about this—to make this happen requires a different culture, though. It requires a different behavior and a different way of working. And I come from the private sector, right? We are very—let’s put it this way: large pharma is very risk-averse, right? So—
Kirkpatrick: Yes. You’re starting to go someplace I was going to push you to go, but keep going there, yes.
Waller: So we’re incredibly risk-averse, and so I think about the things like the Google experiments and the kinds of stuff that Drew does. We’ll toy around with some of those things, but the challenge and where I’m trying to get to with my platforms and as horizontal thinking is to get these groups that historically work in their little silos and keep their information and their knowledge to themselves, because that’s in fact how we reward them as a company, to think about working up and downstream of each other, such so we can actually see these benefits that Drew’s talking about.
Kirkpatrick: Well, that’s why you should describe the Pistoia Alliance briefly, because, you know, you do work in a big company—and Techonomy has, in many, many of our programs, tackled this question of how do the big existing companies survive? I mean, as basic as that. And you have been a pioneer at trying to get them to think differently about collaborating. So talk about that.
Waller: So in 2007, when we started thinking about this—other industries have tackled this problem, right? So semiconductor, right, you think about the semiconductor groups who decided they couldn’t now keep doing the same duplicative research within their walls, so why don’t we do some things together, and the challenge of that becomes defining what is precompetitive in that space, what can we do together, yet still differentiate in the marketplace? Because in fact, we are companies and we have to make money. So when we put the Pistoia Alliance together, it was with that mindset that we’ll bring folks from the IT side largely, because we’re building these technology solutions at, once again, duplicative at the exact same companies, doing exactly the same things, so why couldn’t we get together, talk about these things that we do within the workflows across R&D that are in fact precompetitive, where we should be standing up platform solutions, or at least singular solutions, or at least solutions that can talk to each other. And that’s how we started in 2007, and it’s evolved to the point where—we’re in 2015 now, right—where we really are pushing that notion of platform, just what can we stand up as a precompetitive, if you will, platform, and to Drew’s point, that we can stack together with other platforms to come up with useful solutions?
Kirkpatrick: Right, and this is so interesting and pertinent to this idea, because it is so daunting when you hear something like sequencing every genome of every living thing by 2090, and then we have this economy we live in today and how do we get from here to there, right? So we were talking before about why shouldn’t all the pharma companies essentially really double down on what they actually are best at, which is the science and the scientists they employ and the research they do, and not waste a lot of time and money with systems that should be shared that are way beyond even what you’ve got today. So you know, could we get there?
Waller: I think we’re going to have to get there, right? I mean, one of the challenges is do we have to wait until we melt down, right? So one of the things that we talk about a lot is, what caused other industries to change, right? What forced them to think differently about how we do things? So you think about the financial industry, that had to completely burn to the ground until they realized that data was in fact an asset and they should treat it differently.
So I think we’re trying to learn from those industries and not burn down and leapfrog to this happier space. I think we’re going to have to, right? And we can argue about what pharmaceutical companies really good at. Large pharma companies are really good at sales, small pharma companies are really good at research, right? So I think we probably ought to separate—
Kirkpatrick: It’s the science of the sales is really how you differentiate.
Waller: Yeah, exactly.
Kirkpatrick: And David, you know, one of the interesting things Drew talked about before is this cultural intersections, you know, and Google is a company of the modern economy without any question, right? I mean, you are the ultimate data-centric company, I’d say—you and Facebook, maybe, and Amazon—I don’t know. Facebook and Google in particular, like the ones that really have the giant amount of stuff. And culturally that puts you in a separate category, in a way, so you’re trying to bring that capability to all these other parties, right? Does the cultural challenge of helping them understand what they could do with data occupy a lot of your thinking?
Glazer: It’s not so much the opportunity. I think everyone gets that. I think the people who are excited about the space are excited because they know there’s gold in them thar genes, right? They know that there’s something—
Kirkpatrick: Right, gold in them thar genes, yes.
Glazer: They know there’s something to find and unlock there. I think there are some—when I think about what Google brings, it’s not that we have data. We have no relevant data in this space. We have, like you said, the culture and the expertise. A small example, and it’s around the data side, but also the ecosystem side. So around the same time that Google started working on Google Genomics, the Global Alliance for Genomics and Health was organized. And Global Alliance for Genomics and Health is a couple hundred-plus people now, nonprofit and commercial, working to provide a framework for advancing, accelerating the work in the space.
Part of that, and the part that we at Google has been particularly active in, is the data working group, and what we’ve contributed, as much as any particular expertise, is just the mindset that if you want to have an ecosystem of people working with this kind of information, maybe you ought to define some APIs. Maybe you ought to define a way that you can have interoperability, so somebody building a tool and somebody providing data, any tool can work with any piece of data, right? Henry Ford had this insight a long time ago, but it’s not yet a habit at artisan-scale, because it’s actually slightly counterproductive at artisan-scale. It’s better for you to just make them fit. But if you want to have an ecosystem—we take it for granted at the scale we’re used to work, you have to have that good separation, and that’s an example of where the cultures are actually blending in a very good way by saying here’s an area we have some experience that applies in a new domain.
Kirkpatrick: Okay, I want to throw one question out to all of you, and then I want to start hearing questions and comments from the audience, because I’m sure, given the scale of things that are said up here, a lot of you have a lot of thoughts.
When I said before that Facebook and Google were the two data-centric companies that stand apart, I was thinking after I said that, now, that’s not entirely true. Baidu, Alibaba, and Tencent are three others that have a lot of those capabilities, possibly not at quite the same level, but probably if you were looking at all the companies in the world, they’re like way close compared to most. The question—last year, on this stage, Nancy Kelley, who’s on our program later today, who’s very deeply involved in the genomics community in New York, gave a fairly alarmist talk about US competitiveness and her worries about some of the things that we are not doing in this country that other countries are doing. I’m curious, any of you, when you look at the landscape, do you think about, worry about how the US is going to be able to play in this arena, vis-à-vis other countries? Are we distinctively ahead now?
Actually, one of you said that—I think it was Drew, I think you said we’re definitely ahead now, if I’m not mistaken. But I’m sure—anybody, are we ahead now? Will we stay ahead? Does it matter? Who wants to tackle that?
Jurvetson: I think we’re ahead. Europe, luckily, you know, if you think of it competitively, is crippling themselves with their stance in GMOs. I just don’t really see them leading the charge. The sequencing effort in China at BGI used to be the largest in the world, but both Google—
Kirkpatrick: Beijing Genomics Institute?
Jurvetson: Yeah, exactly. But I know Human Longevity Inc. with Craig Ventor, and probably Google and others, are planning this year to surpass that in their aggregate human genome sequencing, both healthy and cancerous cells, in the case of HLI. And there’s these new machines that just came out from Illumina that are just knocking the socks off the predecessors, so it’s kind of one of these arms races, driven by Moore’s Law, that gives us that potential dataset lead. But I think more importantly is the algorithmic and sort of general software tech strength that we have, as well as some of the interdisciplinary learning that we do here at institutions like Stanford and elsewhere, where, you know, bioengineering as a concept is well understood by fleets of students, and a lot of the software talent—let’s say, take Google’s approach to just about everything, the sort of deep learning, fungible approach to large datasets to find patterns. That is going to be very powerful in this domain once the datasets reach a certain critical threshold, which we’ll probably hit this year, where you can find patterns humans would miss across multiple variables in regulatory genes and things that are just really hard to just see the pattern.
You know, I’m not saying we’re going to have breakthroughs this year, but the methodology will start to bear fruits. You never know. I mean, there are a lot of areas where China is being incredibly creative. I know Microsoft thinks that China teams more creative and productive than their US teams. So our visibility there is less, I’ll have to admit, so there’s an observer bias, but just from what we see here in America, much of the action and much of the excitement is here.
Kirkpatrick: Drew, did you want to chime in on that?
Endy: I just strongly agree, it’s ours to screw up. And to add some more secret ingredients that are in our favor, the Department of Commerce in the United States, with their investments in NIST, National Institute of Standards and Technology, now—you know, yesterday in the federal register, there’s a public meeting on a standards consortium for biotechnology to be held out here at the end of the month. And they build the sewers of science and technology, the things you take for granted. If you want interoperability, you want scaling, you want massive public benefits to result, you need them to be working on our behalf. The Department of Justice, right, and Ed You, who’s here, and his very forward-looking strategic view on how to create constructive culture around biosafety and biosecurity as we head into a second genetic engineering generation.
So you know, there’s just so many things that we have in our favor right now. We could always wish for more, we could always wish for more cluefulness, but some of the secret ingredients we have in the US really leverage the strength of government to convene and promulgate in ways that produce chances at creating exponential returns. I haven’t seen anything like that anywhere.
Kirkpatrick: Okay, interesting. Anybody else?
Waller: I’ll just put my Pistoia hat on for the moment and take the counterpoint and say that Pistoia is necessarily international and interdisciplinary—
Kirkpatrick: You have 70 corporate members now.
Waller: Yeah, 300, 400 individual contributors to the effort. And I think as we think about doing—like you mentioned the supply chain mentality and things like that, and the kinds of software companies that provide solutions in those spaces are not necessarily US companies. So I think there are things that we can learn and transfer from other—from companies outside the US, to help support the revolution that we’re talking about, and that’s something that I am keeping forefront in my mind.
Kirkpatrick: Okay. Who has a comment or a question, in the audience? Who is daunted and amazed by what they heard, or think it’s bullshit? Go ahead. Is that VJ? Oh, good. Well, I always like to hear a question of VJ. Please identify yourself.
A: Yeah, I’m VJ. Drew, I have a question.
Kirkpatrick: Okay, VJ used to run the HP Printer division, which was a $28 billion dollar business. Go ahead, VJ.
A: Yeah, so Drew, can you give us a roadmap of, you know, what will happen in next, let’s say five years when we’re able to grow the cellphone, and what are the key checkpoints on that roadmap?
Kirkpatrick: Great question.
Endy: Yeah, so, hybrid materials, biosynthesis, being able to take metals and other things out of solution and bring them into a structured pattern, you know, there have been demonstrations that that’s possible, but scaling up our capacity to make things that you wouldn’t think of as biological, just connecting biology to many other elements on the periodic table on a systematic basis, that’s one principal component. Second principal component, if you think about software and IT, would be the development of ways of managing complexity associated with growing patterns. Many of our programming languages are one-dimensional, for controlling systems in time. How do you develop multidimensional programming languages? You start checking some things off.
Other dimensions of this would have to do with, you know, do you envision transitioning from a nineteenth century model of trending towards centralized industrialization to a post-industrial twenty-first century model, where we’re flourishing in partnership with natural manufacturing capacities and much more distributed manufacturing, and then you’d apply that to the space and you’d go, okay, well the world of DNA construction, we’ve seen a trend towards centralized DNA construction on the inkjet print heads and other things. You know, what would be a next-generation technology for DNA construction, and would that ever go back to the desktop, right? And so you can begin to roadmap in ways that are aligned with big variables associated with centralization versus distribution.
Kirkpatrick: So wait a minute, will it go to the desktop? I mean, are we going to be building DNA—I mean, am I going to do it in my den?
Endy: There’s reasons that the answer would be yes and reasons that the answer would be no. And it depends who does it.
Kirkpatrick: Wow. Go ahead.
Glazer: My son is in eighth grade. He just did a science fair project transforming the DNA of bacteria. And it was not particularly cutting edge. It was, here’s how you do it, here’s how you get the material, let’s do the work, it worked. It shocked me that that was a good but straightforward eighth grade project.
Jurvetson: Can I ask Drew just a quick follow up?
Jurvetson: So are genetic computers part of your vision in how this happens, the idea of embedding memory registers in logic, or do you think—
Endy: Yeah, I mean, people think about Moore’s Law and how that’s going or where it might not go. But Moore’s Law should be thought of as multidimensional, meaning it can be expanding computing into other spaces where we have no computing yet, and so any improvement, you divide by zero and it’s an infinite advance. And so operating a byte of computing inside every cell in your liver seems like a reasonable advance to me. A massive advance, actually. And obviously, if you’re thinking about controlling patterning in space, you need modest amounts of memory and logic. You also need communication. But I think you’ve said something, that I heard through a different conversation, but you said it again out in public today. This metaphor of parenting, that our future technology, who will be parents of them, I think is right. One of the things is, as we’ve gotten more into growing complicated structures in biology, that I’ve been taught by the biology is, it’s just holistic in its approach for generating complexity. It’s everything from the computing we might control explicitly to the micro-scale geometry or shape of a cell and how that interacts in strange ways over massive light scales. And little changes in that result in a checkerboard or a yin-yang, right? It’s like, what? And so that we’re learning to become parents of these more complicated technologies really hits me in a positive way.
Kirkpatrick: I like that metaphor too, but quickly, Drew—when you say the thing about the liver, are you saying that we could essentially have biological computers inside us, or inside anything, that would essentially analyze and decide and respond to—
Endy: How could I not say that?
Kirkpatrick: Well, we have that, but we didn’t make them. We have them now, right—
Endy: Look at what’s happening with T-cell therapies, right? And you start with something that’s approximately a natural thing, but people are now beginning to wrap control around that, so you can control the proliferation of the T-cell in a very specific way, using a preapproved clinician-administered small molecule. It’s not too many steps further to create some memory and some more sophisticated logic.
When we think about health, and global health in particular, the idea that we’re going to solve it only with software and hardware, and not wetware—like using biology—medicine is limited by our ability to get information into and out from people, and biology is already inside people, right? And so there just seems greenfield opportunities to use biology as platforms for diagnostics and therapy themselves.
Kirkpatrick: Okay. Okay, over here, sorry. Identify yourself.
A: Yeah, Terry Sejnowski, Salk Institute. So we can now do recombinant DNA technology in grade school, I just heard, and we now actually have recombinase technology, this is CRISPR/Cas9, which allows us to do that in humans. And I wonder where that is heading. Recently a group of academics suggested that maybe we should have a moratorium on doing any research like that on humans, and I wonder where are we heading with that?
Kirkpatrick: Thank you for asking that question.
Endy: Yeah, the “New York Times” article, the science editorials from Dave Baltimore and others—it’s interesting how it’s a reprise of the 1975 conversations around recombinant DNA. I’m on a committee with David, and you know, his representation of the puzzle is—and it’s specific to human germline engineering, you know, making changes to us that would then be inherited by future generations, and basically how we’re not in a position to make those decisions for future generations is one of the points he brings forward.
You know, there’s a lot of other things happening specific to these technologies around who invented it, who owns it, who will benefit from it, and so some of these eruptions in the public conversation are coupled to these underlying currents and I haven’t detangled them.
Kirkpatrick: The public conversation is more on the ethics, but you’re saying there’s a lot more commercial weird stuff also underneath.
Jurvetson: I don’t think you would decouple it, by they way—
Kirkpatrick: Say what?
Jurvetson: When you engineer something that fundamental, you also have the capacity to silence it, so you can detach agency from effect, meaning you could give the next generation the opportunity to turn on a silenced chromosome, if they so desire, without making the choice as a parent.
Kirkpatrick: Oh, wow. Any thoughts from Merck on this general—no, you don’t want to get into it.
Kirkpatrick: Okay. Another question or comment here.
A: Hi, my name is Ioana Cozmuta, and I support the space portal here at NASA Ames Research Center, so my question comes from space. And since you’re talking about revolution, specially looking from the perspective of the commercial space and the fact that the government broke the boundaries and translated a lot of capabilities into, again, the commercial and encouraging public-private partnerships—two questions. One is more science-based. Have you thought about gravity-limited phenomena, and what can we do in space, and how can we take that into building, manufacturing, clean manufacturing, that ties to clean-tech, and leveraging upon resources from space so that we create a more environmentally friendly manufacturing on earth?
Kirkpatrick: Was that one question or two? That was one good one, anyway.
Glazer: Before getting to the serious answer, I’m going to repeat the aside we just had. At the very beginning, when you were talking about all the different laws we should discuss, we were joking, well, what about Newton’s Law? Newtown—you thought it was a joke. What about Newton’s Law?
Kirkpatrick: It’s not a joke, right. What about it? Does Merck think about what you could do in space?
Waller: No, we don’t.
Kirkpatrick: But maybe you should. I mean, who knows?
Waller: Perhaps, yeah.
A: But would you be open to suggestions of what Merck could do in space?
Kirkpatrick: I mean, I think that was a great question to ask, and that’s probably all we needed is to hear—
Jurvetson: Well, I mean she knows part of the answer to that question from where she comes, that there’s been a lot of pharmaceutical research that’s made good breakthroughs. You can do some interesting crystallography without gravity interfering with certain things to understand what’s going on in protein. You can discover, in some ways, the effects of gene expression because you have a radical shift in gene expression when you don’t have gravity. Take the human body, enhance some of the impact that astronauts feel. There’s I think over—well, it’s a huge number, it’s like well over 20 or 30 different significant genes that change in the kidney alone in a human when they’re in space, in close to a zero-G environment.
Kirkpatrick: Oh yeah, you’re an investor in Space X, so you know a lot about this stuff.
Jurvetson: Well, not a lot, but I know that there’s a lot of potential research that can be done, perhaps more in life sciences and biology than just about any other field I can think of. Perhaps some material science stuff as well. And we just haven’t done a lot of it, right? I think the space station is just now reaching its relatively golden period of ability to do work, now that construction is slowing and science is going up. But it can do a lot more, and that’s a whole other conversation. You know, I’m fascinated by it, and there’s a whole related field of astrobiology, but that’s slightly different from your question, which, you know, I think also is very important, is how do you—
A: I’m really talking about practical things, like practical things, so can we create a consortium in the Valley, where we put our free-floating laboratory and several parties would explore the next cutting edge for innovation, in terms of what is gravity-limited on earth and then creating superior products that would address a lot of markets and economies?
Endy: So we spent a week talking about space and biotech in the introductory courses on campus, and what comes from that is a lot of interest in learning how to sustain closed ecosystems, right? And the experiment with Biosphere II, re-appreciating that and sort of doubling down on that for the future. The second thing that comes up is in situ resource utilization, so how you could reduce payload requirements if you went to Mars or elsewhere, and you know, for example, if you’re actually living on Mars and you needed a medicine, you’re probably better off if you could make it in fermenter with instructions coming over via radio waves that make the DNA, that program the fungus to make the medicine that you take, as opposed to waiting for I don’t know how many days for the supply ship to come with the medicine.
The question I’d return to you is, you know, you’ve got this amazing agency, NASA, but within it, you’ve got the Planetary Protection Office, and it strikes me as this organization which in its heart and soul is championing the Enlightenment, and is being overseen by the Inquisition. And so many of us here—I don’t want to speak for everybody, but I’ll speak for the generation after me. We kind of want to go somewhere else to learn how to be better here too. Not to leave, but to do it all together, and there’s this really weird gatekeeper inside NASA that is sort an ideologue, destroying that possibility.
A: So I think, since it was question, okay, the answer to that is, again, break that boundary open. What I think we ought to have a discussion with the Planetary Society about the definition of an open system. The universe is an open system, so you cannot—you have to adapt the policies and understand that you cannot create a closed system because that’s physically impossible.
Kirkpatrick: Okay, well that is very interesting.
A: So but I think it’s a good topic for a—
Kirkpatrick: Well, I’m really glad you asked that question. Who’s got another comment or question? Over here.
A: I’m Jeanne Loring, from the Scripps Research Institute in La Jolla, California, and Terry stole my question, but I want to follow up on it. So David, something you said reminded me of the fact that it is extremely easy to do these methods now. So the reprogramming of skin cells into very potent stem cells that can then generate everything, including gametes, is now done by the high school interns in my lab. So this is the kind of thing that people will be able to do in their own homes.
Now, my lab is one of those that could actually do germline changes, and I’m not alone, by any means. So I think what we’ve been talking about among the scientists is that this has probably already been done, probably in China—we blame China for all this kind of stuff, but in fact, they don’t really care about the patents, right? And since it is so easy to be done that I could actually teach a high school student how to change the genome, the germline genome of humans, how do you think we can really keep this genie in the bottle, this Pandora’s Box closed?
Glazer: I’m not sure I heard anyone here say we can.
Kirkpatrick: Is Juan Enriquez in the audience yet? He’s talking later about his book—well, anyway, we’ll get into some of that when we have him onstage later, because he has a new book called “Evolving Ourselves.” And the first conference I ever organized, which was in—what was it, 2000—for “Fortune Magazine,” Juan was there—Steve might have even been there, I don’t know. And Juan was talking about, “Oh yeah, we’re going to have kids putting wings on frogs, I mean, that’s happening,” and he’s really spent the next 10, 12, or 15 years thinking further about what does it mean and where should we go and not go, and basically his book is about, well, we are already in total control, we have to own that.
So I mean I think it’s a classic question for the modern age, which has even bigger ramifications, because we have control now. I mean the stuff that Drew does and talks about so eloquently, holy shit, I mean, we have to think a lot harder about what we want and don’t want to do, and think about the cultural—again, cultural and social and legal and ethical global implications.
Endy: I think Pandora’s Box is pretty interesting, and I need to be a better student of it, but one of the things I’ve learned is that there was something left in the box: hope. And it’s sort of a short way of echoing what you’re saying. We have to—I get strange questions sometimes, like how do I feel about people working in garages, right? And I get this on campus at Stanford, and it’s strange because we build temples to garages, right, and yet somehow in this space, the framing is not the same as when Bill and Dave started H&P, right? The framing is the garages will be occupied by the garagistas. And so what’s up with that? And maybe it’s true, but I haven’t seen it yet.
Jurvetson: We have a commodity GlycosBio in Houston that literally started in a garage, and I have no idea in that—
Endy: Are they garagistas?
Jurvetson: No, they were like, “We literally started in our garage,” like in the old school methodology of, you know, used aquarium equipment and bizarre use of materials, and they’re a thriving company.
Kirkpatrick: When Juan comes we’ll definitely go down this route, but I thought—on the phone when I was talking to him, I was saying some of these questions, and he was saying, “Hey, we’ve already changed the environment of half the planet. Let’s take ownership of that, to start.” I mean, we are already screwing around with all this stuff at massive scale. That’s like climate change, among other things, so—anyway, please.
And I’m not saying that’s a positive or negative statement. It’s just reinforcing this idea that—
A: Hi, I’m Jacob Corn, I’m the scientific director of the Inhibitive Genomics Initiative at UC Berkeley. We actually—so I’m a co-author of this perspectives piece in science that was covered by the “New York Times.” We organized that Napa meeting about bioethics. The intent that we had in this meeting is really not to—it’s really to introduce a pause. It’s not to introduce a moratorium. The idea here is we stand at the brink of being able to change human evolution basically faster than evolution can change it, right? And yet, we don’t understand human evolution to any complete extent. And so I think it’s incredibly exciting what we can do, and the question is not should people be doing this in garages, should bio-hackers be trying this stuff—there’s a lot of really exciting, interesting things, but the question is, if we do this on a systematic level, as a society, as a species, is that right? And I think that’s something—it’s not a question just for scientists and labs and people in garages. I think it really needs to be a question—since it affects all of society and it’s heritable change that we pass down through generations, that’s really a question for society as a whole, and I think that our intent was to open that box up so that it was not a conversation just happening in labs, but a conversation happening on the street corner, people talking to each other about that.
Kirkpatrick: Great, well thank you for saying that, and thank you for being here.
Jurvetson: Can I just say a quick comment?
Kirkpatrick: Yeah, real fast.
Jurvetson: I think part of what’s scary is we want to understand the artifacts, the things that we build, but we can learn more and more about the process of their creation. So you don’t know a lot about being a parent before you start it and the knowledge accumulates slowly and you have a huge impact on downstream generations. You’ve just got to focus the locus of learning from product to process.
Kirkpatrick: Okay, we’ve got six seconds left on the clock, we’ve got two people who want to make statements, so you go—I told him first, and then we’ll get the guy with the orange shirt, okay.
A: So is government regulation going to impede growth? It certainly won’t lead it, but in my view, science is so far ahead of regulation that we’re going to come to a little crossroads and have a bump in the road. Do you think that’s going to be true, in terms of government and other governmental interference with the growth of the science?
Glazer: I think it will do both. I think there are examples of regulatory frameworks encouraging ecosystems to thrive and examples of them slowing it down. Yesterday, the NIH issued a new statement, saying it’s okay to use NIH data in the cloud if you’re careful. Took a little while, but it’s a good thing, right?
Kirkpatrick: Techonomy has a whole conference called Techonomy Policy in Washington on June 9, because we are very concerned about that very question. Next. Please identify yourself and make it fast, sorry.
A: I’m Romie Litrell, I’m at the Tech Museum of Innovation, and speaking about the Pandora’s Box question, we have the other side of the coin is how do we educate and inspire these young people who want to get into this and create spaces where it’s not in the garage, but with this—you’re trying to solve problems for the next few decades, but the topic’s going faster than the curriculums can even catch up. So do you have any models for how we can create spaces where people can do this safely and creatively?
Kirkpatrick: In five seconds?
Endy: Tell stories.
Glazer: Tell stories and share examples.
Endy: GenSpace, FBI, iGEM keep going, going, going. I mean—
Jurvetson: Teenagers are building microbes with—
Endy: Yeah, none of these—all of these are fantastic, none of them are sufficient. They all need to work together, we need a lot more of it. We could do a lot more of it. It’s good bang for buck.
Kirkpatrick: We have a lot of community bio labs, people on the program as the day goes on, so that will re-emerge throughout the day. This was a great panel. You know, when the panel is better than the phone conversations you had preceding it, that’s always a good sign. And the audience is quite impressive too. So thank you all.