Update (June 23): Dario Gil has now been officially appointed to the National Science Board.
The coronavirus pandemic has ushered in new alliances between the tech industry's biggest players and government agencies as the world races to limit the spread of COVID-19 and find a cure. Dario Gil, IBM's research director, has been in the thick of everything.
Gil has been serving on the President's Council of Advisors on Science and Technology since the group was revived in 2019, and he helped launch the High Performance Computing Consortium. The group brought together supercomputing resources from some of the most powerful machines in the world to tackle 51 projects — and counting — aimed at modeling the virus and potential drugs. The experience has led Gil to ponder the broader question of how tech can unite in quieter times, helping the world to prepare for the next disaster more rigorously. "It was wonderful that we could create the [HPCC], but we had to sort of invent it on the fly," Gil said. "Why couldn't we think ahead?"
Now he may be in a position to do just that. Gil has been nominated to serve on the National Science Board and is pending confirmation by the White House. The board oversees the National Science Foundation, the government agency that supports fundamental scientific research in the U.S.
Protocol recently spoke with Gil about what he wants to achieve as a member of the board, including a "science readiness reserve" — like the Army Reserve but for scientists who could be mobilized — and what he sees as the existential crises that research divisions like IBM's should be confronting right now.
This interview has been edited for clarity and length.
What would you like to achieve as part of the National Science Board?
I'm passionate at multiple levels. The National Science Board is a unique institution in the world, in that it has a governance responsibility in the National Science Foundation, which funds most of the basic research right now, in addition to serving as a science board for the United States and helping with policy for both Congress and the federal government. I want to learn, I want to contribute, and help make that a success.
Within the broader context of science, I think there are a lot of urgent tasks ahead. Every time we have had a large crisis in the world, it has been an opportunity, and a necessity, to imagine and sometimes create new institutions. If you look at World War II, it was an unprecedented mobilization of scientists and researchers across the nation — the Manhattan Project is just one example. It led Vannevar Bush at the time to [create] the National Science Foundation. He wrote that famous paper on science, "The Endless Frontier," on how you tap into the R&D talents of the nation in peacetime. I think in this moment we find ourselves in, we should be asking the same question: How are the institutions that we have for science [faring], and what may we need to create as a consequence of the situation?
An initiative I've been very involved in — the creation of the COVID-19 High Performance Computing Consortium — is illustrative of a model. The idea was straightforward. We're trying to compress the discovery process of new treatments and vaccines, using computers. Today, supercomputers are in many institutions. Could we collaborate together, add the compute capacity, make it available to scientists in the United States and all over the world, and accelerate the process of decision-making and matching of the best ideas to suit the supercomputers?
From idea to launch, we did it in five days. We brought together the White House Office of Science and Technology Policy, the Department of Energy, NASA, the private sector — IBM, but also Google, Amazon, Microsoft and others — and many universities. Today, we have 38 consortium members, and we've aggregated 440 petaflops of computing.
So you want to take a similar approach with an idea you have for a "science readiness reserve" — how would that work?
The military engages in planning unwanted scenarios all the time. They also have another institutional capability — they have active troops, but also have reserves and in the case of an emergency, they can mobilize reserves. So then [with Avi Loeb from Harvard, who is also on the president's tech council] we got to thinking what in science is the equivalent? Why couldn't we create a "scientific readiness reserve"?
R&D today is very distributed across the federal government, industry and academia. Of the $600 billion of R&D spent a year in the United States, 70% is in the private sector. It is well-known that the private sector has some of the best talent and resources. So the thought was: In times of crisis, we do come together, but it seems that we're improvising a bit. We have to learn how to do this.
So the idea of the scientific readiness reserve would be to tap into the talent that is distributed. It could be voluntary, of course. We have capabilities at IBM, and I'm sure people at MIT, Google, Microsoft, they could raise their hands and say, "Some of these capabilities, I'm willing to go and make them available." If these situations arose, maybe a few times a year, some of those institutions could come together and plan ahead, and think, "What will we do if this happens?"
Is this one of those things that sounds great in theory, but actually getting people to do the hard work of setting it up would be really tough? Have you spoken to anyone that you'd want to be a part of this yet?
Yes. We put forward this idea just recently, and we have had a number of conversations with tech leaders, in different agencies, across universities, and in the private sector. A lot of people have written to us informally and have been telling us, "Hey, sign me up, I'm interested in this." There's good momentum around that, and Avi and I are going to work on creating working groups to start fleshing this out more.
The issue will be how to make sure that spirit and capability for things that are of the national interest doesn't dissipate, six months or a year from now. I do think that that idea has legs, and perhaps an organizational approach could be to take a cue from the open governance models that you see in open-source [technology]. It's something that feels fluid and is voluntary, yet there is organizational structure and governance around it — it's not a free-for-all.
I've got to believe as well that there must be a group of people that would like to contribute their skills to the country who don't necessarily want to join the military.
A desire to contribute, that's a very natural human spirit thing. I'm 100% in agreement with you, and giving people a way to contribute is powerful. And sometimes you need that institutional aspect to give people that channel.
What sorts of things should the National Science Foundation be recommending that the government invests in that perhaps it isn't right now?
I need to do more homework, but one area I know they're passionate about, and I personally think is very important to continue to double down, is how we can accelerate the process of scientific discovery itself.
We're in a unique moment where there is a real revolution going on in data, AI and quantum. You can take the byproduct of those core investments that the NSF and others have made in decades past, that has manifested itself in technology like AI and quantum, and bring it back to accelerate the process of doing science itself. Let's say we want to create a new material, in the context of a new battery technology or developing new vaccines — things that require modeling the physical world.
Historically, there were two approaches: One is purely experimental — you just try stuff out, and we've been doing that since time immemorial. With the advent of computers, there was always the promise of using [them] to simulate. That's kind of like what HPCC does. But there was an intrinsic problem in that modeling — because nature is quantum mechanical, it's an exponentially difficult problem. The best you can do is to approximate, but it leaves a lot on the table. With the advent of quantum, there's going to be the possibility of creating a whole new generation of instruments and machines that allow us to model nature much more efficiently. If we can build these machines, we're going to be able to drive all sorts of breakthroughs.
I think what is amazing is that each one of those communities has a lot to contribute — the classical computing community, the AI community, and the quantum community — but they're each separate. So what I think we really need to do is not only to excel at each one, but to combine in a philosophy of convergence. I call this bringing together "bits plus neurons plus qubits." The convergence that is happening right now can massively accelerate the rate at which we can do scientific discovery. I'm passionate about that because if we could make that better, it lifts all boats.
How are things at IBM Research? I've got to imagine there's only certain types of research that you can do outside of specialized laboratories. How is work progressing right now?
Overall, we're doing well. IBM has 350,000 employees, and the research division has slightly over 3,000 scientists. Since early March, we had about 95% of IBMers working from home. A subset of them that were site-critical employees have continued to work under very special circumstances and protocols.
Within IBM Research, in the Yorktown facility where I work, we have quantum computers and a lot of experimental facilities. We've had about 200 researchers that have continued to work. Normally in the building, we have 1,600 people. We've done that, with a great deal of care, to maintain critical infrastructure, like critical gases. We have 18 quantum computers on the IBM Cloud, and we've been maintaining them. They're running wonderfully, and we supported the community around that even in the middle of all of this.
We'll go through a very careful process of adding folks that require experimental facilities to be productive. But then there's many others — our communities in AI, cryptography, security, the cloud — they're doing fantastic work from home.
Beyond the current pandemic, what sorts of existential threats is IBM, particularly the research division, thinking about?
Before all of this happened, we had a new strategy within research that we call "impact science," borrowing the term from impact investing.
We've always had an exploratory science agenda, but we're also looking at these changes that are happening in science and this revolution in computation. What is it going to mean for the future of work? What is it going to mean for the future of health? What is it going to mean for the future of climate? We picked those three areas last year, and now in the context of all of this, we're not only doubling down on, but revisiting how we are thinking about the future of health — it's changed in the context of the pandemic. How we were thinking about the future of work has dramatically changed, and it's also changed on climate.
In some ways, I think sometimes society can feel that science is sort of like a pastoral endeavor. I think the pandemic makes it clear that every day matters, and the urgency of science has never been more vital. And I feel very strongly that what we need to do, as a research community, is bring science and the scientific method and scale it out more broadly in society. In these areas of health, climate and work, we need to also elevate scientific thinking, even if it's not being practiced by scientists, into the halls of power.
That must be tough right now. It feels at times as if there's a kind of anti-science sentiment with the current administration.
When you look at the most admired professions or disciplines in the United States, the No. 1 most admired is the military and the second is scientists. What I find endearing about this whole thing is sometimes it can feel like, well, that cannot possibly be true, but actually, there is a lot of broad popular support for scientists as a profession and science as an endeavor.
Social distancing, the economic stimulus packages — all those things are measures to buy us time. Ultimately, we need to get out of this situation through treatments and a cure. To do that, we need to pursue science.