Q27 We will hear oral evidence from Dr Peter Highnam, deputy director at the US Defense Advanced Research Projects Agency; Professor Pierre Azoulay, Massachusetts Institute of Technology; and Dr Regina Dugan, chief executive officer of Wellcome Leap. We have until 3 pm for this session. I call on the witnesses please to introduce themselves for the record.
It is a pleasure to serve under your chairship, Mrs Cummins. May I also echo your thanks to our witnesses for taking the time to join us for this important session? I am in awe of the range and breadth of your experience in innovation and scientific research, and all the more grateful to those of you with experience of DARPA for joining us this afternoon, as the architect of the Advanced Research and Invention Agency, Dominic Cummings, the previous adviser to the Prime Minister, has apparently refused to give evidence to the Committee. We are able to go only by what he has said previously, and there seems to be some confusion as to what ARIA is and whether it should be engaged in cutting-edge research or in the translation of existing research. You might be able to comment on that.Q
Dominic Cummings said:
“The purpose of ARIA ought to be to sample in this broader design space, to do things differently, and to learn from the things that have been super-productive in the past. That means in very simple terms extreme freedom.”
Dr Highnam, does DARPA have “extreme freedom”? What does that mean in cultural terms? Does complying with, for example, US freedom of information laws or procurement regulations—it is proposed that ARIA would be exempt from them—impact on that freedom?
That is a great question. DARPA is an agency in the Department of Defense in the US Government. We have a number of regulations and laws that of course we operate within. We have a number of special authorities that allow us to operate a little faster and with a little more independence, but with oversight. It is a place that moves quickly. As you are probably aware, when you show up at DARPA, you have an expiration date on your badge, as we say, so you move fast and the whole place is geared to do that. The agency now has a record of 63 years of production—again, with oversight at all times. It gets the job done, in that context.
Q I will follow up on that and then bring in the other witnesses. You speak about oversight. Would it be possible to give a bit more detail on that? In particular, the UK Government are currently mired in a cronyism scandal; indeed, that is what is being debated in the House today. DARPA is well known for having exchange between itself and the private sector, but how do you prevent projects or programmes from going to friends, mates and those with, if you like, special interests, without some degree of oversight?
I can speak only to how DARPA operates. We have very rigorous review processes—technical, financial and others. We have conflict of interest rules and so on that we all follow. There are robust processes and independent looks at those processes. Again, we could not operate any other way.
If I might add one element to the question that the hon. Member asked, the programme managers at DARPA and also at ARPA-E—the Advanced Research Projects Agency-Energy—have a fixed expiration date, which means they will need to go back to academia or to the venture capital firm or large firm that they left, and generally they want to do so with their head held high and their reputation intact. I think that that has created over time a norm of correct behaviour, if you will, and the absence of cronyism. That norm element is also very important, in addition to the formal regulations.
Q When they go back, are they allowed, for example, to direct finance at the companies to which they return?
I served as the 19th director of DARPA and echo Peter’s statements that there are indeed oversight and regulations that govern the behaviour at DARPA. We have free and open competitions. One of the things it is important to understand is that part of the reason that innovation is so robust at DARPA is that there is a sense that there is an equal opportunity for many to apply to the programmes and to be fairly judged. As a result, many bring their ideas to DARPA. That is part of the robustness of the ecosystem that has developed around the agency. It is a very important element of the work.
Q Thank you very much for your responses. May ask one more follow-up question? I have spoken about some of the concerns about oversight and so on, but may I ask each of you what you think is the key positive element of culture? You have spoken about the desire to return with your head held high. What should we be looking for in the directors and programme managers as the key positive part of the culture that ARIA should seek to build?
Yes. You join a place like DARPA to change your field and make a difference for defence. We are a defence agency. When you come to DARPA, we give you the lever arm, we help you position the fulcrum, we give you the mass to make things happen, and we give you the processes around you to make sure, as Regina said, you do it fairly, openly and robustly. We do exit interviews when people leave DARPA, and one of my favourite quotes is, “If you don’t invent the internet at DARPA, you get a B.”
I second that entirely, but I would also say credibility in both the scientific world and the business world. It is a relatively rare breed of individuals who have credibility in both domains at the same time, but that is to quite a large extent the X factor in the typical DARPA or ARPA-E programme manager.
Q Thank you very much. Dr Dugan, I will ask you the same question. The tech sector and science and innovation are not known for their diversity, but we have heard that diversity of thought is very important in the agency. As shadow Minister, I would like to see a broad range of diversity in the people who are recruited, in terms of gender, class and race. I am asking this question of you because you are the last person that I came to. How can the culture promote diversity as well as being positive?
Let me take the questions in order. I would add that DARPA and ARPA-like organisations are optimised to create breakthroughs. Those breakthroughs happen at the intersection of some science and engineering that we are pulling forward in service to a new capability or a new problem solution. We design the programmes such that we have a very clear and ambitious goal that is also measurable and testable. Programme directors have a finite period in which they collect a group of performers from a mix of organisations and disciplines in service to that goal, and there is passion, spirit and urgency that comes with that. It cannot be created in the abstract; it has to be real in order to engender the kind of genius and collaboration that is characteristic of these programmes.
The programme directors are themselves scientific or engineering experts. They are great musicians, as you might think, but they are not playing an instrument at the time of conducting the programme; they are rather conducting an orchestra of expert musicians who together make a symphony. That is very important.
What I can tell you about diversity from my own experience, both in Silicon Valley and at DARPA, is that for decades we have known that specificity of goal and outcome is a good way to get more equality and diversity in assessment of ideas and in people conducting or pursuing those ideas. We know that across academic institutions and across companies. One of the things that is important is to set crisp and clear goals, because the ideas are then measured against them, and they can come from many different individuals and organisations. As I said previously, I believe that is central to building that ecosystem out, and for that ecosystem to be diverse and more equitable.
Thank you very much. I know that others have questions to ask so I will leave it there, but I just want to say how inspiring it is to hear such positive reference to the power of public service, science and research, and to oversight as being an enabler rather than a burden.
Q It is a great pleasure to serve under your stewardship, Mrs Cummins. I want to acknowledge the excellence of our witnesses and to thank them. With your permission, I will ask one question to each witness.
I am going to start with Dr Peter Highnam. How do you ensure evaluation and scrutiny of DARPA’s programmes outside what is mandated in legislation? What information do you gather to assess when to start and stop projects and programmes, and how are these decisions made?
That is a surprisingly big question. The p in DARPA stands for “projects”, which is critical for a place like DARPA. We are not doing technology area x or y just because, and we do not do it for the long term. We have projects that are well defined at the beginning. A case has to be made. They are monitored, they have metrics and all manner of independent evaluation associated with them before we go out to find the best teams we can to participate and to be funded to work on that research. Then that project ends. That is very important: things begin, and they end.
To make the case for a project to get off the ground, we use a structure called the Heilmeier questions, named after the DARPA director in the mid-70s, George Heilmeier. They are five very important questions. They look easy, but they are very hard to answer well. In my view, that is the creative act in the DARPA model—to answer those questions well and make that case. Once the project is approved and teams are onboard, you then have regular evaluations. As things change in the world around us, in science and technology, with us in defence, and in other aspects of our environment, they may be overtaken by events. That is very rare, but it would be grounds for no longer continuing. Were we too ambitious in certain aspects of the programme? Do we need to change it or change some of the people participating in the teams? And so on.
This is a constant process. It is not about starting it up and letting it run until it finishes. It takes a lot of effort to make sure you know what you are doing when you start with taxpayer funding and the opportunity cost that comes with that. Then you keep an eye on it, especially during the transition of the results to our national defence.
Q Thank you. That was a very precise answer to what I know was a very large question. I would next like to come to Regina Dugan, if I may. As you well know, Wellcome Leap was created as a separate body from Wellcome Trust. Why was that decision made and how will it help Wellcome Leap achieve its objectives sitting outside of the Wellcome Trust?
The story of Wellcome Leap actually dates back to about 2018, when the Wellcome Trust, from its unique position in the world, asked, “Is there more we could do to have greater impact?” It did a pretty careful analysis of innovation as it happened in larger organisations in the venture world and also at DARPA. The assessment was that in global human health, there is indeed this innovation gap. That innovation gap is characterised by larger programmes with higher risk tolerance, which are not driven by consensus peer review. This is very much the way we conduct programmes at DARPA—the intersection of a goal and the science and engineering that need to be pulled forward in order to attain that goal. That effort—those large programmes—are what Wellcome sought in the formation of Wellcome Leap. What I have observed in the last year of operation is that, in fact, there is this innovation gap in human health. It is same one that was identified after Sputnik that led to the formation of DARPA. The coronavirus is showing us just how much work needs to be done in human health across policy, equity and the economics, but it also shows us the power of a breakthrough and how tough it is to get one.
I was the director of DARPA when the pivotal investments in mRNA vaccines were made. Many others came to the table to create this success for the world in this time, but we need more breakthroughs like that, and we need them faster. That is why Wellcome Leap was formed.
Q Thank you very much. That is really helpful. My final question is to Professor Azoulay. In your paper you mention that organisations’ flexibility is essential to the ARPA model. Can you talk a little bit about what that flexibility involves please?
Absolutely, it is essential and I think it happens at multiple levels. It happens in the relative administrative autonomy that those ARPA-like agencies have, relative to their Government Departments of reference, whether it is the Department of Energy for ARPA-E or the Department of Defense for DARPA.
It definitely also happens at the hiring level and in the fact that one can hire programme managers in ARPA-like agencies from very diverse backgrounds, not necessarily a background in the civil service, and pay them according to rules that might not be those of the traditional civil service.
Focusing on programme managers, that matters because they themselves have quite a bit of autonomy in the way in which they delineate and orchestrate their programme. They have a lot more discretion in choosing what projects to fund and assembling the teams that will perform those projects than would be the case in a traditional science funding agency, such as the National Institutes of Health or the National Science Foundation in the United States context or, I would think, UK Research and Innovation in the British context.
Thank you, Ms Cummins. It is a pleasure to serve under you as Chair. I thank the witnesses for their comprehensive answers so far. They have been incredibly helpful.Q
I would like to pick up on a comment made by Dr Dugan, I think, in respect of the intersection of a goal, and using science and engineering to achieve that goal. It would appear, from looking at what is front of us, that the ARIA Bill does not have a goal. There is no mission or bright light that we are trying to get to. What is your collective view—all three of you—in relation to that? ARIA has no mission: is that a good thing or a bad thing?
Secondly, and hopefully briefly, do you think that the UK needs ARIA in order to compete globally when it comes to science and technology research and development?
Let me start by clarifying. From its beginning, the mission of DARPA has been very simple: to both create and prevent strategic surprise. Its connection to national defence has been important to its success. The particular goals that I spoke about were related to the programmes themselves. The programmes are constructed such that we have a clear way of measuring success or failure for the programme at the end of the programme. It is those two things that fit together: the programmes with individual ambitious measurable and testable goals, in service to the overall mission.
I have said in previous testimony that there is some wisdom in thinking about ARIA as directed to specific areas of interest in the UK; I think that is worthy of some thought. There is a strong base of expertise in the UK related to health and the life sciences. Therefore, that could be an area of focus within the resources that you have available to you.
To answer your second question with respect to the UK on the global stage, I believe that at this moment there is a historic opportunity in front of your Government to take a position on the global leadership stage. My particular area of focus has been in human health over the last year—that might be a way for the UK to come from the perspective of both national efforts and multinational efforts, in service to a global vision for what we want the world to look like post pandemic and post Brexit.
If I may, I would like to answer the first part of the question. I read the Bill carefully, and I too was looking for a mission, because DARPA and ARPA-E are mission-oriented agencies. Having a high-level mission is very important to define the programmes with the specific goals that Dr Dugan was talking about, which will fit in the overall mission. It is entirely possible that ARIA will be something new in the innovation funding landscape—a UK model that will blaze a new trail. But if we compare it explicitly to something such as DARPA or ARPA-E, in its current form it is lacking a high-level mission. To give an example, for ARPA-E that high-level mission is to overcome the long-term and high-risk technological barriers in the development of energy technologies. It is quite high level. Having that front of mind for everyone in the agency channels the energy and lets people animate or catalyse a community to allow the portfolio of projects to be more than just the sum of its constituent parts.
DARPA: defence and national security. Clear mission; clear scope in which to work. Of the ARPA-like entities around that I am aware of, the only one that very closely follows the DARPA model would be the Intelligence Advanced Research Projects Activity in the US intelligence community. When you change what I would regard as the key elements—ephemeral or temporary people, project based, and no fixed assets—that have made DARPA nimble and forward leaning for 63 years now, you get something else. That may be more appropriate for what you need, but if the objective is to mimic or replicate, there is only one example that I know, and there are three key ingredients.
Within that model, DARPA is a very shallow place in the managerial sense. Three layers deep: there is a front office, some tech offices and the programme managers. The overall mission provides the context, but the frequent hiring of office directors and PMs, and front office people too, means that there is always exploration—looking for that advantage. Part of our mission is to impose and avoid technological surprise. That is why we are here. It focuses everything.
It is a pleasure to serve under your leadership this afternoon, Ms Cummins. ThankQ you to our excellent witnesses. I am interested in the practical approach. When you have made some breakthroughs on these various high-risk projects, how do you ensure that the breakthrough reaches its full potential? Is it done through the ARPA model, passing it on to someone else to take it to the next stage? Is it the project manager who has a role in ensuring that it goes into safe hands, or is it the churn of people—the revolving door—that helps transfer that knowledge out an ARPA and into business, so that it can create service and product?
If I may, I will take the first shot at this one. It is the first two: we do not rely on the churn, as you say, of people for transition, but when you show up—when you come here—you come to make a difference. So you are always focused on transitioning the knowledge that is discovered in a more systems-oriented research programme—the thing or the entity—across into service of the nation. It is part of what you do. I think, as someone said earlier, it is that intersection of managerial and technical expertise, and a passion: those are the people you want at DARPA at any given time to frame and to drive—and not just to drive to discover, but to drive to transition as well. We watch that very carefully and the responsibility belongs to all of us in the agency.
We used to say at DARPA—and this is, I think, generally true of most organisations—transition is a full-contact sport, always has been and always will be. It is very difficult. Transitions of breakthroughs that are showing what is now newly possible, or a solution that did not previously exist, require a tremendous amount of effort. I think that it is important to recognise that there are many transition paths that grow out of an organisation that is ARPA-like. Some of the programmes, in the case of DARPA, transition to our military counterparts. Some of them transition to the commercial sector and then are bought back by national security or military. There are many different pathways. In some cases, programme managers go to other Government organisations to help in those transitions. In some cases, they rotate out and go to new things entirely.
It is important to recognise that the breakthrough itself is not sticky through the organisation that it was created in. The breakthrough then gets transitioned to impact and scale in the most suitable organisation in order to create that ultimate impact. I would add, in addition to the passion that many programme managers and directors feel, they are also impact junkies. They really come to make a difference. So the ultimate transition—the ultimate scaling and impact—is the goal. Make the breakthrough, and then transition it to scale.
I want to note that there is a distinction between DARPA and other ARPA-like agencies in different contexts. I am sure Dr Highnam and Dr Dugan will think that it is an oversimplification, but to some extent there is one customer for the projects that come out of DARPA, whereas for something like ARPA-E it is a much more diverse and scattered ecosystem. The breakthrough needs to latch on to the energy system, and there are lots of different actors with lots of different interests. At ARPA-E that has meant that they have created explicitly a tech-to-market group, to try to get ahead of the translation problem of the project that has come out of the agency. I want to say that this is not independent of the mission. To create a good tech-to-market group, you need a certain scale within a certain scope, and to the extent that your projects are too scattered, it is going to be a lot harder to create that scale, and so harder to create the transitions.
Just to follow up briefly, thank you for that; it is comprehensive and helpful. It highlights the fact that you are looking for more than just individuals with some inspiring ideas. They have got to have the ability to own the research and inspire the next stage in its progress. I just think we should put that on record—that programme managers have to be multi-skilled in a number of different areas. So thank you for that.
It is great to serve under your chairmanship today, Ms Cummins. Thank you very much to the witnesses today. It is very enlightening. On the back of the last question with regard to managerial content I really like the idea that the transition is an impact or contact sport. You go in, do your best work and get out with your reputation intact. I have two questions about that. First, how do you reach those people who are not known—who may be working on something very creative but are not well known in the industry? Secondly, you have talked a lot about evaluations. Are they peer-to-peer evaluations, and is that evaluation transparent? Perhaps we will go to Dr Highnam firstQ .
We do—I am very proud of this—full and open competition to the greatest extent possible. The process is approximately like this. A programme manager has framed a programme, using the Heilmeier questions, and received approval to launch. They put out various announcements in different places. They organise industry days—these are more virtual than in person, but we do both. We put it into the various mailing lists in all manner of technical communities. We push it out through small business and make sure the universities and the vice-presidents for research and development are all aware. We make the maximum push that we can, certainly for unclassified activities.
Then, when proposals come in—we are very clear on what we expect to see in a proposal, which is how we then evaluate proposals; we are very transparent on the requirements for that—we take a look and, surprisingly often, to respond to your point, you will find a technology or a small business had an idea that meets the goal. We do not over-engineer the request for proposals. We say, “Here’s what we want to do. Here are the boundaries, if you like, in terms of technical elements we are interested in. It’s up to you guys. Come back with the best team that you can and the best approach that you can for solving this.” And there is always a surprise. From a PM perspective—Regina and I have both been PMs at DARPA—you always find yourself saying, “Oh, I didn’t think of that. That may be the one that actually wins; we don’t know.”
If we want to get down to some specifics, I think it is important to recognise that the evaluation process for us is very much about separating the abstracts or the proposals into two baskets: those that are responsive to the call and could potentially help us to meet the goal; and those that are not. But it is not an explicit, peer-reviewed consensus rank ordering of those proposals, and the reason why we do not do it that way is that rank ordering tends to favour the most conservative of the proposals. What we seek instead is to take those that could contribute to the goal and, from them, construct a programme, with the appropriate pieces, the right risk profile and the right disciplines and mix of organisations, to achieve the goal.
In this respect, I want to be clear. There are practices and principles that we use here. We can write down some of the rules that we use and give you some elements of the playbook, but there is here a certain mastery of practice and principle that it is necessary to understand, and in that respect the programme construction is fair and equitable but also designed to take the elements of the proposer’s work that most substantively contribute to the goal, even if they are potentially high-risk. That is how you construct a programme that is optimised for breakthroughs.
I think that those two modes of funding are complements, not substitutes. It is very important to have an ecosystem of funding. In the US, we are blessed with a very diverse ecosystem. Lots of domains, such as health—there are many, such as agriculture—in some sense are missing the ARPA-like elements, when they have a lot of those other elements.
It is important not to put those two agencies in competition; they both have a role to play. Of course, there is a perfectly legitimate debate about the relative levels of funding, but they would both be doing things that are tremendously important and that would complement each other in the long run.
Pierre makes a very good point. These are important elements of a robust and functioning ecosystem. We talked about the advances in mRNA, which have been so important in the corona pandemic. That relied on basic science, curiosity-driven research that happened mostly through NIH, pivotal investments in this breakthrough mode, this Pasteur’s quadrant style of work that DARPA is famous for, and also the private sector, which was instrumental in bringing it to scale, use and impact.
To Pierre’s point, these things have to fit together in order to create the breakthroughs—that is the innovation gap that is often filled by an ARPA-like organisation—but you must have a foundation of science from which to draw and you must have a mechanism of transitioning to scale, if all of it is going to make sense in impact.
It is very important to understand those things and in appropriate measure. Just to give you a sense of it, DARPA has operated with about 0.5% of the DOD budget for its entire 60-plus-year history. Small investments, relatively speaking, in these breakthrough-focused activities can make a big difference.
Thank you, Ms Cummins. It is a pleasure to serve under your chairmanship. I thank all our witnesses and those who gave evidence to the Science and Technology Committee, which was greatly appreciated.Q
I have two related questions for the panel. First, notwithstanding your responses to Mr Flynn about the need for a mission, which it seems is going to be delivered by the chair and the chief executive of ARIA, how important is it that ARIA remains autonomous and free to pursue whatever its aims are, without interference from Government Ministers?
Secondly, what advice can you give the Committee about the funding methods ARIA might use? The Bill envisages potentially grants, loans, prizes, grant-prize hybrids, investments in companies. Could any of you give us advice on what has worked well in other settings? I would like to start with Dr Highnam, please.
On the funding mechanisms, we are an agency in the Department of Defense in the US Government, and we have a number of options available to us, which we make use of depending on the context. Of those that you listed, the only one that we do not do is take investment positions in companies. That is not what we do. You can make a proposal to us for research. You may offer a cost-share, depending on whether it is a major company and very systems-oriented work, all the way to a standard research grant to a university or small business, or a combination of those things.
We have a number of other options in between, including a modified form of commercial contract called an OTA—other transaction authority. They are referred to as OTs and are a very useful tool. DARPA was the first user of that about 20 some years ago. It is a great way of doing business.
To the first question, we are an agency in the US Government. We work in the Executive branch. We work and deal closely with Congress on all manner of things. We have flexibilities as an agency. We have ways of doing business and we are very careful to make sure that the wins that we achieve are well-known, and that we work within those boundaries. Again, the Administrations and Congress over the years have watched and helped DARPA, and have been incredibly supportive. The agency—Regina and I can both say this—keeps delivering as a culture and a mission place, because back in ’57,’58, they got a good recipe, and that culture persists despite 25% or higher personnel turnover. It is part of the Government, with all the benefits. All the—“constraints” is the wrong word—rules that come with that are there for a reason, and DARPA gets the job done.
I want to attach independence and autonomy to desired goals and outcomes here. The reason the agency sits so independently with respect to its decision making is to find this intersection, and get through the Heilmeier questions, as Peter has talked about. I would often refer to it as figuring out how to get a project inside Pasteur’s quadrant—the idea of having a very specific outcome in mind and having the science and engineering to support the idea that you could achieve it. That is a difficult analysis. That is the creativity that Peter is talking about.
You cannot mandate that from outside the agency. That work happens on the part of the technical teams inside the agency who are assessing the state of the science and the engineering. They are working in service to the mission of the organisation with an understanding of national security goals, and they are finding that intersection. It is the single hardest thing that we do in the agency: forming programmes in that spirit. It is not possible to do that by mandate outside the organisation.
That independence of decision making and the crafting of those programmes in that spirit are coupled, and that is part of the reason why the agency has been so successful over years. I think independence is in service to those outcomes and those breakthrough objectives.
Much as Peter described, we use a variety of strategies. As you may have seen over the last year, one of the things that we did was to build a health breakthrough network, which now has almost 30 signatories on six continents. The goal there is to speed contracting, so that we can move down to days or weeks what would more typically be months or even as long as a year in contracting. The particular way that we work is through contracts; we do not actually do grants. I also think that this position of not taking equity is important, because the non-profit element of it is part of the differentiation, and we have an entire commercial sector that is good at assessing value and figuring out return on investment. That is not what is pivotal or differentiating for the organisation—neither for Wellcome Leap nor for DARPA.
Yes, absolutely. First, I second what my colleagues have said and agree wholeheartedly. I would say that in terms of the modes of investment, the track record of Government agencies taking investment positions in companies is not very good, to put it mildly. It is interesting that it is something that neither ARPA-E nor DARPA actually does.
At the same time, it is important to point out that one way for an ARPA project to transition is to give birth to a start-up company. I know for a fact that in the context of ARPA-E, at least, that is something that is happening on a fairly regular basis, and that is actually tracked as one of the outcomes that one could look like, in addition to maybe much more traditional intermediate outcomes such as scientific papers and patents.
The more general point about autonomy is very important. It is really difficult. It requires forbearance on your part because the kinds of missions and impacts that you are trying to achieve at a very high level are long-term goals fundamentally. I might be overdoing it, but I have a sense that if you start ARIA today, you will not know if it has actually fulfilled its high-level mission for at least 15 years, and that might even be too optimistic.
It is a pleasure to serve under your chairmanship, Ms Cummins. I thank the witnesses for what has been an extremely interesting insight so far. I represent the city of Cambridge, where, as you can imagine, many of these issues are a frequent source of discussion. From what I have heard so far, I take it there is not one model of ARPA. You have a number of models, depending upon the different sectors. I heard Dr Highnam say earlier that DARPA had not been replicated anywhere else, yet we are trying to model our system on what you have got. How much does it depend upon the context, not just within the sector and whether it is ARPA-E, ARPA-H or DARPA, but in the wider system? We heard from witnesses earlier today about other aspects of the ecosystem—public procurement, things such as the Small Business Research Initiative, which I understand we have a much less successful version of in this country, and having a single client or defence system. If you do not have clarity on any of that, what would be the consequences? Perhaps I could start with Dr Highnam.Q
I said in my previous comment that I am aware of only one example that replicated DARPA intact, and that was the intelligence ARPA in the US, where I served for about six and a half years. It is very true to DARPA as it stands. Others depend on context, which includes the context of discussions like this one where there is certainly the framing of an organisation. It is being pulled and pushed and moulded by many different forces and interests. What you get coming out will, I am afraid, naturally reflect that. In intelligence here, it was a straightforward thing. We wanted something very similar to DARPA. A number of us had come from DARPA and knew what that was.
ARPA-E is not identical to DARPA, but we certainly try to inspire it to a very large extent. I think the difficulty here is that it is a tight bundle of practices that fit together, so one open question is to what extent can you pick and choose in terms of the menu of practices? What can you undo until you in some sense undo the entire model? It is important for us to level with you that we do not really know the answer to this question, because fundamentally there has been one DARPA, and that is the one we have been able to see for 60 years. One possibility that we might want to have in mind is that it does not take a lot of changes in the model to undo its effectiveness.
I agree with what Pierre just said. I might use an analogy if you will permit me. I think most would agree that Guardiola is a great coach. We could ask him how he has achieved the track record of wins and successes that he has. How has he envisioned a new style of play, constructed a team, coached the players, made decisions on the way in? He could write down some of the principles associated with that. On a day-to-day basis and across the duration of a season, he makes countless decisions, which are in service to these basic principles that create such a winning team. It is those detailed decisions that come from intuition and experience—the mastery of the practices and principles as Pierre would say—that are important to success.
At Wellcome Leap, for example, our first rule is to make as few rules as possible. Part of that is recognising that we have these practices and principles and we need to adjust a lot as we go along in the process. In setting up Wellcome Leap, I think Wellcome did a very good job of saying, “We are going to do the few things that we think are central. We need independence and governance. We need an experienced team to lead it. We need to free it from a profit motive, and then we need to let it do what it does.”
So there is this combination of a few principles that we can write down for you and then many other things that are about the practice of that come from the intuition and experience of leading these types of programmes to breakthroughs.
Q Just to come back on one point there, you say free it from the profit motive. I was quite struck through all of this by this stress on public service and honour, which is wonderful—perhaps we are struggling a bit with that here at the moment. Who owns the intellectual property on all of this? It is a complicated question. Is there a simple answer?
The organisations that create the breakthroughs own the intellectual property in the case of Wellcome Leap, and that is usually the case in DARPA. Now we usually also have a backstop, which says we have march-in rights if the entity either chooses not to commercialise it or to transition it to impact. Then we would go and say, “We need to take this in service of national security,” but at its core the intellectual property belongs to the inventor of the breakthroughs.
One addendum to that is that we have a notion here of Government purpose rights. Yes, the invention is owned by the creator, but if you receive DARPA funding and the appropriate terms are in the paperwork and the arrangement that we have with you then there are limited rights available to the US Government for those inventions.
We are clearly taking a huge amount of inspiration from DARPA in the creation of ARIA. It is an organisation that has many strengths, but every organisation has weaknesses as well. I would like to give you the opportunity, given your huge knowledge and experience of DARPA, to give us advice on what to avoid. To put it more diplomatically, how can we improve on some of the DARPA processes?Q
One can look at any set of processes and ask, “Are they optimised for the outcomes?” I think ARPA-like organisations are very much optimised for the outcome, which is to catalyse breakthroughs. It is not optimised, as my colleagues have said, for multi-decade-type funding that supports basic research that is foundational and builds a body of knowledge and extends incrementally our understanding of the world. Neither is it optimised for commercial success. I think those things are okay, and there are other organisations and other funding mechanisms that are optimised for those types of activities.
Part of what we see is that the programmes very much take on the character of the programme directors. That is good from the perspective of speed, agility and getting the work done. Sometimes people do not agree with all the things the programme director says. That is the nature of the type of work we do, which is high-risk and breakthrough-oriented. We used to say that the good and the bad of DARPA is that it has no institutional knowledge, which means that we can take a shot at something that has been tried before, and most of the people who tried it before are no longer at DARPA. That is good, as it gives us multiple shots on goal in a changing science and engineering landscape.
Q Finally, to Professor Azoulay, DARPA clearly recognises the benefit of greater integration between the public sector and the private sector, but inevitably that exposes civil servants to political accusations of cronyism. How do you protect DARPA and DARPA servants from those kinds of attacks?
I think there are two elements. One is rules—conflict of interest rules are very important in this regard—and the second, which I mentioned at the beginning, is norms. It is a lot about whom you choose to put in those roles. They typically have credibility and a reputation that is established in the world that they come from—it could be academia or the private sector. Serving as a programme manager at DARPA or ARPA-E is a wonderful opportunity to have an impact—
Order. I am afraid that brings us to the end of this session, I am sorry. It is a perfect end to the session, but it is the end of the time allocated to the Committee to ask these questions. I thank our witnesses on behalf of the Committee for that evidence. Thank you very much.