Techno-economic Modeling and Why it Matters for Invention with Jesse Lou

In our latest episode of the Climate Biotech Podcast, we explore where science meets business with Jesse Lou, the CEO of Conductor Labs. 

Jesse shares his unique insights into the indispensable role of Techno-Economic Analysis (TEA) in guiding the commercialization of climate technologies. We then explore a particular use case — biomining — with  Jayme Feyhl-Buska of Homeworld Collective.

As innovators in biotech, it's paramount to understand that while groundbreaking ideas hold great promise, success hinges on solid economic foundations. Failure to integrate economic considerations early in the innovation process often leads to avoidable pitfalls. By adopting a proactive approach to TEAs, scientists can identify crucial economic factors, optimize resource allocation, and pivot when necessary, thus bridging the gap between technical advancement and market reality.

In this episode, we discuss how TEAs not only function as internal decision-making tools but also as compelling narratives used to engage investors and the wider community. A well-structured TEA can convincingly present a technology's real-world value, portraying its scientific merit and expected economic returns. 

Join us for this enlightening discussion to understand how embedding economic thinking into scientific endeavors is not just advisable but essential for driving impactful climate solutions. 

(00:00) Introduction to Climate Biotech Podcast
(00:48) Guest Introduction: Jesse Lou
(02:10) Jesse's Early Life and Education
(03:31) Career Journey: From Engineering to Climate Biotech
(04:48) The Importance of Techno-Economic Analysis (TEA)
(10:45) Challenges and Misconceptions in TEA
(18:34) Current Work and Future Directions
(23:35) Introducing Jayme Feyhl-Buska
(24:08) Introduction to Mining and Bioleaching
(24:30) Challenges in Mining Technology Adoption
(25:06) Learning and Implementing TEAs
(25:58) Importance of Process Flows and Communication
(27:37) Insights from Mining Conferences
(28:59) The Role of TEAs in Decision Making
(30:51) Focus on Copper Bioleaching
(32:03) Challenges in Copper Heap Leaching
(34:32) Modeling and Parameterizing TEAs
(37:06) Communicating Error and Risk in TEAs
(42:04) Rapid Fire Questions with Jesse
(45:33) Conclusion and Final Thoughts

Transcript

[00:00:00] Jesse Lou: I've spent a lot more time trying to use this as a tool to mitigate risk and the front load risk so that when we do have more technologies that are deploying to market, can we increase the quality?

Or can we increase the sophistication of these technologies so we can have more qualified shots on goal to fight climate? 

[00:00:22] Daniel Goodwin: Welcome to the climate biotech podcast, where we explore the most important problems at the intersection of climate and biology, and most importantly, how we can solve them. I'm Dan Goodwin, a technologist who spent years transitioning from software and neuroscience to a career in climate biotechnology.

As your host, I will interview our sector's most creative voices from scientists and entrepreneurs to policymakers and investors. 

I'd love to just say a couple of words about Jesse. This is really fun for me because at this point we've known people for many years now. We've had this Slack and Jesse was one of the, like maybe first 20 people on it, so we've had a long time of brainstorming together.

Today, Jesse is a co founder and CEO of Conductor Labs, whose mission is to provide tailored commercialization expertise to all innovators developing technologies for energy transition. He conformed Conductor's thesis as a fellow at Breakthrough Energy, where he helped scientific founders build their techno economic analysis.

Now, TEA is one of those things that you just don't often think about in many different aspects of science. When I was a programmer we never thought about TEA analysis, techno economic analysis, because you could just put something on a server and it would scale up infinitely thanks to Amazon.

Or when you work in bio and you do stuff in therapeutics, you never really think about the techno economics on the science side, because when you actually have a product that you can put into people, the margins are so wild that nobody needs to really care about the economics. But when you talk about climate biotech, you're talking about doing stuff at industry scale, often in commodity markets.

We're getting the percentages and the financial models right is a difference of a dumb idea or one that actually has potential. And Jesse from my humble opinion is one of the absolute best in the space and has just been a really wonderful thought partner to me and to Homeworld and to a lot of people in this community.

So I'm really looking forward to this conversation. And so Jesse would love to put you on the spot and introduce you. Who are you? Where did you grow up?

[00:02:11] Jesse Lou: Yeah, thanks, Dan. My name is Jesse Liu calling from Brooklyn.

In terms of who I am, like and where I grew up. So I was born in China in Suzhou, outside of Shanghai. And then it came to the U. S. when I was young, but grew up in Texas. My family is really more of the engineering and sciencey kind of vibe. And so I grew up very much more technical, science, engineering so I love math and science. I went to basically like nerd high school in Texas. There's a school called the Texas Academy of math and science, and it's embedded in the university of North Texas. And part of the value prop for going to the school is do a lot of math and science.

And then you also get to do research while you're in school during the summer part time. And so that was my first exposure to deep tech, hard tech, stuff like that. I was working in the Brostel lab, so they were doing material science research. And I got to work for this postdoc who was really kind and guided me.

And I didn't really know what I was doing being in high school, but I got to run a lot of her experiments and stuff. So yeah, so doing kind of the research really fine. I was working on like different rubbers and adding in ceramic particles to make it more industrial resilient but also cheap material I think it was fun, but then I also just got really bored.

I couldn't sit still while these tests were running, like this mill doing tracks on this rubber, like for 2000 times. I ended up going to school for engineering. I was at Columbia. I started off applied physics, but then very quickly went into more of the industrial engineering which is like a blend of math, stats and CS for me I really became more attracted to the system level of thinking.

I think also tying into TEAs as well rather than just thinking about a technology in a vacuum, like, where does it really sit, or where will it live, within a system, within a supply chain. And getting to go through the different levels of abstraction from micro to macro, or technical and business, has always been really fun for me.

[00:04:03] Daniel Goodwin: It's a great story and it's important. 

[00:04:04] Jesse Lou: Yeah. Industrial engineering was fun. I ended up on McKinsey afterward where I was in Houston and I did a lot of industrials work. So chemicals, oil and gas agriculture, mining, and really great experience to be able to just be exposed to how things really operate at scale with some kind of fortune 500 companies.

Building a lot of Excel models. And that's where I really honed this. Like it was like bootcamp for Excel and PowerPoint making slides as like a consultant. And I was what people would call an Excel monkey. I could just, build anything into Excel. And I think that was actually really fun. I'm thinking about Excel as a tool to basically abstract ideas, process flows, everything into this way that you can set up Excel an infinitely scalable canvas to make decisions.

But then in 2016, when AlphaGo was beating Isadora in Go and this kind of reinforcement learning, deep learning AI wave was coming. I think that really caught my imagination of this is also where like software is in the world and things like that. So I ended up leaving McKinsey to go to the Bay Area where I was a product manager building AI enabled software, actuarial modeling tools for the insurance industry.

The startup ended up getting acquired a couple of years after I joined. And so that immediately for me, it was like, great, this is like an opportunity to think about what's next. And at that point in my career, I was thinking about what do I want my life's work to be like, or what do I want my career to be like, and that then led me to climate and thinking about these broader spaces.

So I'll stop there and turn it back to you, Dan.

[00:05:34] Daniel Goodwin: Yeah as someone that loves martial arts and the belt system, I have seen you use Excel and it is absolutely black belt. And should you ever want a side gig, you could run a twitch of just doing Excel and it is magnetic. But if we riff on that, I love this idea of just did you ever see this?

And so did you always know that you'd be a thought leader in techno economic analysis?

[00:05:57] Jesse Lou: Yeah, I couldn't say that. I knew that as I thought about what I wanted to do for my career, I wanted to be an expert in something and really, focus my energies and build towards that. I think you can think about it as like a function and then as like a topic, I think functionally being able to bridge the business and then the scientific stuff and then industry and topic wise.

I think climate as just a, as an area that is going to be increasingly more and more important. There's something that kind of I focused on and then. It fast forwarding and through the last call, like six, seven years of this exploration in climate, I think that's really where I discovered that techno economics is super important and a really big gap.

And so what that journey looked like for me was, I went and ended up going to business school and when I was in school, I was working with a postdoc out of the church lab and also Nate we had overlap there as well, working with Max, but we were looking at photosynthetic microbes cyanobacteria and this, there was this idea of Hey, if we could take this strain of cyanobacteria that grew faster than basically anyone ever found could that be the basis for a really amazing biomanufacturing company?

And so for me, as like a commercial co founder my job was to figure out like, is there a business here? And what would it look like? And so what we were particularly interested in was the ability for cyanobacteria to produce your commodity market products, like your sugars and oils that are also super low.

Price and super low cost. And if we could have a line of sight there, that would be a really compelling part of the broader storytelling. And so I tried to build basically a TEA from that. You can find TAs online, like NREL has TAs for, microalgae and, but they used Aspen plus, a really.

Intense simulation software. They built this TA and Aspen plus they had a really long a hundred page report that I've now read like probably 20 times. And I tried to rebuild their TA and Excel and basically like hunt for numbers to find. And as part of that exercise, we found that there were some physical limitations.

We can grow the algae really quickly, but we might be running out of basically photons per square meter to power photosynthesis to be able to hit a price point, particularly for those really large markets 

and so for me, that was like a big aha moment of T. A. Yeah, we don't have to necessarily build out the big farms. We talked to a lot of people who had built algae farms before. Could we try to distill that into an Excel model and just try to figure out what are the couple of big rocks that you can move up or down and try to identify what are the big drivers of that After finishing business school, and after realizing that PicoGreens as a biomanufacturing platform wasn't probably going to be the next step I ended up going to, Breakthrough Energy worked with a lot of early stage startups who are spinning out technologies from their PhDs.

And I think similarly, pretty shocking that, There's all these teams that are out there raising money, moving far along. But I think the degree to which I would have expected teams to really understand and have a really clear idea of what their economics could look like at scale, not even talking like fancy models.

I thought there was a big gap there. And so work with a lot of these teams to try to, put down their technical ideas on paper, plus connecting it to like the potential market reality. And just seeing how things shook out. And so that's why I've spent a lot more time trying to use this as a tool to mitigate risk and the front load risk so that when we do have more technologies that are deploying to market, can we increase the quality?

Or can we increase kind of the sophistication of these technologies so we can have more qualified kind of shots on goal to fight climate?

[00:09:26] Daniel Goodwin: I think this is great. One of the, this is a funny language thing where I think investors speak in risk. And I think a lot of innovators and scientists speak in prototypes but they're really two different sides of the same coin. The way I would play back your story is that you did a back of the envelope prototype for an algae based company, right?

And build the back of the Excel envelope. said that, the numbers would be unreasonable to ever expect that to be a company. And so you made the decision just by a prototype, a financial model prototype, not to pursue a technology. And then you took that to work at probably one of the premier hubs of startups from breakthrough energy ventures.

And you probably had a lot of perspective on fantastic companies coming through that program. And it's really fun to hear your stories and hands on experience from world class teams who get in and can put millions of dollars in the bank. Before really having a baked idea of what their technology would look like as a financial model.

And so I'd love to push on this because if I was in the audience and I'd never done a TEA. I think maybe lab first, I would really think that, gosh, this is a dumb topic. This is so vague if I can do all this fancy biology stuff, I'm surely able to do a spreadsheet, 

and balance dollars, like that all sounds obvious. at the same time, really smart people get millions of dollars in and then they screw things up. And so I would love to just push on your experience here. I'm just gonna shake your tree of knowledge here, which is that from your experience, what does scientists do wrong when it comes to techno economic analysis, TEAs?

Not doing them until they're too late is an obvious starting spot.

[00:10:57] Jesse Lou: Yeah. I think part of the, so what of this, if I'm trying to understand like TAs early on are so rough and uncertain anyway. And I've talked to investors and startups who feel like, Oh, what's the point of doing it? I think there is a broad, North star that I would say, which is how do you just take everything that, it could potentially know, put some error bars around it and then see where you land.

And I think that's kind of the core value proposition of a TEA. It could be rough, but. It at least tells you direction where you land and then you can work backward from that. And , it feels obvious for sure. Have like really smart experienced people building these TEAs for multi million dollar projects and a time and time again.

And there's a great book, how like big things get built. Like people are always off by two or three X, the projected, budgeted costs and how much they end up paying. And so like, why do we do this ourselves? Time and time again, and could we just, account for some of these pieces?

Obviously, it's easy to say we should just build models. It's actually much harder to actually build them accurately. And so how do we to take and adopt that like mindset and bring it as early as possible when there is actually a lot more agency and flexibility and things that scientists could do early on to try to steer their direction?

I think tactically, like some things that I've seen that are coming. misconceptions or challenges. One is not comparing apples to apples. A lot of times scientists will be working on a particular part of a broader process flow, and they might, impact one of seven or eight or 10 different unit operations as part of an industrial scale supply chain.

I think a lot of times the challenge is they'll really zoom in and index on, okay, how much does it cost to produce the output directly after their unit operation, but not account for downstream processing. And so they'll take however much it costs to produce the widget coming out of their process, compare it to the commodity price and be like, wow, like we can produce this five times cheaper.

And then you realize that where it makes up the five times is actually the downstream processing step. So I think comparing apples to apples, second, and actually, as you said a lot of times scientists from the R and D scientific mindset are like, I'm working on something that is the frontier breakthrough technology that's never been done before.

It must mean something for commercial viability, right? I must be able to improve on the status quo and yes, but I think the real question, the operative question here is by how much? And, you might be pushing the frontier of something that impacts the unit operation that contributes like 1 percent of the resulting cost.

And so even if you were to double the improvement, you maybe can save 50 cents on that, 100. And so then the order of magnitude also really matters in terms of which problem in which like target market product market fit problem you're solving. And then lastly, I'd say it's. Like being too precise too early on I think finding the right level of abstraction for what a TA should look like and being able to find like a good enough end to end answer without also spending way too much time getting in the weeds.

And also getting enough depth so it tells you a good enough story or it builds, helps you build enough conviction is like really hard to do. I think scientists definitely, because they know the space super well, there's a bias to just adding more and more, knowledge and like data and then naturally just gets really deep.

But I think you end up. Trading off like granularity or detail separately, and even if you could build a lot of detail across all stages, you're just delaying decision making as a startup. You don't have infinite time or money, and so you have to make decisions with a lot of uncertainty.

There was another podcast where Dr Vanessa Chan, who used to be the director of OTT, she made a comment about how. And she was also a ex McKinsey partner. People from the private sector have an 80 20 approach to these things. Whereas I think from the science side, like everything has to be like really precise.

And I'm generalizing here and sorry, I don't have a PhD. And that's the high level generalization. But I think there's really something there where. Being 8020 is good enough. And I think for me too, coming at TEAs from non like chemical engineering PhD perspective, I'm also just trying to capture like, how do we get 8020, mole TEAs, like really personalize each team that I work with to solve particular questions and then not have to like have the same level of depth and granularity across the whole thing.

[00:15:08] Daniel Goodwin: To play off what you're talking about, I see two roles of TEAs. I see the first one, like when I meet teams all the time. The first role of their TEA like thinking is that they pass the initial sanity check.

And the easiest person to roast is myself, so I can tell a quick story on that, which is the first thing I did in climate biotech, is I thought, let me make a protein that binds carbon dioxide. And I met Klaus Lackner, the, one of the godfathers of carbon capture, and in one conversation, he said that the costs would never, ever make sense.

It was like five orders of magnitude off. I'm like, I was trying to solve a thing that was very cheap by doing something very expensive, and who cared about the precision? So for me, I was like, I just saved five years of my life. Quote unquote, prototyping an idea by just killing it on the finances alone.

But I think the second thing is, this is what you're talking about in the apples to apples, is it's a team showing that they actually understand the full end to end, right? So it's one thing to compare the prices, but it's also to say I'm building this thing in mining and I can bind a metal with a protein better.

But then that goes into this thing and just, it's an element of homework. This is maybe what Vanessa Chan was saying is we did the parameters don't need to be perfect, but you need to demonstrate the structure of what a deployment would look like does that sound kind of running with what you're saying?

[00:16:18] Jesse Lou: Yeah. I have a couple of comments on what you just said too. And maybe working backward. I think part of the value of TEAs is to show the, Hey, you've thought about everything. And part of it is that identifying gotchas in the process, or I like to think that like every TEA has seven to 10 big rocks or like big levers that you're moving around.

And then a lot of the other stuff is just details that kind of connect all the pieces together, but. For any like electrolysis system, you have big rops, big rocks around like Faraday deficiency and electricity prices and stuff. And so being able to make sure that you have all the gotchas is important.

And then but also I think people don't appreciate that. Like the more that you understand about the rest of the process, the more that you can also try to find better product market fit or improve your own value proposition. If you think about a particular. Like supply chain and indexed on one particular unit operation.

If you realize that maybe a waste stream from your process is actually super important for a later on, unit operation, then there's real unique value propositions that it's not, I'm just replacing the existing petrochemicals, which is super cheap solution, right?

But, and I think that's like what I push teams more to do and I think especially now more than ever in the political climate with climate in general now more than ever, green premiums are like, you really can't bake in an account for green premium premiums.

I think everyone knew that already, but you just have to build commercially viable businesses you're building commercial enterprise. And so you have to have a really clear line of sight on what you're selling. And it'd be able to create value, right?

And I think that's another unintended part of the strategic part of understanding the full broad end to end, because again, I think as a scientist, you're working on a particular technology. You don't appreciate what all the different problems are throughout that supply chain. So the more you know about it in the course of building a T E A, the more you can also integrate that into your decision making, whether it's on the R and D or on the go to market side, which I think is super powerful.

[00:18:08] Daniel Goodwin: I love this. I think it's just such an important empowerment skill, you are any career scientist, what are the skills that you need? You need to be able to write, you need to be able to communicate. I think you also need to know what the deployments look like.

And so I think this is really important thinking because the classic failure mode is someone says, this is a really cool thing I can do at the bench. Obviously, it's going to be a startup and then it goes nowhere. So , I think what the work you do is so important. That's why we're so happy to have this conversation.

What I'd love to do is just spend some time hearing about what you're doing now. You're doing really cool stuff with AI and leveraging that in a great way. And then I'd love to loop in Jamie from Homeworld, who's our geo biotechnology lead that we can talk about some work we're doing together.

But first let's really make space to hear about with all the background that you have and all your deep understanding with TEAs and all the work you've done with Breakthrough and other fantastic hubs of startups. What are you doing now, Jesse?

[00:18:57] Jesse Lou: So conductor labs is basically the, this the kind of the manifestation of a thesis that I was developing at breakthrough, which is that commercialization support for deep tech companies is super important, it is also fundamentally. And maybe has to be personalized to every single technology and the team the leaders their expertise and also the market.

So it's very hard to just create temple of tiesable stuff and scale. There's like tens of thousands of early stage, either proto companies or pre seed startups that I think need help. And there's just not enough like breakthrough energy fellows, RP, T2M. Activate support for these early stage teams.

And so that's part of the thesis is that can we support all these teams in a very personalized way? And in a non hype way, I think AI and like large language models can help us get there. And we also. Not to get, not, we don't have to dive too deep into this, but I think part of this was actually started right based on one of our early partnerships and projects back in Boston, hanging out at the home world office and like GPT had come out and we're like, man, it's going to be cool to use GPT to do some of the work that I was doing with these early stage teams.

And so also working on some of those projects and that project with you and the team that we had put together and the Dan, you presenting that under the Novo Nordisk. Conference, I think is really cool. So just taking that forward. 

[00:20:18] Daniel Goodwin: Yeah. And I don't think everyone has context there. So the idea is that you can have AI do. A T E A, right? But you're saying it's not a professional level, 50, 000 contract. It's just I think the very early level thinking, like how would you describe it?

[00:20:34] Jesse Lou: yeah. I'll start with the actual like manual use case, is when I work with teams to build TAs, there's TA consultants out there who help teams build TAs, but oftentimes I think the challenge was, and I faced this too, I'd give teams a TEA, but they wouldn't really understand like how it worked.

They'd be like, Oh, can I, hey, can I add a row here or a column there? I think the moment they ask that question, I'm like, Oh shoot, like they actually won't be able to use this TEA to actually make decisions or adapt it and, as a company evolves. And working with Fico Bloom, one of the teams that is also in our community John, the CEO was like, Jesse, can you just sit with me on zoom while I rebuild a TA from scratch.

And I was like, yeah that'd be great. And then that really helped John. And then that helped me build this thesis that if teams actually build their own TEA, they actually understand what everything is going on in the TEA. That then helps them build intuition.

It helps build trust to then actually use the TEA to drive decisions, which is the ultimate goal there. And so I think with AI, like what you can do is take a similar kind of semi Socratic kind of curriculum based approach of coaching teams. I like to think that we're like driving school where if you're in driving school, you're the driver's seat.

You have like your instructor next to you. Maybe they have another set of brakes, but you're the one going through the motion and like learning, and I think that's the best way to learn by doing. And so in a way, can we use AI to not give you the answers and AI is going to be probabilistic.

And so can we use AI. For the purpose of a language conversion of like unstructured or structured data. There's a lot of information about TEAs and about industrial scale stuff that's out there already publicly. If we can collect all that information in one place, validate it, have these like structured reference classes that you're just trying to map an innovator's context to existing data that is also validated using these LLMs as just the interface between that.

I think there's an opportunity there and that's what we're building at Conductor Labs.

[00:22:25] Daniel Goodwin: I love it. I'm super into this just because I see so many people come in, with just ideas that after five minutes, I just want to say, I think, like I support the ambition, but you're going in the wrong direction, and the authoritative way to say that is to think. In the industrial sense first 

[00:22:40] Jesse Lou: and one thing I would just want to add too is I think we we also feel strongly about these tools as publicly available goods. Like I think early stage startups are also not great SAS customers. Like your LTV is also not very high. And so if we can create publicly available goods, that really helps democratize access to commercialization.

And so in that, in the vein of that, like we're also fundraising from philanthropies to be able to fund this, to create public tooling to be able to really democratize this 

[00:23:05] Daniel Goodwin: It's so important. That's why Combinator worked, right? Because then they created Hacker News and they had the safe note and they had all these common resources, basically paper resources that all the other teams could run with. So I think that work that you're doing in that spirit is a hundred percent needed.

Cause when you talk about climate technologies, really, you're talking about large scale industry. And large scale industry is commodities, energy, chemicals, agriculture, and mining. And I think to make any innovation that matters, you really need to understand the field.

And so on the homeworld side, have chosen geo biotechnology as a high priority area. And we're so excited to have Jamie join us as our GeoPi technology program lead. Jamie, this is going to be a woefully inadequate introduction because we're going to have you as the core guest for another conversation.

But I think this is a great time to bring you in. And I'd love to just have you introduce yourself for a little bit, just briefly, and then I think you can set up some of the conversations and the work that you've been doing with Jesse exactly on this idea of open source TEAs.

[00:24:04] Jayme Feyhl-Buska: Yeah, thanks, Dan. I really appreciate that. My name is Jamie Falbuska. I have worked in mining for a couple of years now. I started working in mining around copper bioleaching. So trying to use microbes to be able to leach copper from ore. And when I got into this field, A lot of my friends who had PhDs same thing as me, geomicrobiology, started also looking into mining applications for their research around the same time.

And the really horrifying part to me was they had brilliant ideas and absolutely no idea how mining works. None whatsoever. And they would bring these technologies to me and they would say, Can you ask a mining professional to take my technology into their mine? And I'd go, No, absolutely not. You are so far from where you need to be in terms of understanding how your technology is going to work, but also understanding how the mind itself works.

This was something that I dealt with for years, working in this application, and I started to learn about TAs myself during that time, because I had to employ them as part of my own work. I was working for a startup called Endolith. During that time, I learned how to lay out the basics of a TEA, but when I finally met Jesse, coming here and working at Homeworld, trying to build out what this program was going to look like and what resources we could offer the community, that's when I finally learned how to really make TEAs, because Jesse, And his Excel expertise sat me down and showed me a whole entire new world.

It was like that Disney song, he really has made a difference in my understanding of how these TAs work. One thing that I would love to talk with Jesse about here is what we're trying to do in terms of changing how the community thinks about their process flows. And then also how they think about approaching VCs approaching mining companies, approaching all the people that they need to talk to in order to get their technology adopted.

That's one of the first things that I'd like to discuss here. Jesse, would you like to chat about the process flow work that we've been doing in this specific area in which we're working?

[00:26:05] Jesse Lou: Yeah. And yeah, Jamie, too kind. I think, and and maybe to also give a little bit of background to what it is that we're doing to make it a little bit clearer, one of the projects that we're working on is to create, these publicly available resources and TAs and tools for the ecosystem.

We don't need AI to be doing the TA for every single industry. We can just start doing it kind of piece by piece. And so that's what Jamie's describing here. Then I think one of the challenges following the conversation with Dan as well, as a lot of times technologists or innovators or scientists will think about a particular part of that belly chain, but not appreciate a lot of the other downstream impacts.

And what I've also learned from you is you can, there's directly as it relates to like bioleaching and mining operations. What I didn't appreciate is that if you have really harsh chemicals, you're using. To do purification or cleaning or processing later on that also kills microbes.

And so those are like non obvious insights that I think you get by being able to work with someone like you. We talk as in our working team, like we're really just mind melding. And my job is really just to, take Jamie's brain and then put it into Excel. And however that process needs to look like to make it, relevant for the very particular style of TEAs for early stage teams is this effort.

And I think that the really cool thing is, once we can get into an Excel, then we can scale Jamie infinitely to anyone who wants to be able to, use and play with the Excel. That really accounts for her depth of experience in understanding a lot of these different interactions. I think that's like the really fun part.

[00:27:37] Jayme Feyhl-Buska: One of the other things that I did want to mention here is to go back to what I was mentioning with the VCs and the mining companies, because I have spent this last week at a mining conference here in Denver. That's where I'm based. And at this mining conference, when I meet.

With mining companies, one of the very first things that they want to see in anybody that brings the technology to them is that process flow. And then also a really good understanding of how a product that a biotechnologist is developing is going to impact the other parts of the process flow. So that's one of those things that you build an understanding of that through this discussion.

And then the other question they're going to ask. Is this actually going to make a difference in our long term budget for the mine? So they don't want someone that has a loose sense of this. They want a very defined number. And it's okay in the beginning for a company to say these are our predictions right now, but they still need to see that.

So if you come to a mining company trying to get your technology into that company, You need to have your TEA, and ideally your LCA, your Life Cycle Assessment, too. Because otherwise, the company is not going to be very interested in your technology. This is something that I've been hearing again and again at this conference.

And then Jesse and I also have been meeting with a few VCs just to discuss the same thing. And we hear the exact same story. So I can't emphasize enough how important it is.

[00:28:59] Jesse Lou: I think one of the things you're touching on too is like TEA, we've been talking about it more as like a decision making tool ah, do I focus on this market? Do I focus on that market? Or what should I do on the R& D side? But I think just as important as TEA as a communication tool, which is exactly your point.

Like how do you. Demonstrate all the cool stuff that you're doing in the lab and then communicated in a language or in terms that someone on the commercial side who doesn't know the science really understands what's the impact on my all in costs or all in sustained costs for my whatever pound of copper.

And I think another piece is too, is being able to communicate that you are sophisticated, et cetera. I want to throw in a really quick piece here, just a very common thing that I hear teams. Say is Hey, Jesse, VC is asking me for a TDA. Can you help me build one? And that's totally the tail wagging the dog.

Like you should have a TDA because it needs to make sense to you. So you can build conviction that it's worth spending the next five, 10 years of your life building this thing. VCs are secondary. They have a portfolio approach. There also have slightly different incentives,

but I think the being able to have a tool. For decision making, it is super important, but then secondarily, able to communicate here's what we're doing, why, here's how we raise money, or here's what we need to raise money to be able to find the next set of experiments, because that will help us get to the next milestone.

A lot of teams will build these analyses for VCs, just like one off, like different Excel spreadsheets, but all of it should actually just come from your TEA. And should just be things that come out of like the broad decision making that you have.

[00:30:24] Daniel Goodwin: it's so important. This is the vision I have to with TEA level thinking is that part of PhD level bio training or any hard science training needs to have this element of pricing for anyone who wants to be applied. It has to have some element of pricing and deploying. And as you're saying, I don't think it needs to be an MBA.

I think it just needs to be a very hands on example, seeing what it might look like and help people pick different markets to work in on that note. When we talk about mining and geo biotechnology, that's a huge space with each one is its own huge vertical. So with the work you guys are doing now, which is the two, what are the two areas that you're choosing to work on?

And why is it not easy? Jamie I'll toss that to you to set it up and then we'll send it to Jesse after that.

[00:31:08] Jayme Feyhl-Buska: Yeah, thanks dad. So where we're starting is in the thing that I'm the most familiar with. So we're starting in copper. Copper is a very important metal to work on because we are starting to run out of copper based off of what we are planning to use copper for in electronics, electric cars, and We have declining ore grades where one of the best ways to access copper is through microbes.

So we've actually been doing that for a long time as humans, but we need to do it better. And there's all sorts of possibilities for how we can do that better. And one of the reasons that Jesse and I wanted to start focusing on it is chalcopyrite, which contains copper, has canonically been very hard to get the copper out of, especially at those very low ore grades.

So we wanted to see if you try different biotechnologies. What's going to make you the most money, because that's going to be what the mine will actually use. We are working on copper chalcopyrite heap leaching. In this field, you build this gigantic mountain, and that mountain's made out of ore, and you traditionally would put sulfuric acid on the top of that, it would percolate through, and out on the bottom, you would end up with your copper in the solution.

You can add microbes into that, you can add proteins into that, you can add nutrients into that, you can add all sorts of different things into that. Possibilities to increase the amount of copper that you'd bring out because the mechanism by which that whole process works is that microbes are actually making the necessary lexivians, that's the thing that does the leaching in that heap leach.

What we wanted to do is provide with all these new researchers getting into the field a tool to see if their technology is going to work. And the challenge here is that the mining companies are all very different. They have different ores that they're working with. We have different secondary mineralogies to them.

They have different considerations for when they build a mine. How far out do you have to bring your electricity? How much does it cost to run a haul truck in that area? And poor Jesse had to spend weeks working with me trying to figure out all of these numbers because this is really hard information to aggregate together, but also to make.

Somewhat generalizable, considering that we are mostly giving people a tool to start, not a tool to build their own mind. Jesse, if you want to talk about the challenges associated with that.

[00:33:28] Jesse Lou: Yeah. I think this is like the mapping of Jamie's brain into Excel. And how do we find like the relative impacts? And also I think there's also insights that like maybe things we're not sure if it really matters, let's put it all on paper and then let's build a model and then you can look back and see which ones like have the biggest impact and it's iterative.

And so there's probably some discovery there as well, but I think that's definitely the hard part. Like we've had conversations about how. You can turn on and turn off heap leaches and while it's still cooking and you can save money, save on your sulfuric acid or like the heap temperature really matters, based on pyrite that's in the ore.

And it's basically a prioritization exercise, right? Of okay what do you want to account for? How do you like bake all these things in? And a meaningful way. And so I think that is probably at a meta level. The hardest part is like, how do we build something that is 80, 20, but also doesn't miss any of the big rocks, no pun intended, and then also just is easy to use and also can also move quickly to get this out there because the longer that we take to build this there's a very large opportunity cost of time here too.

[00:34:32] Daniel Goodwin: Can you give us a brief overview and a teaser of what the other topic you've been looking at in biomining is?

[00:34:38] Jesse Lou: I think there's going to be many different potential targets of all the different types of important, critical minerals that are going to be out there.

[00:34:45] Daniel Goodwin: What I think this wraps up is that there's this idea of a combinatorial explosion, right? And if you're any sort of large process that has the heap and then the moving and then the heating and the adding acid and whatever that is.

Each one of those has a variant, right? And so you can go from what sounds like one way to do things can very quickly explode to a million variants. And I think that's why the T. E. A. work that you do and are building for the community is so important. So as we conclude here, I'm going to open up to some questions we've had and one of them is a pretty interesting meta question, which is that I think we've talked a lot about, TEAs to model an innovation going into an existing market.

But is there a way to think about how innovations can change the market when you're doing your TEA?

[00:35:27] Jesse Lou: Yeah, I think great questions. I think there's like a simplistic answer here, which is if you think about things as uncertainty bounds, if you're trying to predict what will happen in the future, you're just going to have larger uncertainty bounds on those things. And so like almost like the most simplistic way is trying to be really conservative on your projections for TEA.

One way I think about that is let's just assume status quo into the structurally is there a market here? Great. There's like, not a lot of what would you have to believes? You can also try to project or predict like pricing or costs. You can project those. And I would project those conservatively, basically against the direction that you want to go in.

For example, electricity prices, everyone thought it was going to go down like monotonically, but then, AI data centers and then also clean energy slowdowns kind of policy wise, I think electricity might not be going down monotonically. I think another related thing is you can count the number of miracles or what would you have to believe between today's status quo and the future.

There's already going to be, one miracle you have to believe in, which is your technology scaling and being deployed. If you add more what ifs. I think it just increases the uncertainty bounce. Same thing with TEA is I, when I work with teams, it's like very attractive to find like the highest efficiency thing in the literature and say great, we're going to like, with, if we use that part for our downstream processing, it's kind of like increased efficiency so and so, but then like, you're having to stack two of these miracles together, which I think is like exponentially more difficult to, bring to reality investment partner, investors, deployment partners don't want to see that. You want to find like the unsexy at scale, hundreds of thousands of hours in deployment to be using as assumptions for the rest of your process flow.

[00:37:05] Daniel Goodwin: I think that's great. I'm going to move on to the next question, which I think has an, a kind of a semantically linked idea, which is how do you, so you make a TEA for an innovation. How do you then approach communicating the error or the risk in the early stage TEA? Does this vary for different like different process types?

The example is, error in biological production and downstream processing. Is different for, a chemical extraction leaching process. So how do you approach communicating error or risk in early stage TEAs?

[00:37:37] Jesse Lou: Yeah, I think I would approach it similar to like how call it like late stage finance consulting approaches things, which is like finding comps and I say the late stage finance consulting thing is because I think we maybe have this like assumption that like people should things really figure it out like at later stages and it's like super well oiled and if you look under the hood of like how I used to build models at McKinsey.

Like we're like this seems similar to that. This seems similar to that. Let's use those as comps. And in as an example, like trying to find error bars, you can basically try to look at what are similar comparable processes and what are their efficiencies? If it's lower, if another one is higher, those are like rough estimates for you.

I think there's a couple of levels to this as well. There's like individual like parametric. Uncertainty, like for a specific data point, whether it's costs or efficiency or whatever, you can have error bars on that. I think it's hard to actually conceptualize. If you were to like aggregate all the error and the uncertainty at the large scale, people like to use, I think very standard tornado charts to show like a unit variant sensitivity.

If one thing toggles between low base high, like what the impact is, it's harder to understand anything is at a big level. And I would say the kind of the high level uncertainty is are you roughly in the ballpark? Can the sum of the different bars and your tornado get you there either into the money or out of the money?

And then. Not part of the question, but the last thing I'll say is a lot of times teams will end up with an unprofitable TEA when they first build it. And they'll be shocked and I think it's okay. If you're two or three acts off, I think it's fine.

If you're like 10 X, there's probably no business here. If I'm 2x off, there's enough uncertainty. There's things that I could potentially do. I could tell the story. It's a non starter to give like an unprofitable TEA to a VC, but there's enough like wiggle room and fudge factor to get you in the money, but that would be like a cause for concern if you were like still like 3x off.

[00:39:23] Daniel Goodwin: And the operative word there is story, right? I think one thing that scientists miss is that a lot of early stage investing is buying in on a story. And the TEA is an asset to substantiate the story. But I think if anyone's just trying to multiply all the, all the probabilities of risk and blah, blah, blah.

And pick your statistical method. I think they're missing the point that. You've thought through the things, you've put the error bars, you've put your reasonable best guesses, and then a person's either going to buy into that story or not.

[00:39:49] Jesse Lou: And to double, it's also really important storytelling for yourself, right? If you're the founder, forget other people can you buy into the story of Hey, this will be a thing or not. And maybe your error bar or your like uncertainty, like your risk tolerance a little bit higher or maybe lower, but you're single tracking right now building this company.

What is your error risk tolerance on this thing? And do you buy into the story or not of your TDA?

[00:40:11] Daniel Goodwin: I love it. I think one extreme example, and this isn't strictly bio, but I think Casey Hanmer at TerraScale Industries. wrote a huge white paper about why he was making his company and he was very explicit with his expectation of solar costs, right? And some people look to that and they say, oh, that's too ambitious and blah, blah, blah, but people bought in on that.

They said, okay, if I believe that story, then I'm excited about his company and it did really well. So I think that's a good example. There's also a sub question here, which is what about things that are extrinsic risk.

So how do you price an uncertain event like tariffs or, suddenly Canadian trade to American trade has a, as a hitch or is there a way that you can add those in? Does it become a different scenario? Or what are some of the tactics you might have to suggest there?

[00:40:51] Jesse Lou: I think this is related to the follow up point that I was going to make, which is I think the TA is really just a tool to help you help investors pick like what bets they want to make. Like I'm putting, I'm making a bet that there will not be tariffs or I'm making a bet that like electricity will be a certain price.

the more that you can almost parameterize your technology in the context of the Go to market commercialization. If you can parameterize into a set of bets, imagine like a waterfall chart. Here's the target. And there's like chunks, right? That bring you from like status quo to like price parity.

Each of those chunks is a separate. Bet. And if you can parameterize as like tariffs or one as like electricity, if you know what those are, then like an investor might say yeah I know you're not there yet, but I bet that that things will move it in a certain direction.

Everyone's making bets at the end of the day. And so That helps with that storytelling and decision making and same with tariffs, like with green premiums or whatever, you should bake in, if something is like zero, if something is high and whatever, 

 There's a lot of cases where there is no business without something, then you're just overloading on that risk as part of your path. But I wouldn't be too fancy about it. I just had different scenarios, like no tariffs. However many tariffs and you can try to like have some probabilistic curve to it, but I would just keep it like discreet.

[00:42:04] Daniel Goodwin: What I'm going to do here is I'm going to wrap up with some fun rapid fire questions for Jesse,, put him on the spot, 

The first one is What is a single book, paper, art piece, or idea that blew your mind and shaped your development?

[00:42:16] Jesse Lou: Thinking fast and slow. I think that like impacts professionally, personally understanding people, the behavioral psychology is is super interesting.

[00:42:26] Daniel Goodwin: Perfect. Best advice line that a mentor gave you.

[00:42:29] Jesse Lou: Taking ownership I, when I first started McKinsey, I was a terrible consultant and I would just do things that I was told and then I think really building this ownership mindset of seeing kind of the full problem solving was, yeah, and then I got to be a better consultant after that.

[00:42:45] Daniel Goodwin: If you had a magic wand to get more attention or resources into one part of biology, what would it be?

[00:42:51] Jesse Lou: I would say TEA. And I think data on and so there's a couple of efforts. There's like the ABPDU, the Advanced Biofuels Processing Development Group in the Bay Area. I think one of the big problems with bio is that there aren't a lot of like big bio projects beyond like ethanol, for example.

I think getting more funding and attention to creating Public data sources and like consortiums of data, it could be anonymized. We're trying to do some of that. A conductor is trying to like pull public data, but I think the more that we could have a clear line of sight into what scaling looks like, I think that's a big challenge.

So the more that we have data that we can share it would be really unlocking.

[00:43:28] Daniel Goodwin: I love asking senior people what they think are overlooked skills for people who are developing. And the obvious answer would be techno economic analyses, but I'm going to remove that from the consideration set. And so from your perspective, as someone who's seen a lot of startups and a lot of people be successful or not do you have off the top of your head?

Like an important skill that you think people might overlook when they're developing as a scientist or an entrepreneur.

[00:43:51] Jesse Lou: Yeah, I'd say I think a similar feedback that I got was like to think bigger picture systems level, like trying to understand the end game or the end use case. I think a lot of entrepreneurs are really in their own heads. And I think the more that you understand what customer needs are and like why you're doing things and connecting the dots there, what your manager needs, where things are going, I think that is really important. 

[00:44:14] Daniel Goodwin: I would love to just have you share with people, two parts. One, how do people find you? And then also if you have some general pointers of how people get started, maybe in thinking about their own TEAs.

[00:44:26] Jesse Lou: Yeah my email, jesse at conductor dash labs. com. But LinkedIn is just as easy or like on the slack if you're in it. And I would love to chat I know there was a lot of questions that we didn't get to or like other conversations. And so we'd love to follow up. I think for you to get started this is actually related to Lipsa's question a little bit.

I think you just use GPT, go chat GPT and be like, hey, help me build a TEA for something. And walk through like the steps and try to get started. I think there's a really. The cold start problem of facing Excel spreadsheet that is empty and like having to build a TA is like this massive, like mental burden task.

And so just get help. You can use GPT today to build your TA. I think the more that we can like validate, cause GPT numbers are unvalidated, the more we can validate and have a structured process around it, the better, but just get the piece of the process flows, get it end to end.

I tell teams just write down stuff in English and then give that to GPT and they can help you create a process flow diagram. There's a lot that you can be doing already with these tools. Just don't put a ton of credence in the numbers they give you. But I think that's if you were asking me two years ago, it'd be a totally different answer.

But I think today if you don't have anyone near you to help you, that's like a great place to start.

[00:45:32] Daniel Goodwin: Fantastic. So Jesse Liu, thank you so much for coming on. Really enjoyed this conversation and all our conversations we've had so far.

[00:45:38] Jesse Lou: Yeah. Thanks, Dan. Thanks, Jamie. Thanks, everyone. 

[00:45:42] Daniel Goodwin: Thank you so much for tuning into this episode of the climate biotech podcast. We hope this has been educational, inspirational, and fun for you as you navigate your own journey and bring the best of biotech into planetary scale solutions, we'll be back with another one soon.

And in the meantime, stay in touch with homeworld on LinkedIn, Twitter, or blue sky. Links are all in the show notes. Huge thanks to our producer, Dave Clark, and operations lead Paul Himmelstein for making these episodes happen