Left Of Greg 090 Greg and Brian have really smart friends.

Hey, everyone, thanks for tuning in. I'm Brian, I'm the host of The Human Behavior Podcast. You're going to be watching the video version of our audio podcast. Please, guys, if you like the video, like it, subscribe to the channel. There's going to be more content on there if you're already a subscriber, and a better way for us to get you guys some more stuff. If you have any questions or comments, go ahead, leave them below. Check out our links down below to get a hold of us and to actually find out more places where you can get more information about this. Please like and subscribe, follow us on Facebook at HBPRNA. Remember, all these cases that we discuss, and all these discussions that we have, are through the lenses of what we call Human Behavior Pattern Recognition and Analysis. So please, like it, share it, tell your friends about it, and we hope you enjoy the show. Thanks. I'm good.

Alright, so you guys are good. Alright, so we'll go ahead and get started. Today we have a special guest on the show, our long-time friend and scientist, information theorist, scientist, aviator, all kinds of things: Mr. Mike Syracuse. So Mike, thank you so much for joining us today.

Gentlemen, thank you for having me. It's my privilege.

So we're going to kind of jump into a whole bunch of different topics, but I kind of wanted to give a quick little background on you. I know everyone just kind of listened to a little intro for those listening to the audio version. But one of the things to kind of introduce some of the topics we're going to talk about is specifically when we get into information theory and decision theory and all this stuff, you have a strong kind of military slash DoD (Department of Defense) background, both as first enlisted Navy, then you were a pilot in the Navy, and then you did a whole bunch of work as well with DARPA (Defense Advanced Research Projects Agency) and then what became, right after 9/11, which was TIA (Terrorism Information Awareness), which eventually they called Terrorism Information Awareness. And you were actually one of the people involved in the project of building what became basically the largest mass surveillance program in the history of the world at the time, I guess.

And so that's kind of where you come from on all this. So, as much as you can kind of give a little bit about maybe that background and what that is because for a lot of people, especially people listening who don't know about it, they would look into that and go, "Oh my God, what is this? This is some secret magical thing." And it's a lot more kind of dry, sciency, really hard, critical thinking than magic or anything else. I'll let you kind of maybe talk a little bit about that as a segue into what we're going to discuss.

Yeah, that's an interesting segue. I think when I look at this, I consider myself right now, and a lot of people, we're sitting on the shoulders of giants. And the explanation of what happened after 9/11 was very nuanced, but it has its history way back in the beginning of when we started to address information theory and information science and the heroes: George Boole in 1850, Claude Shannon in 1947 with information theory, Vannevar Bush.

And I think that gets to the primary point that I really want to make today. Vannevar Bush, in 1945, wrote an article that says, "As We May Think," which formed the foundation of really the modern analytic stage. And he was truly a second-generation information scientist. Boole, Turing to a degree, first generation. And then Bush and Shannon. And then really, after that, then started this kind of culmination after this horrific event of 9/11 and says, "Hey, how do we better detect these signals?" So much like in your guys' world, it became just at scale a signal-to-noise ratio problem. And the fundamental underlying truth, which kind of segues into what I've been kind of doing for the last three to four years, is that you kind of have to know what you're hunting for. And the bottom-line problem was is that certain communities didn't really understand how to hunt things, and certain people with military backgrounds knew how to hunt things like submarines and other things. And we just kind of parlayed that into a set of methods that would serve us better post-9/11.

So flash to today, my move in 2014, 2015 was to say, "Hey, if we can hunt submarines, and if we can do these types of huntings, we can hunt these fires too because they're going to get massive." And for the very simple reason that it's an easier problem to a certain degree, is that they're not trying to hide on you. So my background has been air traffic control, aviation, and I ended up getting into analytics and I just naturally feel comfortable there. But I guess what I don't leave that whole segment is that when we look at decision-making today, and the combination of decision-making, information theory, and learning theory, which have to come together to deliver these services, there's been a lateral expansion of the 2D web. This vertical expansion, extension now, of the 3D web is the same problem that we face in the military. It's the same problem that a lot of police forces face. The strategic and the operational and the tactical are now compressed.

Go ahead, real quick, Mike, no, because you brought up a bunch of great things that I want to hit real quick because you just talked about three topics: one, we talked about 9/11 and hunting terrorists, you as a pilot hunting submarines, and then you just mentioned hunting forest fires. So, I think that's a good analogy, big picture, of what you do, meaning what information theory is, and what you and all these brilliant minds had to do after 9/11 was go, "Okay, we've got all of this information, all of this stuff out there, so how do we categorize it? How do we connect it? How do we use it? How do we operationalize it? What is it?" So, these are all questions.

So just for everyone listening, folks like you and Greg and I, we look at all of these problems as that, at an information level. And something you brought up as a signal-to-noise ratio. So can you give maybe some sort of almost like a street definition of what you mean by a signal-to-noise ratio? Because that has a specific meaning to it, but it's also a great analogy for even just scrolling through your Facebook feed, right? So if you give like kind of a street definition of signal-to-noise ratio.

I'll give it my best shot. Let me just correct one thing that you said: some very brilliant minds, and me at the end of it. Well, I categorize you 20 years ago, me at the end of a firehose. But yeah, I guess my point before I go any further is that the stuff that was learned back then is so critical to today. To me, it's because I believe we're in the age of systems engineering. So basically, a signal-to-noise ratio is the classic needle in a haystack. And wherever we're working, or whatever we're doing, or whatever we're looking for, it's a signal in the digital age because, let's face it, everything's zeros and ones, or letters and numbers, and everything's transactional. So what you want to do is try to find that needle in that haystack as quickly and as efficiently and as authoritatively as you can.

And what we kind of came to understand is if you're going to find a needle in a haystack, get rid of most of the hay. And tune, like you guys do at the tactical level, tune your sensors. And in this world, everything's a sensor because again, we're at the operational layer here. And a lot of people are playing instruments, doing this type of sensor, or this type of surveillance. And somewhere like a composer of a musical score, it has to be orchestrated in real time because everyone has their own experience. So that signal-to-noise ratio is basically finding it. It comes down to clarifying the signal and then empowering the response of the people, whether it's a fire intelligence signal, or whether it's a national intelligence signal, or a business intelligence signal.

I've come to understand that underneath that is a common dynamics and a common method that can be used regardless of the domain. And, you know, we obviously right now focus on threat, but in keeping that into the level where it has to go, it's really about strengths, weaknesses, opportunities, and threats in concert with each other because all of those things kind of mesh together. Not only project management type of perspective, but in a real type of perspective. So I guess in that nutshell, that is just trying to find that thing that you're interested in, that bit of information, or more importantly, the aggregation of 15 different information services into one service that you can make sense of.

And Brian, if you think back to the time probably in 2005 or 2006, I'd spoken at Hampton Roads a number of times and I was meeting with some very smart people trying to get the Combat Hunter stood up. And the idea there was that I kept getting told from the different services and the different leaders that every Marine was a collector and every soldier was a sensor. And it just wasn't true. And so the idea is that I listened to this guy across the room briefing, and he gave basically a synopsis of getting rid of the hay and the signal-to-noise ratio. And I go, "Oh my God, this guy gets what we're trying to do. We're trying to establish context, and we're trying to put meaning to context." And he just does it in a different milieu. So if we can get—and by the way, my dear friend Mike Syracuse was the guy speaking.

And remember, Alan Turing was castrated, Mike. But the idea was that I was sitting there in the audience going, "This guy gets it." And I'm proud, Brian, to say that we've been friends with Mike ever since. And Brian, Mike, and I, folks that are listening and watching, we're all involved in the Future Immersive Training Environment Joint Capability Technology Demonstration, also known as the FIGHT JCTD. And on the shoulders of giants doesn't even... Folks, if you've looked for Schrödinger's cat, it's on your screen. There, under Mike.

Yeah, Mike's petting it.

Yeah, so it's just, it's just great to have you here. And Mike, what I'm interested in is your views on context, obviously, and getting the band back together.

Absolutely.

Yeah, so that's, I appreciate the kind of definition of signal-to-noise and what that means and looking for a sensor. And I think sometimes we forget that, and I'm trying to take it from just my perspective of my little street view that I have as a knuckle-dragging Marine, of constantly being bombarded by information, right? I know you kind of delete—you use the term data and information a little bit differently, and I think there is some confusion on what data is and what information is. But in general, we would say that everything we take in all day long is little bits of information. Everything from literally my body just sensing the ambient temperature of the room, to something I'm reading, to something I overheard someone say. And what a lot of times we don't have, and this is what kind of Greg and I provide really much at a very tactical and operational level, is that gating mechanism for one, a lexicon, how do we articulate this? How do we organize this? And then most importantly, so what? How do we operationalize that, right? And you do that at an even kind of, I would say, even a grander scale overall. So can you, if I can really...

Go ahead, I mean, it gets back to when I talk to coders, they go, "Well, you know, you don't code." And I say, "Yeah, but we have a coding practice. We're just at the operational layer. There's a bunch of people playing instruments, saying, 'In the COVID world now, okay, but are you all in tune and is it making a score that is going to provide actionable information, insight, hopefully after inquiry, to people that need it?'" And I think to get to your point, it becomes pretty clear to me that one of the big problems that we have, and that's why I think it's important to distinguish between data and information, is when we import data, or we import a source, as a human we do it quite naturally because we associate. Well, that's not what we do with machines. And that's why the signal-to-noise ratio is really important because it gives us a provable method that has worked in ASW (Anti-Submarine Warfare) and throughout all that type of stuff. And let's face it, there's a science that's required in information at the operational layer.

So that coding method, that methodology approach, is more in line with what kind of Jobs would do when he said, "I didn't play an instrument, I played the orchestra." Well, it's the same thing, but we're not Steve Jobs, we're just a bunch of guys and gals that understand the operational world and we want to make it more operationally efficient through what I'm now calling the information aggregation services. So it's not about—because we're stripping away, the bottom line is, you import data, if you strip away all the context when it gets to the stage, say in the outer loop, where it's observe and orient, let's just say it's import data. And then what's the importance of that? And if there's layers upon layers of statistical engineering and indexing done on top of it, well, at the data level, we're losing something. I think it's become pretty clear that we've stripped the context out of a lot of our sources and we, I believe, we need a method and a theory that can be provable to a science to say, "Hey, can we do this better?" Not only at the tactical level where you guys live, but you also live the tactical operational buffer too, but at the operational layer, but incorporating incumbent the strategic layer because to me again, that's the 3D internet now. And it's really personal.

Right. And so, I don't, Mike, this is really good stuff and you said something, you know, when he talks about we talk about gathering information, whether that's on a machine or a human, you said humans associate, machines don't. And that kind of leads to context, right? Meaning we, we as a human, I have the ability to take in a piece of information and apply some sort of context. It's not a zero or one for me, it's a little bit more descriptive like, well, you almost ingest it in context.

Right, right, right. You're recording it in context, which is brilliant, right? Because there's no stripping of the experience if our mind is clear enough to.

And right, and that's kind of the difference between, you said that's what a human can do, a machine can't really do that yet. So I know when we all get obsessed with the future of AI and all this and everything and I love always, you and I have had some great conversations over the years, because I always call you and go, "Hey, what about this?" And then you're the guy who goes, "Yeah, yeah, but not really." You know what I mean? You're the one who goes, "Yeah, that's not quite, but yeah, I know what Elon Musk is saying, but we're nowhere near that yet." But I think that's an important...

There's an AI hard problem, right? It's the Turing test and it's tough, it's hard. And it again, back to what we talked about earlier, yeah, we can focus on threats and that's important, but when we're building this machine that can be like a human, strengths, weaknesses, opportunity, threats, always in balance, always kind of responding to each other, but you got to start somewhere.

Well, that applies. And then yeah, you're right. And that is what, when you said, "Hey, a lot of this information we take in gets stripped of its context," because of that, when I put everything down to a zero or a one, or a very roadmap or when this happens, I do this. Well, that lacks context and the subtlety and the nuance of everything that we deal with, especially information, is it needs context. So I know we always talk about it at a different, you know, at a basic level, we've done a whole podcast on context just meaning how you observe something or where you observe, and the context is what gives that information relevance. So when we strip too much of that away, that information can become less relevant, right?

Well, I think possibly a better word is experience. Okay, because what we're trying to do, say in the, because when I looked at you guys' website, at the bottom right, and I saw about, we talked about the environment, and then when we talked about electronic communication, and then through that to human contact and then delivering these capabilities. And that's what the issue is. I don't want to get too weird out here, but at quantum physics level it says, "Hey, things can exist in two places at the same time, at the quantum level." Well, we go into the real world, say, "That's not true." Not really, but it really is. Information is that thing that can exist in two places at the same time, if you have real-time full awareness of something. It's almost like you're there. But with that comes the host of problems that we're talking about because that direct human contact is not there anymore. You're relying on an abstraction and an aggregation of those abstractions to inform you about what's important.

And it may be really important, especially because the internet's about one thing, right? It's about the depth of distance. Well, that distance may be four blocks if you're in a wildfire situation. You get the right information at the right time, it could be 40 seconds if you're trying to predict a grand mal seizure on someone's arm. It could be instantaneous real time if you're a Marine on the ground seeing something in real time and you've got a half a second. But again, underneath that is human analytics. And that point about association is that our machines, I mean, it's always been the hope. I mean, back to Vannevar Bush, Vannevar Bush wrote in 1945, "The day that we can," and I un-paraphrase, "and that we can master our indexing in an associative manner the way that the human mind does, then we'll be there." So it really gets to that point you made earlier about AI: are we really after artificial intelligence, or are we just still after augmenting human intelligence?

Well, that, you know, that's, I'm always like, "Hey, before we start coming up with new problems, why don't we solve the ones we have right now?"

Right now, the final first two, definitely you guys get in there and you say, "Here's the hard problem. Now, why does it matter? What's our time frame? And how are we going to fix it? Where are we going to do it? Who are we going to do it with? What's our intent for doing it?" And that same passion and approach needs to happen at the operational layer. We, when we know this in the military, we know how to do the operational layer of warts, why, you know, we have acts and everything else. But hence the civilian world and services that possibly will serve all of us back to Vannevar Bush, it's time to transition not only MRAPs (Mine-Resistant Ambush Protected vehicles) and everything else to police forces, but it's time to deliver tried and true methods of detection to the nation, to states, so we can start to get our arms around in the case of these massive firestorms. It's the same thing: we've got to take what we know, our art and science, and apply it. So I guess that's kind of what drives me as an entrepreneur and still drives me living in the state and in this country that we've got a lot of big hard problems, but they're not unsolvable.

We're going to have to do climate change unbelievably so. You know, if I could take a second, it may incorporate certain things that we may not be comfortable with in wildfires, such as, "Hey, if General Atomics can build us a..." Sorry to use a specific name, but if a big manufacturer can build us a drone to do certain things, then we can build another drone at 50 an hour that can carry wingtip foam and weaponized foam and water that can get these fires at half an acre. And of course, and it's all very similar, we have to prioritize and do everything else but context.

Greg, did you want to...? Yeah, so right now, just for everybody that's watching and listening, welcome to our world every day when we're dealing with a brain as big as Mike Syracuse. Yeah, so everybody, I want you to that's in the audience, I want you to imagine leaning back or laying flat and looking up at the sky at night on a beautiful starry night with no other light. What you're seeing is what Mike is talking about and he's also saying as nebulous as that is, it's solvable. And what I would like to do now is I would like to change the focus to a straw and be looking just at a straw, at one star, and give you an example of something like police work. Because we have somebody, the we has decided that the threat and at "bang" are the nexus of where the problem exists. And we're only going to look at the threat, then a term like de-escalation and psychological de-escalation loses all of its relevance and meaning.

The context of the argument is that we can keep conflating the actual problem. We haven't defined the problem fully and therefore what's happening is we're spread-loading our energy, whether that's, you know, mental or physical or money, fiduciary, monetary. And what's happening is we're not solving any problems. And so as an information scientist and a decision theorist, people wonder why we talk about decision theory, Brian. And the simple answer is this: if we empower those personnel that are closest, and I'm talking tactical, Mike, if we empower those personnel that are closest to understand what signals we're looking for in their environment and which of those signals likely lead to these outcomes, then we can use, in the police work context, a psychological de-escalation approach that'll work, that'll improve training, that'll impact the bottom line and address the situation that's going on in the United States right now. But that self-same assessment and data analytics would work for COVID, would work for future problems, would work for a zompoc (zombie apocalypse) or an invasion. So for anybody that's listening that's thinking, "Hey, I don't think..." I think at that high level now, what Mike is doing is Mike is looking at a map on the risk board and deciding where to put his troops. That's all he's doing. Mike, is that a fair assessment? I know I kind of brought it down, but is that a fair assessment of what you're saying?

Oh, that is a fair assessment. You know, because even when we look at any of the issues that we have, what is the hard problem and why are we trying to solve it? We say that we have this problem in the police force, but it's not simple as that. Every day there's tens of thousands—I grew up with the cops family, I'm a cops kid, and a lot of good people get good things done and there's some others that don't. And transparency. So what's key, I guess, is that everybody at the tactical level likes top cover. And that's really what we're talking about: what's that operational information that I can count on? Because right now I'm getting bombarded with stuff. You know, and you say it as an infantry guy, it's, "Yeah, it's nice to know that you've got that top cover." And I think it's the same thing in that, if, because let's face it, in human nature, if you're kind to somebody, I guess it was Madison used to say to be kind, "Everybody you meet, them plan to kill them." But it's tough to be kind because people perceive that as a weakness. And so again, it's that constant, we're asking a lot of people on the tactical level, of the tactical level, whether it's cops on the street because they're being...

You bring up a bunch of good points and Greg, I appreciate kind of narrowing that straw a little bit for us because, you know, that's what we mean. And Greg alluded to it when we say, "Hey, we got to stop thinking at 'bang'." And then everyone goes, "Well, yeah, no, now we're here." It's like, "No, no, I get that you're a few minutes before that, but you're still looking at the problem at when it presents itself." As we're meaning, we're not clearly defining the problem sometimes.

Can I add one thing? And just because I'm a little bit about it because it's related to fires and everything else, yeah, I'm, and I did jury duty too and it became an issue: the weaponization, if you will, if you can allow me that statement, of 911 calls by the person making the call is a dangerous situation from an information theorist scientist perspective. Because you have all the problem, the context is stripped away or is defined by the person calling because the person who may be on the other end of that may be the innocent one and we, and then it gets abstracted away to the police officer. Now, I know there's a lot of, but that, that, that cycle right there. I think it was Patriot, it's a great show and it said, "The straightest way to get from point A to point B is a circle." And it's so true in what we're talking about in analytics, and, you know, yeah, it's easy to be black and white and say, "I'm either going to do this or that." And for a lot of years at the infantry level that was okay. Now, that was okay because we didn't have the top player information, a lot like a sailor going out to sea, and having his own ship. Yeah, that's what you had, you know, you got some orders, you got commander's intent, go fill it out.

But Mike, you, and I apologize, Brian, but Mike, you've just hit on something that was like a core argument and FIGHT JCTD. And I go back to the, "Every Marine a collector and every soldier a sensor." We cannot say that this, this young person that's about to be deployed to combat or this officer that's newly minted from the academy can go out there and have all of the metadata and all of the analytical background and all of the tacit and experiential knowledge that they need to operate on the road. That's, it's absolutely impossible. And then going further to say that a computer will be able to do that. No, no, it's not going to. We have to reorient where that data is fed, where it's taken and what the intent of the information is. If you don't do that, if we continue to look backwards, if we continue to use the lenses of our ancestors, we're going to miss the point completely. And the idea is that we have to adapt and overcome using all of their genius injects that we've forgotten that we've got to blow the dust off of all the stuff that's in there and actually utilize it to to meet our goal. And we're not doing, we're not there yet.

Yeah, I, I want to roll back on something because I think it's important when we talk about, you know, at all layers, right? Like there are decision-makers at all layers, you know. So they're all kind of tactical in their own regard. They just have different time and space issues to concern themselves with because there's a real human at the strategic level sitting there making a decision. They may be on a five-year timeline, but they fully understand that there's a five-minute timeline too that makes that, right? So what I want to kind of get is that it, I'll throw, you know, that fractal nature, that self-reporting nature that humans are humans making decisions.

So as part of that, as we talked about the signal-to-noise, it's really, really important to recognize that it isn't an either/or. When you're hunting a submarine, it's a, when you get in the lips that's in the middle of the ocean, it's a possible submarine. You know, with some certainty that 15 hours ago within these 100 so miles there's an 82% probability there was a sub. And then you take an asset like the P3, the Mighty P3 Orion, you fly it out there, you drop some buoys in the water, and you try to take that, that possible signal and turn it into a probable signal because you're building this heuristic. You know, it's not either/or. Just like when you're an infantry guy, it's possible, it's probable, and then at some point it becomes certain. So I think it's really, really important at the tactical, the operational, strategic level to understand, as the humans, we kind of go, in the machines, we have to go through that process: possible, probable, certain. Now, when you're hunting things that don't want to be found, there really is no certain. I mean, it's kind of like the movie where they were hunting the, with the pods in the water, you know, there is no certain. They haven't done anything yet, that was the problem we had prior to 9/11. But we do have a pattern in the history and things that we can fall back on, what you guys do naturally at the tactical level, to inviting people that we have history and we understand patterns and really that's the key of it. And, you know, if you want to, if you want to make an inference or find a causality in a pattern, you have to make sure that you go through a rigor of possible, probable, certain, as best as I can be, that I believe this.

Well, and the great thing, we only have to achieve the level of likely reasonable suspicion, probable cause, only assumes likelihood, Brian, and artifacts and evidence to support a reasonable conclusion. That's the beauty of this.

Well, no, and that, that, that also is what makes these hard problems. And you both kind of alluded to it before, and Mike even said it earlier, like we're putting, specifically for like a law enforcement example, and what's going on in our country, I don't think sometimes people realize, not even police officers themselves, not even administration, not even anyone protesting or either for them or against them or wherever you stand. I don't think they've clearly defined the problem where they realize how complicated these situations can be. Now I say complicated because we have answers, we have solutions, it's not we, it's solvable, it's not some weird complex nebulous thing. But, but it is a lot and, and so like, you, you brought it up because, you know, that went into, you know, the the, the Strategic Corporal concept and the four block war and all that stuff is that a 19-year-old kid, because that's what you are at 19. I don't care if you're the military or not, right? 18, 19-year-old kid can make a decision at a very tactical level that will have massive, massive strategic influence. That's no different than a police officer at a tactical level makes a decision that has overwhelming strategic consequences. So you and then it's no different than from the other end of it, from a, from a threat perspective, that someone might weaponize a COVID, one person, and introduce it in such a way or start a fire like that in Big Sur.

You know, it's the butterfly effect, right? I mean, it's the, but it's not, I guess the classic butterfly, it's not that the butterfly is going to create a tornado, the butterfly is going to create something directly related to it. The question becomes, can we, in my mind, the question becomes, can we provide the same level of time and space command and control for these other problems much like we do in the military without being onerous and overbearing, but letting that 19-year-old kid again, or that 25-year-old hop on the street, or the firefighter, whatever the case may be, that they have top cover and that that information that they're getting is as pure as it can be? And I know "pure" is a tough word, but that's why the signal-to-noise ratio thing's so important, trying to put math against this problem of context.

And that's what you mean. Tell me, correct me everyone, but that's what you mean by a signal, right? A signal can be a number of things, but your signal, the information you get from a signal is only as good as your signal, right? So now you get into false positives, false negatives, meaning...

Absolutely. If I, if I think I have a good signal, like you always give, which I love, is, because it's a very simple example of artificial intelligence that everyone understands, is the smoke detector in your house. It's got a very, very simple job: it detects smoke. When it detects smoke, it beeps really loudly. When its battery gets low, it is that annoying beep where you can't tell which one it is in the house, right? So we've got it standalone, it doesn't really even need, it could have a battery or backup power.

Yep. And, and so, so that's a classic sensor to shooter.

That's what made the P3 such a mighty airplane. Another greater, if you can sense something both, yes. The informative strategic kind of run the operational and then be there at the tactical. And so, yeah, and I just want to use that as a very simple sensor that everyone gets. Okay, hey, that's a very simple, it's very effective, right?

And, and, you know, obviously you're concerned about a false negative with that one, right? Where it, it doesn't go off. And false positive...

It's always about false positive, false state. You have a false positive, somebody jumps out of bed and they break their leg. It's like, "Yes."

Exactly. So you, so you're guarding against that. I just think that's a, that's a good example and we often don't, you know, and what you talk about, you give the analogy because it's perfect from your anti-submarine warfare background and what you talk about flying a plane and then you're going to drop buoys, which is a sensor. So you're dropping this instead of smoke detector, the sub detector, right? In the water. But, but even if we have the best sensor, and hey, this is really great and it's very simple and it works very well and it's extremely accurate, it has a fault low level of false positive and false negatives. Well, I could have the greatest sensor in the world, but if I don't put it in the right spot, if I don't use it correctly, if I don't implement it right, well then, then I get nothing.

And you also, but again, we go back to the composer and the conductor of an orchestra. You've got multiple sensors too, right? Like, and this is very relevant if anybody's interested in the autonomous vehicle world, or even the UAS (Unmanned Aircraft Systems), UAM (Urban Air Mobility), AAM (Advanced Air Mobility) world, and even at the tactical level, right? Like it's important to know that if you're going to be using these systems, what the logic is that's providing the information that you're going to be taking actions on. So it is relevant to everybody that's going to be playing with these toys and involve themselves in this, in this kind of world. But again, that, that, that concept of aggregating sensors is everything. Because when you're hunting a sub, it's not just a buoy, it's a camera, it's a set of eyes, right? It's Doppler, Doppler never lies.

And the only thing that I know of, and it's, because every time I talk about the AI at the executive level, I say there's a hurt factor involved, then it requires full transparency. And a hurt factor can be an algorithm or set of whatever they call them, do what they do to decide if someone has bail or not. That needs to be fully transparent. So it's not the sausage factory, you can just do some statistical engineering and throw something out the other end. No, people need to see that because people's livelihood and their lives are on the line. And that becomes obviously much more important. We start talking about weapons release authority, which in and of itself is not complicated, it's complex because one simple loss of contextual situational awareness and you could make a decision, especially if you're hunting something in time. Because the reality is these signals, all of them, if you close your eyes and close your ears, you may think they're not moving, but they are. I mean, that's the, you know what I mean, that's the big challenge whether it's at the tactical level, the operational level, whatever your time frame is. You know, you just be an old saying in ASW, "Just because you don't hear them doesn't mean they're not doing, I mean, they're not there." Got it.

Yeah. Well, let's tie that into the gift of time and distance. And let's go back to our straw for just a second, Brian. So you guys both brought up some incredible points about the smoke detector. So smoke detectors work. Some are hardwired, some should be battery operated. You should change them every year on a specific date. Now, we augment—huge important word—we augment that by having the local fire department send a person through and that person comes through and conducts an audit, a vulnerability assessment of our facility and says, "Don't block that door. Open that, that's a fire door, that'll create a vent." We go downstairs and say, "Look, all these boxes, all these old newspapers, we have to move them away from the the water heater because that's a fire source." What we do is we identify the problem and then we try to move as far left of the problem as possible and augment the technology with virtual reality, with augmented reality, with with human sensors that receive training. So, so people sometimes ask, "Why do we push training so much?" Because without training, you won't know if you have a fire extinguisher. A fire extinguisher is at bang. If you have a smoke detector, a smoke detector is at bang. Those things are wonderful, but it's like the police officer relying on the weapon or the less than lethal force. We've already crossed the threshold of use. Mike, your asymmetric submarine warfare and your anti-submarine warfare was designed to give an advantage, a technical and tactical advantage of time and distance, so you wouldn't get smoke-checked, but so they couldn't launch missiles, let's say, against the coast. All of those things presuppose that here's bang and there's all of this time and distance over here to the left of bang where we could make a big difference. Isn't that a fair assessment?

Oh, it's totally fair. It really gets to like my, my bigger passion, which is, you know, I, I, I like to look past theology, philosophy and science and say, "Where do they all agree?" You know, and my only, the only conclusion that I've come to is the best place to find that answer is in music. You know, it's like Zappa said, "All that other stuff, music's the best because there's something that happens at a vibrational level." But what you said was, it was spot on. Whether that left of the bang is a business intelligence signal that's going to hurt your company, or whether it's a fire signal that's going to hurt your community, or something that's going to hurt your body. Again, that, that constant balancing of strengths, weaknesses, opportunities and threats at all three layers, teams working there to figure out the physical things and then understanding that the physical things are this. Then at the operational layer, we augment those physical deficiencies with actionable information based on the patterns and the things that we know are going to happen. And that's why I go back to the, going past theology, philosophy because it's like a story. What we're creating here, whether you're creating a three-minute, 30-second song, or a movie, or a digital service, we're creating an experience. And that word right there is the most important word that I can think of that ties together, you know, when you're looking at that straw and saying, "Okay, but I want to look at the whole sky, what are the things I should think about?" Experience. And because that's what we tried to recreate for people, say, "Hey, you're here and the internet's about the death of distance and I want to make you informed about something here because I need you to do something here that'll preclude something from affecting you over here." Well, there needs to be a composing and an orchestration, or composing and conducting and an orchestration beforehand of those events in such a way that we understand them before we code to them.

Brian always talks about that, Mike. Brian, Ryan always had that moment where he looks and he, he grabs his chin, he goes, "Hmm, I see what you've done here." And we all laugh at the Mike Myers because I think, Brian, I think Mike's analogy of music is spot on. Because the variety of music, have you ever listened like you're listening to Tchaikovsky for example and somebody's going, "Chai-hoo!" And then all of a sudden you hear Levon Helm and The Band and, and then you're changing the channel. See, I can't do that. I can't remember the the thing but it's like Paragon, it starts with a P where all of the music in a theme come up. It's a program, Pandora.

Pandora, I'm sorry, because they don't listen to social media. Right, you want different varieties. Yeah, I heard about this company called Apple.

No, no, but I hear they're... you get my meaning that certain music appeals to certain parts of our brain and and that stimulation that external stimulus that that...

But I do understand is that the left side of the brain and the right side of the brain and then music kind of miss in the middle and I don't know the official names and all that, but you know, and I want to roll back on something because I think it's important when we when we listen to music, we bring it's an experience. Whether you're sitting on a couch or whether you're, you know, you're sitting in a big sur, it's an experience that you've lived through for those three and a half minutes if you're engaged. But rolling back to the FIGHT JCTD and what's important, say, to street cops, whether let's just say like there's some work that we want to do with the Naval Postgraduate School focused on on on how do we make ethical use of unmanned systems for our police forces? You know, we've got a hard problem. And let's just say the whimsical thing that a lot of people believe the hard problem is, we don't want warriors, we want guardians. Okay, whether we buy into that or not. But there is a movement towards de-escalation and all the stuff that you do. I'm sorry, I just lost my track. You got Naval Postgraduate School...

You said something...

No, but it goes back to the FIGHT, that's right. What we're trying to do, what we're trying to create is an experience that gives them an inoculation. One of the things that we recognized early on in the FIGHT that we were doing an inoculation, all training is kind of an inoculation. I find still surprising, even after five years and then, you know, being just distanced from it, there's still a problem in the DoD and the fact that, "Hey, there's only one reason we educate and train, that's because somebody's putting a uniform on, putting their f***ing on the line. Let's never lose focus of that." And then my focus now, and I know your guys', is to say, "Okay, how most efficiently and again, like in this global consciousness that we talk about the internet being one synapse for this group of people, how do we do that as authoritatively as possible, right? And as authentically as possible to give them the experience that they need to make the proper decision."

No, no, and I, I, Brian, here's another lexicon argument. The authenticated, authentication and the level of experience and having an aggregate but a pure signal, that's all important. Because what's happening is there's some people that are listening and watching and going, "Oh, I think he means that the user experience should be good." That's not what he's saying at all. What he's saying is that we've got to try to get the most pure information possible at the beginning to have the outcome be something that's very close to what we're trying to accomplish. And defining that that problem is as important as defining that goal. So that's a great...

Absolutely. It, like when you're when you're doing a, an ATO (Air Tasking Order), an air tasking order, right? There's a bunch of people sitting in a room making a whole bunch of decisions. But every one of them that, you know, has the value of knowing what they're doing there, understands that there's a kid turning a wrench. And yeah, you have to concern yourself with that. And that gets into the whole thing about supply chain management post-COVID, but it's all pretty much the same physics that work. I guess that's my main point of physics at work.

So, so go back to your orchestration. And I think of The Band as well. Brian, I think one of the points that that you and I see only because we might, we get flooded sometimes when people send us articles about what's happening right now in in the state of police work. And I would turn that around and say, "Listen, this has always been happening. You now just opened your emperor." Yeah, and, and you want to make an uninformed decision based on on whatever criterion that you've decided. And you're not looking at the context and you're not weighing each individual situation. Isn't that the same problem you're talking about, Mike?

Yeah, I mean, it gets to, I mean, I don't like to say it's a pet peeve, but it's a general problem that seems readily apparent, is that when we as humans tend to learn something new, it's brand new to us. It's, yeah, and it's determined. It's working the problem. It's the Charles Manson quote. "Yeah, hey, if I haven't seen it before, it's new to me for me." So, and then they get to that with, somebody did some studying about flat Earthers, and if I offend anybody, well, I don't care. So get off.

No, we, we rip on them all the time on here. We really do.

But what they tend to find out with with flat Earthers, and it kind of expresses itself in other ways, that some people if they don't experience it, they don't believe it's true. So that, that's another, I can sit there and say, "Hey, the Earth is round," because I've flown over the North Pole at 200 feet. It's round because every direction is south. But unless they've done that, they're not going to believe it. And it's the same thing here. So I guess what I would advocate back to our, one of the themes that we talked about earlier, sitting on the shoulder of giants. We, I'm at the point right now where I want to to build the bands together that we can stand on those shoulders. And what I've kind of considered is for the last five years and definitely the last nine months, I've had to crawl up the back, you know, after working with these guys, crawl up their back to get just the ability to sit on their shoulders. And that's how hard it's going to be to stand on those shoulders because this stuff is hard.

Well, that then, that that's kind of goes back to what I was saying before. That is a big problem. Greg brought up, you just brought up a great example of flat Earthers, you know, unless I've experienced it, it isn't real or all this stuff. And this is that's with any what has become known as conspiracy theory or something that becomes this emotional reaction. Everyone's all of a sudden, "Did you know this?" And I'm going like, because that's what's going on. There's another one with this whole everyone's a pedophile. I don't know. Apparently there's this huge... And I'm like, are you just becoming aware of crimes that are hap that have been happening in society for a very, very, very, very long time? Because that's what it seems like to me.

A lot of them are societal crimes too. Systematic, right?

And, and so trying to convince millions of, oh, I won't go there. But millions of people that have been systematically abused to trust systems nowadays is really hard.

Yeah, oh no, and they're afraid and they're and they're and they're yelling. You and you brought up kind of systems management and like there's not a lot of trust in our systems because of that. And now all of a sudden there's this, and this also gets into information theory of...

Obamacare so much.

Totally. It's totally information learning and decision theory. Every time. It's that analogy, "We don't do one billion transactions, we do a billion transactions one at a time."

Right. And that's a good way to look at it because I always, you know, what we always try to do or always try to tell listeners or people I know is like, look, one, if you're having an emotional reaction about this, I, I, it's odd to me when it doesn't pertain to you or you have no actual connection to it, right? So like, you know, Greg will send me something like about another thing that happens in law enforcement, this, and I get it because, you know what, Greg worked really hard for a really long time as a police officer and seeing the change and seeing how things work. So, so I understand the emotional reaction, you know, when I see something happening with the military, this, I might get that, but but I have a connection there. So there's, there's that issue right there that I'm a little confused. But we don't have to get into that. But, but the big thing is this kind of that information overload of people don't realize how difficult and nuanced a lot of these problems are. And we oversimplify the problem, which means we oversimplify the solution, which means we oversimplify how we legislate that, how we decide, how we vote. And then, and so what am I supposed to do with all this information that I get? How am I supposed to know what is what and what's a good signal? Like, like the pure signal you talk about, that's what I want to get my information from. But how do I know if I have a pure signal? And these are these are difficult problems. I'm not expecting an answer out of either of you. I'm, I'm just kind of more on it.

Not so much an answer, you know, it's like, okay, number one, are we asking the right questions, right? And I think we are, you know, as a group, are we asking the right questions? And then number two, recognize, you know, what is, in the, in the theory, in the field of information theory, what is the hard problem? Well, the hard problem is, right? It's like a general intelligence for a machine. Well, okay, Alan Turing, we know, bad apple and everything else. It's a shame that a man that literally saved the West from Germany with his work and information theory and developing Christopher. And it was what I thought most interesting about that, if you take Christopher the first computer and now you take this today's chip, some silicon, the exact same structure. They're just now the size of, you know, half a Connecticut if you exposed about to Christopher's machine. But I guess to my point is that, okay, what are the, what's the general hard problem? And again, I fall back on, we've got as a company and as an entrepreneur, as, you know, a theorist, we've got to clarify these signals. And those signals are much like in program management for modeling and simulation, we can do anything, but you can't do everything. If you want to solve the hard problem of detecting grand mal seizures on a human being, that's all you can do. Now, you may be able to aggregate those things later on, but but it gets back to what we talked about earlier. It's a whole different ingestion, you know.

When we talk about, if I've got five or if I've got, let's just say I've got a hundred books that I'm really interested in that I think are hold the essence of my future and I'm 20 years old and I want to sit 12 hours with that and do an associative exploration of that, well, yeah, you can do that with, you know, elastic search engine and some micro services. Or you can change your ingestion method, your import capability. And that's kind of where I'm at, you know, because it isn't so much information overload, I think that's happening with people, it's information confusion, it's conflicting, you know. The first word is always wrong, but is it? Sometimes the greatest insights on the first raw thing that you get. It's not everything, right? Put in the context of someone else's experience and that's semantic and that ontology starts to really matter when we start to talk about listening, for instance, something a lot of people are familiar with is the OODA loop: you observe, you orient, you decide and act. And I don't really, it's not that I don't support that as a Marine on the ground or in certain contexts as an aviator, but it doesn't work at the operational level because we sense things, whether that sense is from a human or whether it's from a database or whether it's from an unintended ground sensor. Don't really care because it's specific intent. And then we orient, or to put it the other way, we import stuff and then we describe, we try to understand its importance. And that to me is universal, whether whether you're a cop on the street, whether you're a Marine, whether you're a boardroom guy. There, there's a universal physics, I believe, that may be a counter...

It's a great unifying theory, Brian, to tie in what you said. Unified theory, yes, it really is. To tie in what you said, Brian, and a great point and dovetail on yours and tie in what Mike's saying for our viewers and listeners, when you take a look back when I was growing up, and I'm older than both you guys, back when I was growing up, on Thursday night there was a TV show and at the beginning you had a choice, you didn't know what it was going to be: was it going to be McMillan & Wife or Cannon or McCloud? And I showed a guy with a flashlight going through and there was a song playing in the background. It was a misty day and he was walking around the flashlight. At the end you figured out what you were going to get. That's what search engines are like. There's so much information and there's not detailed information being onboarded. So you have this vast array of signals, we're right back to it, and most of those signals end up being noise because they're not the specific thing you're looking for. It's not intuitive, so you have to create that that ability to be intuitive. Isn't that right?

Yeah, I mean, to me, when I, when I think about it, I call it analytic engines now, just to have a simple building block, right? So let's just say that search, which is now past 22 years old, operated on metadata and it's current construction and it's got problems, right? I mean, search engine optimization corruption, all of it. You know, but it's still amazingly a great capability. Second-generation analytic engines is what I'm focused on right now. And what I'm looking at is the art and the science of an inquiry, assuming that a person has a knowledge and they're waiting on some type of trigger or an event or something to actualize that knowledge into what they're getting paid to do, which is make inquiries about things that are being made so they could gather an insight to make a decision. So that's what I believe are second-generation. I don't want to call it a search engine because the human mind doesn't search for memories, it selects memories. You know, like, and others have said, you know, we have to change our import mechanisms in such a way that we can maintain the importance of that import without stripping it away and then reestablishing it. So you can only go so far, I guess, is my major point. One of our networking air traffic control system, by the way, one of our good friends, he's only think that resiliency so far.

So Mike, I want to throw this in here just because it actually, you were just talking about it, so it's a perfect spot, is that one of our good friends of the program, Ryan Shea, who's brilliant, brilliant guy, who's been in the intelligence community for a long time and he's just really, really smart guy. You know, he kind of just wanted almost a drag, he kind of brought this up and said, "You know, can AI, an AI system, be trained to be more analytical by having skilled trained analysts feed data in the machine learning process? You know, find the threat tied to the classic 'so what'?" And my immediate answer, he actually kind of said it, is in the, "Can an AI system be trained?" Because that's what we call it, rather than machine learning, you know, we call it like, "Hey, it's got to be machine training." And you have to kind of be manage your expectations on what you want it to do, and the more focused it is, the easier it is to accomplish. Would that kind of make sense? Or, or, or...

Yeah, if I may want to flash back a little bit here. Let's go post-9/11 or pre-9/11. The big problem was the FBI was in the business of proving things happen and the CIA was in the business of predicting things would happen. And no one was in the business of hunting, which incorporates that actual on the ground and prediction. So it's all of it. So I think that same approach to analytics matters. It can't be just broken up into segments. Same with the fire world. Yeah, you want to educate, yeah, you want to detect, and you want to put them out, but those aren't separate. They shouldn't be driven by the funding sources, and in certain ways they become in competition with each other. So again, back to aggregating information in a way that matters, it takes that discipline at that level. So when I look at the intelligence world for example, what do we need machines to do better? You know, if you look at it from just a classic sense, we collect information that we assess it and then we respond or we make a decision about it. Humans are great at analysis. So if we have spent less time collecting, which we spend way too much time even today doing, and we'd spend way too much time reporting, and if that's a bell curve and a graph, you look at it, you go, "We want to flip it upside down. We want to do have machines right now." So you train them to, let's train them to be an inductive sensor for us, to feed our observations in such a way that we can give them some autonomy to do that. But the analysis is what humans do.

DARPA did this work in 2006, 2007, 2008 where they said, "Hey, we want to take these DIMEs (Diplomacy, Information, Military, Economic) - we want to take diplomacy, information, or military and economic models and put them all together." And after a couple years they said, "Well, the only thing that can do that is is the inches that live between the ears in the human mind." Well, okay, so let's always remember that human in the loop is going to be there for the foreseeable future as part of that education, training, and subsequent learning for machines. And I take a little bit of issue at times with machine learning before there's training because, yeah, kind of whimsically saying, "Well, it's going to train itself." It's like, no, it's not. No, you don't want anyone.

But it will. But not the way that may be favorable that you can track or be transparent with. Well, I know, I, I do simple terms, it's like I use the, I always use the gym example, like what am I going to do better, where am I going to get a better training program and workout, is if I just go come up with something on my own that I think I need, or I go to an expert and say, "Hey, you tell me what I'm supposed to do." And that, that's the concept there. I don't...

Go ahead, Greg. No, no, you're right on, Brian. And what you just did is made my point for me. Listen, Mike, I take umbrage to when I hear people say, "human in the loop," because I changed everything when we were designing supervised autonomy to be "computer in the loop," because I want the computer to give me the cues and and set it out for me, "Well, left of 'bang,' that this is what's likely." And using your analogy for fire, using your analogy for health systems, using that same analogy for COVID, it all works. You think of the the Will Smith movie, I, Robot, and I think the actor is James Cromwell, which is talking about asking the right questions. Listen, we could not get the designers of huge programs to understand that a person laying face down in the prone is more significant than a person standing and walking around or reading the newspaper. And then if we took just that simple face down prone position and we put that person in a geographically significant area and then added orientation towards a potential target, you've got the mother lode of information of that the supervised autonomy the leader needs to know. But we're not at that stage yet because everybody's still fighting with bandwidth and frequency and and they haven't solved the base problem that certain human factors lead one to a reasonable conclusion that this event is about to occur.

I mean, Greg, that was what you just did though, is, is you, that's a perfect example of providing context, right? So that, you know what I mean, like, okay, a person laying down, you're a bot, you're a drone, whatever you want to call it, some autonomous asset or or unmanned vehicle, and you're filming something or you're picking up on something, you can see into an area. A guy laying on the ground means nothing. A guy laying on the ground, in negative space where he can't be seen from anyone, oriented toward a significant intersection where there's a, you know, US convoy coming through. Well, all of that context...

Exactly.

That's what, what makes that observation significant, right? And that's the point of of we're not there yet. We humans can barely do that looking at the screen, let alone have have an action.

That's what the operational layer has always been about, right? Yes, you know, because I look at it that, you know, its strategy is about the what and the why, right? It is that, it's the defining the semantic, the ontology. And as we look at machines to do these things, it becomes really important to have this type of, you know, discussion, just simply. But, you know, it's the what and the why, the operational is about the when and the how, and then the tactical, or the functional in this case for technology development, but the tactical in real world is the where and the who. You know, and you look at major efforts all over the government, you know, for the FAA (Federal Aviation Administration) and NASA (National Aeronautics and Space Administration) looking at, you know, what, what does the unmanned traffic management system look at? It's like, "Well, okay, those are all great questions and we've got plenty of great examples of what to build on if we're willing to be humble enough to recognize that there's three things, in my opinion, that always, always, always matter." So when I look at designing information aggregation service to clarify the signal and power the response, I mean, it universally. But, but the things that matter are always upfront is safety, security, and access, always there. There's no getting around, you know, if you take the drone world, you say, "We're going to do, we're going to do all this unmanned traffic management over the internet." Like, "Okay, access, how are they going to do that? Is everybody have to have a cell phone on the drone?" You know, like, and then you've got all these seams and, you know, and then because when you start talking about technical terms like remote ID and all these things, these same underlying principles always, always, always show up. And it fundamentally comes down to a very simple thing, what you talked about earlier and that's why the signal-to-noise ratio thing is so important: false positives, false negatives. Man, it's the only thing that matters. If like, take your your prone, you know, discussion and take it to another level, right? You're, you're a first responder and and you have an ability through a house as a smart house and it's able to give you because it has a smart drone in it to tell you that a person that should be prone in this area and they're, you know, 82 years old is now laying on the ground and the position that we see them in says that they're in distress. And oh, by the way, they're wearing an Apple Watch. But if you take that same information and you don't have the context, you've got a bunch of signals, right?

And then you end up with, and then if you put that with bad policy, you end up with things like Breonna Taylor and and other things because there's aggregation of information that's just because we and it gets back to what you guys have talked about plenty of times. People for the most part, we see what we want to see. And that was one of the big problems and that's what you can't have like an ASW or time critical targets or, you know, anti-terror. You can't like, I was an accident investigator in the number one rule aviation accident and the number one rule you learn is if you step into an accident scene, don't have any assumptions. Observe what the world is telling you, have a methodology, have a rigor when you're done in that mode, but you've got to be able to see what the information at hand is telling you without that bias. And I guess what I would leave that with is there's a very fine line between confirmation bias and what are called selected priors.

Yeah. Now I want my my neuro vascular surgeon to have a wealth of selected priors, right? What I don't, what I'm doing is operating on a 52-year-old man from last week when he's operating on me.

Yeah, well that's a fine line, right? It is. It, it is. And it goes Greg and I just did a conversation about this because that goes on both ends of the experience spectrum, right? So you just talked about a surgeon and and yeah, can they have so much experience that instead of using those selective priors like you said from gathering from what they know, they're utilizing confirmation bias and make a mistake? Well, yeah, absolutely. Can someone with very little experience fall into that same thing because they have no selective priors in that particular area and now they they have this confirmation bias? Yeah, I mean, so, so that's the issue.

We're going to get a bigger problem earlier. Yeah, I'll be a little because, you know, it's my group. When you don't have a really sturdy set of selective priors and you want to answer big problems, some people tend to become insufferably arrogant along with their confirmation. But I just keep building that narrative. Our machines to do that is my point, right? Do not.

Yeah. So one of the simple things about the algorithm for for HBPRNA, for Human Behavior Pattern Recognition Analysis, is it's not an oversimplification. Baseline plus anomaly equals decision. It's not. It's the most simple code that you can have. The product of training in my estimation, Mike, is that we deliver decision theory, among others facets and skills and and knowledge that we transmit. But it's up to the agency that hires us, the group, the company, the HR (Human Resources) personnel, the factory, to become the decision strategist. They have to lay out, "These are three or less decisions that we'll ultimately have to make." And then the implementation is through their academies and their internal training and their T3 (Train the Trainer) process. And we've lost that ability. We no longer have the skills passed on from a journeyman. We no longer have those skills for the experienced vet passing on. We codified it and said for cops, for example, we have the Field Training Officer program that makes sure that all these checks are in the boxes. And that'll never go away. But that doesn't build critical thinking skills. That doesn't build the experiential level or the person's ability to to create an explanatory storyline. We're not there.

Well, right. I mean, my background and I know, you know, Brian's in the, definitely yours, Greg, is coaching is everything, right? If we want to, if we want to, if we want to transform in military training, then set the conditions and enable the coaches out there and everything else to rest for a year, go out there and see. Because like, the reality is like the air traffic control system, most people don't recognize that 90% of the time when people are talking on the radios, it's a trainee because they have to train in real time. You know, so that, that reality, that what you're doing in real time, but you're still training, same with flight school, right? Like people say, "Oh, you know, training's training," but it's like, no, you're still training an airplane, but there's a big thing there, there's an impact. If you mess it up, you and your instructor, you're going to be a smoking hole. And and getting back to what we do in the FIGHT is that what's what you try to do with that inoculation. And so I guess if I would tie it together, what that pattern is, the pattern recognition, getting left of the, of the bang as you talk about it, Greg, it's just taking experience and taking rigor and balance and saying, "This is the narrative that we're trying to master, we're trying to achieve."

Exactly.

You know, so much like a story. You know, when you look at a great movie, there's, there's like three sine waves, you know, and, you know, it's the character, it's the story of the plotline and they they kind of intersect at plot points and at the end they're together, right? And that experience, and that's the same thing we want with our technical services. You know, if we provide a technical information aggregation service to a street cop, when he walks away from that experience, not only is he empowered now to share that with others, but he feels empowered to take the next step to make that information service maybe closer to autonomy if that's the case, but make it serve us better. And that, that After Action Review, which is so inherent, you know, the machines have to be held at the same level. So something that we want them to do everything, we have to be very specific about what we want these technologies to do for us, never losing, never losing focus on what the problem is and why it matters.

So, so that kind of brings up or, you know, leads into you, you talked about it a number of times, but what is that information aggregation service? What do you mean by that? What does that look like? Because, you know, there's different types of examples of an information aggregation service. So like a Google search is one, right? I can get on Google, I look for cute videos of puppies or something like that and it'll give me something there, right? Whatever it is I'm, I'm searching.

Yeah, that's what you're searching for. It's well, search engine, if I want to die, this, if anyone wants really, really good information, go to Google Analytics and find out what people are Googling because it's about, I think the most honest a human will ever be is, is what they're Googling because you'll, you'll say anything on there that you won't say to another human being. But any, but that, but what do you mean by...

I would rather, to my queen. And I guess that's what, you know, 20 years ago everything was metadata. Yeah, the focus then was to how to automatically create metadata because metadata is still, is predominantly the primary map that creates the map and the signatures of what we navigate. What I'm advocating is for very specific use cases and hopefully generally overall, is that we need different navigation maps. If I have a machine that can read 60 or 100, back to what we were talking about earlier, if I'm a 19-year-old kid and I'm trying to define my life and I say, and somebody says, "Here's a hundred books," and I've got a machine that can read them, that can listen to the a thousand podcasts or a million podcasts, and that you could sit down at a simple query, so let's roll back to, back to the simple fire detector. I'm not interested in saying, "Oh, I want to build, you know, this service that can do all 100 things better than a fire detector." I just want it to do one thing better, you know, if it can alert and and it could be, it can be, it triggers, it says, "I got a problem," either going towards the house, like, "Hey, there's a fire," or, "There's this coming," or, "This coming," or, "Out of the house," you know, with everything that happens with that. But that, but that trigger mechanism.

So what I believe is most important is going back to context. How do we import information sources as clearly and as cleanly as we can while maintaining their context in a way that will provide people actionable insight? So when I think of a simple, let's just call it a selection engine, I call it an inquiry engine. Now you sit down, just like you would at Google, and you type, "You know, I'm from Buffalo, New York and I was born in 1960 and I love red roses." And now instead of getting a list of things, you're literally getting a representation of that, that corpus of information that you're interested in because your machines now are literally working at the associative level. They're taking subject-verb-object triplets and they're creating them as, and everything is now treated as this discrete entity. So much like the human mind. Now I can put in these stupid terms together, but I'm not going to get relevant things. And it could be Alice in Wonderland, or something completely. And then you start discovering that. Now that's not for everybody, right? But I think it naturally gets at the way back to what we talked about, data, information, knowledge, insight, right? And that pyramid is real and there's some other pyramids that matter. The question becomes what sits between those? Well, clearly there's one thing that sits between those and its context. If you want to understand data, understand its context. Somebody say, "59 degrees." Okay, there's a data point. But if I'm sitting there in my room and I'm looking at 59 degrees and I'm standing in Monterey, California, it's not data, it's actionable information. Now I know what the temperature is. I can go to other things and I can get weather review, all that stuff.

And I think that's the, the important distinction right there where it becomes actionable information, right? The, the all the data and information in the world because we, we like you mentioned earlier as we're constantly collecting and reporting down to everything down to I can have a watch on me that tells me all kinds of different things about how I slept and this and my resting heart rate and what I need to do if I set goals there and all. And it's like, well, what, what, what am I using that for? So, so what? You know, I always make the joke on here of the this sleep study and different this whole study I was part of and they sent me the Fitbit watch and they tell you this and it's like, "Oh yeah, you know, it showed this and it showed that you didn't have good sleep here." And I was like, "Yeah, I knew that. I didn't have good sleep. I already knew that." Like that, that's, that's, that's not the issue like you're just describing something that I already knew. You're just giving me a little bit more detail about it. And, and you're just, you're giving it like a, a number or a classification. But, but so what? That doesn't help me sleep at night.

Answer, though, that watch, because that's one thing I've been focused on just the last few days. Like, if if we want to predict a grand mal seizure, we can do that. Dogs can do it, right? What we need is multiple sensors, an information aggregation service that unfortunately has a low profit model on the other end because it needs to be done at around five dollars a month, right? And when you look at the scale, it's like, and let's face it, even after 1968, 1950, whatever, the biggest problem in the IT world is finding the business model afterwards, the tail. And that's what we're seeing now as we move CAPEX (Capital Expenditure) into OPEX (Operational Expenditure) money, right? Capital expense and operational expense. It's like, there's no tail end anymore. So what do the professional services firms do now, right? Well, and that's kind of what we, I think we're on the cusp of.

One of the things that I'm really interested in when I talk about building a band is, okay, what are these big problems? And let's say, like a guy that I highly admire, Steve Lukasik, one of my mentors, he literally VC (Venture Capital) funded the internet when he funded the ARPANET (Advanced Research Projects Agency Network). And he was always really clear, Steve would say, "You don't want to get more than four people in a room doing anything." I agree with you. You write four people, if you get an information theorist, you get a decision theorist, you know, you get a learning theorist, because none of these things are science until they're applied, right? Against a specific problem. And then you have a producer, you know, someone that gets the big picture and you're going after a very specific problem. Let's say that specific problem is, you know, COVID type awareness and buildings for this massive person that owns millions of hotels. I don't know, whatever. And then you put a team together and say, "Okay, what, you know, bringing in decision theorists like you guys, bringing information theory so we understand the strategic to tactical to operational and we do it for three or four months and we define the problem. And then you figure out a point where you can hand that over." But that upfront work is everything. Yes, it's really hard as an entrepreneur. Right now, you also have to back in because we can get exhausted as human beings if we don't automate our own services, which may be information-based too. Everybody talks about automating machine services. Well, you automate human services too, so I can be in, you know, I can represent 50 people. It's the core of capitalism, right? It's like, but my point is without that rigor and without that structure, it's really hard to get to that end state, I guess.

Brian, do you ever, I don't know if you don't have a title for the episode, but the one that's been batting around in my mind is, "Greg and Brian have really smart friends." And then the subtext on that would be, "And we're willing to use them." Okay? Mike, I, I would love to be able to leverage your knowledge and the knowledge and skills and and and that Brian brings to bear and that our team brings to bear and go after some of these really hard problems. I love that.

I mean, that's, that's what people like us love to do. Yeah, it's like putting a band together, man. It's not like, it may be a new band because part of us we're defining, you know, new genres of music here too. It's like, you know, you can only do one problem, we'll do it well. Like back to your point about that sleep habits. When you're trying to predict a grand mal seizure, if you've got blood oxygen and you've got some type of indication from an EEG (Electroencephalogram) and then you have three days of sleep history. Oh, you're really starting to get pretty left to that bang now, right? And I just want to make one point when we talk about an analytic, it's really important to understand that predictive analytics do not stand on their own. Yeah, no analytic stands on its own. And there's a repeatable, again, we can do it a billion times but it's still one thing. You need a descriptive first, "This body is experiencing this and I can get that and I can trust those signals." And then you have a diagnostic that says, "Okay, that body is doing this." Now I'm diagnosing it. And then on top of those two, you build a predictive, you know, you build on your predictive in the hopes that the fourth mechanism, or the fourth modular part of your mechanism, is a is a prescriptive that says, "Okay, I've sensed your body, I've done some diagnostics, I'm predicting you're going to have a seizure in five minutes, now I suggest you go do this: some biofeedback, energy distribution, or just yell at the top of your lungs something." But every analytic requires that same discipline. I see so many times, like my favorite one to jump on is Watson. It's like, "Oh, Watson, Watson." It's like, Watson's object-oriented and it's, it's, it's literally applying its knowledge into an object, it buys an entity and this new internet. But I'm calling the 3D internet. It requires an individuality that can't be compromised. So if I would say in 2010 we were in the age of, possibly the cloud, and by 2015 we're in the age of possibly the platforms. Now we're in the age of systems engineering, decision engineering, learning engineering, information engineering. Come together against every problem every time. And we can't kick the can any further in my opinion. And all we have to do is look around and see where our information aggregation services are failing. It always comes down to hardware, software, and wetware. You know, where's the human condition? Those can't be compromised and they have to be concerned and understood. So back to your point, Greg, really focused groups, small groups of teams to attack hard problems. I, I agree with that.

And Brian, quickly, I want our viewers and listeners to know, the person that coined the term "3D internet" is is the guy with the ponytail sitting there in front of the mirror with us right now. So he, you know, he's one of the big tankers. Brian, actually, a lot of people have crafted the 3D, but it was done with VR and AR. And it was like, Greg, give a, give a real quick explanation of what you mean, what was the 1D, what's 2D, what's 3D?

Well, the one, I always think the 1D internet is was the telephone system. It's been around 100 years. The 2D internet was the World Wide Web, term, better as Lee saying, "Hey!" And it was this two-dimensional internet. The three-dimensional internet is what we've been talking about. Yes, it's the telephone system and this ability to communicate in real time, which is really amazing if you think about it. It's quan, again, let's talk about earlier, it's quantum physics that is best when you're on the telephone. You know, so it's the merging of those three things, and then being able to look at those three dimensions in a vertical expanse as well as the lateral one that says integrating the strategic, the operational and the tactical. That's part of the three-dimensional aspect of it. And then the other part of it, I see, is being able to have a method that is capable of capturing all three levels at the same time. You know, back to the simple thing about about, you know, capturing a seizure. You'll see that's a pretty tactical thing. It's like, no, it's not really, it's all of them. You know what I mean? You've got to have a human loop system that understands paying attention to other people that are having grand mal seizures and that may go automated but they're sharing that information in real time based on what they're seeing and they may be monitoring that person, you know, in a way that we can monitor 100 signals now, we used to only monitor one. That makes no...

And, and I think, but that looking at the 3D internet is just saying that, hey, it's all local now, like the drone work, UAS work I'm doing is I, I have a very simple thing: community-led innovation. If you don't have the community involved in the innovation, good luck, because they're the ones getting affected by these things. And if you don't, that's weaknesses, opportunity, threats from their perspective, they're going to fail.

And I think that's also how a lot of this technological advancement has opened things up to look at these differently. You can look at it from a community approach. I mean, given, look at, look at even just how things are happening right now in the US and how different it is across different cities and states and everywhere. And, and the point of it is is that we're getting away from this, "Hey, we need this centralized model or this is how we're going to do things" to just what you said, that community led. Like, I've got some successful business leaders here that can work with the city and the people in the community and this to go, "Hey, we can find our own solution for this area." And now if someone wants to scale that either like you said, horizontally or vertically, that's great. But, but these, these, these...

Well, I think the scale, the scale issue is that you're combining, we talk about scale and we go lateral, but really, you can't go lateral until you have that, like for a city, there's a vertical story out there that can't be abstracted away. It has to be connected.

Well, yeah, and the point of it too is is taking that, you know, what works in Chicago isn't going to work in Portland and that's not going to work in Miami. But but not, not, not, not a template problem. Go, "Hey, they did this here, why don't we just apply that, that solution over here typically?"

But you can't do it, you know, a template match. Right. We can... Well, the word "template," I would say, let's go deeper than that. And, and what we're really talking about here in my opinion is the physics of information, right? And that we can discern that is rather universal. I don't care what the signals are, right? I mean, what's the physics of the information? And every time we have to, and it's really important because in my experience, the best information scientists I've known, the top three out of four that I've known were physicists. Yeah, because they get to the core of the problem. It's like a great screenwriter, you know, you've got to get to the core of the problem, what's the story of for this?

I'll ask for your opinion, Mike, and yours, Greg, as well. What I always tell people is, you know, when you know, we talk about human behavior and information theory, decision theory and how these things work and how we interact and I always tell people when they ask, "Hey, what book should I read or what should I study?" I said, "If you really want to understand human behavior, I would start with physics." Meaning I think it's, you know, I'd love to get you guys' opinions, but physics to me is one of the most accurate and best descriptors, right? It's a great, great descriptive of of how everything works and then everything can layer on top of that. But but meaning when you break it down to literally the level of physics, I, I don't have my thumb on the scale then, right? It's not being influenced by anything other than particles interacting with other with with waves and particles. Right. So if you can do it at that level, at least have an understanding, I think that's, I wouldn't, I don't know what you guys would think about that, but that's kind of been my what I've found.

Brian, I think it's smart and I'll start because mine's going to be brief. I completely agree with you and you know, I'm a pop culture kind of guy so I'll go to the the TV show The Big Bang Theory and somebody was asking about something on a car and he goes, "Are you asking that..." Sheldon character says, "Are you asking about the internal combustion engine? I'm a physicist, I understand everything." And the great thing is that if your model, if your background is in physics, then you can build on the sociological, the psychological, the physiological, but you have to have an architecture. And I think physics lends us that architecture. Mike?

I mean, yeah, architecture, architecture, architecture, architecture, right? Um, again, it's, it's, it's the physics of information. It doesn't matter the use case. What we're trying to get at is the the heart of the story, you know, the narrative, the the experience. And then, you know, instead of, it was really popular to say that we're connecting dots. Don't connect dots. Work with somebody that knows how to hunt dots. So you're seeing the dots in real time present themselves because if you're waiting to connect them, you're going to lose. It's just the reality. That's what we're kind of doing in the fire world sometimes. We're just connecting dots. It's like, no, you got to get ahead of that dot. And part of that is having the humans in place that say, "Okay, I'm interested. I've got this billion-dollar problem, right? This company has a billion-dollar problem. There's a signal there." Alright, well, how do I build the dots for your systems that they can see that as we talk about going from a possible to a probable to a certain, right? That that same thing has to happen as we're connecting, as we're collecting, and then making sense of these information sources as they present themselves because, let's face it, until something happens, it doesn't happen. And it's an approximation. And some things you just can't wait. You know, 80% is good enough sometimes. We may take a mitigation effort. You know, we have probable cause and we think we're at 60%. Yeah, I'll set another sensor on it or I'll reappropriate something, I'm going, "Well, now I'm at 90%." And narrative saying what it is because Campbell and Young told us, you know, "The common story is a common story." Everybody wants to be the hero at the end. Their means may be different than yours and their needs. But they want somebody to carry on that message.

And I guess it's the same way here, is where I'll kind of leave it, is when someone is using an information service, again, it's a, it's an experience. And in order to completely understand that experience every time because these things have to be 99.9% good all the time. Hence back to why we only go one step past a fire detector, because it just gets too hard. Right now, we've got to crawl, walk, run, just like we do in training, you know. And just to summarize it all up, I don't not believe that metadata is the proper mechanism for moving forward. I'm interested in in creating a set of meta context as a new way of navigating. If we can keep machines to read and compile, let's just teach them to treat and compile different mechanisms of mapping. This is no different than exploration we've been doing for, I mean, you know, if eventually I would have said, "Well, I'm not going to do anything because, you know, what's his face did it with the Apple." You know, we're not doing that. So again, after crawling up the backs of giants, which I know you guys have too, as we're sitting here, what I know like back to to a screenplay, you don't get 20 people in there to write a script. No, most you'll get is a couple, three, four because it's a tough problem. But, but they're still serving an architecture up. I mean, that's what a screenplay is for example. It's an architecture that other people can build on. And I guess that's what I'm advocating here. What is the architecture of the operational layer that ties together the strategic and the tactical in a in a way that is verified, validated, and accredited to provide accurate aggregated information sources in a timely manner? And my bottom line is, I think we need to start over again to a certain degree in the best way. And you know, there's an over-the-top revolution happening. You asked earlier about what the first 3D, 1D, 2D, 3D internet was: the telephone system, the World Wide Web, and now this. And it's going to be different. And we need to, because when I look at the world right now, our country, I would say very simple, and I'll be done: overwhelming capacity and underwhelming capability. We have all the capacity we need to build any number of services, and our capabilities are, I believe, in many aspects worse than they were 20 years ago.

No, I, I, it's actually, that's actually a really good spot to kind of, kind of wrap on for this discussion. The idea of, you know, overwhelming capacity, underwhelming capability. I think that that really drives the point home. That wraps literally a lot of the points that we brought up in terms of of defining a problem and then defining what the solution is and how we start with it. And, and I think the the, you know, the the physics of it all, you know, is is amazing. We talked on a whole bunch of different topics, Mike, and Greg has brought up some stuff and I, I would love to, you know, maybe deep dive some of these, maybe one at a time. We, we stick to something and do kind of like a separate type series. I think it would be fun to do some episodes of each one of those and get into great detail about specifically what you mean and that way you can define it how you define it. Greg will define it. I'll, I'll go, "Hey, what does that mean?" My big thing is always with anything and especially with information theory is what does that mean for the person who who is, gets their information scrolling through a feed on Facebook? Because that's a dangerous place and it's, it's not good. Most of what I've seen on there, it's really not. But, but it's, he says while we're on Facebook. But, but, right? Of course. But, but it can be, is the thing. Is my point is it doesn't have to be that way. It's, it's, I'm looking at it, I'm eternally optimistic because otherwise what the f*** is this all for if it's not for a greater cause or we're going to do better or things are going to get better then then f*** it, why are we having these difficult conversations? Why try to solve a hard problem if it can't be solved? And that's my thing with all of these issues that we face is they are solvable. Alright, we got to get rid of the the hay as you say, right? Get rid of that hay to get to that needle. What's an accurate sensor? What do we need to pay attention to? What should we stop paying attention to? What do we concern ourselves with? And that's everything we discussed in different areas on this. So Mike, I appreciate you coming on and and it was an awesome conversation. We got some great feedback here and I, I just, I would love to go in deeper on all these topics if you guys want to.

Oh, I would love it.

Yeah, Brian, I, I think you're onto something. I, I devoted my life to to giving people the confidence and competence to make the right decision at the right time for the right reason. And I think this is a great step into a new direction that that we definitely have to drill down into. So I'm ready, Mike. I know you're ready.

Yeah. If I just think one thing, only because he's so underrecognized: George Boole, who all of this is dependent on, set out to prove God, not as an entity, but as a force. That's what all of his logic and math is predicated on. And he proved it because he said, "God is real, not as an entity, whatever mechanism, whatever," but he says, "God was real because suffering was real." And all the logic that came from that is what defines everything that we do. So I think the point that you made earlier is when we look at this overwhelming capacity and this underwhelming capability, what is the biggest, what are the set of needs that need to be prioritized and why do they matter? Because we can do anything, back to the M&S (Modeling and Simulation) world, but we can't do everything. Right now, it's like, "Okay, we want to do everything, but we're really not doing what we need to do," in my opinion. We can look around, right? And again, clarify the signals and power the response. But yeah, I would love it. And, you know, at some point, hopefully, you know, we can start bringing other members of the band in and saying, "Okay, here's a systems engineering expert, here's a learning theory expert, here's some counter to their decision theory experts and say, 'What would this band look like and what would it start doing and and how do we go about, you know, approaching these big problems?'" Because we can talk about it all day long, but when we have the methods and the models and the mechanisms and the metric in place to do this through a method, then we can start putting bands together and, you know, much like Berry Gordy did with Motown, man. Yeah, he said, "Here's the source product," and he fixed it. And then he created brilliance and magic.

And it's not jazz. I don't trust jazz. And now no one likes jazz, they just pretend to like it. It's just noise.

So I meet with a guy tomorrow who absolutely loves jazz and he's a brilliant jazz musician. I can't say that to him.

Well, I can't play it, but I could certainly talk crap about it.

No, I joke about the conductor who's the least talented guy in the band or in the orchestra.

Well, it's not the guy with two strings, it's a guy with one string. All right, one. Yeah, right. We've got the orchestra.

Yeah, I know that's a great point. And but but guess what, you that that that guy or girl needs to be there, so you need one, they hear everything, that's the point. They got to hear it all, right? Um, again, my hair takes a rare mind which you guys have to to see the obvious. I've been told "rare mind" before and it was not a positive comment towards me. I'll agree.

Alright, I appreciate you guys coming on. I think that's a good spot to to wrap on and I always end it with to everyone, don't forget that training changes behavior.