Mosaic Warfare and Human–Machine Symbiosis | Bullaki Science Podcast with Timothy Grayson
![Mosaic Warfare and Human–Machine Symbiosis | Bullaki Science Podcast with Timothy Grayson Mosaic Warfare and Human–Machine Symbiosis | Bullaki Science Podcast with Timothy Grayson](/pic/mosaic_warfare_and_human_machine_symbiosis_bullaki_science_podcast_with_timothy_grayson/XzVNa1hEX202UWM_.jpeg)
hi the following is a conversation with dr timothy grayson as the director of the strategic technology office at the defense advanced research projects agency or darpa timothy leads the office in the development of breakthrough technologies to enable war fighters to field operate and adapt the distributed joint multi-domain combat capabilities at continued speed he is also founder and president of fortitude mission research llc and spent several years as a senior intelligence officer with the cia here illustrates the concept of mosaic warfare in which individual warfighting platforms just like ceramic tiles in a mosaic are placed together to make a larger picture this philosophy can be applied to tackle a variety of human challenges including natural disasters disruption of supply chains climate change pandemics and so on he also discusses why super ai won't represent an existential threat in the foreseeable future but rather an opportunity for an effective division of labor between humans and machines in what he calls human machines abuses dr grayson thank you very much for doing this pleasure to be with you so i wanted to start maybe with some introduction if you could introduce a bit uh what's darpa and what's your role at darpa sure absolutely well thank you and thanks again for the opportunity to talk with you here today so first of all let's start with uh what darpa is so so darpa is uh stands for defense advanced research projects agency and so we are considered sort of like the lead r d arm of the us department of defense uh we were created uh in response uh way back in the 19 late 1950s by the eisenhower administration in response to the united states waking up to sputnik and uh president eisenhower in 1958 said uh we don't ever want to be technically surprised and darpa it was actually arpa at the time the d got added later to distinguish it from other arpa kinds of organizations around the government was created to make sure that we always at least keep abreast of if not are leading in technology we like to say the best way to avoid surprises to create surprise so we like to stay out there at the cutting edge of r d and and we can talk more as we go through the discussion but we've got a lot of different ways of how we do innovation i run what's called the strategic technology office within darpa and there are six technology offices uh of various different levels of of research and types of technology maturity uh from very basic research to more applied research so so my office is at that end of the more applied uh types of research areas uh along with another office uh tactical technology office that mostly looks at weapons and platforms so so the more physical material kinds of things we work with what are called mission systems so a lot of communications sensors things of that nature and but look at them again from a more mature research and a very systems uh type of view and for the youtube watchers and the robot enthusiasts i mean darpa collaborated with the boston dynamics in the development of the famous big dog and little dope dog robots and as you mentioned you said darpa was initially called arpa and in 1969 from right they came up with the arpanet which was the first wide area packet switching network with distributed control and one of the first networks to implement the tcp protocol suite so basically you guys invented the internet so i was wondering if you could tell me what are the top greatest darpa inventions that change the world so i i certainly think that internet is is one of the things that gets raised there uh near the top another one is uh stealth technology most of the original prototypes of stealth technology uh came out of arpa at the time uh there have been a myriad of other things from the f-16 rifle during vietnam uh to self-driving cars uh in the in the darpa grand challenge and what's interesting when you look across that array of different big breakthroughs it highlights i think a couple of things about the agency that are that are fairly unique you know first of all we're often contrarians and stealth is a great example the air force at the time was all about making planes go faster uh you know supersonics and and uh you know and just go faster darpa out there pushing prototypes for stealth was really trying to open up not just new technology but a new way of thinking about what the mission was and have conducted that mission and oh maybe you don't have to go so fast if if it's really hard for a radar to see you that's one thing we do and you know so we don't we we like to say we don't do requirements and i'll put that in in air quotes you know we do like to solve problems we're focused really mission focused on solving problems but we don't wait for someone to tell us what to do you know otherwise we would have been sitting there in the 1970s saying let's figure out how to make an airplane go faster you know instead it's like yeah we understand what your fundamental problems are air force let's see if we can think of a way that technology can solve those problems entirely differently and then the other interesting thing about you know where our breakthroughs have been i mean the internet that you mentioned is certainly one of them but you know talking about the grand challenge and the self-driving cars a lot of times what we do because we're out there being contrarians we don't necessarily see the one-for-one immediate impact of what we create that original arpanet as you pointed out was late 60s it really didn't start getting converted to widespread even use within the research community until the 80s i had my own uunet account and some other things based upon early tcp as a grad student but um you know it wasn't until uh world wide web and in the 90s and then all of a sudden this boom of commercialization never a thought when the arpanet was created uh and i think we've seen a similar thing with self-driving cars i was a one of the judges during that grand challenge and you know no one there at that first grand challenge would have ever thought that there would be a whole industry of self-driving cars but yet when you look at a lot of the winners uh the people who completed the grand challenge they're they're all now the teams at the forefront in the commercial and academic world uh really advancing uh you know what will likely be a a global commercial market so we work in strange and mysterious ways yeah and the grand challenge was the one uh the long drive right that's right yeah it was uh the whole question was could a vehicle drive by itself on very rugged terrain across the desert and it was fascinating i was a judge for the first one and the farthest vehicle made it about seven miles and it you know people kind of looked at it and chuckled a little bit and said wow what are you crazy darpa people doing and but but darpa didn't give up and did a second grand challenge and just a year later without really uh the government providing the upfront funding this was done as a challenge people were building their own teams and raising their own money for it a year later uh i think it was five teams that created this entire race uh and tremendous advances on for incredibly complex terrain in the course of just one year and what's a darpa hard problem and why do you need an organization like darpa to tackle it sort of as i was pointing out the first part of a so-called darpa hard problem it's something that lends itself to this contrarian alternative view if there's a nice clear technology roadmap here's where the research community is and here's the next logical step yeah we don't we don't tend to get involved in those you know we look at things first of all in a very different way so is there a different way that technology can help solve this particular problem but then the other part of it is you know we're known as a a risk-taking agency but i like to characterize how we do things fairly uniquely as uh smart risk taking you know i put i put it the two extremes you know one is i don't want any risk at all i want you to sort of prove to me all the technology is going to work before i go try something and then i'll do little incremental baby steps there's the other extreme which which i call you know hope is a strategy you know someone comes forward with an innovative sounding idea and a pretty cartoon uh powerpoint chart and say wow you're onto something that that's a really clever idea say how are you going to do it and it's like i don't know we'll hire smart people and they'll think about it you know we start with problems where we understand enough about the problem that we know what the risks are and we have some way to rationalize that we have a chance of being successful and then we can build programs around that that are focused at retiring those biggest risks first and then we can move on to building bigger systems and doing more exotic programs and that gets to one other thing i think is sort of magic about our model versus a lot of other r d organizations we do have latitude for a certain amount of of of curiosity driven work and we give a lot of flexibility and latitude to our program managers uh it really is bottoms up based upon those program managers but we're very problem-centric we like to say we're mission focused you know so pretty much everything we do starts with what problem are you trying to solve and then from there we can explore you know one of the different technology opportunities and have an extremely wide aperture but it's i think it's that problem focus that also creates a lot of ability for darpa to move things quickly uh especially when you mix it with the the the risk acceptance culture yeah and so you did a phd in physics where you worked on quantum optics before it became quantum information and i guess people now are familiar with the the possibility to build a quantum computer i can open new new fields especially for things like cryptography or serving quantum systems and so on and then then you moved to darpa as a where you worked as a scientist so i was wondering if you could tell us a bit about this journey and how that influenced your sort of approach towards problems yeah absolutely yeah great great great question so yeah i like to to joke around that i did i did uh quantum computing before it was cool i was in a lab a very very prominent pioneer in the field of quantum optics professor leonard mandel unfortunately passed away a number of years ago but yeah he was one of the pioneers of quantum optics and really laid the foundation and and we were we were a lab group that was uh doing experimental uh realizations of of quantum optics and it was interesting it was a fascinating you know time for a phd and i love the research work at the time most of what we were doing uh was frankly experimental demonstrations of copenhagen theory and and various aspects of fundamental quantum mechanics and uh you know it was it was exciting stuff it was really fun research uh at the same time there was sort of this practical side uh nagging at me that that said you know everyone already believes this uh and accepts it we just haven't demonstrated these particular principles and it was interesting i was writing my dissertation the year that peter shore uh published his factoring algorithm which i would argue sort of kicked off the whole you know quantum computing quantum information area so but i had already made the decision that i was going to go do something more quote practical uh then continue this quantum optic stuff so i went i went to work uh as a postdoc for the air force using a lot of the tools of what we had done uh in the lab for instance to do our research uh you know my dissertation was was doing quantum optics based upon uh a two-photon entanglement and and so to implement that we were doing a lot of work with non-linear optics and and uh uh spontaneous down conversion when i started working for the air force lab uh out at wright patterson and dayton ohio they were interested in in laser sources based upon non-linear optics so it got into the business of developing different types of laser and other optical uh sources and doing research in uh non-linear optics but i think that's where again this this sort of practical side kind of kind of tugged at me you know because it was really interesting research going on there at the device and and the system level and yeah i kept saying but but what's this good for how are we ultimately going to be using it and and tried to look at the full system level problem you know if we could do this radical new laser source how would it help us build a sensor and oh what would i need to go along with it in terms of a new type of camera uh is is there software and different types of algorithmic processing that might make it a more practical capability if a lot of these lidar types of uh you know called beeps and squiggles uh you know weren't intuitive to to a human uh looking at a picture and so that led me to sort of this the system level kind of thinking i i would call it and you know i i guess to some basic researchers that sounds like you know might be boring or a little bit of a sellout but you know i i found it fascinating to think about how can we take really fundamental science but apply it in this problem-centric sort of way and that really excited me and i had the opportunity in in the mid 90s uh to move to the dc area and start to work for darpa uh first as a support uh contractor uh doing a lot of the technical analysis to help with execution of programs and then a couple years later they yanked me over to the government side as a program manager and and that was a big jump also because up until that point i was doing uh very hands-on research i had my own lab my own lab group some graduate students i was advising uh publishing papers all those good things uh that i'm sure your viewers their normal life as a person at darpa um i i like to to joke somewhat tongue-in-cheek we don't actually do any real work um uh you know we just make powerpoint and shovel money uh it there are no there are no darpa labs uh despite what you know is in the tom clancy video games and some of the movies all of the actual research we do is extramural we we have our success based upon the team of contractors and university people and some of the other government labs that actually do the the real hands-on research but at the same time we demand that the people who come to darpa be highly technical they need to be leaders in their field they need to have done hands-on research so that uh you know they're not they're not bureaucrats they're not paper pushers they they drive the vision uh you know they're helping to create the vision back to that smart risk-taking that i said they know enough themselves that they they know what the technical risks are they know how to structure their programs around those risks and then even though the work's being done extramurally by this group of contractors and university performers they can provide the due diligence and again a common sense way you know a lot of parts of the government that don't have the same level of technical expertise and their program managers have to resort to to checklists and requirements and things like that our our men and women are so uh experienced and and acknowledged in their fields that that they they intuitively know what the issues are and can dig into them and ask the hard questions and then and then based upon the response then that they're empowered to make decisions and pivot quickly as their research evolves yeah that's why i was going to ask you how important it is that leaders know and understand what's happening in the labs because sometimes you have managers people coming from business environments or managerial careers that have no idea about the technical sides they don't have any background so how important is i mean we have the biggest example probably is elon musk right it knows every single thing probably that's that's the stuff but i i imagine that's correct and that's how we run successfully it is a multiple companies so how important is is this this kind of aspect for you it it's very it's very important um but it's it takes a certain breed of researcher so you know that you have to have been technical coming in uh interestingly enough we don't we don't have any firm requirements on credentials you know while the majority of people at darpa have phds in a technical field it's not it's not a hard mandate but we do look for people who are technically accomplished who have actually done their own research the model i like to say and even this is not hard and fast we like to say that you've got to be you know a mile deep in some technical area just so you have that that experience base and you know what it's like to do research but you also have to be inch deep at least across a very wide range that's i think one thing that differentiates a darpa program manager from a lot of other very accomplished researchers you could have someone who is you know one of the the world leading researcher in whatever their academic area is and they work in it their entire career but but you ask that person to step outside that lane and they get very uncomfortable darpa pms have to also be very fast studies i think that maybe the biggest characteristic i never really thought about this but it's academic curiosity it's someone who has already proven that they can go deep that they've got the technical chops but then it's also augmented by that technical curiosity oh a new challenge comes up and someone presents them a new idea i can be a quick study and i'm not going to be the expert who's going to go toe-to-toe doing the research but like you said with elon musk he's not building spacex rockets or he's not built you know designing batteries himself but he's a quick enough study that he can ask the right questions and and make informed decisions and that's kind of the model that we look at for our program managers so what kind of organizational structure do you have there is it a flat or is it tall i mean is it uh is the management command oriented or is it flat it's very flat it's very flat so to my point about you know we don't have darpa labs it pretty much begins and ends with the program managers you know i mentioned we've got these six technical offices but uh it's they're populated by program managers i don't know the exact number off the top my head but it's somewhere around a hundred or so program managers and everything begins and ends with them you know they generate the ideas uh they execute their program activities overseeing them but again the actual research work is conducted extramurally as a result you know we don't have a structure uh below them and and above them it's basically folks at my level at the office management level and then uh the agency uh director and deputy so uh you know within execution of an actual research portfolio uh the the pms have have more or less total autonomy and freedom on how to execute the programs now if any of them listen to this podcast some of them will probably uh you know grimace when i say this but you know i uh we at the office level have very little control uh or oversight of what they do or even when they start their programs the ideas generate them we do provide at the office level uh what's called an office strategy and there's a similar strategy at the agency level but those are just general guidelines you know to lay down these are the types of problems we're interested in uh and then it's ultimately up to them to generate the ideas and then execute those ideas on their own okay and how do you protect uh the intellectual property there am i how do you ensure that the things don't get stolen by whatever it always happens i mean it happens in companies it happens everywhere so what are the best strategies to protect intellectual property so so i'll say there's there's lowercase intellectual property and uppercase uh intellectual property um the thing you're referring to are actual you know secrets and things of that nature and yeah i can't go into a lot of that but we are a government agency we are part of the department of defense uh and so uh a good chunk of our work is is classified and it's protected through all of the you know the actual security and and classification measures there on the though capital uppercase intellectual property you know what's interesting about the way darpa functions we we sort of stand with a foot in both camps you know so we do a lot of we do a lot of classified work that never sees the light of day but we understand that a lot of the technology is out there uh in the commercial and the academic world and you know if we do everything you know lurking in the shadows you know we can't engage with those communities and we we risk you know just sort of you know being insular so you know we we do our best to while still protecting all the the rules and constraints of security to try to engage with those communities the fact the fact that i can be here talking to your podcast is an example of how we try to be transparent and open that does create a different kind of ip challenge you know because we want to be able to engage with startups you know we want to be able to engage with commercial companies that you know have some i you know legal ip and and so it's interesting one of the things that also is nice about darpa is we've got even though we're a government agency uh we have a lot of flexibility that other agencies don't have we've been able to get a lot of special authorities and a good example is on contracting because that's where a lot of the ip constraints pop up in a in a traditional u.s government research contract the default is what's called government purpose rights that basically says hey you know company or university we're paying for your research and as a result we want to be able to get access to it we don't have to pay for it twice you know wants to fund your research and then wants to have to buy a license back from you but that being said we understand we want to partner with people that have done a lot on their own and have put a lot of their own investment or a vcs investment or whatever and we respect that so we will uh occasionally enter into special contracting relationships where uh there's a policy within the u.s government called other transaction uh for prototyping and darpa makes a lot of use of that as an example so that we can engage in more of a traditional type of almost business type of contract with people as opposed to a strict government contract so a lot a lot of different ways that we we try to engage and protect people's ip do you know what's the percentage of darpa projects that pay off and how many develop like unexpected spin-offs so that's a that's a really tough question because as i was saying before um most of our technologies don't go immediately into use you know we're usually working things that you know by their nature or contrarian or thinking about a problem differently so we i don't know the exact numbers uh i would say the percentage that go directly into you know say a big military production program or something like that are actually uh really pretty small i don't know i'll make up a number probably somewhere around 10 percent uh maybe even less i don't know but that's almost by design you know because if if we were doing things that were so well aligned with the production program you know we're probably not out there taking enough risks and we're probably not being contrarian enough so i would say the majority of our efforts do transition but transition indirectly and i would say there are three big ways they transition probably the majority fall into this category where someone is going to pick them up to do more research and it could be another government lab or it could be a company that chooses to do it on their own awful lot of our capability does come back and get used through these indirect paths you know we were talking before about uh you know arpanet you know arpanet in 1968 or whatever it was certainly didn't didn't go anywhere right away but you know 20 years later it started being used academically another 10 years after that it changed the world uh you know so and i guess the other thing a lot of our technology especially that some of the more fundamental research either goes uh into commercialization you know where it might spin out uh you know be matured and then shows up in all kinds of different products that come back and benefit the government and the military but also the rest of the world something that i think your listeners that you've covered before uh mems microelectronic mechanical systems that was something that that a lot of the early research was done by darpa and none of it went directly into a dod product you know but it the the component level things get matured and then they start showing up in all kinds of products from ejection seats and military aircraft to your cell phone accelerometer and actually it's interesting this is outside the field in my office but a lot of the big current push for vaccines for the kova 19 pandemic darpa did not directly fund any of that vaccine development but some darpa research about a decade ago led to the original research into messenger rna okay based vaccines and the fact that they were able to develop those vaccines so quickly it wasn't just about governments throwing a lot of money at it it was that there was this new technological foundation that enabled them to look at vaccine development in a different way we we transition things in a lot of strange different ways and i think a majority do have an impact in some regard but just not in the way that people like to think in terms of that direct one for one transition so i'd like to change a little bit uh topic so the thing i wanted to ask you is related to war and i was wondering if you think there is any link between conflicts and the emergence of new ideas so do you think extreme competition can lead to more innovation i mean we've seen plenty of innovation during the second world war uh even during the cold war so so i think the short answer is yes um i i think the thing that is interesting to think about in the 21st century is what does competition mean uh so you know i i i surely to goodness and this is the whole reason that you know it sounds like an oxymoron but the reason the department of defense exists is to create peace and stability you know so we don't want to see a war we don't want to go create a war for purposes of creating uh you know innovation but uh you know independent of whether there's a war there's always competition uh and i don't care if it occasionally is some military competition bumping up against each other or you know unfortunately when we see regional conflicts pop up because of uh natural instabilities there but there's always some degree of global competition i think one of the things that really has changed the innovation landscape and and this is something that we think a lot about within darpa is competition largely is even driven in the general economy uh and in commercial competition uh and and you could think of that as a form of warfare uh look at the disruption that's happened over the last decade or two we had all of these household name global industrial corporations that have tumbled and that's a form of of economic warfare if you like so and i was just reading an article uh this morning over my breakfast over over the you know predictions for 2021 and and has the pandemic shakes itself out you know that's that's not an intentional warfare uh you know that's a naturally occurring uh disruption but nevertheless you know pandemic is a form of disruption and and uh you know one of the terms used in the in the startup world is is creative destruction you know in any kind of disruption it creates uh it creates dis misfortune it creates discomfort but it opens the door for for new opportunities that emerge uh you know both and i think there are two things that drive that uh one you know there's the old saying necessity is the mother of invention and and it goes back a little bit to what i mentioned with darpa's model about being mission centered if you've got a really clear tangible problem that motivates people that gets them focused you know it's not technology or investment looking for a problem hey it's right there in front of you so i think that's one reason that conflict does accelerate innovation that demand signal that focus the other is that it it removes a lot of the barriers um you know with it within government we love to you know complain about all of the bureaucracy and process and procedure you know when you're faced with a serious conflict or any kind of disaster you know all of a sudden a lot of those those various processes and checks and such become a lot less important and if people can go try something and realize oh it wasn't the end of the world when we tried this thing a different way then after the conflict is over people say well why were we doing it that way before you know why why can't we just keep doing it the way that worked effectively during this conflict so i think it's both that demand signal but then also reducing barriers that the competition and and any kind of disruption does lead to innovation so how would you describe the world of warfare i mean if i try to describe a system we have different parameters we need to talk about domains dimensions rulers and protractors or sensors players laws so what are all these aspects i mean how do you accurately model this word what what do you need to take into account to make some good model of the world of warfare so so let me let me pull two threads here uh the first one is sort of uh you know i'll say a bounding framework for what you describe there's something that's a uh sort of a driving uh trend right now within the us dod uh it's got a couple of different names but but for purpose of this discussion we'll use one of the names that goes under a joint all domain operations or join all domain command and control so we'll just call it jdo and the the key thing there is all domain uh you know let's talk a little bit about what a domain is historically you know you got an army that fights on the land you got an air force that fights in the air you got a navy that fights at sea you know so we talk about those physical domains uh there's been a lot of press over the past year we just passed the first anniversary of the u.s space force that's an acknowledgment that space is a domain people talk a lot about cyber or the electromagnetic spectrum those are all domains i've even heard of a domain referenced before when we start talking about um you know things like uh information or you know i'll call it the cognitive domain uh you know people refer to hybrid warfare or gray zone warfare uh hearts and minds things of that nature so all of those are different domains and and one of the big trends right now is a realization that you really put yourself at a disadvantage if you look at only one domain or even you you look at multiple domains but each individually you know because the reality is they're all of it's very very interdependent and codependent the second big uh factor uh that ends up popping up and they're both directly connected and and the the punch line here i i want to come back to the question of complexity uh and dimensionality in all of this the other one is speed yeah time yep time and and one of the terms that we've been starting to use around darpa uh in fact i'll give credit to our our director uh for inspiring this term victoria coleman we're calling it time compression uh it's how can we make time speed up essentially uh how can we it if simplistic levels how can we do things faster you know but but there is this traditional determinism you know it particularly within within dod it's like to say okay i want to study things and and try to forecast you know what the future is going to be and in the process of doing that forecasting come up with every possible contingency and then i'm gonna go build one you know heck of a powerful system that is either you know has a high enough performance level or or is adaptable enough that it's going to address every single one of those contingencies the reality is we are now in such a complex multi-dimensional world that that all all posit that this forecast model doesn't carry you know so instead we have to be in a mode of rapid responsiveness you know how can i just sort of assume that that i i can't you know how can i acknowledge not just assume how can i acknowledge that i can't predict the future you know with any degree of accuracy and there are going to be contingencies that occur that i just haven't forecasted so how can i be in a rapid responsive mode that's time compression and and so a lot of what we're looking at right now is both of those big themes pulled together how can we be all domain and and recognize all the independent dependencies of this very high dimensionality space and how at the same time can we do it with incredible speed in this notion of time compression you know there's an interesting little aside if if i can i love i love this story uh because it shines a light on this issue of time compression uh and it's about toilet paper of all things okay uh it's it's uh this was a this was inspired by an article i read in uh uh fortune magazine back early in the pandemic yeah and and they were talking about the the shock to the supply chain for consumer goods when everyone had the run on stores and uh and the shortage of toilet paper and they said look if you look in the business world particularly in manufacturing they're also a very forecast centric uh you know type of world they do a great job at data analytics you know a company that makes things will forecast out almost you know to i don't know how many decimal places what they think their sales are going to be and it's all about managing that inventory you know it's efficiency and one of the things they said in this article is that one of the reasons the shelves ended up empty of toilet paper is they were already running at something like 93 percent capacity just so they were making sure they had no inefficiencies that's great from an efficiency and effectiveness standpoint but it leaves no resilience that's a that's a very brittle system and you know when all of a sudden there's this contingency disruption of a pandemic and and people doing a run on stores uh it doesn't have the latitude to respond to that and you think okay i'll take a legacy forecast approach and says well i need to open up my air bars you know i need to account for well that's not practical you know because that would say okay i need to have three times the quantity of milling well if the typical toilet paper manufacturer you know went and bought you know three milling machines for everyone they have today they'd go out of business yeah they can't afford that level over provisioning so the answer is what are things i can do that can be rapidly adapted how can i measure disruption and then how can what are the knobs i can turn to get to good enough it's not going to be optimal but how can i get to good enough in a rapidly responsive manner and that's that's what i see is the the big trend going forward and that's certainly where my my organization has been focused i think the problem with the toilet paper was that toilet paper takes place so when people buy let's say five or six packs of it they just uh then the next customer will see an empty shelf so they will think they're running out of toilet paper so you could think about compressing it even more so it looks small and you can put more and more so that doesn't that's you you you hit the nail on the head in fact in this article they talk about that okay and that's exactly that's exactly what one of these i'll say soft knobs is you know they couldn't go whip up a new milling machine in a day but what they could do is reprogram their production line so that they could start packaging uh you know packages at smaller packages fewer roles vacuum pack the toilet paper yeah well yeah in this case i was just doing four roll packs as opposed to 24. but yeah it's the same idea yeah yeah and and there were other things too like you know if you know some of the shortages were caused by disruptions of the supply chain yeah and and uh you know so oh can i have what amounts to a vendor you know radar for news sources of pulp and a new process for vetting and validating those providers in a faster manner so it it's a great example of taking a very system architecture level thinking of these complex problems as opposed to the obvious solution which is build more milling machines yeah yeah it might be the obvious solution but it's not a very effective one in the long run i mean i was reading about uh uh you know uh alphago and all these uh these ai systems where they take into account uh like certain games and so on and then they try to build these uh super expert systems that can uh fight against or play against humans but when you model a game for example uh chess the sort of game three complexities 10 to the 120 or if it's go is 10 to the 700 but when i think about a war game that must be like thousands of orders of magnitude more complex than chess and go so that's why i was thinking what are the approaches to achieve an acceptable level of modeling without without something that works but that is not too much complex because in theory if you want to like model things you could start from the wave function and the schrodinger equation that's the most ridiculous thing that someone can do but so what what's the sort of modeling that you guys do if i can ask this question yeah no that's a great great question and something i personally think a lot about and and i'll even i'll even toss it out to your uh your viewers if anyone has any uh ideas on how to to to do this kind of uh you know framework analysis uh in a quantitative way i'd i'd be interested in hearing your ideas and who knows maybe get a project out of it the fundamental way we're looking at it is is managing uh in some ways this is borrowed straight out of uh network theory uh it's it's uh uh looking at how do we manage complexity and dimensionality by breaking things apart into scale and and so you know if we think about some of the dimensions you mentioned i mean that's those are consistent with some of the things we see at the decision support level inside an individual platform or a payload and and one of the things i think we'll probably end up talking a little bit more about is our alpha dog fight but if we have a you know ai flying an aircraft or we have an ai controlling uh you know a sensor payload you know those kinds of decision processes are on the order of you're saying maybe you know 10 to the 200th or so now imagine you know that platform or that payload is part of one mission unit and that in turn is part of you know whole squadrons or forces and then i'm talking all domain and all these other types of things so each of those is taking about that same dimensionality and now growing that geometrically so people who think that they can take that kind of endeavor and as as you know one of my bosses likes to say sprinkle some ai pixie dust uh on it and and just assume that the the uh that some algorithmic approach is going to discover a way to manage you know that level of dimensionality is is just really not practical so instead we say how can we break up those those decision layers uh i'll think of it in terms of decisions that have to be made how can we partition those into a manageable degree of complexity or a manageable dimensionality and then still create optionality by being able to abstract those interfaces and those boundaries so it's really about partitioning and abstraction and if you look about it the original inspiration for this this was before my time as office director but um a number of years ago uh maybe 2015 i want to say 2016. darpa sponsored a conference uh called uh wait what and one of the speakers there uh was a chair professor from uc berkeley and he was one of the creators of the original design tools for semiconductors and and he really presented this notion better than anyone i've ever seen you know because it's it's intrinsic you mentioned knowing the wave function well you know at one level if you want to do a really good semiconductor it would be great to know the wave function of every you know gate uh you know and sink in a semiconductor but you know like you said not practical do you need that no probably not yeah that's right so we've got we've got moore's law because of this notion of being able to manage scale by partitioning an abstraction and then once you've abstracted things being able to do composition with those abstracted elements and that's how we're approaching this this all domain uh you know this all domain uh warfare challenge i i will say that where we're living right now in a very challenging but also interesting intersection between technology and culture because to live with that model of abstraction and composition drives a certain amount of acceptance of uncertainty you know i don't know what's happening underneath the the hood underneath the abstracted boundary of that next module uh but but trust me it's going to satisfy some function well usually if you're talking to people in the military they don't like an answer to just says trust me you know they want determinism they want a certain number of decimal places of certitude and but i argue that that is committing the statistics 101 fallacy of mistaking precision for accuracy you know we're living in such a complex dynamic world that we have to be able to back to the toilet paper live with disruption live with uncertainty and the only way you do that is by you know taking a more stochastic kind of model toward things and so let's say that you have these these models of the world of uh warfare and what can you use them for how are these models deployed is it for training or actually fighting real wars or how does it work yeah so it's so it's it's all across the board so so uh within my office our our guiding uh portfolio for these kinds of things we're calling mosaic warfare and uh you know you can see the uh our little logo here behind me uh the the whole metaphor of mosaic warfare is you know if you look uh as a contrast a jigsaw puzzle that's a highly engineered architecture you know i'm creating an overarching effect with a composition of existing pieces but every single one of those pieces is very carefully engineered to how it's going to fit into that that broader picture they're difficult to put together they are they are very brittle and fragile once you've created them and they're not flexible at all so the mosaic analogy is to say i want to again stochastic kind of model i'm going to have a bag of tiles so to speak that are some arbitrary perhaps even opportunistic distribution you know i might have some control over the statistical distribution of those tiles but i'm not going to specify exactly what any one tile should be but but i i'm going to have confidence that i can piece those tiles together in some way that can still produce an overall picture and and oh by the way you know it's it's not just completely arbitrarily i've still got some kind of substrate i've got some kind of you know adhesive or mortar so i've still got a framework i'm working within but i'm living with that that uncertainty that's too this casticity of a bag of tiles but now it gives me an incredible ability to to flex and adapt and in principle i can create that adaptation with much less difficulty than designing a new jigsaw puzzle and oh by the way it's it's much more resilient to disruption and to uncertainty because i can lose a tile i can you know find one similar and throw it back in so when we when you talk about modeling you know we have to do modeling across the board with those kind of things so some of our modeling goes into just how do we plan the mosaic you know i can go back to the toilet paper example and say okay yeah i know i've got a problem i can't keep the shelves stocked you know i can i use modeling tools to figure out where within my process my supply chain is breaking and and be able to experiment with what are options to be able to create a more resilient supply chain you know so we've got a program for example called proteus that on the surface looks almost like a video game and it what it's actually doing though is allowing people at things like uh you know military universities to explore new types of force structure and to create you know notion of of hybrid military units uh that could be designed much more uh finely and specifically to to give admission needs so that's that's an example of using this kind of modeling uh at a planning kind of stage we've got a fair amount of modeling that goes on uh in how we design the the networks we need although that's a little bit less about modeling than it is uh about new networking constructs and and uh communications uh virtualization and interoperability kind of things and i'm happy to talk a little bit about that more later if you're interested but the the third place where the modeling really becomes important and this is where some of it actually gets used uh in operations it gets back to how can we how can we simplify the problems that human beings have to deal with in this kind of highly networked very fluid uh you know dynamic kind of architecture uh you know if you're responsible for being one of those tiles yourself how do you know what your role is supposed to be in this mosaic and that and that's where a lot of our work in ai has come in but in the spirit of abstraction and composition it's all up and down these various different degrees or levels of a of a uh called a decision making stack so we've got some technology that is at a very high level think of that as the mosaic artist who's saying you know i've got a certain function i want to have happen in the battle space and i want to use a new collection of tiles to go conduct that function what are the best set of tiles to use at this moment in time and so it becomes automated modeling to make those high level decisions to your point about dimensionality and complexity that decision maker using that model doesn't have to know anything about how to actually use that tile or how to you know what what other calculations might have to go in it just sort of votes the tile itself is working a lower level type of modeling again managed dimensionality where it comes in and says okay someone asked me to serve this role i don't know why i don't have understanding of the whole battle space or commander's intent again that level of complexity has been stripped away i just know as a tile i've been asked to do something can i do it what calculations and modeling do i need so that might be for example for an aircraft just something as simple as route planning you know or or we've got one program that's been looking at for collections of of air platforms how do you deconflict the airspace in this this very dynamic manner again none of those functions need to know that high level uh awareness and then another layer of abstraction the things that are doing things like that air space planning those modeling tools don't have to know how to actually actuate the platform so there's a different level of ai that can worry about actuating platforms actuating payloads so each of these layers and this is where my challenge problem comes in if someone has a great you know model who's an expert out there on network theory it'd be really interesting to say how can we actually define a scaling law you know based upon a certain number of you know interfaces divisions and boundaries and this notion of abstraction and and composition yeah and so you came up with this idea of mosaic warfare as opposed to monolithic uh architectures right i mean there is a huge rapture with the past you gave a talk what was it in 2018 where you spoke about the problem of uh dominance you have a couple questions buried in there so let me talk about the dominance one first it actually it in some ways i've really touched on it already if we just change the terms a little bit dominance the way certainly the u.s
military has thought about it is back to that very deterministic forecast based model you know i i want to you know do lots of studies to try to predict what the future mission and the future threat is going to be and i'm going to go ahead and and you know by doing that right just design something that is big enough and bad enough and high enough performance that it can accommodate all of those possible contingencies that's that's the dominance mental model and and so what we're talking about instead of of dominance is this notion of what we really need to be focused on is how how do we achieve our objectives you know whatever those objectives might be it you know and frankly i think this mindset can apply again outside the military you know back to how can we make sure that people can buy toilet paper yeah uh it's it's it really is how do i get away from this i've got to forecast everything and i've got to provision for all possible contingencies and instead time compression how can i be rapidly responsive and adaptive regardless of whatever the opportunity or the disruption is so that's that's really what we were getting at with this notion of we've got to move away from the dominance mindset because otherwise you're you're stuck in this classic cat and mouse problem it's counter counter counter counter you know as soon as you think you've built the biggest and baddest system someone is going to now focus on how to either build the bigger and better system you know or just some counter measure that directly you know negates that capability you know so instead of this dominance approach we're trying another analogy i use you know apology i like metaphors but it but it's it's like trying to pop a balloon with one finger you know it's it's sitting there bouncing on your finger and you can't get any leverage against it because however you push it's just going to squeeze out somewhere else that's that's the the anti-dominance kind of approach you know dominance would say i just want to you know make that as hard as a block of granite yeah but then someone comes along with a chisel and a hammer and your granite's no good you know so i'd rather be the balloon than the granite how is this mosaic philosophy being implemented in phases uh recently i've been referring to what we're calling three waves of mosaic and and the reason for laying it out this way is that as i mentioned before we're right at the intersection between technology but also culture and organization and process and again i i don't think what we're doing is unique to the military or dod you know i've read a lot of articles about in fact this is going on right now in the commercial world uh you know the there's there's the gartner hype curve that you're probably familiar with or your viewers may have seen you know new technology comes along and there's first this huge excitement over its adoption and then people look at how it's being used and all of a sudden they're not seeing quite the outcome that all of the hype seem to justify and then you get the opposite response they call the trough of despair you know and then ultimately you know people who stick with it slowly claw their way out of that trough of despair and you find out well what really is this useful for and then you do see a little bit less hype but adoption and real impact a lot of what's going on in the ai world is similar right now you know it's like okay ai going gonna change the future and companies that have tried to adopt it are like okay we're spending a lot of money buying whatever this ai stuff is where's my return you know they're not seeing it a reason for it is the exact reason why we've got three waves of mosaic and it's that you can't get for really disruptive technology you can get some marginal improvement if you just sprinkle in the technology but you can't get the orders of magnitude kinds of improvements if you aren't simultaneously challenging your processes and your structures to go along with it a business has to truly change its workflows and how it thinks about executing its business to get best advantage from automation dod not only is no different dod has an even bigger challenge because it's so you know locked in rigorously to doctrine and structure and tradition in a very disciplined manner so what i'm seeing in these three waves wave one is really to a large degree outside what darpa's doing right now although i i'd like to say no different than the internet taking a couple decades to catch on we've been working system of systems architectures uh you know at least going back to the late 90s when i was a program manager and and so we've been pushing this for a long time and as recently as about five years ago or so my office uh now has was still working system of systems and people would look at us like we were ogres with two heads what is this system systems thing uh give me my next fighter aircraft so so the fact that there's this whole joint all domain push within the department is incredibly exciting you know uh i i also use the term monolith busting uh sometimes when i describe mosaic so system of systems is busting up monolithic platforms where i've got to have you know the sensor the weapon and the decider all programmatically vertically integrated and technology integrated into one platform so that's that's good and so wave one is where the big military is right now in starting to implement and experiment with system of systems and so you got to start somewhere so it's exciting to see this happening put some markers down try some pilot projects provide some tangible concrete examples of how you can get advantage by disaggregating capabilities distributing capabilities so that's that's wave one and there's goodness there the challenge is we risk replacing monolithic platforms with monolithic architectures and in other words jigsaw puzzles as opposed to mosaics and a lot of the wave one joint all domain activities are jigsaw puzzles you know they want to study things they want to figure out what's the mission going to be what exactly is the set of stuff that i want to go wire together to conduct that mission and then how am i going to manually go integrate all of those and again i'm not knocking it you got to start somewhere but if we stop at that point and say okay we're just going to replace the whole dod with these tailored architectures i would argue what we've done is replaced vertical stove pipes you know the platform-centric monoliths with horizontal stove pipes the architecture centered monoliths and and that that frightens me if we get to that point because as anyone who's tried to do system architecting knows uh and you brought it up in your question about complexity the the more things we put together that complexity grows geometrically building system systems architectures is hard so if we try to build the whole dod with system systems architectures it has a real risk of you know just collapsing under its own complexity so wave two is where we're really focused right now what we want to be able to do is to enable a military operator out in the field to say yeah i've got a bunch of stuff out there that buy it itself maybe has a an existing function that it was designed to go do it in and of itself is a useful standalone capability but now i've got a problem facing me and it could be a new mission it could be a new adversary you know it could be a new environmental problem or whatever how can i take what i've got and take this architectural mindset and build a bespoke to this problem facing me today how can i build an architecture that addresses that problem facing me today with whatever i've got so it's it's it's this notion of it's not everything wired together in a big mesh it's looking for a federated approach you know in this limited set of capabilities on a focused problem there are these you know subset of things that have to work together and maybe they weren't designed to work together but i've got a way to make them more interoperable on the spot and and so that's what we see as wave two i like to describe it as letting the warfighter do system architecting without realizing they're actually doing a technical act they just think they're doing mission planning but there happens to be all these technical wiring diagrams going on behind them the challenge there uh for for those of your your viewers uh who are old enough to remember the days of uh you know like windows 95. you know by the time we got to the 1990s with a personal computer you know it was pretty cool because you had a computer at home you could configure it the way you wanted if you wanted a new printer you didn't have to go buy a whole new computer just to get a new printer so it was thing there were things that were tailorable architectures designed to whatever your need was however if you remember those days adding a printer to your computer was not for the faint of heart you know you had to tear open the box pop in a card you know put in a floppy disk and do a manual installation of a bunch of specialized software that usually didn't quite work right and so that's where we are with wave 2. you know the the technology that's coming out of darpa uh is is you know think of us almost as dell uh you know we want to create the environment for the warfighter where they can again focused on need be able to piece together architectures just like in 1990 we would have put together our home computer as an aside one of the organizational things that i've been out there pounding the pavement about is we need a new function within the military that for lack of a better term i'm calling a combat support geek squad yeah again to use that personal computer analogy you know for someone who didn't have the uh the technical fortitude to rip open their computer to install their printer they could call up geek squad and so who is the the support organization that can be out there 24 7 supporting operators it's not like going back to one of the vendors or program office but yet they're more technically skilled than your typical flight line mechanic so that's wave two um and that's the next logical step because we want to be able to demonstrate that we truly can flex to need wave three and i'm not you know as is as much of a evangelist as i am about these things i'm not willing to push too hard on wave three yet but if we can prove and we convince uh even ourselves let alone the warfighter that they really can operate in this very fluid more stochastic build me a capability to need it is going to change how we think about new systems you know so think about how we're going to populate that palette with new tiles the way we're thinking about wave 2 is the tiles you know the platforms the weapons the sensors are all going to be more or less the current things we've got today when we buy the replacement for those in the future if we're wildly successful with mosaic you know imagine i go buy a new sensor and i say you know what i'm not going to tell you uh what the performance requirements for this sensor should be um i'm not even going to tell you quite what all the functions have to be and i'm not going to tell you how it's going to be used but it's great trust me you know it's it's but we're trying to enable that kind of a future where you could have let innovators come in and say i think i've got this great new radar and i know it's got to be good for something and you throw it into that palette of tiles one of the interesting things you asked about modeling earlier you know one of the things we're even exploring with are sort of game theoretic models as a way to score tiles uh you know in this future vision you know we don't want to dictate how a tile is going to be used how a particular you know capability is going to be used but we still you know dod has a finite amount of money it's not truly an open market so we can't just let the market decide the way you know a commercial environment would so how do we figure out how do we want to spend taxpayer money on these capabilities so you know we've contemplated using modeling tools against game theory to say okay someone brings me a new tile it can be used in the course of this very fluid mosaic does it move the needle you know does this tile actually improve things does it make no difference hopefully it doesn't make things worse but as a mechanism for deciding where to put future investment so that's way down the road we're doing some more fundamental research trying to understand how we would do that kind of modeling and evaluation in the future but but right now we're we just want to prove that we can be good geeks and and uh you know help the warfighter build their windows 95 computer okay so now talking about technical challenges i mean i can think about uh what what what we do in the labs for example when we process multi-domain data and it happens sometimes that you have a sample a
2021-01-31 19:37