The Deep View: Conversations

Ian Krietzberg:
Welcome back to the Deep View Conversations. I'm your host, Ian Kreitzberg, and today we're talking about programmable plants. As you can see, if you're coming in here on YouTube, I have dressed for the occasion, I'm wearing green. But my guest today is Dr. Brad Zanft. Now, Brad is the CEO and founder of Heritable Agriculture. Heritable recently spun out of Google's X, it's Moonshot Factory. with a basic mission, vision, goal of leveraging artificial intelligence and machine learning systems as a means of processing vast amounts of biological data which it then can use to program plants. The idea is the combination of these methods can be used both to increase resiliency which increases food availability, while also boosting the sustainability of the agricultural sector, which is notoriously not a very sustainable sector. It's early stages, but the impact, as always, is vast. Brad, thanks so much for joining me today. Wonderful to be here. Yeah, I'm really excited to talk to you. I know you and I have had the pleasure of connecting before we actually jumped on to the call here today, but I came across you guys, I came across Heritable Agriculture probably a few weeks ago at this point. I was working on building out the newsletter and found an option for my AI for Good section, which is something I try and do. And I was just very intrigued. Um, so that led us here today and, and I want to start that there's, there's so much to talk about as always. Um, you're doing really interesting work. That's very scientific and, you know, deep into biology and machine learning models and all these other things. But I want to start with the journey. Like this, this has been a long process for you. to get to the point where you have this company that is starting to enter into operations. And so can you walk me through from inception of idea to where you are now, what it's been like to put this together and what you're trying to accomplish?

Brad Zamft:
Just generally, I've sort of always been passionate about both technology and sustainability, you know, how we can use technology to have a better use of our planet and be in more sync with our planet. And the journey that I suppose I'll tell you right now is really about figuring out that agriculture is the biggest lever for that. So my academic training is In synthetic biology, I sort of had a front seat to the early 2000s when folks like Drew Endy and Jay Kiesling and George Church were formulating, you know, the kind of biology as a computer vision of synthetic biology. And it was really fun and really exciting and super technical, inspiring. And in a lot of ways, a lot of the things that I was a graduate student at the time, a lot of those concepts and the technologies developed therein led to some pretty big advances in the industrial biotechnology space. So using microbes mostly to generate new molecules and have big, impact, sustainability impact, in the way we produce materials, if you will. So then I went and did a postdoc with George Church and continued that fascinating journey on thinking about biological systems as effectively computers. I had always sort of been into politics and policy, just understanding how science goes from the act of doing science through to impact and actuation. So I got a fellowship to go work at the Advanced Research Projects Agency for Energy's Department of Energy, the DARPA for the Department of Energy. And there I helped run grant programs on biofuels and on automated phenotyping. So making fuels in the tissues of plants themselves as well as using robots and drones and whatnot to measure those plants. And that's where the light bulb sort of struck that my passion for building a sustainable civilization and my interest in biology Really, the place to have the biggest lever is in agriculture. Agriculture – plants are grown on most of the terrestrial landmass. Agriculture feeds us. It provides – it can provide fuel. It gives us fiber. And the potential for agriculture is massive, but actually the way our agricultural system works today, largely because of the stresses placed upon it, it actually is a large amount, about a quarter of anthropogenic GHG emissions. It uses most of the water. Again, it's a major driver of land use change. So again, there was that light bulb moment where if you're going to have impact in biology, agriculture is one of the greatest places to do it. I then went ahead and did, I worked at the Gates Foundation in infectious diseases and then I worked at a startup in cattle breeding, so that got me the entrepreneurial bug. After that, I went to Google X, where I was given the freedom to work on whatever I want. The only requirement was to make the next Google-sized business, or at least incept a business that could scale to Google. So that's a audacious, almost unimaginable goal if you think Google has $100 billion in revenue. So you really have to think completely different, and no surprise. Because agriculture works at such a massive scale and is at the center of water, carbon, land, food, nutrition, that idea sort of hit resonance. It got traction within the organization. Along that time, I invested a little bit in the tools to be able to do higher throughput experiments. And I used what I learned at ARPA-E, actually, in the government as a way to try to enable that innovation. I went to all of the leaders in the world, people that worked at core facilities and universities and famous professors, on how to manipulate plants to do experiments on them, and I gave them an audacious challenge, which is again what I learned at ARPA-E, which is make me like 10,000 independent plant experiments by one person in one year. That is so many orders of magnitude off of what can currently happen that it causes people to think, not trying to extend or improve on current processes, but try to come up with new processes. And that's where I met Dan Vojtas, who's been a leader in gene editing in plants since before there was CRISPR. And we've had such a wonderful relationship that he's a co-founder of Inheritable, the company that I'm the CEO of right now. Again, investing in experimental, running field trials, really focusing on getting the data and being able to run experiments in plant biology. I met Davide Soso, who is a serial entrepreneur in agriculture, worked at one of the juggernaut companies of plant biology called Inari Agriculture. So he joined me while we were at X and then has now been a co-founder And then finally, understanding, and I'm sure we'll get into this, the more holistic view, the breeding view, not as much thinking about gene editing, but thinking about how to incorporate environment and entire genomes into crop improvement. We recognize that we were missing that part. There's a synthetic biology view of plant biology, and then there's the larger scale breeding view of plant biology. So that's where we met Tim Bisinger, our final co-founder, who then he leads our modeling team. And so that formed the pieces of Heritable at X, along with a group of many, dozens of people that have just been absolutely brilliant, from leaders in ML at Alphabet to people that had really deep product expertise to everything in between. And then here we are. We just recently spun out of X. Now we have access to more of the investment ecosystem and to the markets, and we're running.

Ian Krietzberg:
You're running. Yeah, so you spin out of X, right, but at the kind of culmination of, I guess, stage one, right, of this culmination of work that you've been pursuing for such a long time. And now you're at a point where a different kind of work starts happening. And before we dive into the details of what that what that looks like, you know, and you kind of danced around this, right? We have agriculture is an enormous, it's enormously important, and it's also an enormous strain on our ecosystem. And that's kind of fundamental dichotomy of we need it. How can we make it better? And, and, you know, you're talking about gene editing and stuff. And that's the fundamental pitch, right of what heritable offers, what you're trying to do is this is going to be a massive oversimplification, but the idea of programmable plants. And I want to start with that concept, and why that is the solution to this problem, or why that is the important solution that you're pursuing as a means of addressing this kind of crisis that you identified.

Brad Zamft:
Well, let me clarify. This is a really good opportunity. Yes, my roots are in synthetic biology, so I kind of saw that world as, you know, Again, as the biological systems as computers, gene editing, all of the fancy tools being the disruption. Over the years, we've really learned that, again, there's much, there's a lot of need for just making breeding more efficient and simulating what nature does, has been doing for two billion years since sexual reproduction was invented. doing a better job of simulating that. So yes, we do some CRISPR stuff. To be honest, we see much more pull on the product side. We see much more pull on what we do in breeding. We use CRISPR now mostly for model validation. It's a very powerful way to, within months, know that the outputs of your model worked. It allows you to generate data that would be very difficult to generate. You'd really have to go and probe a lot of nature to find the exact changes in order to train your models. So we've actually shifted slightly away from gene editing. I think it is a part of this programmable plants vision that we have, and it will play a role in it. But there's a lot of opportunity on the breeding side, first and in parallel. But let's back up for a second. Again, I think you're kind of asking about the vision generally. And yes, our vision is that the biggest lever to making a sustainable civilization is through plants. Plants are these miraculous, carbon negative, solar powered, self-assembling machines. If you get the genome right, you've made a seed that you can then amplify and spread throughout the world. They assemble themselves. They feed on sunlight and carbon. And so we want to make the ability to use them easier and use it for any crop, not just the crops that have been the focus of the traditional focus of agriculture. for geographies that can help not necessarily just the American Midwest, et cetera, and for applications of plant biology that have been traditionally underserved. So a large amount of the focus of the agricultural system is on calories. By design, because our population has been growing and the agricultural industry has been serving that growing population in the best way it can, which is by just massive focus on improving yields and calories. Now, through a few of innovations, four real major innovations that have happened technically over the last couple of decades, we really do have the opportunity to rethink all of that and develop crops that maybe have not had the emphasis or the attention that they deserve, apply to geographies that have not gotten the attention that they deserve, and endowing those plants with the capabilities to solve problems that they're capable of solving, but we just have not put the focus on it. Does that make sense? That's the whole concept of the programmable plant. How do we lower the cost, accelerate the time to market to be able to expand the agricultural industry and its applications to solving civilization-scale problems.

Ian Krietzberg:
scale in a sustainable way. Now, I mean, it's interesting the place you play in, because it's technological, but also it's biological. And the biological part, I imagine, presents some challenges. And I want to start there with those challenges. Just to gather the data necessary to do the work that you're trying to do, you're talking about gathering biological data. and then analyzing that in a lab, but biology is different in nature versus in a lab. So what's the challenge just on the data collection side of making sure that we have usable information that we can then work with?

Brad Zamft:
I mentioned that there's been four major revolutions that we leverage that lead us to believe that now is the time that we can actually enable this kind of step change in agriculture. One of them is data. So it's a good point. We do live in an era of unprecedented data for both biological systems and our planet. So folks may have heard about the massive drop in the cost of DNA sequencing. This really enabled the Human Genome Project. But even more, now we sequence thousands of genomes, not one genome. We do metagenomes. We look into entire communities, microbial communities, for example, and we sequence thousands of genomes at the same time. That has been an absolute breakthrough in society, a huge momentous occasion in biology. And there's been concomitant advances, well, Sequencing RNA is virtually the same technology, but also we've really had major advances in our ability to measure proteins, the chemicals inside cells known as metabolites, the modifications to DNA known as epigenetics. So we live now in an era of an unprecedented amount of information about biological systems. Likewise, there's been a remote sensing revolution, which is the second disruption, in my opinion, where we have an absolute unprecedented amount of information about our planet, weather, soil, climate, down to very, very high resolution, both temporally and spatially. The problem with all of that data is it's complicated, it's messy, biological systems don't follow. There's not a user's manual to how a cell works. Integrating that data is very, very complicated and unclear. And that's where, of course, the third revolution comes in, our ability to simulate systems and to integrate different data streams, artificial intelligence. Now, getting that data, you're absolutely right. In some cases, that's hard. And that's actually what we spent a lot of the time while we were at X doing. For specific applications, for specific products that we have, you really do need this multi-omic data. You really do need RNA metabolites, et cetera. And that has traditionally not been that available. The majority of the agricultural system focuses on taking DNA and correlating it to how a plant is going to exist in the world. One of our breakthroughs is taking the entire beautiful diversity of molecular information enabled by that revolution that I spoke about before. It has meant that we've had to go out and run field trials. Oftentimes we advise the partner who has the expertise in that particular crop, in that particular application, in that particular geography. They have all of the growing operations, the relationships with the supply chain and the customers. Many times it takes the form of us advising them on how to build the trial and how to do the sampling. Sometimes we do it ourselves. What that has done, that effort that we've done over the last five years has enabled us to build the bona fide AI, right, because we have the AI ready data sets that have come through great investment by us in time and money to be able to develop the AI that we believe can really have an impact in the industry.

Ian Krietzberg:
The that's the third revolution, right? That you mentioned the, the AI, the artificial intelligence, the ability to process, uh, all of that data. Now you spent the past five years, as you just mentioned, building it, collecting that data, doing the field tests. Once you started assembling it, how do you go about validating a model like this? That is, you know, it's output is biological predictions.

Brad Zamft:
Wow, I promise we did not practice, but you led me right into the fourth revolution. So good call. The biotechnology revolution, I mean, CRISPR is absolutely famous. It's one of the biggest discoveries, biggest changes of our time, innovations of our time. And I think the biggest impact it's having right now is on the validation side. CRISPR allows you to go to the exact place that you think is conferring the advantage to the plant and modify it and to see what it does. Traditionally to do that with breeding you would have to try to find that difference in nature and you would want to cross it with the elite lines that you already have to see how that confers. because DNA molecules, chromosomes, are super long, you end up bringing all of the other parts of the chromosome along, so it's hard to disentangle, not all of them, but large, large parts of that chromosome. So it ends up being very, very difficult to disentangle. This is the thing that we think is conferring this advantage to this plant and actually validating it. CRISPR skips all of that. And what we've seen is that we can go from running the field trial to get the molecular data, to identifying the specific gene that is conferring some amount of control over that property of the plant. We can do that in a year, 14 months. Traditional timelines for that kind of validation is closer to a decade, five to 10 years, than on the order of a year. What I'm most excited about in CRISPR is not about developing CRISPR-based products. I think that will come eventually. It's not about, yeah, CRISPR-based products. It's actually about speeding up our learning and our understanding about the biological systems, about plants, being able to generate the data to retrain our models, and again, accelerated validation.

Ian Krietzberg:
So those four revolutions all kind of roll into each other and you're able to vastly speed up a process that normally would have taken a very long time, which is big. So what happens next? Like if you're, if I was a partner, right. If you're walking me through it, if I grow whatever plant, um, uh, tomatoes, maybe tomatoes are really hard in New Jersey. Uh, but, uh, if I was growing that plant, right. And we go through that, that, that phase, we use all those different revolutions. We identify the gene for control that we're looking for. we are able to validate it quickly, then what happens? How do we physically go in and breed for, if that's what we're talking about, breed for that specific thing? How do you, how does the physical activity based on the technological information that you derive, how does that work?

Brad Zamft:
It's going to take me a little while to explain this. Give me a moment because there's different problems in optimizing plant genomes, if you will, making programmable plants. They operate at different scales, they have different data needs, and they have different approaches, including actually the making of the product. This will take me a second. I hope I will inform some folks on biology. If you want to optimize a plant, there's a couple scales. There's the whole genome scale, moving large fragments of the genome around. And that's important because nature has done a lot of the work. A lot of these traits, properties of plants, have thousands of genes that are controlling them. And focusing on a couple of genes is not going to have substantive impact. The most famous trait that this is a property, that has this property, is yield, which is the trait that basically underlies all of the investment in the agricultural system. Most of the investment in the agricultural system. So there's thousands of genes that control yield. They're all over the genome. You kind of need to do whole scale movements of parts of the genome around. So this is kind of the genome scale. I liken it to If you wanted to make the best magical realism novel, you would be borrowing things from Gabriel Garcia Marquez and Salman Rushdie, entire chapters, ideas, you'd be moving these things around. That, you can do, you can get a lot of work done by just knowing, by making a unique identifier of the plant, saying this plant has this kind of genome, I don't even need to know the details of the genome, this other plant has a different kind of genome. When we put the two together, we get another thing. That's breeding. And actually, the big disruption there is on the environment. Integrating how those two genomes come together or combinations of the genomes with how the environment plays into it is a grand challenge in computational biology and is one of the products that we provide. Data needs, again, we can do it with unique identifiers of plants, not even knowing the genome sequences. It's better if we know the genome sequences because that allows us to do breeding. And the other molecular modalities are less important there. Now to answer your question there, Then how do you instantiate that? It can be as simple as moving genetics around. Do we have the genetics that already exist? Companies already have large libraries of different types of plants, different lines for their particular crop of interest. Are they putting them in the right place? So our ability to scalably integrate the soil and weather information down to 10 meter resolution is the major disruption there. And so to answer your question, What do we do there? We help companies either place their existing genetics better or initiate new breeding programs for new applications and new geographies. Knowing the holistic view of the genome, being able to put Salman Rushdie and Gabriel Garcia Marquez together to make the perfect or the better magical realism novel gets you far. The individual sentences in those novels matter, and the nouns in those sentences matter. What confers the meaning in the book? What confers the meaning in the sentence? It's the nouns. If we have a sentence, I want I like coffee. The coffee is the gene. The coffee is the thing that confers the effect on me, in this case, me being awake. And that's gene identification, a target identification, some people call it. There we use the breadth of molecular information that we can, which is Again, I think the major disruption here, historically folks have tried to, again, correlate pieces of the DNA information to that functionality. Embracing all of the different molecular tools and information has allowed us to do that better and faster and cheaper. So now what do you do with that information when you've identified a few genes or a couple of nouns in the book that really confer huge amounts of meaning? You could do gene editing to manipulate those genes, or once again, those can serve as markers, regions that you focus on. Now you can go out into nature and find those specific places that you need to incorporate, or you can target your breeding to better, quicker, faster incorporate those specific regions of the genome. That's not even enough, right? The fact that I want coffee, yeah, that coffee is operative, but if I said to you, I want more coffee in winter and I want more iced tea in summer, it's the grammar, it's the adjectives and the adverbs that matter there. More is important. Summer is important. So analogously, If you know the genes that confer a property, that helps. But do you want to have more of the gene or less of the gene? Do you want to have more of the gene in winter or when there's a drought? Do you want to have less of it when there's a drought? Do you want to have more of that gene in the leaf or in the root? It's the context, the grammar that matters there. And this is where My strained analogy towards human language, you can see the transition that's happening, right? We have a revolution in how we model human language and the grammar of human language, which by proxy is human intelligence and human logic. Likewise, this large language model revolution that's happening when applied to the language of DNA is allowing us to understand the grammar, when to express the genes, where to express the genes, under what stress conditions like drought or heat, etc. And that is literally finding the exact places on the genome that tell you this gene is going to go up, it's going to go up in drought. Just like I want more coffee in winter, I want more of this gene when there's drought. That has a very different data requirements. We actually don't need to grow huge amounts of plants in fields and get either line IDs or genomes or transcriptomes or metabolomes. We only need to grow a few plants, but we need to do very, very deep molecular characterization of them. Once we have that, This, once again, gives you the exact places in the genome to focus. That can be focused through CRISPR. It can be focused by looking for that in natural populations and targeting your breeding programs to integrate those exact changes where you want them.

Ian Krietzberg:
That stuff is just, this is not the first time I've heard this from you and at both times it's mind-blowing. It's just kind of crazy. And you're dealing with such complex biology. So what happens when you transfer into, okay, we've identified the stuff, we're growing the plants, you know, like you mentioned, we have 2 billion years of evolutionary biology at work here in natural environments that there is We have impacted it, but there is a natural ecosystem. Are there times when your specific identification of how to express this gene fails in unexpected ways because it doesn't always work in nature because you're adjusting the way nature designed this plant?

Brad Zamft:
Well, breeding, crop improvement, computational biology, this is a numbers game, right? We're talking about shifting our accuracy, shifting the distribution towards success. There's no, we're nowhere near. I mean, we are decades, it's impossible to even predict how far we are to having a deterministic model of any biological system, even a virus. Forget about a multicellular plant that's growing in a complicated environment. And breeding is just about a numbers game. It's about finding what you think are the best plants to then propagate into the next generation. Nature does this, by the way. This is what natural selection is. It creates, you know, there's random mutations that happen because DNA replication is error-prone, because we have cosmic rays that hit the surface of the Earth, because there are mutagens, natural mutagens in society, in the environment. So you get random diversity, and then what nature does through survival of the fittest is select for the fittest things that can fit that particular ecological niche. And as you alluded to, it's actually communities. The diversity matters. It's communities of organisms that support each other that are the most robust communities. What we're effectively doing through computational biology, through AI, is doing that selection in the computer millions of times faster than what nature can do. It's the same natural process. It's just simulating in the computer and then making the decisions there. But you're always playing a numbers game. You never truly know what's going to work. A lot of stuff will fail. And by fail, Usually what happens, actually the most likely thing that happens is you made something that's less fit and it gets outcompeted in the environment. It's very, very hard to make something that is better. Nature is so complex and complicated that it usually outcompetes us. This is actually the reason why we have monoculture. Because we need to have very coddled, delicate fields that grow one plant. Oftentimes we have to use herbicides and pesticides. Because if we let those plants out into the natural ecosystems, they would be out-competed by the natural ecosystems. emphasizing the fact that we're still in really early days. It's still hard for us to simulate a single cell, let alone a multicellular organism. And then when you add on to that ecosystems of multiple organisms interacting with each other, most of the time we actually make things that are less fit, not more fit.

Ian Krietzberg:
And I definitely want to dive into the monoculture point at some point, but before we get into that side, some specifics around the sustainability that we can achieve with something like this, what you were just talking about with the analogy of a large language model that speaks DNA and that enables us to understand it, as with large language models themselves, which can be and have been and are being misused by bad actors or misused by accident, right? I wonder what the risk profile of something like this is and whether, as you were saying, the kind of mitigation for it is simply that biology is really, really complex. Um, we don't have deterministic, uh, you know, solutions here where you can just type something in and have it spit you out a perfect, you know, X, Y, Z biological organism.

Brad Zamft:
Yeah, I mean, that is one of the major points. Absolutely. Biology is so complicated. We know so little about it that the risk is lower, especially in plants. I think, I mean, the reason why we feel that we can have really major impact, one of the reasons, one of the reasons is because there's so much room for improvement. You know, we've been breeding corn for decades, we've gotten a very consistent increase in the yield in corn. It doesn't seem we haven't hit the limit yet. And just imagine if you could democratize that for all the other crops, for all the other applications and all of the other geographies. What's holding us back is that plants take a really long time. It's very expensive, very complicated. So if you were to try to be a malevolent actor, Biology is a terrible place to start, first of all. I would definitely go into LLMs, into social engineering. And then plant biology is even worse, because we can grow 10,000 plants, let's say, in a year, grow and test 10,000 plants in a year. You can do a billion independent experiments in bacteria in a day. The level of orders of magnitude of your learning rate is astounding. And let's also, let's remember how unique this large language model revolution is. It was trained on the entire internet, right? And which is a pretty good sample of all of human language, which is a nice proxy of human intelligence. So I can't think of a better way better data set to do AI. And that's the reason why these things have moved so quickly is because the data was just sitting there waiting for us as a consequence of the internet revolution. That is not the case with biology at all. I cannot imagine a world where we have even close to the amount of data of biological systems as we do the data corpus of all of human language or most of human language that is captured on the internet. So we have a really long way to go. Now, that doesn't absolve us of any sort of ethical, we have, as people who work on technology, every technology can be used for good or for bad. And as the ones developing the technology, we have an absolute moral imperative to be considering it at the outset. I do think the risk of not doing this is worse than the risk of doing it. Because again, our goal at Heritable is to democratize the use of plants, to make a more diverse and expansive agricultural system. And that will come in the form of technological solutions for civilization scale problems that are natural and carbon negative and green, as opposed to concrete and steel and emitting GHG emissions as opposed to sequestering GHG emissions. So yes, there might be some people that use it for bad, but by having a democratized system, the opportunity space for good is way bigger than the opportunity space for bad. And we just have to maintain our ethical focus. We have to make sure that the good folks are working on this more than the bad folks, just like with any technology. And actually, the moral imperative is that we have to do this. We have to enable a more diversified agricultural system as opposed to being protectionist and maintaining the status quo of strong focus on a few crops and a few traits and a few geographies.

Ian Krietzberg:
Yeah. I mean, the need for an adjustment of the system, I think, is pretty clear. You're kind of facing two challenges here and they're related, but they're not quite the same. And, you know, you mentioned monoculture earlier, monocropping, you know, the one challenge and Maybe this one's a little more pressing than the other one, but again, they're related, is resiliency to a changing climate. We have increasing temperatures, and we have storms as well that are increasing. These are things that we're seeing. that puts our agricultural ecosystem at risk if these plants can't adapt to be able to produce what we need them to produce in more challenging environments. And then related to that, to a degree, as a result of a monoculture that we have been living in for a long time, is you have weaker soil. And the idea of how can we make agriculture itself more sustainable? Because if it is done more sustainably, if we keep the soil healthier, if we employ these other methods, like regenerative agriculture, of working more in tune with the land, then it snowballs in a positive direction where these things could all help us together address the the pressing problem for this species, the human species, which is we might have a food crisis if the food can't adapt to growing in climate change. And so you've got those kind of two ends of climate resiliency and then overall sustainability. And obviously, it's a massive thing that I just brought up, right? But when you're identifying these genes, when you're examining these plants, How are you working with those two things in mind to address The nearer term but also the bigger term problems to make sure that we're not just increasing yield We're doing so in a way that you know will help us down the line rather than the other way around I mean those things are are nearly one in the same climate resiliency

Brad Zamft:
It's called abiotic and biotic stress resilience. So there's the pathogens that are coming across ecological barriers as the climate changes, and there's the climate change itself, the drought and the heat, and other, and floods, et cetera. So there's that. There's diversity, biodiversity, and productivity, and those are all The same is the wrong word, but they all play into each other and there's a mutual benefit to, we really do have a beautiful synergy here. And the concept of regenerative ag, it's all part of the same thing. How do we lower the barriers, accelerate the time to market, to diversify our agricultural system to more crops and more traits and more geographies? That will reduce monoculture, i.e. increase biodiversity, even at the field level, which in turn makes a more resilient ecosystem. resilient to both these abiotic and biotic stresses, which in turn increases yield. It might not be yield at this individual plant level, but it's a systems level yield. You lose yield when you have a climate shock. You lose yield when you have a pest invasion, and having a diversified system mitigates that risk and overall would increase the average yield over time. Let me give a really nice example of this about how focusing on a new crop and a new trait really does have all of these benefits. One of the most famous and successful stories recently in agriculture is the company Covercress. So they took an esoteric species, pennycress. It's an oil seed, so it makes oils. Very small plant. part of the brassica family, which is cauliflower and kale, very diverse family. But it dropped its seeds, you know, kind of like randomly or over a long time period, a trait called shattering. And by them investing in this crop that had no, basically no market share, in this trait that was not directly a yield trait, they enabled intercropping of this this crop, pennycress, in between the rows of corn and soy. So what does that do? It's a doubling of biodiversity. You've gone from monoculture to biculture, if you will. But actually, when you look into the soil, when you look underneath the pennycress, you have a lot more bugs, a lot more diverse ecosystem because you're increasing the diversity of the entire microbiome and the ecosystem. It adds value for the farmer because those oil seeds can either be used to make sustainable aviation fuel or feed that has the fatty acids that animals need. It puts a cover on the soil so then the soil degrades less, you don't have as much runoff, you don't have as much carbon emissions from the outgassing of the soil. So this is just the first step going from monoculture to biculture and it has disproportionate impacts From a regenerative agriculture standpoint, from a value standpoint, from a biodiversity standpoint, and from a climate resilience standpoint, that ecosystem is now more robust to pathogen invasion and climate shocks, abiotic shocks. And this is just the beginning. We can do this over and over again. There are so many species, so many traits, so many places. that can use this kind of optimization. And we're just in the very early days.

Ian Krietzberg:
So it's not about just optimizing a specific plant for growth. You could also optimize plants that could help that plant grow, but also aid the ecosystem. And I guess that's the big challenge and opportunity of where you're working, where it is ecosystem plays. It's how do we impact and improve the wider environment than just one specific thing, because this is where this thing exists.

Brad Zamft:
Yes, yes, and. Focusing, just shifting the focus into more crops, more traits, more locations, even if it's you only focus on pennycress, or even if it's you only focus on camelina, or you only focus on forestry, just shifting the focus or actually broadening the focus, adding to the, expanding the agriculture industry, even that will then enable these kind of ecosystem services that you're talking about. I just want to caution, we can't let the perfect get in the way of the good. If we focus on engineering entire ecosystems at the outset, that's an entire other level of complexity that I'm not sure, you know, it's still hard to just even focus on Pennycrest, let alone the interactions of the ecosystems. But you get these things as it comes for free. You get the additional ecosystem services. You get the additional biodiversity by expanding your focus.

Ian Krietzberg:
Now are you seeing, as you're starting to bring this out to people and working with farmers, are you seeing an immediate kind of interest in this approach of advanced breeding through these kind of technological methods? Is this something that people have been looking for and excited to have? Is there a barrier to understanding what this can enable?

Brad Zamft:
I mean, there's plenty of barriers. oversell it, this is a complicated, difficult industry, but where we have the feedback has been resounding and our commercial pipeline is full. There are so many folks that produce food, feed, and fuel that do not have the modern computational techniques. They do not have the attention that the folks that work in major seed companies in corn and soy have. And it's especially true when it comes to breeding and the incorporation of environment and climate into breeding pipelines, breeding and product placement pipelines. So yes, we found just an absolute resounding need here. And that's what we're chasing down and being selective on who we work with. We're really focused on the folks that want to expand the system, folks that want to improve, that want to enable the step change, and that want to do it hand-in-hand with us. I mean, this means really, really significant partnerships, not services, folks that don't see us as a service for their operation, but one that will allow them expand their market, move into new markets, new crops, new geographies, and do that in a collaborative partnership kind of way. So we're getting a lot of traction and a lot of resonance. It's a change in the way the agricultural industry works, and that's part of the fun.

Ian Krietzberg:
Is there anything you can tell me at this stage, I know it's early, about those specific partnerships that you're entering into, the work that you've started to undertake on that front?

Brad Zamft:
The one that's public is our partnership with ArborGen, which is a nursery, a tree breeding company. You want to talk about impact on the planet, forestry is where it's at. I mean, that is where A lot of the carbon is sequestered. It's where most of the carbon credits are issued. It's responsible for the production of an amazing carbon negative material, functional material that we call wood, that we've been using for thousands of years to improve society, and it happens to sequester carbon as well. So that's the big one, and think about the opportunity in forestry. It really, the increasing biodiversity of forests, increasing the sustainability of our forestry industry, the potential is astounding. The other ones, we're not ready to announce yet, but you'll be seeing a few press releases and some press coming out over the next couple of months. Again, what you'll be seeing is that we're working in expansive portions of the industry. How do we increase the market share? How do we bring new traits, customer traits like flavor and nutrition to our food system? That's where we are, these partnerships. We did some partnerships while we were at X. Those were more R&D partnerships. Now we will be having announcements on commercial partnerships where we provide real impact to the world over the next couple months.

Ian Krietzberg:
For the tree one, I'm going to circle back to that. I love trees. It's the highlight. I go on a walk every day and I just love trees, kind of apropos of nothing. You were talking about some of the traits that you're looking at specifically in some of these future partnerships for nutrients and taste and these other things. When you're working on trees, what are the traits that you're looking at to improve or adjust?

Brad Zamft:
This is an interesting question. I mean, yield matters, right? How much of this miraculous carbon negative material called wood, lumber, can you get out of a particular amount of space of land? It's more complicated than that because The trees have to be straight, and the material properties of the wood matter, and all of these kind of things. So without going too deep into the weeds, actually even simple things that you think about in terms of productivity of the land end up being kind of the sum of many, many components. I think what's more important about the forestry side is that there's just so much opportunity, and it can be such a beneficial thing for the planet. Because right now, we grow the private lumber industry doesn't necessarily grow the perfect tree for the perfect location. Nature has figured that out. It does. Over billions of years, it has selected for ecological niches. And I think we can just do a lot better. And again, both have a more productive forestry industry, generate a lot of value, while at the same time increasing ecological services and the sustainability of our planet by better placing tree genetics around the world.

Ian Krietzberg:
And that's the balance, right? You know, even on the taste and nutrient side, right? The idea, I suppose, would be that these optimized traits are kind of byproducts of working on, like if you approach it from a sustainability perspective, if the soil is more diverse, there's more nutrients in the soil, then you're optimizing for having a more nutritious, you know, carrot. And if it's more nutritious, it's going to taste better. I guess what I'm asking is which side is the kind of free byproduct of optimizing for the other thing? Is it sustainable because you optimize for taste and nutrition or is it tasty and nutritious because you're optimizing for sustainability? Uh, does it matter?

Brad Zamft:
I don't, I don't know if it matters. I mean, this is maybe a point for me to interject one thing, just that that photosynthesis is somewhere around 1% efficient. in intercepting photons or in sequestering, capturing CO2 molecules. I don't know if I'm answering your question, but I just want to put it out there. We can have our cake and eat it too. We can sequester our carbon and eat it too. This is not a zero-sum game. We can have increased nutrition. We can have increased biodiversity. We can have increased yields. And by the way, One of the biggest, maybe the biggest climate trait, one of the biggest equity traits, if you care about social equity, is yield. The more productivity you can get off of a square meter of land with the same inputs conditioned on that, the less land use change you will have, which is a major driver of GHG emissions and other negative environmental consequences, and the more you'll be able to feed the planet. So this is a case where if we get this right, we really can optimize and improve lives, nutrition, sustainability, all at the same time. There's nothing, we don't have to give up anything here, in my opinion.

Ian Krietzberg:
Yeah, that would be the ideal world. And I guess that kind of brings us nicely to the last point here that I'd like to end on, right? You guys are still early stage, but you're starting to do the work. You're seeing the reaction to it, and you're also, you know, as you mentioned, the four major revolutions that we were talking about at the top, they are here. And the public understanding of what we can do with these kind of major revolutions, the data and the ability to process that data, I think is expanding. And with that all in mind, I wonder how you look at the next several years, five or ten years, where we are at a point with climate change and where sustainability efforts are going or are not going. that you could call it like a critical inflection point where we're either going to get our stuff together or not. And working on the agricultural side of things is definitely a huge factor in getting our stuff together to try and aid what's going on with the planet. And so what do you see happening? Are you optimistic that we can leverage these tools and address all these problems or some of them?

Brad Zamft:
I'm optimistic. I think markets are responding. You can do what you want with policy. We can debate the existence of anthropogenic climate change. Farmers know that the climate is changing because they are planting crops of different maturity grades than they used to plant. So the maturity zones of plants have moved. That's not it. There's no opinion there. That's a fact. Farmers are buying different maturity grades than they were a decade ago. And what that means is that seed companies are breeding for new environments now because that's what's needed, not because of some climate change model, not because of some ethical impetus to make a more sustainable planet, but because that's where the market is going. As these shocks become more apparent, we will be accelerating our breeding. We will be incorporating climate models into our breeding pipelines. And then there will come a day when even more dramatic interventions, like what happened with papaya in Hawaii, where there was such a dramatic pathogen invasion that they were forced to adopt this GMO trade. There will be isolated incidents where there's just no solution but to innovate. So I'm optimistic. I'm a techno-optimist. I think we'll rise to the occasion because the market is going to demand it more than because of some policy shift.

Ian Krietzberg:
Yeah. I hope we will. And I guess we'll have to see. Brad, I really appreciate your time. This was a lot of fun.

Brad Zamft:
This has been a blast. Thank you so much for inviting me on this.

Ian Krietzberg:
Yeah.