Charles Clancy is the CTO of MITRE,
an American not-for-profit organization managing federally funded
research and development centers (FFRDCs) supporting various US
government agencies in defense, healthcare, national security, and
cybersecurity fields, among others.
In this interview, we discuss:
What is MITRE and how does it support national science & technology strategy
How China threatens America’s infrastructure and university R&D
The cyber workforce gap and how AI could fill it
Finding mission-driven work for highly skilled technologists
How the ecosystem of S&T and R&D funding evolved through the 20th century to today
20th-century Research & Development
Jordan Schneider:
Charles Clancy, CTO of MITRE. We're going to talk about what MITRE is,
how it's organized, what it does, as well as broad questions about
critical emerging technologies and what a national science and
technology strategy should look like.
“A
new federal agenda for promoting S&T innovation must address itself
to critical market failures. To do this effectively, what is needed is a
national level effort, a synergy between government, industry, and
academic activities to holistically address our nation's most critical
science and technology priorities.”
Tall order!
Charles, let's start with the priority-picking process for the critical
emerging technologies that the US government should even think about
starting to play in. What are some interesting methodological frameworks
for how to allocate MITRE's resources, and more broadly, the
government’s at large?
Charles Clancy: First
off, I think the starting point would be that the US tech economy has
really evolved organically over the last 50 years or longer. We've
placed certain bets throughout history. The Manhattan Project was a
critical national need. We needed nuclear weapons and nuclear power. We
were going to create the national lab system and we were going to go do
something big.
After
that, it was the Apollo program. We were going to send someone to the
moon. That is our grand challenge problem. It was going to unify much of
the science, technology, and industrial base in the country to do
something big on behalf of the country.
Since
the Apollo program however, as a country we haven't had a big goal for
what we want to do. So we've just been free-wheel, organically letting
the market decide what the US is going to be the best at in terms of
technology. What happened was globalization. As a result, much of the
low-end manufacturing work moved overseas, moved to China.
Here
in the US, what emerged was Silicon Valley and a focus on software, the
highest margin. We built an entire tech economy based on software
driving the value in the market. Arguably it worked well for us. In the
last 15 years, we've seen the ascendance or the birth of the largest
companies in the world. They're all based here in the US and built on
top of that foundation of software.
What
happened is the fundamental premise upon which we built that, these
globalized supply chains, began to erode as we decided that China wasn't
our friend anymore. We had built this globalized tech economy on some
faulty national security assumptions.
We
had to back up. We can't trust China to be our manufacturing base and
drive all the low-margin hardware innovation that the world needs. As a
country, we want to be able to do that ourselves, not only for ourselves
but for the world. That necessitates a complete change in how we think
about and invest in science and technology.
A lot of that led up to those 10 technology areas being legislated as part of the CHIPS and Science Act.
Jordan Schneider: Let's do a little more 20th-century
science and technology history. The most salient ones that you pointed
out are the flashy mega-projects which got a lot of political focus and
cover.
We
have had a National Science Foundation spending billions of dollars
every year and we elided over private philanthropic support which ended
up creating modern biology. We had the Human Genome Project in the early
2000s. Why do those other models of national and philanthropic
investments in basic science not cut it today?
Charles Clancy: You have to look at who's spending the money and the underlying economics. Back in the era of 20th-century
S&T innovation, you had a combination of industrial monopolies like
AT&T, and consequently Bell Labs, and major federal programs
driving innovation. The corporate spending on R&D
outside those big monopolies was very small. The federal government was
driving more than half of the R&D economy of the US.
That math has changed though as we zoom forward to the 21st
century. We no longer have the big industrialized monopolies that were
investing heavily in research and development back in the 1980s. The university stepped up.
Before the 1980s, universities were not as involved in driving the
national R&D agenda as they are today. The NSF and NIH became
conduits for investing.
Certain
sectors figured out how to take advantage of the new world order, in
particular bio and biopharma. They built the necessary relationships
with the universities. They had enough scale to invest corporately in
research and development, particularly for the next drug. The NIH was a good complement to that ecosystem. That ecosystem matured and went off and did great.
Led
by Stanford and Silicon Valley, in software you got this symbiosis
between tech investors and software startups with university programs.
That ecosystem really came together and as a nation we did well with it.
Now comes hardware. Many
of the technology areas that are listed in in the CHIPS and Science
Act, like quantum and semiconductors, are not areas where we have
well-developed symbiotic ecosystems. The goal as we look forward into the 21st
century and beyond is to make sure that we have that thriving
combination of academics, venture, startup and federal investments.
How
do we steer the investments that we're making now in order to help
create those ecosystems that aren't as mature? In the last five years,
we've seen the birth of this whole notion of a deep tech investor,
startup companies that claim to have patient capital who are looking to
invest in these new environments. That's a positive development. We
still have a ways to go to truly see the return on investment in that
new ecosystem.
Technology Areas & Cybersecurity
Jordan Schneider: Enough teasing. We've got:
artificial intelligence (AI)
high-performance computing (HPC)
quantum, advanced manufacturing
advanced communication
cyber
biotech
advanced energy efficiency
material science
disaster protection
They had to get to 10 so they threw in “disaster protection.” We're not going to talk about that today.
Why don't you pick one and illustrate where the gap is in the current model of how money flows through the system?
Charles Clancy: For
several of them we've got strong existing ecosystems. For example,
Silicon Valley and the venture community have rallied around AI as the
next big thing that's going to make them all plenty of money. We're
seeing that ecosystem come together.
There are issues in AI. An example would be this whole notion of AI assurance, AI safety.
There's
a big debate that's happening right now about regulating AI and how we
do that. There are opportunities for federal investment in the pieces of
that ecosystem that aren't seeing as much attention from industry and
the venture world.
There
are other areas like microelectronics. The economics of trying to build
the next generation of semiconductor fabs in the US just haven’t been
there. The “chips” half of the CHIPS and Science Act is really trying to
double down on changing those economics.
While the $39 billion in fab subsidies are helpful, it won't be sustainable unless we can build a real closed-loop economy. We need to be able to take the successes from the first decade and reinvest them so that they don't need constant subsidy.
Jordan Schneider:
Cybersecurity seems like an area where many large and very well-funded
private sector organizations have a pretty clear market force to get
them to invest. What kind of things fit into your schematic of “too
applied for basic research, but too upstream for immediate commercial
marketization” that would potentially apply to a cyber context?
Charles Clancy: Post-quantum cryptography is a good one. We have a looming concern of adversaries building a quantum computer that can crack the encryption that we use today on the internet. There are efforts led by NIST underway to develop the next generation of standards.
The
US government is already moving out mandates saying that we need to
adopt those standards. There's a lot of opportunity to accelerate how
those ecosystems come together because we believe the transition is
going to be a lot harder than people realize.
Here
at MITRE we've created a coalition of major tech companies who are
interested in figuring out what are the hard problems at a systems level
that we’re not thinking about in transitioning to post-quantum
cryptography. That’s an example in cybersecurity.
Cyber Threat from China
There
are some other examples that are driven by the threat. We have a
particular threat environment right now that we assume we are operating
under. That includes China trying to hack into our research enterprise
to steal our intellectual property.
China
is trying to hack into our critical infrastructure to gain a strategic
advantage, but without actually pulling the trigger to break anything.
You've got Russia doing the same. You've got Russian organized crime
going crazy with ransomware and trying to make money off cyber
insecurity.
The
cybersecurity industry and economy have converged to a point where
we're doing okay at managing that threat. At MITRE there’s another kind
of threat that we’re particularly concerned about.
What
happens in three years if China invades Taiwan and they're not content
to just hang out in our critical infrastructure networks? What happens
if they want to start blowing things up? That changes the model that
we've been operating under and necessitates potentially different
technical solutions to be able to address that at scale.
Jordan Schneider: Microsoft's cyber division isn't necessarily spending all of their resources planning for the worst-case scenario contingency.
Charles Clancy: They're
focused on the threat of today and the needs their customers have
today. I'm sure they're looking from a research perspective on how to
make the situation better.
However,
industry isn't paying attention to the fact that Russia, and China in
particular, have had underlying doctrinal changes that cyber is not a
tactical resource but a strategic resource. It's up there with nuclear
weapons in terms of the way you can use it to project power and
influence nation states.
That
threat environment that we're seeing change right now is not something
for which the collective cybersecurity vendor community has solutions,
to protect not just individual organizations but countries as a whole.
Jordan Schneider:
That's really interesting. One of the market failures across all of our
critical emerging technologies is the tail end of horrific hot war
between peer nations. That’s something that the government ultimately is
going to be leading the research charge on.
Charles Clancy: I
agree. It doesn't necessarily make sense for industry to be investing
as heavily there. The set of presumptions necessary to make that the
right investment would undermine many of the business models that
they're operating under today. Many just don't think that China's going
to pull the trigger.
I
was just in India last week meeting with the cybersecurity industry
there. They view China as a threat. They see China getting into the
critical infrastructure, but they don't think China will ever pull the
trigger.
If
you don't think they're ever going to pull the trigger, then you're
content with trying to find them and kick them out of your networks.
You’re not preparing to deal with the fact that you could have serious
consequences if China does decide to go destructive.
Jordan Schneider: You mentioned this question of industrial espionage. A few days ago we had a not very sharp Google employee caught uploading TPU schematics
onto his personal cloud and booking a one-way flight to Beijing.
Technological diffusion, broadly defined, happens in legal and more
nefarious ways across all of our emerging and not-so-emerging
technologies.
Research Security & Universities
Charles,
what are the long-term defensible, competitive advantages when it comes
to technologies broadly defined? What are the sorts of things that you
have to assume are going to end up diffused out to your adversary,
regardless of just how much you invest and try to keep your secrets?
Charles Clancy: This is an important question as we talk about things like research security. Professors
at universities are not overly concerned about some foreign grad
student working in their lab and stealing the information to send off to
China because their intent is to publish the results of their findings. It's not a threat that they're concerned about because their intent is to disseminate their research.
You
have to look at the technology readiness landscape. What does
technology transfer mean? For basic research, technology transfer means
publishing a paper. That's contributing to the body of knowledge
globally in a particular technology domain. For more applied research,
technology transfer perhaps means a patentable prototype that is reduced
to practice specific ideas that someone might want to spend money on.
It
might be relevant to the US government for a particular national
security application. That's the space where we have things like export
control today. We have other protections to try and keep that technology
from going out the backdoor to our adversaries and that’s where we're
putting a lot more emphasis on trying to protect.
The challenge is that we've had a very bifurcated ecosystem, particularly with universities.
You have basic research that's fundamental and free to publish to the
world. On the other side you have restricted research, which either has
some sort of export control or something else that limits it so that
only US citizens can work on it.
You
probably can't publish those results. This creates risk. Many
professors don't want to do that kind of work because it limits their
ability to recruit students and publish papers. However, it's also the
kind of stuff that our adversaries are most interested in.
We've
also divided the space into only two buckets. What's missing is
something in the middle. With all these critical and emerging technology
areas, we are concerned about our adversaries getting a hold of them.
We want to be more thoughtful about the sorts of researchers that are
involved.
There was at the tail end of the Trump administration, National Security Presidential Memorandum 33 (NSPM-33),
which defines this ecosystem where you have to report the individuals
involved if you're working in critical and emerging technology areas at a
research institution.
If
they're foreign, you have to have their information. You have to have
basic cybersecurity protections on your networks. Over the next two
years, this is something that universities are going to be figuring out
how to implement as all the new requirements come into play.
There's
always this balance with openness and realizing that much of the
innovation in this country comes from immigration. How do you balance
that with the need to prevent intellectual property from going out the
backdoor to China?
If
the US is going to invest $50 billion in semiconductors, how do we make
sure that value isn't just immediately accruing to China and not to us?
Those are the sorts of trades that we have to figure out how to make
and lines we need to figure out how to draw.
Jordan Schneider:
It's really interesting to think about Open AI. Out of their top 100 AI
researchers, I would assume 75% of them don't have US passports, are
recent immigrants, or have family abroad and are not getting clearances
anytime soon.
On
the one hand, you could not let those people touch this stuff. All of a
sudden the company would fall apart. What may be the most important
firm for America’s future AI greatness would cease to exist or have to
run a lot slower.
On
the other hand, these important foundational models are clearly very
powerful dual-use technologies which, if not today, in the very near
term are definitely going to have really important implications.
Charles Clancy: NSPM-33 is trying to walk the balance between the two by saying it's
okay to involve foreign researchers on your projects. You just have to
report them so that there's a mechanism that can look from a
counterintelligence perspective at those risks. They can
know the individuals involved in order to better understand if our
adversaries are doing something that seeks to undermine our intellectual
property.
Jordan Schneider:
There's actually an aspect of this where, from the perspective of some
of those researchers, they want to keep living in the US and not live under pressure of foreign governments. When this stuff comes up, all you hear from academia is basically that it’s racist or unacceptable.
There's
a piece of it where you can actually feel more protected as opposed to
living on an island where all of a sudden you start getting weird emails
and your family members start saying odd things to you. To know that
there is this institutional framework can help provide you with some
buffer for that stuff.
That
seems to me like a net positive actually. Inevitably if your tech is
important enough, whether or not there is some new regulatory framework,
someone's going to find you and try to do something with what’s in your
brain or on your hard drive.
Charles Clancy: Here’s
a perfect example. Before joining MITRE, I was a professor at Virginia
Tech and I led the university's National Security Research Institute. We
had a FBI and NCIS team down in Blacksburg, Virginia that were part of
the security apparatus trying to help us defend against many of these
counterintelligence threats.
I
remember having a meeting with the FBI team where they informed me that
I was the professor on campus most targeted by foreign adversaries.
About once a week I was reporting some fully-funded Chinese grad
student, or postdoc, or professor who was going to come work in my lab
for free. Almost every single one of them was determined to be an actual
intelligence threat, someone trying to get into our lab to access the
technology we were building.
I
also had a ton of foreign grad students working on a lot of our
unrestricted projects. They all came here for a better life. They came
here because they wanted to get their PhD. They wanted to get involved
in US academia and research organizations, and move on from the lives
that they had in other places in the world.
You're
dealing with two very different realities. For many of the students
that are here to change their lives and their futures, they can be
hugely important to our economy. There are also those that are
legitimately trying to steal our stuff and represent real intelligence
threats.
Jordan Schneider:
It's interesting because I think the discourse on this topic is
particularly toxic. You may be the first person I've come across who's
done “on the one side, on the other side” within 30 seconds of each
other. What is it about these two worlds that they have such a hard time
seeing the flip side of these questions?
Charles Clancy: It
gets down to immigration policy. It becomes identity politics. It gets
into a lot of things that have become polarized in the country. As with
many of these things, there's truth to both sides.
We need a policy somewhere in the middle that can balance these things.
Finding
policies in the middle that are acceptable to either side becomes
challenging in the sort of increasingly polarized political environment
that we're in.
How MITRE Works
Jordan Schneider:
Charles, let's talk about MITRE. What's the 101? What is an FFRDC?
Whenever I talk to MITRE folks, they’re like, “oh RAND doesn’t do real
work, we're the folks who actually create and implement this stuff.”
Charles Clancy: MITRE
is a nonprofit. We were chartered 65 years ago to help spin off this
air defense program from MIT at the time. Since then we've grown
considerably.
We
do operate what are known as Federally Funded Research and Development
Centers (FFRDCs) on behalf of the US government. We currently operate
six of them. The oldest and largest is for the Department of Defense. We
have a smattering of others that work across the civilian agencies.
The
key attribute of being an operator of FFRDCs is that we are a
nonprofit. We are apolitical, neutral, and conflict-free. We're really
here to operate in the best interests of the agencies that we support
and the nation.
That
gives us an interesting relationship with industry, academia, and
government. We sit in the middle as this connector between all of them.
In many cases we find ourselves as the neutral ground between the
equities of industry and the equities of government and vice versa.
It's
a really unique platform because we're able to really understand the
policies and programs that the US government is trying to execute. One of the differentiators at MITRE is that we view ourselves as a deeply technical workforce. The R&D centers that we operate are either focused on research and development or on systems engineering.
We
have this really technical view that we bring to the programs that we
support on behalf of the US government. In this era when science and
technology are instruments of national statecraft, we find that
technical depth is increasingly important to providing sound inputs to
many US government stakeholders.
Jordan Schneider:
Let's stay on this technical depth question and then get to your role
in policy stuff. How do you think about picking projects and structuring
teams? What are the institutional incentives that you have in relation
to your clients in the public sector?
Charles Clancy: Part
of it is just how we're structured organizationally. If you look at the
national lab system, each one is its own little silo with its own
technical staff focused on the mission of that specific national lab.
At
MITRE, we're operating as a matrix. We have half the company organized
by technology area and technical discipline, almost like departments at a
university. Whereas the other half of our company is organized by
agency and department and mission area.
The
cool thing about how we operate is that for any project that we do for
the US government, we have this 50/50 blend of the mission experts and
the technical experts. That gives the mission experts the ability to see
across lots of technology areas and really be the technology
generalists but mission specialists. It allows our technologists to see
across a lot of different missions and be the mission generalists but
the technology specialists.
That's
uniquely important. In areas like AI, or cyber, or microelectronics, or
healthcare, we have our technical experts who are supporting many
agencies and that allows us to bring the technical glue between
agencies. They're leveraging common solutions in the way that they
execute their policies and programs.
Jordan Schneider:
That's the dream at least. Taking one technology and working across the
whole of government to apply it is a really interesting mission and
challenge.
You had the AI executive order
where every department has to do something. Every department over the
next 10 years is going to have to figure out how to put it into their
work streams just like with computerization over the 80s and 90s.
Why
is this so hard? What is MITRE and what is MITRE doing to help the
federal government tackle these cross-agency tech issues in a more
holistic way?
Charles Clancy: There's
a couple different meta issues that come up. How can a particular
agency adopt that technology in order to improve the way that they do
their mission or run their enterprise? AI is going to
have a huge impact on the IT systems of every one of these agencies.
They’ll figure out how to increase the productivity of their workforce
and deploy ChatGPT as a tool for employees.
There's the basic enterprise enabling functions that MITRE supports. You mentioned computerization. We've helped almost every agency figure out their cloud strategy and migration to the cloud.
Cybersecurity is another big area where we've helped them all with
that. This is really a CIO-type function, trying to take this new
technology and sort it in an enterprise.
Then there's taking a new technology and sorting it into the mission aspect of the agency. With the AI example it could be, how
does the DoD responsibly deploy AI in military systems and weapons
platforms? How does the IRS deploy AI as part of its audit
infrastructure? There’s the healthcare ecosystem, VA, Medicare. How do
they think about using AI to improve the quality of the support they
provide to the healthcare ecosystems? That’s the mission angle of pushing AI into supporting those agencies.
Then
there's a broader tech competitiveness aspect. As a nation, how should
we be investing in AI? How should the Department of Commerce be
investing in AI? How should the National Science Foundation be investing
in AI? What are the common platforms that we need as a country that are
going to bring people together? What are the new standards that we need
that are going to bring industries together?
Each
of those different layers are things that we're able to help the
different agencies support. We've got the AI expertise and we’re able to
apply that to enterprise systems, to mission systems, to broader
S&T competitiveness and R&D policy.
That
spectrum is really important. It allows us to horizontally integrate
across agencies and missions and vertically integrate across the
functions that those agencies are executing.
Jordan Schneider:
You mentioned your two types of employees, the technologists who work
across missions and the mission-focused folks who work across
technologies.
What
advice or scaffolding material do you give folks who have to bone up on
lots of different tech topics? What are the types of people who end up
really succeeding in those roles and why?
Charles Clancy: Many
of the mission experts are people either coming out of government or
coming out of industry ecosystems that support those missions. For
example, in our DoD spaces, you'll find more former military.
It's
the people who understand what those agencies need to accomplish and
the unique challenges, opportunities, nuances of the different
bureaucracies that you find all over the US government in terms of how
they work and how they operate. Otherwise if you try to bring solutions
packaged in a certain way to particular agencies, the patient rejects
the organ transplant.
Having
those mission experts who are helping define the envelope of what's
needed and how to accomplish it is extremely important. They will bring a
sense of urgency for that mission and the need for progress to help
those agencies be successful.
Fortunately,
they can rely on much of the technical expertise in the company. They
don't have to become experts in all these different areas. Certainly we
try to provide as many forms as we can for them to plug in and learn
what we're doing. We have all kinds of different technical exchanges
that we do across the company to try and get the two sides of the matrix
interfaced and connected and working well together.
How Charles Works
Jordan Schneider:
Maybe this was a more selfish question than I let on. As someone who
writes and thinks and talks about these sorts of questions across a
broad range of technology issues, it's always a challenge in my head.
Should I read another paper on AI? Should I read another paper on
biotech?
You
have to manage all these folks. How do you think about the marginal
investment time involved in getting a deeper understanding of the
technological or market dynamics for X, Y, or Z? Maybe you can give some
broad advice for people whose roles end up spanning these different
domains.
Charles Clancy: I don’t have a silver bullet.
When
I was a professor, I would always leverage my students. If my students
are really excited in a particular area, I would have them go off and
learn it all and then I'd have them teach it to me. That's the sort of
secret of being a professor. You learn more from your students than the
reverse.
Here at MITRE, I personally have the advantage of leading our fellows program.
I've got at my disposal our top technologists who are world-class
experts in these areas. I can rely on them to help me get up to speed on
lots of different areas.
Coming
into the CTO role at MITRE, I'm not a biologist so all this biotech
stuff I've had to learn. I was fortunate enough to have a team of people
who could help me understand the deep issues and figure out how to
navigate in that very technical discipline.
It’s
the same thing with quantum and microelectronics. My background is in
electrical engineering but my personal expertise is all in
telecommunications, wireless, cyber and all that. These chips were not
something I spent a lot of time thinking about coming up through my
personal training.
I'm
able to leverage a lot of the in-house expertise that we have to even
know where to start. If you're trying to dive into the current
literature on archive.org for some field, it can be difficult to know even where to start.
Jordan Schneider:
I've found that loading the archive paper into ChatGPT and asking it to
explain it to you step-by-step is not quite the same as having 50 PhDs
who happen to report to you. That was one hack. You can also just start a
podcast and make people explain it to you on air.
The Cyber Workforce Gap
Speaking
of cyber, there's this puzzle I don't really understand. You see these
articles talking about how there are like 20 million unfilled cyber
jobs. There is a market and if you raise the price on the salary of an
open position, eventually you will fill it. What are the implications of
all these jobs on monster.com that don't have people sitting in them?
Charles Clancy: The most authoritative source for the cyber workforce gap is CyberSeek. You can see by metro area, by state, what the current vacancy rate looks like. Last time I looked, we had somewhere around a million total cyber jobs here in the US and we had 35-40% vacant.
Yes,
you can raise the salaries but you still only have 600,000 people for a
million jobs. You're basically escalating the money those 600,000
people are making. The effect that you have on attracting
another 400,000 into the market only works if you've got the pipeline,
from an education and training perspective, to get them into that
ecosystem in the first place.
When
I was a professor at Virginia Tech, I was involved with this
Commonwealth Cyber Initiative. It was a statewide initiative to really
figure out how to ratchet up the supply of cybersecurity talent going
into the D.C. economy. What we found was that there wasn’t really a way
to close the gap with the current computer science programs at the
higher ed in the state.
If
you went crazy in terms of investing – in new faculty, new programs,
new buildings, new classrooms, new labs – you could maybe double the
supply of cyber security talent coming through four year colleges into
the market. That would not even close the deficit that we had. We weren't going to solve the problem by just scaling up our universities.
What
do you do? You can try and focus on community colleges. You can focus
on trying to attract people who don't have a computer science background
into cyber. All of these ideas have been tried. There's all kinds of
programs out there that have been trying to chip away at this.
AI
is going to help us automate many of these jobs. That's the only way
that we're going to be able to effectively close this massive and
growing workforce gap in cyber.
Jordan Schneider:
Regarding the 40% of unfilled jobs, does that mean that companies are
at a 5 in terms of security instead of being at an 8? Does that mean
they just hire third parties? Are people trying to do things in-house
that they shouldn't be doing in the first place? What do all those
unfilled seats mean for industry and from a national security
perspective?
Charles Clancy: At
the end of the day, it means that we're less secure than we aspire to
be. It's not as though it means they're getting it from some sort of
managed security service provider. This is the total market. We're
looking at it in aggregate across the entire US population. There's no
secret way to fill those with other people that aren't being counted in
the mix.
We
are not as secure as we would like to be. There are many more routine,
maintenance kinds of cyber jobs that are not filled, and maintenance,
routine cyber functions that are not getting executed at the pace that
they should be. Again, there's opportunity for automation to improve the situation.
Finding Mission-driven Tech Work
Jordan Schneider:
From a technologist's perspective, what are the trade-offs involved
with going to work at MITRE versus some place in the private sector that
would also use their skill set?
Charles Clancy: MITRE
is a really interesting place to come to get a sense of some of the
national-level challenges that we face as a country. I had a student of
mine who was trying to contemplate his career after finishing his PhD.
He
was saying how he could go work at a big tech company on the west
coast. They would take his AI skills and put them to use selling better
ads. They would pay him a lot of money but he didn’t feel
like that was the best way for his skills to help the country. He was
very much looking at going into a national security technology area at
an organization like MITRE.
That's
an important trade-off. If you look at what much of the tech sector is
doing, many of those jobs are really about applying emerging tech to
slightly improve the features of these massive technology platforms and
ultimately sell better ads. There's a certain lack of purpose in some of
those positions.
I
would hope that a place like MITRE could provide a richer experience in
terms of the impact that you can have on not just national security but
on other major domestic programs as well.
Jordan Schneider:
Has the rise of companies like Anduril and Palantir made it more
difficult to keep talent that's already bought into the mission? I’m
talking about people who wouldn't have been that excited to go work at
Lockheed. What has that done in terms of the talent pool?
Charles Clancy: This whole rise of the defense technology unicorn is fascinating. I happen to be the founder of one of them, HawkEye 360. I've been actively contributing to this new business model for how defense can work in the future.
Shifting much of the defense industrial base is probably not happening anytime soon. I’m
really excited about the opportunity in certain areas and missions, to
show how the defense industrial base can buy things as products rather
than by the hour.
That
change means that you're shifting much of the innovation
responsibility, and also a lot of the risk, to the companies that are
developing those products. The flip side is that the US government has to realize that the product they buy may only meet 80%of the requirements.
However, it's going to cost dramatically less than if they had put together a traditional acquisition.
I'm very excited about the opportunity for these defense tech unicorns
to really help shape the future of how the government as a whole
acquires what it needs to do its job.
It
also provides some interesting diversity in the labor market for people
looking to contribute to national security. Still, I would imagine
between all of those companies they probably have 2000-3000 employees.
MITRE's got 10,000 employees, hiring 800 people every year in the
technical areas. There's a long way to go before I'm worried about
competitive labor market pressures from some of those companies.
Communicating MITRE’s Work
Jordan Schneider:
MITRE used to be really quiet. I was talking to your comms person just
recently, Linda. She mentioned that five years ago they had a comms
person whose job it was to say, “no comment.” Yet here you are on
ChinaTalk today.
I've
been reading papers coming out of your organization over the past few
years which are providing a useful technologist-informed perspective on a
lot of these policy questions. They’re a whole lot more salient than
they were 10-15 years ago.
What
are your thoughts on MITRE's evolution in the broader discourse? What
do you think this discussion needs in order for it to be a productive
one in the coming years?
Charles Clancy: Government
has changed significantly. In terms of how the agencies that support
our R&D centers treated us and how they expected us to act, it was
very insular. We were their FFRDC to do their mission. We, as a company,
didn't have an independent opinion or point of view. It was our job to
help inform the government on what they should be doing and what their
opinion or point of view should be.
Certainly
there's still lots of that. Given the trusted relationship that we have
with our sponsors, we still hold many of those attributes in the work
that we do. However, many of our sponsors are also
realizing that they need to be a lot more open in what their needs are.
By extension, they need us to be more vocal and engaged in the rest of
the world.
We were talking about this world of defense technology unicorns.
The
Department of Defense can't acquire some of those product-based
solutions if there isn't a much more collaborative ecosystem. We need to
talk about the needs of the government so that investors can understand
where to put money into companies that are going to build those
products.
We
have been asked by many of our sponsors to help be part of that
connective tissue and foster some of these ecosystems. That's only
really possible if we're willing to say something other than “no
comment” when people want to have us on a podcast.
So
we've got to be out there. Our sponsors expect us to be out there. That
doesn't undermine the trusted and proprietary relationship that we have
with many of our sponsors and their programs. It reflects a shift in
what's expected of us and what's needed in order to drive the broader
federal economy forward.
Let’s
go back to the national competitiveness perspective. We do have this
deeply technical workforce and expertise within the organization and
we're trying to solve problems for the whole of the nation. So there
isn't just one federal agency that we can go tell our thoughts about
semiconductors to, for example.
It's
really the need to tell everyone and tell the White House and tell
Congress. We have to take these technically informed approaches to
policy and share them more broadly. That's a lot of what you're seeing
more of from us, in terms of engaging on tech policy. It's
because the only way to affect things on the level of government and
the whole nation is to be able to tell the whole of government and the
whole nation, rather than just an individual sponsor who might be paying
for a single study.
Jordan Schneider:
With the rise of great power competition and the war in Ukraine over
the past few years, are you guys getting more job applications than you
were before? How are you seeing the appetite wax and wane for these
mission-driven national security technologist jobs over the past few
years?
Charles Clancy: There's
always interest in places where you can do real bench research, like
with scientific work. If you're looking at aggregate applications to
MITRE, that's probably more affected by the macro tech sector economy.
Big tech companies all have a bunch of layoffs and suddenly the volume
of applications goes up or goes down. It's hard to separate some of the
national security things from the macro tech market.
We
see a constant interest in being able to apply scientific research to
impactful things and being able to do that in an environment that's not
necessarily a university. Having spent 10 years as a professor, I certainly understand that world and realize that it's not for everyone.
ChinaTalk
is a reader-supported publication. To receive new posts and support my
work, consider becoming a free or paid subscriber.