An Interview with Jano Costard, Head of Challenges at SPRIN-D

The Spotify version of today’s episode can be found at this link.

Today’s piece is an interview with Jano Costard, Head of Challenges at SPRIN-D — which some call “a German ARPA.” The interview will give readers (or listeners) a peek inside the SPRIN-D experiment, from the perspective of one of its leaders. It should be a particularly interesting listen for those who aspire to be PMs of ARPA-style organizations, procedural nerds in the R&D community, and those who simply want to understand what is going on at ARPA-like organizations around the world.

I’d love to highlight two things before diving into the interview. First, one piece of the interview that fascinated me was Jano’s obsessions with SPRIN-D’s administrative processes, such as their speed. “Fourteen days time-to-money,” the amount of time from a research group submitting a bid to receiving a decision and the funds hitting their account, is exceptionally impressive on its own. But it’s also the sort of improvement which can have fascinating knock-on effects when applied to large chunks of an R&D ecosystem. Speed and operational excellence on that level not only save existing researchers substantial amounts of time, but can also bring entirely new sets of people and organizations to the table, ones that might have little willingness to tolerate typical government R&D funders.

Second, regarding today’s interview, I had to put it together in a bit of a rush. I had a death in the family followed by a work trip. Things happen, of course. I’d love to have taken the time to read more and interview additional SPRIN-D employees to better understand SPRIN-D’s processes, scientific perspectives, contractor ecosystem, etc. But the folks at SPRIN-D and I were very excited to put this interview out in public quickly. The reason why is simple: SPRIN-D is looking for aspiring program managers, and there’s an application deadline coming up in two weeks. (More details below)

So, while I’d love to do my usual and spend 100 more hours understanding SPRIN-D before publishing a piece on them, this is meant to be a practical Substack. And, for today, what’s practical is sharing this (relatively in-depth) interview on the unfolding SPRIN-D experiment with the ambitious researchers and engineers who read FreakTakes — many of whom might want to become program managers, at some point.

I really hope that any of you who are interested will throw your hat in the ring not just for this program, but other opportunities like it. Having the curiosity to both spend a large portion of your life becoming world-class in a technical area and read something like FreakTakes in your spare time, simply out of curiosity, makes many of you precisely the kind of people places like SPRIN-D want. If you’re interested, please know there’s a good chance they’ll be interested in you.

Enjoy:)

The cleaned up transcript — courtesy of my wonderful colleague Ales and my lovely girlfriend Katherine, which I did not have any time to work on myself for the reasons listed above — is below.

---

Big if True Science Accelerator. Applications for a program which trains and mentors bright researchers and engineers on how to design and pitch effective ARPA/SPRIN-D-style programs close August 1. To learn more about the program, called BiTS, check out this link. My colleagues at RenPhil are helping run it. I will not be involved in running it, but I had a fun time writing some of the reading materials for the participants.

---

The Interview

Eric: Welcome to another edition of The FreakTakes Podcast. Thanks for being here. Today we have Jano Costard, who is the head of challenges at SPRIN-D, which some people call “a German DARPA.” But, of course, as with ARIA from the last FreakTakes podcast, it's a metascience experiment all its own. And [I’m] excited to talk to Jano about it today.

So, Jano, do you want to give us a bit about you first, before I jump into the questions?

Jano: Yeah, absolutely. As you mentioned, I'm head of challenges at SPRIN-D. That means I came to SPRIN-D, which is the German version of an ARPA, to build a challenge program. I'm an economist by training, and I have been dealing with competition and innovation. I now have the chance to actually run and design innovation competitions, which is really exciting for me.

Eric: And how did you fall into this, exactly? Because some people would think of an economist as a natural pick for this, and some people might find it odd. So how did it come to be?

Jano: So in my previous post, I was with a consulting body of the German government, with the Commission of Experts for Research and Innovation.
And our job was to consult the German government with regard to policies of science and technology. And that body was one of the leading actors that demanded that something like SPRIN-D should be created. And then SPRIN-D was created! And I thought it would be really interesting to move from just consulting the government to actually go to implementation, and try to do something differently.

Eric: And when there was this push internally in Germany for something like SPRIN-D, what was the exact reason? Oftentimes, there's some founding story or quick tagline that informs…the very complicated…it's the simplified form of a very complicated goal of the org. So, like, you'll hear DARPA people repeat and repeat, "to create and prevent strategic surprise." Like ARIA people talk about standards of living and growth in the UK. Is there something analogous for SPRIN-D?

Jano: So I think, in general, different cultures are motivated in different ways. One might say that the German way of being motivated is being pessimistic. [laughter] And so one of the perspectives was that Germany used to be a place where new technologies came into being that ultimately created whole industries. If you think of the automotive industries, chemistry, pharmaceuticals, and things like that. But that was like 140 years ago.

I think Germany is still living pretty well off of that. But I think we're moving in the direction where we need new technologies, where we see that those industries might not sustain our wealth or sense of living for the decades to come.

We need to find new ways of creating new industries, to create the surroundings for a wealthy society, for a healthy and happy society — as well as create solutions that tackle the grand challenges of our time. I think these two things really belong together.

But looking at how we did in that respect over the last decades, I think our record is not that stellar. If you think about the breakthroughs that Germany created over the last decades, I think not many come to mind.

Eric: And I guess for the listeners, me being history-brained and 140 years not being very long to me, I still look at Germany with a lot of awe. For example, when you read a ton of the — we'll call them the MIT founding fathers, like the early MIT professors and presidents — they all would continuously take tours of German universities, and just copiously take notes. Because in the very beginning of that institution, all they could think is, “If we could create something as good as a German technical institute, what a success that would be!”

And you see similar things with a lot of the early industrial R&D labs, like them really admiring the German chemical industry and things like that. And so it's interesting that you talk about German pessimism, because you have an optimistic mission that they've laid out for you.

Has the pessimism found its way into the structures of how this [SPRIN-D] has been laid out in some way? Because different cultures go at these things differently, and there are strengths, there are weaknesses, et cetera. But I'm just curious, in that, [with] a German ARPA, even if you wanted to create something identical, culture will seep in.
So I'm very curious about how the culture has seeped in as you've been there, from planning to where you are today.

Jano: So the motto of Rafael Laguna, our Director, is "Pessimism is a waste of time." And that really is the SPRIN-D spirit.

In that sense, it somewhat breaks with that German pessimism. He is a software entrepreneur, not by training but by practice. He’s been an investor for many years, spent loads of time in the US, and therefore went back and forth between Germany and the US quite a bit. He has been educated and shaped in his experiences quite a bit by both continents.

And I think that really shows with SPRIN-D, with how SPRIN-D operates, what our values are. We really are techno-optimists in the sense of believing and truly believing that technology can be a great source for finding solutions for the grand challenges of our time, and that that can be created from within Germany and Europe. It's not that we have a pessimism that says, "Okay, we didn't perform as well as we wanted to. And unfortunately, that's how it goes, downhill from here." I think the foundation is there to create outstanding things out of Europe.

Eric: And what is SPRIN-D in a nutshell? And also talk a little bit about how people should think about SPRIN-D programs and why you've gravitated towards the structures that you have.

Jano: Yeah, SPRIN-D's mission is to create what we in Germany call Sprunginnovation, or breakthrough innovation. And that is something that we define as innovation that really changes the lives of people for the better, in contrast to incremental innovation that really takes what we know and can do, and makes it a little bit better. And with that, we share many things with other ARPA-style institutions in the sense of, you know, we are willing and able to take a lot of technological risk in trying to achieve these great breakthroughs, even if that means that most of our projects will fail.

That is something that does not come naturally to public institutions, at least in a setting that has not been exposed that much to ARPA or DARPA-style institutions such as the US. But I think it was really, really important when SPRIN-D was set up that we gained that freedom from typical ways of doing science and innovation funding.

Eric: And can you talk about one or two failed projects that you're very happy with? SPRIN-D has done 10 challenges, something like that so far?

Jano: Yeah, I think we are at 12 challenges right now. So one of the challenges that failed in the sense of… so every challenge, typically, runs in multiple stages. It might very much differ how long those stages are, how many there are. It's completely different between different fields. One that we ran was on tissue engineering. What we wanted to see was complex tissues, artificially generated, that not only reach a certain size, but also a certain functionality at a level that has not been demonstrated before.

And we stopped that challenge after stage one, because we didn't see the progress that made us believe that in that setting we would reach the goal of that challenge. In that sense, it was not successful in reaching those milestones. But what we still saw in all those teams is that they really made good use of that money and our support in the sense of… They built stronger teams. They hired new people that eventually made them progress further, acquired other funding, and raised funding rounds. And so even though we didn't meet the goal of that challenge, I think it was really beneficial for the teams themselves and would give us another shot as a founder at SPRIN-D or other funders to really take that further again.

Eric: And could you talk a little bit about the types of teams that were participants in that challenge?

Jano: So it's typically, as so often in our challenges, it's a mixture between small startups and academic institutions. So universities, non-university research centers, and the team I was referring to, that was a small startup. They raised a seed round before. And with our money — that was really an amount of money that made us one of their primary funders in that startup. That's also something where we therefore typically have a very close relationship with the teams. Not only because we are really focusing on observing the progress that they have, but also because it is a close relationship once you are really the primary funder, and ultimately are essential for that startup either living or not living.

Eric: And I'll ask you more about financial instruments in a few rounds of questions, but for now, I'd love to stick on structure. In one of the pieces of writing you sent over to me, you talked about being agnostic — the challenge model being agnostic towards which approach is used to pursue some challenge, some goal. But some might say prizes are even more agnostic. Can we chat about that a bit? Because, practically, I think I know why you all chose the challenge model, but I'd love to give you first crack at addressing that argument.

Jano: Yeah. Just to differentiate between challenges and prizes — or sometimes it's more confusingly called “challenge prizes,” and it's completely unclear what the difference might be.

So in our case, a SPRIN-D challenge is a setup where we formulate a highly ambitious goal where we think that if it could be achieved, it would be highly impactful — a breakthrough where most people would agree that would really be a very beneficial breakthrough. But equally, many people would agree that it's currently probably impossible to actually achieve that.

And then we fund a range of different teams, with different approaches, in parallel to try to achieve that goal. Do the R&D. Do everything that's around that to achieve that goal. Whereas with a prize competition, you equally formulate a very ambitious goal, but then set up a prize, in the sense that whoever proves first that he or she reached that milestone, they will receive that cash prize, or whatever that prize might be.

The difference being that, for one, we fund teams along the way for the R&D work, and everything else that they need to have to sustain that work towards that goal. This is typically not the case in prize competitions.

Another difference being that prize competitions are typically more open to a wider set of contestants, in the sense of potentially not limiting entry at all. So, anyone could submit any kind of solution and proof of that solution to claim that prize. Whereas we start with a call, and we might select an initial set of teams. And then we continue to work with that set of teams.

And I think there is certainly something that speaks to both approaches. Looking at our settings, the reason why we fund R&D along the way primarily is because there is a lack of funders that would support that R&D until they reach that milestone for the prize. So I think we need to…

Eric: I remember one time I was talking to somebody who did — it was an economist who did market shaping. It was in the life sciences, where a lot of your contestants would be the Mercks, the Pfizers of the world, with a big war chest.

And when I raised the issue of — I think a lot about small R&D groups that have difficulty getting funded but are very resonant, like BBNs, which I've written a lot about a lot on the Substack. And the economist kind of just looked at me weird and said, “Why would you want to condition on or incentivize something that's not the outcome of note?” Which, in a purely theoretical way, I guess is true, if you have a big enough war chest or you make the outcome prize big enough where other people will come. But practically, there are a lot of times where I think of valleys of death as just areas where there are not a lot of incentives or funding. So, to me, it makes a lot more practical sense if you're dealing with small, resource-constrained R&D groups to take this [challenge] approach.

Jano: And I'm very sympathetic to that prize idea in general. But if you talk about the Mercks and Pfizers of this world, you would need to have a substantial prize to make them move. And even then, I have been working on market shaping in drug development, on infectious disease antiviral therapeutics, in particular. And that is an area where I think market shaping, for example, in the form of advanced market commitments, is really something that is necessary to make sure that new technologies actually reach the market.

But in the conversations I had with pharma, for example, it became very clear that even if there is a cash prize, however large it might be, that would not be sufficient. Because what they are looking for are natural markets, or natural markets that might occur due to those instruments. Because any kind of cash prize would need to compete with the natural markets they are in: in oncology, in other blockbusters that give them way more return on investment. And competing against that with a one-time payment will be very hard.

Eric: Yeah. And can you talk about how you think about follow-on funders when you're coming up with challenges? Like SPRIN-D can't fund these things forever. Usually, there's some end state, whether it's a VC-funded company or whatever. So how do you give those instructions, or what line do you toe internally when it comes to follow-on funders?

Jano: So I think the most important thing is that we think about that, right from the start. And within SPRIN-D, you know, we don't have a transfer department. We don't do our R&D funding, and then the project, and then push it to the next building, or something, to other folks who work on transfer.

It goes so far that our challenges, and the goals of our challenges, are shaped by how we might think about commercialization. One example of that is we have been active in a space that is called carbon dioxide removal. That is about removing CO2 from the atmosphere.

That is something that — from all these climate institutions — we know we need to do if we want to reach our climate goals. From an economist's perspective, it's a difficult thing to do, because removing CO2 has large societal value, but pretty much no private value.

So if you are a company like Climeworks, working in Iceland and pumping CO2 underground that mineralizes there in the basalt, there is no private value for a company to do that. They don't have a product that they could sell. So, currently, that is very much sustained by voluntary contributions by industry, but that is not a condition that is ultimately scalable at a global level.

So what we tried to do, and funded in that challenge, were approaches that remove CO2 from the atmosphere, and do that in close relation to existing product markets. For example — including that carbon in long-lasting and valuable products. Think about carbon-negative concrete, for example. Concrete has an established product market. It is one of the largest flows of man-made material on earth. And if you were able to develop carbon-negative concrete, it would not only avoid emissions, but due to that negative effect, it would essentially remove carbon from the atmosphere, in principle, the more you build.

And that is where we start. And then it is something where we very closely look at how we select teams — whether they might have the capability to eventually go to market. And that's also something that we focus a lot on when we mentor and support teams.

Eric: And can you talk a bit about SPRIN-D — is it ‘challenge leads’ in your department, and do they go by' program managers or something elsewhere? Or how do you all name them? The equivalent of DARPA program managers.

Jano: So, at SPRIN-D, we call them innovation managers.

Eric: Innovation managers, yes.

Jano: Yeah. And we have a group of innovation managers in all different kinds of fields. And so they have a high degree of freedom on which problems to tackle, and how to tackle them. So it's not that we have specific innovation managers that only do challenges, or only do whatever kind of instrument. Because I think what's really worthwhile, with a set of financial instruments that is as varied as ours, is that we think good and hard about how we best combine them, to get the most out of the potential that we have there for developing new technologies.

Eric: And where have you historically found your innovation managers? And can you talk a bit about how you're looking for innovation managers currently? Can you tell us about that?

Jano: It's a rare type of people, let's say that! Because what we are looking for is a very interesting combination of deep technological and scientific knowledge and skill, but that combined with what I sometimes call multi-sector experience. So we wouldn't want to have someone that only spent his time in his academic lab all his life. We would want to look for someone who has additional experience from industry, from founding their own companies, from being an investor, because these are all aspects that matter when we eventually want to commercialize those technologies.

And as I mentioned, it's not something — that commercialization aspect — something that we do at the very end of our funding process. But we want to bake it in right from the start, do it in a very integrated manner. We need to have people who not only have that experience, but also have the perspective of what it means to cross those different sectors, which is not easy.

Eric: And are there fields where you've had an easier — they're hard to find — are there fields where you've had an easier time finding innovation managers? Just because some circumstances of the field have made it that more of these are likely? For example, when I talk with some venture capitalists who do deep tech VC, some tend to have a pretty good time with electrical engineers. There's something about the field and the circumstances and the history where they go, "Oh yeah, we actually find quite a lot of founders in this area." And they have more difficulties in some other areas. So I'd love to hear your experiences on that.

Jano: I think there was no field where it was easy. Obviously, one thing that clearly plays a role is competition out there, for certain talents. In the field of AI, getting world-class people is not easy. And I think that's also something that's very important that we think about as a public institution. How do we attract the best talent, both for our employees, but also out there in the teams that we fund?

For some reason — I actually cannot say for sure why it happens — a large share of our innovation managers have a background in physics. Even though they might have done something else in the meantime. That just happens to be the case. I don't know whether it is random or whether there's a larger trend behind that.

Eric: And how should potential innovation managers conceptualize the opportunity? Oftentimes, they might have a vague sense of what's “ARPA-shaped,” what's “ARIA-shaped,” what's “SPRIN-D-shaped,” etc. But you've probably done a lot of thinking about this, and can make it quite clear for them! If you wanted to tee it up for a thousand potential innovation managers, how they should think about the opportunity, and why this is a chance even more exciting than founding a lot of companies to gain a foothold and do scalable innovation management.

Jano: Yeah, I think the kind of theoretical aspect of that is that, in pretty much all that we do, just having one shot might be so risky that if you really want to change the world, you better have multiple shots. And you probably have a hard time having multiple shots if you just found a startup. Then you do your one shot.

If you come to SPRIN-D, or any other ARPA, and run programs, you have multiple shots on goal, and therefore also increase your probability of actually having success.

Plus, I think specifically for SPRIN-D, if I look at concrete examples, one being one of our innovation managers, Patrick Rose. He used to be Chief Scientist at the Office of Naval Research Global, based in London. And he came to SPRIN-D because we have a toolset and an environment in Europe that was unique for what he wanted to achieve. For example, he came to run one challenge, that is still running, which is called Circular Biomanufacturing. And that started at a point where we observed that industrial biotech is currently good at producing niche products with large margins.
And it does that built on feedstocks such as highly refined sugars. The societal impact of that is somewhat slim. What these processes cannot do is do bulk products — so, mass markets that typically have very low margins — and do that based on other feedstocks.

What he came up with was that challenge, where the goal is to produce these kinds of low-margin, high-volume products based on real waste streams. And I'm talking about the real messy things, like supermarket waste, food waste, whatever it is, whatever goes in the garbage bin, take it out and produce new materials from that.

And so that is something where — the way we work at SPRIN-D — was something where he thought it would give really the best chance to actually achieve those goals. As well as having the know-how on bioprocess design, as well as fermentation infrastructure in Europe, was a combination that he had been searching for and found here.

And that is something where I think also, for future innovation managers or challenge leads, that is their chance to use the framework that we have at SPRIN-D and leverage the know-how, the technical know-how that we have across Europe. And I think that is something where we can really push the boundary of what's possible.
And for me, it's really about matching talent and opportunities in the best way possible.

Eric: And if I can give you a slate of three or four questions, and you could probably answer them all at once, but I'll put them all on your plate. So what, are there any particular types of innovation managers you're currently looking for? Are any technical areas in versus out? Is anything too technically risky? And also, how many are you looking to hire or onboard? What are the timelines in which you do that? And I'll stop there.

Jano: Yeah. So, very good questions, because I think we might be operating a bit differently from other ARPA's in that regard, that it’s relevant for those questions.
The one thing is, in terms of topics and fields that we cover, I would say we are somewhat opportunistic in the sense of… we don't have siloed budgets. We don't have, I don't know, a rule where we spend a third on biotech, a third on AI, and a third on computers or whatever it is. We try to run the best programs we can, even if that would mean that for a year or two we spend a very large fraction on life sciences, and then you know, that weight shifts over time. It really comes down to finding the right people for the right program at the right time. We're willing to back and implement whatever we think has the most impact.

Also in that regard, it's hard to put a number on how many people we would want to hire. We just want to see the best people and then we'll onboard them, because it's also the case that the size of our programs varies quite a bit. So, just for challenges, the smallest challenge might be just a million euro in total. Whereas other challenges are larger, tens of millions.

And you could think of challenges, and we're actually having some in the pipeline, where it will be more about the hundreds of millions for a challenge. Obviously, that has an impact on how many of those you could run and how many people you onboard for that.

That is the pragmatism that we have. We are trying to find the best people, and then we'll find a way to work with them.

Eric: And let's say five really interesting innovation managers saw this podcast and reached out and they had programs in the 10 to 30 million range. Is that the kind of thing where if they were all interesting, you have the capacity to bring on that many, or is that kind of reaching the breaking point where you'd go, we'd hire some and we'd maintain relationships with others?

Jano: That's the fun part about being in a public institution currently. Especially, we are not in the same situation in the US and Germany, but we are certainly in a situation where we don't have an approved federal budget in Germany as of now. That means we actually don't know what our budget is this year and we don't know what it's going to be next year. Economists will know the debate about the debt brake in Germany, that it's limited how much debt Germany can make. That is a huge discussion currently.

But what we are currently creating is an exception from that debt brake for investments. So there is something where there's going to be investments in the hundreds of billions. And part of that will go into research and development and innovation. But we are currently at a point where we don't know what our budget exactly will be next year. So therefore, I think it is a great time and a great opportunity to think about the question of whether SPRIN-D might be a good place to really push those programs because we might really be able to implement them and run them.

Eric: And in terms of growing the supply of interesting innovation managers with actionable ideas, do you want to talk about the partnership you have spun up with my colleagues at Renaissance Philanthropy? Because of course, the program they've spun up comes from this line of thinking — that of the number of really interesting technical minds who could do the technical side of being an ARPA-style PM, a lot of them don't learn to think appropriately in terms of servicing a follow-on funder, thinking in demand from day one of an ambitious program. Because programs at SPRIN-D, or ARIA, or ARPA are applied and ambitious and it has been hard to marry the two. So can you chat a little bit about that and what you're doing with BITS?

Jano: Absolutely. BITS, standing for Big If True Science — it's a program by Renaissance Philanthropy, a US-based philanthropy.

And the purpose of that BITS program, that accelerator program, is to take on people who have a clear vision for future breakthroughs, about capabilities that really might change the lives of the people, and how technology could support us getting there. And the mentoring program supports them in designing R&D programs that enable firms, any kind of institution, to really take those steps to achieve that goal.

As we mentioned, it's a rare type of person who has the skillset that is necessary to lead such a program. Therefore, it's absolutely necessary to look for those people globally. And that's what we do. We look for the best people globally. It is really also a special skill and knowledge set, designing those programs, because that is not something you would typically do at a startup or in an academic institution or as a policymaker or in any other kind of position. Plus, the number of such programs that are run are — I would say — very limited globally. Therefore, there are not a ton of people who have very good experience and insights into what works, what does not work. So therefore, combining our experience at SPRIN-D with the experience and expertise of Renaissance Philanthropy, which can really draw strongly on experience from experienced DARPA program managers, for example, or other people who used to work at different ARPAs — that is something that I think is super critical in designing the best programs. Because in my perspective, designing the best programs matters strongly for our actual success probability in the end.

Eric: Yeah, and we're going to talk a lot about know-how as we get into the nerdy metascience bits of this conversation, including in a couple of minutes.
But essentially, if somebody technically thinks they have what it takes to be something like a SPRIN-D innovation manager, it's a kind of program meant to help you get up to speed on the know-how. And if you're interested, you can email Joshua, JP, or Ales at bits@renphil.org.

But now to get into some very nerdy bits that you know a lot about and SPRIN-D has been very impressive on. Do you want to talk about the range of financial instruments you've come up with, and why they're a big deal in your eyes, in unlocking a lot of the upside that an organization like SPRIN-D is meant to go after?

And I have it printed out here too, in case we miss anything.

Image courtesy of SPRIN-D

Jano: Absolutely. In contrast to many other ARPA-style institutions, we have a very broad mandate in the sense of we might be working on life sciences, drug development, biotech, computing — in principle, everything that has civilian or dual-use cases.

That means we could potentially tap into every kind of field of technology or application. And I think there are fundamental differences between different fields of technology that we need to take into account when we want to design effective programs. Plus, we might start funding different kinds of institutions at different TRL levels, at different levels of maturity of that technology.

Taking all that into account, for me, means there is not a one-size-fits-all, "we just hand you over cash" kind of way of doing things. And therefore, I think we are pretty happy that we got a whole range of different financial instruments. So, different kinds of grants, different kinds of R&D contracts, equity investments, convertible loans, and I think that is something where we can really tailor our support to the needs of that particular team. An academic institution will have completely different needs than a startup raising a Series A or a Series B, and if we think they need our support, then we must be able to provide them the best support possible.

Eric: Could I run a couple of hypotheticals by you and you walk me through if they're possible at SPRIN-D and what the R&D piping to make it happen would look like? The first one is somewhat niche — but I think important at an organization like SPRIN-D.

Let's say there's a university with a postdoc and they’re in some niche area that's exceptionally important to an area you'd like to run a challenge in. And this postdoc, or them and a group of their colleagues, are just the best people. And what would make the most sense is for them and a team of two or three to leave and set up shop as a brand new company or nonprofit, and it would be inordinately important to the challenge being viable.

But what it might take is a half-million euro grant to this org that, while these people are proven commodities as postdocs — this is a day-zero org. It has no track record, it's never filed taxes. Can you do that? How would you do that?

Jano: So we can absolutely do that. That's not an issue. It's something that we see very often — that when teams apply with us, they might actually still be a postdoc, at whatever kind of institution, at an academic institution.

But they tell us, if you fund us, we are going to create a spin-out, and then that funding should go to the spin-out. So we can fund those kinds of no-track-record startups, institutions, people that have very little track record or no track record in that kind of sense. So technically, that's no issue at all.

What we would really focus on is a question of whether these people are actually willing to do it. And really, the part of being willing is way more important than the question of whether they're currently capable — in the sense of, when you have people who have spent their whole career so far in academia, they will not have a clue on how to build a company. Which is okay. That is to be expected.

What we really would need to see are signs that they're coachable, that they have potential and can develop. And that is certainly also something that is very crucial for us when we think about follow-on funding.

Eric: And so to turn two knobs a bit and see where this becomes difficult.
Is there an amount of seed funds where you'd say, “Oh, that's going to be difficult. We can do that, but doing it with 500,000 euros versus X are just qualitatively different things.” Like, where does it stop?

Jano: That's a very difficult question. I think it, again, it's also very case-dependent. So if we talk about nuclear fusion, there's probably nothing very cheap to be gained there. You would probably also be looking at very technical people that you would need to support there.

And so it would be a combination of the technical profile of the people and the amount of money that you would probably not be willing to spend on — for example, a software project where you have a purely academic person and hand them millions as funding — that's probably something that we'd never do.

And so I think, as an economist, as an academic, it depends.

Eric: Yeah. And is there, do they have to be in Germany? Do they have to be in the EU? Is the UK okay? Is the US okay?

Jano: We have a broad understanding of Europe and we fund all across Europe. So the UK, from my perspective, and from a personal perspective, I consider the UK still a part of Europe.

We fund in non-EU member states as well, including the UK, Switzerland, Norway, countries like that. Also including Israel. So for a technical definition of that geography — it's EU member states, states of the European Free Trade Area, plus the UK and Israel. That's our typical geography in which we fund.

Occasionally, we go beyond that, especially also when we run challenges where we have what we call a non-funding track. So sometimes, we would open up the challenge for teams that we don't fund but might still want to integrate.

So for example, we had a challenge on, let's say, digital infrastructure, digital proofs of identity. And we had Samsung and Google Android as participants, but we didn't fund them. So that was one of the cases where we extended our geography through that non-funding track.

Eric: And when you're selecting, let's say, postdocs who want to start a new organization, you brought up "do they want to leave?" as one big question.
And practically, of course, that's one of the big questions. But when VCs are funding firms, they also ask, “Are you a founder? Do you have the stuff?” And at VC firms, they're remarkably unstructured a lot of the time. Like even at the biggest firms, they make a judgment call on that. A lot of people maybe think a judgment call is the best way to do that, but they also don't think a random person making a judgment call is synonymous with how government typically does things.

So can you give us a sense of how this goes internally at SPRIN-D?

Jano: Yeah. So if I'm focusing in on challenges — what I really think is super important about challenges is that stage-based approach, that gives us the opportunity to take on a range of different teams where we might be uncertain about some of their qualities.

Obviously, there needs to be something in them that makes us really want to give them a chance. But we try to be very explicit on what it is that we give them a chance on. And that gives us a chance to — either we see someone who has an amazing track record in actually executing stuff, multiple companies founded, brought things to market, but might be somewhat lacking in the tech aspect. Or the other way around — maybe technically super strong, but no track record whatsoever in terms of actually implementing things that can go to market.

If we see the potential in those blind spots, we give them the chance, take them on, fund them for a bit, and give them the chance to demonstrate that they can do what we're unsure about their qualities. And then, at the end of the stage, during that stage, we might be able to pull the plug and say, okay, either they didn't perform as well as we hoped for, and then that journey is over, or they really outperformed.

And there might really be very positive surprises, because every once in a while we get those people who come out of academia, who have never been in any kind of other role than being a postdoc, PI, whatever. And suddenly, it's an awakening and they really have a very strong development up to the point where they say, “If I would have known earlier what I know today, I would've founded five companies before.”

Giving those chances through the challenge-based approach where we can just say, okay, we admit 10 and we see how it goes. It's something that I very much like — and I think it's very important if we want to be able to solve problems that apparently are so hard that no one else has solved them.

Eric: And in terms of the makeup of the performer pool, there's one — like a piece of what some people would think of as operational or contracting minutiae that to me is at the core of the SPRIN-D model. Can you talk a bit about output monitoring, not input monitoring? This is particularly fascinating to me, because in reading a bunch of the early DARPA oral histories, right around the 1980s, you start to get gripes from a lot of the long-time contractors, places like BBN, where essentially they have long complaints about the nature of procurement they’re changing and “there being more accountability for the pennies than accountability for the results.”

So this is precisely what you're going after with output monitoring. Can you talk a bit about how it works, why you thought it was essential, and how it changes the makeup of your contractor ecosystem?

Jano: Yeah. So I think if you look at how science and innovation funding in Germany would classically look like, especially from public institutions — you’d have some kind of call, you write a proposal. In that proposal, you write what you're going to do and what you're going to spend your money on, that’s typically in a very detailed manner. You would typically be required to follow that plan, and it might be a three- or five-year plan, a very detailed one.

And — I'm not kidding — you have to tell the funder how much your train tickets will cost three, four years from now. [laughter] That’s just weird. But that also means that, for one, you're very much focused on what your plan initially has been.

In theory, you might be able to change it, but then you have to reapply and ask for permission. It also means that you have to prove all your expenses were in accordance with that plan you originally submitted. And that is just super cumbersome. Everyone who has filed travel expenses or had that kind of grant funding knows how cumbersome that is.

And typically it involves people who would be better off working on innovation. I think it's really mind-boggling that we keep the best people of our generation — and I think that is the people that we want to address — busy with sorting receipts and doing those expense claims.

It’s highly inefficient for a project that is run. And I think it has a very strong selection outcome in the sense of, you know, certain types of people might not expose themselves to those mechanisms.

For example, I was very curious to see how our challenges would run in the field of AI, because people there just have very strong outside options. They really don't need funding from any kind of public institutions, unless it's really in the hundreds of millions and might be something different. But at our typical ticket size, it's nothing that they would require. I was talking to one of the teams that we fund in one of our challenges called Composite Learning. They are Y Combinator alumni. They are venture-funded. They don't need us.

And what they told me is that they specifically made a change in their policy to not apply for public funds because they saw that we act differently. That we don't waste the time of their most precious resource, which is the people. But we really support them in a way that makes sense. I think what's interesting is that we have to square that with our responsibility towards taxpayers. We cannot just get rid of input monitoring and then say, “Here's the money, have fun,” because that would be neither legal nor responsible.

So what we do is focus on what you mentioned, which is output monitoring. We are very close to the teams, so we have regular update calls, meetings, we visit them. And what we care about is new data being generated, demonstrators being built — something that is tangible, where we can see actual progress. Most of that will mean learning things that don't work, which is a great way to actually know what progress can eventually look like.

And by replacing that input monitoring with output monitoring, we actually make that monitoring mechanism worthwhile for the teams as well. Because if we are in a stage where we have people at SPRIN-D that are so good that the teams expect us to actually be helpful in overcoming their hurdles, they will voluntarily share with us what went wrong, because we might help them to actually solve those things.

That gives us a transparency that I think is way, way higher than if we would just look at how much you spent on a computer or on copy paper or whatever. I think it's way more productive.

Eric: And if sometimes groups will enter a challenge competition and they pick one technical path to pursue, and the path itself proves to be unfruitful — often for no fault of the team's own, other than a not-perfect initial hypothesis. But that's why you get in this game, is to test the hypothesis, not to be right off the bat.

If they proved to be an exceptional R&D team who just got unlucky with the hypothesis they explored, do you all have a way to keep them around and throw them at further stages of the challenge, if you thought that they proved themselves to be the right team to do that, and they just need to build off another team's learnings or something?

Jano: Obviously, it somewhat depends on the extent of that. But to give you a concrete example, in one of our challenges in drug development for antiviral therapeutics, one of the teams was expecting to use a specific type of delivery mechanism. It's one of the essential building blocks of the final product that will have huge consequences on whether it might be effective, what side effects might look like, things like that.
It's a really crucial part. And they changed — completely changed — what their delivery mechanism will look like mid-stage. I don't remember whether it was stage one or stage two of that challenge. And the process for that, if you can call it a process, is they call me or the innovation manager running that challenge and say, "Look, that's the data we generated. We think we have a better solution. We need a switch." And then we say, "Of course, you need to. Let's do it."

And it's the other way around [as well]. If, at the end of a stage, they would write, as they do, a written report and they pitch, and if they come to us and say, "You know what, months ago we figured XY is not working. Can we change it?" And then we would ask them, "What did you waste your time on, the last couple of months?" Because once you figure it out, that's the time when you need to change it.

Eric: And the government should care, 'cause that's time you paid for, too.

Jano: Absolutely. Yep. But it's not the typical logic. The typical logic is, we need to be able to make sure that money is spent the way we allowed it to be spent. And that keeps you on your path even if it might be completely bonkers. And that's the rule, not the exception, in our business, because I'm 100% sure that none of the plans of people that apply to us and get into a challenge can follow that path 100%. 'Cause technological uncertainty is just so high that it's impossible that everything will work out the way they planned. If that would be the case, they either lied to me or they weren't ambitious enough.

Eric: Yeah. And this is an approach that sounds common sense, reasonable, to most engineers and researchers. But also the economists love measuring things, and you're at least somewhat sympathetic to this.

You can imagine one variable, like an outcome metric of success here being, "Do you all get an inordinate percentage of people who are talented and are able to win a grant who have never won an EU grant before?" Because there's a certain bit of know-how involved in winning and maintaining an EU grant in the same way that, similar to SBIR grants in the US, where a lot of the researchers will roll their eyes and they'll say winning that is its own skill, and being good at research is just one thing they're meant to do, but it's probably not the primary skill.

So is this something you all track or would be interested in tracking or could care less about because you're busy with other things?

Jano: You mean in terms of it being a positive KPI if we attract people who never got an EU grant, because then they're the right type of people?

Eric: Yeah. Or something of that sort! You might think of a better one right now. I'm sure you will.

Jano: The difficulty there is, it's not a skill issue — for the pure fact that there are consultancies out there who write that application for EU funding for you. In fact, there has just been a study that was published on evaluating EU innovation funding.
One thing they found is that the effect of that innovation funding was very small, if at all positive, at those institutions that they funded. The largest effect that they saw was in consultancies that helped those companies and institutions get those grants in the first place.

And that's pretty much the only significant effect that they found on firm success — those consultancies.

Eric: So is something related to a consultancy popping up or being the major coefficient in your paper, I guess, a metric you would be interested in, then?

[laughter]

Jano: No, it's really bad in the sense of — a large share of funds that were supposed to be for innovation actually end up in consultancies, because it's so complicated to apply to most EU grants that you've got to have a consultancy that does it for you.

And I think that is a very bad state to be in.

Eric: Yeah. And can you talk a little bit about 14-day time-to-money, which is something you're very proud of and should be?

Jano: So 14-day time-to-money refers to how long it takes us in SPRIN-D challenges to go from the deadline of application to having teams selected and having contracting done.

If I zoom out a little bit, going from the ultimate goal we have, which is fostering breakthrough innovation — we care about what are success factors. Some of those success factors are attracting the best teams out there. And then, as we talked about, that might involve working in a way that is attractive to the best performers out there. I think therefore, time-to-money or speed of processes, lean processes, no unnecessary bureaucracy, is really crucial to address the best performers.

I think it is something that far too little attention goes into. It's also something that is not straightforward, in the sense that certain aspects of what is required to go at that speed are formalized. So for example — what are the legal requirements in procurement law or state law or whatever, what do our contracts look like? That is something that we worked out, that we can share, and that in principle every other institution could do.

What I think is the way more tricky bit, is doing it hands-on and focusing on the processes that are necessary to arrive at that goal. And that goes very deep down to what is the operating system of the organization that you are.

To give one example, how do we arrive at those 14 days? For a two-stage selection process that we typically have, it means that within a week we need to be able to analyze all applications that come in. That means I need to have a team, and typically it's a SPRIN-D internal team, part of our technical team at SPRIN-D, who have to be able to analyze all applications within a week. That means, for example, I need to be able to block their capacity so that four or five people, however many I need, are really there full-time, no nine-to-five, and work on those applications.

It also means I need to have in place a jury, because final selection at challenges is done with juries. I need to have that in place beforehand. I need to have those jury members, they need to be informed about what's coming, what the setting is, what we are looking for. You know, it's up until event organization. 'Cause at the end of that second week, we have pitch events where if we have 20 teams on the shortlist, we have an event where 20 teams are pitching over the course of two days.

Then we'll need to have a way of arriving at a funding decision at the end of that second day, at the end of those two weeks. Plus, we are looking at contracting as well. I need to be able to do contracting, in principle, overnight. That means at the end of the selection process, my team in the SPRIN-D office is preparing the contracts, sends them over to our directors, and they really sign those — I don't know, 15 contracts or however many teams we take on — overnight. That means the teams then the next morning have those signed contracts in their inbox.

That is something where, for example, as a counterfactual, I would have a very hierarchical organization with very strict limits. I would not be able to tell my team of analysts, "That's a week that you have to block full-time to do analysis." I would not be able to tell my directors, "You need to be there to sign 15 documents overnight." That very much goes to what I would call the operating system of that institution. It's first the ability and then also the shared perspective that this is something that is important.

Eric: And can you talk about your partnership with NSF and how you imagine transferring know-how on these matters back and forth?

Jano: I think that is something I would call tacit knowledge. Those formalized aspects, what are the legal underpinnings of what we do? What are the contracts? That is something that we frequently share. And anyone who wants to see our contracts can do so. In fact, we publish them all the time. But I think that for pretty much most institutions, it would be impossible to copy the results. Just because you have the legal basis and the contract, you will not be able to run this process in 14 days, or a week, or three weeks.

I think that requires very much hands-on process knowledge, experience, and knowing the limits of your organization. And that is also something where I think very little knowledge is shared, just because it is tacit knowledge. Also because such programs are not run that often. I've been running 12 challenges. I was talking to a DARPA program manager who said he was running a challenge and once he did that, he was trying to find someone at DARPA who used to run a challenge before.

But if running programs is so rare that you have to struggle to find someone who did it before, I think it's a very bad starting point for learning at a very professional level. And that's why, for example, in our partnership with NSF, what that encompasses is that we jointly run a challenge. So it will be run in that way that NSF funds teams in the US, SPRIN-D funds the European teams, but we will very strongly integrate all our operations down to the process level. Because I think that is one of the ways in which we can really share that tacit knowledge, the process knowledge, the know-how, and really understand how the other side ticks, what works for them, what might be applicable in each other's settings, and how we can actually improve on that.

That is also an aspect that I really love about the BITS program, because that is another way of really deeply integrating experiences, sharing experiences down to that level of operations that we otherwise very rarely have.

Eric: And are there any genres of, or sorts of innovation managers who you haven't had at SPRIN-D yet? For whatever reason, you've just failed to get a certain archetype that you would really like to come through that BITS program or hire as innovation managers at some point? I'm imagining, let's say, a potential innovation manager reading your site and saying, "Oh, they've never had anybody who looks like me before. This is not meant for me." I'm trying to give you a chance to flag that you would love X sort of person or Y sort of person.

Jano: You know, I wouldn't want to focus on one particular profile because I'd rather want to put it the other way around. I think there are potentially many people out there who might not know that they might be a great fit.

And they might come in various different shapes and forms. What is, I think, really crucial is that they have a strong commitment to change a field, that they have a strong conviction that what most people think would be impossible or might not even have thought about approaching or trying to deliver, that that is something we would really need to have.

In that sense, I really don't care whether it's in AI, computing, life sciences, or any other field. It is a strong conviction. And what I want to see as a track record is the ability to execute.

Pure theory will bring us, at least in this business, nowhere.

Eric: Yeah. And in selecting people like this, you do not have quite as finite an amount of money as maybe something like DARPA's BTO, which probably gets handed a specific budget, but the funds are somewhat finite.

While the specific number is hard to pin down and you have to make really difficult judgment calls between fields — like computer science and electrical engineering versus quantum biology, or how to generate more life sciences data. How do you go about making those judgments? What does that look like internally at SPRIN-D?

Jano: It's interesting in the way that I've been talking about the pragmatism that we try to have, and that has been supported, in very practical terms, by SPRIN-D's budget increasing from year to year. We are a young institution. And I think so far, for most programs we really wanted to run, we eventually were able to find the funds.

You know, we started out with a budget that came solely from our Federal Ministry for Science and Technology. What happened in the meantime was that, for one, more ministries from the federal German government saw what we're able to do, and that they're not able to do the same R&D funding, and gave us extra amounts of money to be active in certain fields. As well as other governments starting to approach us to run programs, just like the NSF collaboration was born out of the fact that NSF observed that we are able to run programs in a manner that's way different from how NSF would typically run programs.

That is something that we see increasingly — that we have other governments, mostly from across Europe, but also the US, trying to work with us. That also gives us the chance to have a much broader portfolio of programs that we can support.

Therefore, I really have yet to see the case where we say, "Okay, that's a great program, a great innovation manager, but unfortunately we cannot fund it." So far, in pretty much all the cases, we've found the funding. Because hopefully, if you can really talk about the impact that a successful program, a successful challenge in that field would have, and can show some of the track record that we have in the sense of — we really manage to bring people into the game that typically would not be addressed by public funding, and that really giving us an increased probability of success.

That is something that's very convincing to policymakers or funders — to actually go after that combination of topic and method.

Eric: So is there any really germane constraint that you would want innovation managers to know about before reaching out?

Like whether it's — is there a TRL you work in less, or is the answer, "Oh, your program needs to be a bridge to something interesting and it could be lower or higher TRL as long as it's a good bridge." Is there any quick framework that you think should shape their thinking?

Jano: Yeah. A very short answer would be, we do TRL three to seven, and that's it. But that's not doing justice to the complexity that we are in. When we talked about the difference between challenges and prizes — in prizes, you have that big honeypot at the end that might be very tempting, and we don't have that.

But our perspective is that we try to design challenges such that if a team is successful over the course of that challenge, then that should unlock follow-on funding. That might mean different things in different fields. It might mean, in certain fields, that point is coming earlier.

In biotech and AI, for example, there might be cases depending also on one aspect that is market risk, that we have talked about, but I think it's very crucial to differentiate between technological risk and market risk in how you run programs.
We try to think about that kind of handover point before we start a program, knowing that if we learn that other funders would not be able to take over funding in cases where we see great potential, then we would have the means, the instruments, and the budget to further support those programs, those teams. That's also what we did do in the past. For example, you might see teams coming out of challenges but having trouble raising the necessary funds due to excessive market risk. And then we will be able to catalyze funding rounds getting together by being able to fund actually up to 70% of a funding round.

Eric: And I'd love to ask you — take the time and ask you three niche metascience questions that you could have quick answers to. Or you could dive into them if you think they're important! And so the first one would be, in understanding the opportunities and constraints of the innovation manager position, are there scenarios in which they might really want to do something where somebody will stand in their way, or them wanting to do it is not sufficient for SPRIN-D to do it?

Jano: That might be the case. For example, what we have in our group of innovation managers is very strong experience and knowledge from the investment community.
There might be innovation managers out there or at SPRIN-D who have less experience in that regard when it comes to structuring funding rounds, clearing up cap tables, and things like that. So there might be a case where we go into that discussion and decide to make sure that we need to see certain requirements fulfilled before we can do follow-on funding, for example.

And that is something that, due to the experience that we have in our group of innovation managers, we also can very much lean on. Therefore, I'd rather see it as arriving at a better result due to our internal processes, rather than forbidding a program manager to do something.

Eric: And who picks the juries that decide on program applications? Who makes up the juries? Are they looking for unanimous approval from the jury? Or can high-variance reviews on applications be good?

Jano: Juries are one thing that we only do in challenges. And it's typically a combination of the innovation manager and me looking at jury members and selecting them.

What is very important to me in that decision-making process is, for one, that juries are somewhat small, meaning typically around six to eight people. What I tried to do when we set that up was try to combine a broad set of perspectives and experiences, both in tech and commercialization, while avoiding having a large committee.

Because large committees typically tend to agree on the smallest common denominator. And that is exactly not what we want. We want to take those bets — those high-risk bets that might actually be able to really pay off. Therefore, the decision-making process within that jury very much embraces that variance in opinion.

It also very much depends on the field, on how much variance we might have in there. One measure that we take to embrace that variance is to have a system of golden tickets. So, for example, for the initial selection for stage one teams, the average jury member has a golden ticket, and if he or she thinks that team X is really so exciting in one category, “I really want to see them perform, I want to see that thing demonstrated” — then we fund them.

We do not have a majority vote typically for the decision-making where we say, “Okay, the majority is for the team, we take it on, and the majority is against, we don't.” We try to arrive at a consensus within the jury, where we have a shared opinion on the strengths and weaknesses and risks of a certain team, and decide why we want to give that team the chance to perform or why that's not the case.

Because that also gives us a basis to evaluate them against our decision, over the course of that stage.

Eric: And who gets final say, once you incorporate all of that and you say, "Oh, we've generally frozen on an agreed-upon list of strengths and weaknesses," whose job is it to look at that and say, "We're going to do X"?

Jano: It is the jury. The jury will arrive at a consensus, even if it means to agree to disagree on certain aspects of a team. It is my job to moderate these juries to arrive at that consensus. And it's really something where, from afar, it might look somewhat weird.

For example, I was in the US, I was talking to a former office director of DARPA, and I was telling her about our selection procedure. And she was like, "Juries? That's not how it's done," coming from a DARPA background. Then she and several other people joined us for a certain jury session.

And she was in a jury session, saw the discussion, saw how we arrived at that consensus, and saw how we tried to aggregate that information that each and every single jury member brings, and build a consensus about what kind of the best-performing teams might be down the road.

And then came to the conclusion that actually, that is a very nice process.

Eric: And do you have something written down on what you would do in the event of a hung jury?

Jano: Never happened. I don't know. I was like in like — I don't know, like 30 jury sessions — never happened.

Eric: This is fascinating. This is also one of those areas where I can imagine there being cultural differences. Like when I learned how Swiss children vote on what field trips to go on, even from exceptionally young ages. It just wouldn't work in the US. [laughter]

Jano: I don't know. But I mean, we recruit jury members internationally.
It's not just Germans sitting there and agreeing with each other. So we have international juries. European, North American, Asian, it's really diverse. And that is exactly what I mean with tacit knowledge and something that you might only be able to fully grasp with all its pros and cons when you are in the process, because it's highly context-dependent.

And that is something where, from my perspective, we don't have — so far — a very good way of learning. And it's also something where we typically don't have the scale to do proper learning. I was talking about the amount of times that we run challenges, other institutions run challenges.

If you think about the fact that we run or start a new challenge maybe on average three times a year, what is the process where you would think that, on a professional level, you get really really good at things that you only do three times a year? And I think it's something that we do more often than other institutions in the field.

Therefore, I'm really concerned about that. How do we make sure we learn across institutions, especially those aspects of tacit knowledge? Because I think these really matter. All that I said about how the operating system of a funding institution might have a critical influence on success factors such as speed — it's nothing that I think we have a good way of learning from one another.

If I look at other fields where you have a strong dependence on tacit knowledge —think about the culinary world, restaurants. That is something where I would like to compare the way we structure processes to. There is this French term called mise en place that describes the situation where chefs would prep their kitchen and station beforehand, to be able to execute when service time is and execute both with quality and speed. It's a very deliberate process about what you prepare beforehand, what you focus on in the moment, to achieve the highest quality at the highest speed. And that is something that we try to do when we try to achieve those 14 days time-to-money. And it's something that you cannot easily copy.

But in that culinary world, with all that tacit knowledge and hands-on experience, we have systems like apprenticeships, mentorships, where you go through stations, you go from restaurant to restaurant to learn all that tacit knowledge. I think we don't have that to that degree in our world.

That's also part of why I think that collaboration with the BITS program, with NSF, to really go down to that level of operations is really exciting. Because that might be one step of trying to do that learning and then transferring of insights on those elements of tacit knowledge.

Eric: Yeah. And so, I'm going to move us along a little, skip a few questions because we're running out of time, but I have a few I really want to get to with you. So one of them would be — we're going to have a lot of West Coast listeners in the US, and Boston and the UK, but I have a pretty West Coast center of gravity. And they're going to have a lot of curiosity about the German R&D ecosystem in general. And also, you might want innovation managers and things in cohorts coming from American institutions and things like that! So I would love to tee you up with something and I would love to force you to step out of your own skin slightly here.

So — I get the sense that Germans don't really brag enough. For example, I read the Fraunhofer-Gesellschaft annual reports that come out, and as an American, I see a lot in there that makes me jealous. And I think there are things in there that make the UK folks jealous as well. Like, I think Catapults and initiatives like that were overtly Fraunhofer-based.

But what I've brought up Fraunhofer to — and for those who don't know, Fraunhofer is the applied version of the Max Planck Institutes, in very simplified form, in Germany, and they have a whole ecosystem of them. But when I bring up Fraunhofer to Germans, and like really fantastic metrics from annual reports, like they have an 80 or 90% spin-off survival rate after five or 10 years, German folks often bring up to me all the things Fraunhofer's not doing or could be doing better. But if I was going to force you to put on your West Coast of the US hat and brag for a bit, could I hear your most optimistic pitch for the sorts of ARPA-like programs that are possible in Germany, that are difficult elsewhere?

'Cause you have a different set of institutions, often with very different focuses or that exist in places that are the valley of death in the US in certain fields.

Jano: Yeah, if you think about the innovation that led to what now is ASML and EUV lithography — that is a degree of engineering that I think is very hard to find elsewhere. If you think about the fact that you have countless technologies at the highest level of engineering, that you then need to integrate to build something like the lithography machines of ASML, that is super hard. If you think about the precision that is necessary to shoot those lasers at those tin droplets that create the radiation that is used for that lithography process, it was probably a 10-year R&D effort of major companies and institutions involved at an unprecedented level. And I think that is something where you really need those sustained levels of engineering at the highest quality in areas that have never been tried before. I think that is where we really have great institutions in Europe.

Eric: And would you care to, for the listeners who won't know precisely what you want to talk about, can you share 3, 5, 10 — whether it's areas, specific centers of engineering excellence — where you would just want people to know if they're thinking of anything in that area? Like a strong argument that Germany is the best game in town, sometimes the only world-class game in town, to do work in that specific area. Like Americans might off the top of their head think of a company like Zeiss and understand that we don't have things like that, but there's probably all sorts of not-famous firms in areas that you think Germany is world-class at. And people don't force you to brag on a podcast like I am right now, often enough to talk about that.

Jano: I think it's really not about those singular, stellar companies or institutions. I think the real strength of Europe is its very high level of quality across the board. It's the diversity that we have with very strong centers of excellence.

For example, life sciences in Denmark and Copenhagen, engineering in Germany, places like Munich and Berlin, AI also in France as well as the southwest of Germany. And I think the real strength is that all these places are really, from American standards, not that far away from one another.

Plus you have a really strong industrial base that could take on things and really scale. And I think that is a thing that we talk not enough about. As I mentioned, our innovation manager, Patrick Rose, came to Europe to run challenges with SPRIN-D because we have that infrastructure in biotech that really enables us to not only develop, but scale things.

Eric: Okay, fantastic. Thank you for that. I have a bunch of questions that you could probably run through each of them in under a minute that I would love to get you on the record talking about, but none of them is that important. This is the list of things I wanted to catch at the end here.

And so one of them was, when you're putting together challenges, you can understand why people think, "Oh, you don't want to condition not on the outcome or constrain things in that way," but it's often essential for practical reasons. Is coming up with those intermediate benchmarks painful, or is it actually often quite straightforward and forces you to think hard about it?

Can you talk about the trade-offs and the emotions involved in coming up with those?

Jano: It's very difficult. Very hard to do well. And from my perspective, it cannot be done purely based on quantifiable metrics and KPIs, because the technological uncertainty is too high to be able to exactly pinpoint both the exact KPIs that you want to use and the right threshold levels. But it also brings in a factor of appreciating all the work that the teams do in different dimensions. The combination of seeing teams in parallel gives us then the opportunity to not just like a team because we like the people, but really compare that to competitors and really show who is ahead.

Eric: And so I talked with Ilan Gur about the idea of one ARIA program leading to another, and he said, "Oh, I actually really hope we do something like that." Is that a tool that SPRIN-D thinks about, or is that a bit far away? 'Cause you're giving yourself a lot of leeway to not think in translation.

I'd love to know how you consider that, 'cause there are obviously pros and cons to the approach.

Jano: Yeah, no, absolutely. And we do sequels. For example, when we look at autonomous systems research and development, we did a sequel to a challenge that we ran last year and run version two this year. Not unlike the famous DARPA Grand Challenge, where pretty much everyone failed in the first attempt, in the first challenge, and then you have a shot at it again in a slightly different way. But no, that's certainly something that we do.

Eric: And so to sit on that DARPA Grand Challenge point, that's an area where DARPA was a workhorse funder of autonomous vehicles from the early eighties up until the Grand Challenges. So that's almost two decades where they had a typical like, "gardening, we don't know what is there in this space yet, we have practical outcomes we're looking for, but also we're developing a tech branch here," that ended in a challenge.

If you had an area that felt like the one approach was ideal for one decade, and then a completely different approach would be ideal a decade or two from then, how might SPRIN-D go about attacking that field in terms of its program structure? If you were going to do sequels, what would that look like, if you all were trying to do the same thing?

Jano: You're talking to an institution that is five years old. [laughter] So I think most of it is still to be developed. Let me give you another example, and that is nuclear fusion. We have been funding technologies for nuclear fusion for a couple of years now, and that certainly will not be finished next year or the year afterwards.

So we continuously evaluate what the right next step is. That depends on tech development. That depends on the whole ecosystem developing, how other funding efforts are structured. And that is certainly something where we think a certain field of technology is both super important and would have a great impact, but wouldn't be able to move along without us.

Then we are certainly willing to continuously support that development.

Eric: And SPRIN-D exists in part to fund stuff that kind of presently ends up falling into the valley of death more than it should. Let's say there was some contractor where it maybe wasn't worth any given program or challenge officer funding the org into existence, but SPRIN-D itself would benefit from some contractor existing and being good for the rest of the ecosystem.

And a specific example from DARPA history — I think of early MOSIS, where DARPA kind of created their own silicon broker at the USC Information Sciences Institute. What they did was enable like really small chip runs for all of DARPA's researchers. I don't know if you had to spend that out of your program, if you would do it as a program manager, but like from the office level, it makes sense. Structurally, how would you all go about doing that?

Is there some straightforward mechanism? For example, ARIA has activation partnerships to do things like that. How would you all handle it?

Jano: So it would be something that we would steer from within SPRIN-D. We don't rely as much on activation partners as ARIA does. We do more of that work SPRIN-D internally, but it certainly would be something that we would look at and would be willing to do.

When we figure out that there is a certain bottleneck that is in the way of achieving an important breakthrough, then we would address that bottleneck. Even if it would mean that that work in isolation would not be a breakthrough. We're just aware that oftentimes those breakthroughs are different puzzle pieces fitting together, things building on top of each other.

It's not just the one problem that we solve and then it's a straight line to a breakthrough.

Eric: And if SPRIN-D's budget 10x'd, would we see 10 times as many challenges and programs, or is it not so linear?

Jano: So I think the one question is, do our ticket sizes get bigger? And the other question is, how broad are we going to be? I think there is a limited scope for having bigger ticket sizes. The amount of funding, the ticket sizes in a challenge — I think that will not change a lot. What we will look at is, if we have a 10x budget, are there technologies out there that need bigger ticket sizes?

Because, for example, we might need to fund phase two clinical trials in certain cases because no one else is going to do that, participate in Series B rounds because they would otherwise not come together. Really, I think there is a natural limit of what certain institutions or team startups at a given time can productively work through in terms of money. Having money is necessary, but it can make people lazy and slow if they have too much money at the wrong point in time. So it would rather be probably a combination where we have more programs and are able to address more challenges.

Eric: And if SPRIN-D was the only group funding some area for X amount of time, how long can X be before you start getting uncomfortable and saying, "Are we just making ourselves feel good here, and nobody actually values this area as much as we do?"

Jano: I think that can be a long time, especially when…

Eric: Is it five years? Is it 50 years?

Jano: Asking a five-year-old institution about what we are going to do when we run things for 50 years is like looking into a crystal ball.

[laughter]

But I think we have the commitment and the willingness to really follow paths as long as we are convinced that's the right way to do it. Because we start programs when everyone else would be against that. So we can continue programs when everyone else would be against that.

Eric: Yeah. And okay, just three more and then I'll let you go. And thank you for all the time you've spent with me today! First one, if you have a really wildly successful challenge, whether you had a huge technical success or you learned something that showed you the promise in this area and you said, "Oh, the expected value is even higher than we thought," or we got a good outcome — to what extent, what does doubling down for SPRIN-D look like?

And at what point is some percent of your budget maybe too much? Like I constantly bring up this Warren Weaver at the Rockefeller Foundation example, because it was 80% of Rockefeller's natural sciences budget. And I think a lot of orgs would say, "Oh, that's a bit too much." But I would love to know where expansion is natural and good if you're doing a good job, and where it might be too much of your budget or there's a sort of soft threshold that you wouldn't want to pass.

Jano: Yeah. That's a very tough question. I mean, if we spend a budget share of like 80% on a given technology, it would really warrant a question of why is no one else committing that amount of money into that project? I think it's very important to think about opportunity costs because spending 80% of money on a given project means not spending that money on a range of other topics and teams.

And I think that is something that we should be aware of and where we should have somewhat of a balance, especially given the broad mandate that we have. I think it's important to be aware that when we spend 10 million on contributing to some Series A, this might mean not funding 10 teams in any other fashion.

Eric: Second to last question. SPRIN-D is a metascience experiment in and of itself. And that being the case, I'm sure you look at all sorts of hyperparameters you can tune or things that could be improved. Are you all currently dealing with any bottlenecks for which you'd love some kind of organizational innovation to address? Or maybe you have one in mind and you're about to test it?

Jano: What I'm really concerned with is really the learning of that tacit knowledge and setting up ways to do that across institutions. I think that is something that really has great, untapped potential. Continuing that work with other ARPAs out there is really super important and I think can be greatly beneficial to the work that we all do.

Eric: Okay. And last question. So SPRIN-D as an organizational experiment, a government R&D experiment, and a scientific experiment — what's an example of a success of each where if in 20 years X happened, you would say, “We've succeeded. This has been a great experiment. I'm so happy I spent that portion of my life working on that, or this whole group of people spent this portion of their life working on that.”

Jano: I think one thing would be if we still exist and still work differently than the rest of government, that would be one, in terms of that government dimension. One thing that I start to see is that people change, for example, other funders changed their reporting requirements because teams that we fund would go to them and say, "SPRIN-D is not requiring that. Apparently, it's the German government, that seems legit, so why do you require it?" Therefore we see other funders in other countries lower their reporting burden on teams, especially that input monitoring.

That is something — when we can really have that push in the ecosystem, in the funders' ecosystem — that would be something where I would say that experiment went well, when we have that push to change how we operate, also beyond the ARPAs.

Eric: Yeah. There are a lot of scientists in the US who desperately hope you're right and that does come to pass. And then on the technical end, is it a Turing award? Is it a Nobel Prize? Is it another ASML?

Jano: It certainly would rather be the ASML than the Nobel Prize and the Turing Award. You know, one example that might be world-changing if it happens is, if any one of our teams from our very first challenge actually succeeds. Because that is about creating a new technological paradigm to treat virus infections. That's something that is really striking me in the sense of, we apparently just gave up on the fact and don't even notice it, that we are typically sick for half the year in the cold season due to some kind of virus infection. And we do not have the tools in antiviral therapeutics as we have with bacterial infections where we're just used to having broad-spectrum antibiotics, but we don't have that same thing for viruses.

We're just used to being sick, having a cold, having the flu. Not to speak of pandemics. And if we could solve that, really make those new technologies in that field successful, we would not only be able to avoid future pandemics but also do away with the flu, with the cold season.

Not suffer from not being at full performance for a couple of weeks in winter. And I think that would be something that might really be world-changing.

Eric: All right. Jano Costard, thank you so much. And I'll stop the recording there.

Jano: Thank you.