No Way Out
Welcome to the No Way Out podcast where we examine the variety of domains and disciplines behind John R. Boyd’s OODA sketch and why, today, more than ever, it is an imperative to understand Boyd’s axiomatic sketch of how organisms, individuals, teams, corporations, and governments comprehend, shape, and adapt in our VUCA world.
No Way Out
The AI Edge: Why OODA, Cybernetics, and Neuroscience Matter with Sean Manion
In this timely episode of No Way Out, neuroscientist and systems thinker Sean Manion joins Ponch and Mark to explore the untapped potential of the OODA Loop beyond its military origins. Sean brings over 25 years of experience in neuroscience, clinical research, and health technology to an expansive discussion that redefines the OODA Loop through the lens of natural sciences, cybernetics, complexity theory, and neuroscience.
The conversation dives into how the OODA Loop’s foundation in feedback systems aligns with modern advances in AI, particularly reinforcement learning and active inference models. Sean explains how concepts from cybernetics and information theory can help build more adaptive, energy-efficient AI systems that reflect the dynamics of living systems rather than rigid, linear processing.
Listeners will discover:
- How cybernetics, the science of communication and control in living and machine systems, influenced both John Boyd’s thinking and the evolution of AI.
- The critical role of neuroscience in developing more explainable and trustworthy AI, moving beyond today’s opaque, energy-intensive models.
- How feedback loops, adaptive systems, and even epigenetics play into decision-making processes in humans decision making.
- The intersection of the OODA Loop with symbolic AI, neuromorphic computing, and complex adaptive systems, showing pathways to a more holistic, biologically inspired AI.
Sean challenges listeners to rethink intelligence itself, arguing that modern AI tools like large language models are not true intelligence but instead tools that mimic limited aspects of human cognition. Drawing on Boyd’s interdisciplinary approach, Sean makes the case for a more integrated, systems-oriented framework to guide AI innovation.
Sean on Substack
Sean on X: @TheUnjournaling
Sean on LinkedIn
AGLX Confidence in Complexity short commercial
Stay in the Loop. Don't have time to listen to the podcast? Want to make some snowmobiles? Subscribe to our weekly newsletter to receive deeper insights on current and past episodes.
Find us on X. @NoWayOutcast
Substack: The Whirl of ReOrientation
Want to develop your organization’s capacity for free and independent action (Organic Success)? Learn more and follow us at:
https://www.aglx.com/
https://www.youtube.com/@AGLXConsulting
https://www.linkedin.com/company/aglx-consulting-llc/
https://www.linkedin.com/in/briandrivera
https://www.linkedin.com/in/markjmcgrath1
https://www.linkedin.com/in/stevemccrone
Recent podcasts where you’ll also find Mark and Ponch:
Hey, moose, I want to kick this one off, moose. And here's why. Here's my frustration with the way a lot of folks go forward with the OODA loop. They start with the military, they always start with the military, make a connection, say it's a decision or tactical decision-making process.
Brian "Ponch" Rivera:What I've discovered is, had Boyd been a football coach, had he been a musician, he could have still come up with the same OODA loop sketch. And here's why. It's because the natural sciences, things like cybernetics, biology, evolutionary biology, plexadaptive systems, neuroscience these things matter more than his background in the military Unfortunately everybody gets fixated on matter more than his background in the military Unfortunately everybody gets fixated on that. They get fixated on it, they go. It's all about warfare. I'm like bullshit. It's about how living systems persist through time. And if you start there, you start to see the real beauty of the OODA loop. And I imagine Sean and I think you know Sean, so I want you to introduce Sean for our audience. I think Sean may have started in that position of hey, what is this thing? And, based on the work that I've read from Sean, he's been following this path of cybernetics, information theory, connections to Von Neumann, connections to, I think, who else?
Mark McGrath:Ashby Pitts Bateson.
Brian "Ponch" Rivera:Yeah, that's the pathway you should start with when you look at the OODA loop and unfortunately I think a lot of military folks start with the military connection and I just like, well, that's only going to drive you into the linear orient or, excuse me, observe orient.
Sean Manion:Plus, when you Google it, it's going to send you right there, anyway, yeah, yeah. So, hey, could you introduce our?
Brian "Ponch" Rivera:guests. I just wanted to start there because I think we have an amazing show set up today. So, Mark, can you introduce?
Sean Manion:us. It is a small world and we talk about at VUCA. We talk about complexity, and one of the things is the connectivity between people. Sean and I this is not our first meeting, so Sean and I are both graduates of Central Catholic High School in Pittsburgh, pa. He was a senior, I was a freshman, and Sean's first cousin, mike, is one of my best friends and Mike is also a brother Marine of mine. We went to the recruiting office together. So Sean and I this is a really interesting circle to bring us together. You're the first Central Catholic graduate that we've had as a guest. Dan Marino hasn't returned any of our calls, but absolutely so we're glad to have you so well. Why don't you break it down for us? So you know we were talking about before. You know, before we hit record, how you stumbled into this stuff. As you know, planche and I have military backgrounds and you don't. So give us the you know, give us the story of how you got there.
Mark McGrath:Certainly, and thanks for having me on here. I Give us the story of how you got there, certainly, and thanks for having me on here. I've been excited about the work you've been doing and the deep dives you've been going into with where Boyd and his work can stretch, and not just the OODA loop but beyond that. Sean Mannion, neuroscientist by training, spent a lot of time supporting the military health system, moved into tech and health tech several years back and focused on that area right now. But kind of get my hands in a lot of different pots. So I've never served in the military but I did end up going as a civilian, starting in 99, to Uniformed Services University of the Health Sciences to do my graduate work in neuroscience. Started doing PTSD research in the psychiatry department there in 99. So this is pre 9-11. We had to. We had to struggle a little bit to justify our funding sometimes because they were they were thinking that's a veterans problem.
Mark McGrath:I mean, you know it's. It's after a long stretch without a whole lot of hot wars going on and then 9-11 happens and everything kind of switches. And you know the research we're doing is in demand. One of the reasons I went to uniformed services as opposed to Georgetown or somewhere else was that the military has an appreciation for the acceleration of science into practical, evidence-based medicine and there was a guy who kind of the captain Will Watson, who kind of introduced me to that concept and it was what drew me into military medicine to start. But as you know we're, you know I'm seeing people I know go off to the Middle East. You know we have OEF, oif kickoff and I'm trying to understand what's going on and you know it's early 2000s.
Mark McGrath:I'm learning on the Internet a little bit but I'm mostly just kind of reading newspapers. I was taking the train between Baltimore and DC so I was like a five newspaper a day habit and I keep seeing I start going on the internet. I keep seeing terms like fourth generation warfare and there's the name that pops up over and over again and it's Boyd, and I don't have a background, I didn't know who he was, I didn't understand the context of a lot of it. So it was 2003, 2004. I'm learning about this and I don't know. I was reading a lot of the war bloggers back then either Richard Fernandez, it probably was John Robb who you've had on the show that started to impress upon me the importance of this one individual and his ideas, and right around that time a couple of biographies came out. So I started diving deep and I was amazed, I was fascinated.
Mark McGrath:And I'm busy, I'm doing work. I'm 80 to 100 hours a week as a grad student but I have a lot of downtime time to read and so I start reading things about him, looking at some of the stuff that was thin but available on the internet. It was really when Ossinga's thesis turned book came in that I looked at it and I looked at all the stuff Boyd was reading and it was a cross section of the stuff that I had been studying as biochemistry undergrad, interest in neuroscience, philosophy of mind. Definitely. He was moving into a direction in the later years where he was expanding his ideas into places that I knew well. Now, as we all know, and sometimes is a little bit of a misdirection, he's got his application training as a fighter pilot in the military, so that's where he took these ideas as they grew, and gave examples, but it was definitely the case that I saw a connection between him and all these other areas that I didn't really think about as being connected and I think, punch, that's what you were talking about.
Brian "Ponch" Rivera:Spot on. That's what I was bringing up. There is it isn't until you start looking at the Toyota production system and learning how control is outside and bottom up and you start looking at the parallels between that and then maybe what you're seeing in neuroscience and complex adaptive systems thinking and you think about how those evolved independently. Perhaps you know Shingo and Ono I don't think were studying as far as I know, they weren't studying complex adaptive systems and neuroscience to come up with the Toyota production system. So you get this consilience that John Boyd. It emerged from the interdisciplinary approach to how he looked at the world, what he studied, his curiosity, always challenging assumptions. I didn't know a lot about cybernetics until I picked up I still don't know, by the way, but I wasn't aware of it until I picked up Asinga's book and I was like what is this thing? I thought it was like something that you did in college for weightlifting you took some pills and you worked out funny.
Sean Manion:You know what I mean? Cybergenics, oh yeah, there you go. Wow, that's a Gen X reference, yeah.
Brian "Ponch" Rivera:So it was a Singha's book that started taking me down the path of hey, let's start looking at complex adaptive systems, and then now we're looking into neuroscience your field. And hey, let's start looking at complex adaptive systems. And then now we're looking into neuroscience your field. And then I was reading through some of your material over the last few days and the connection to Brownian motion, claude Shannon, information theory there may be some loose connections of Popper and all you know going back to the 30s and 40s and what was going on then. So just absolutely thrilled that you're here to kind of walk us through your journey and then kind of have a back and forth conversation about what is this thing that John Boyd gave us, you know, and what could it be, and how can you apply it to artificial intelligence. So thanks for being here, sean.
Mark McGrath:No, I appreciate you having me and I love seeing this constant ongoing conversation in parallel tracks across different ideas and different industries tied back to Boyd, because I think he would have appreciated the vastness that you're touching.
Sean Manion:I mean Ponch and I have spent days in the archives down in Quantico and it's amazing to see the diversity of books, all these books that he read. But, as you were pointing out, the interdisciplinary nature of it, and it's interesting, when we talk about theorists, that we bring up a lot like John Boyd, marshall McLuhan and some others. I mean these were well-read people across and down. I mean these were deep thinkers across many, many different disciplines and you're not the first neuroscientist we've had. And I think that's feedback that we also get is like, well, you guys have a lot of people that are not in the military on your show or were never in the military, and I think that's feedback that we also get is like, well, you guys have a lot of people that are not in the military on your show or were never in the military. And I'm like, yeah, and we bring on people that know a lot about John Boyd.
Sean Manion:In your case, you do know a lot about John Boyd. Sometimes we say so you know everything about John Boyd, but you know nothing about John Boyd, just the nature of his work and what he was actually doing. I was having a conversation with somebody. What you think of this, you know, when we Google Boyd or you go looking, it does reduce it to sort of that military, maybe that linear depiction of the OODA loop. You end up getting like maybe 1% of what the guy was actually trying to do. And then when you dig into the Ocinga book, as you mentioned, you realize that well, there's 99% of this guy's thinking that no one's talking about and that's that's kind of what. That's why you're here, that's what we're, that's what we're trying to do.
Brian "Ponch" Rivera:Yeah, we've had a Bobby Azarian on to talk a little bit about cybernetics and the connection, you know, to what was happening in the forties and how it was suppressed by potentially suppressed by our government. You know, going coming out of the forties and the fifties. But I'd like to unpack that for folks to understand the basics of cybernetics and how it influenced John Boyd's thinking and how it even influences today's AI, the race for agentic AI and there's so many different names for it generative AI, active inference AI and, of course, mental health. So can you help us unpack what is cybernetics, what are the controversies around it, who are the key figures and how does it fit into John Boyd's observed oriented side activity?
Mark McGrath:Decide, acclimate, sure, sure. And, to be upfront, I mean I appreciate you saying I know a lot about Boyd, but I feel like an amateur. I've had a 20-year run of looking at his stuff but it's like it fades to the background. It's definitely influenced me, but then it comes back again and I think this is sort of my yeah.
Sean Manion:Yeah, see, that's proof that you know a lot about Boyd, because we would say the same thing. We're constant students of this. You have to take that amateur approach to keep going at it, at it, at it, because he even told us the minute you call yourself an expert, you're screwed. We don't want to be the experts, we want to be the students and the teachers of it.
Mark McGrath:And with the cybernetics I feel the same way. And with the cybernetics I feel the same way. I mean I am three or four years into a pretty steady dive into the history of cybernetics. I mean, if I was like you, you asked me five years ago what's cybernetics? I was like, ah, you know the cybersecurity and cyberspace and cyberpunk and whatever internet stuff you're associating it with. But I didn't really have a concept of what this was historically and honestly, it was a blind spot to me.
Mark McGrath:I read Ocinga and there's cybernetics in it, but I didn't know what it was at the time so it didn't really resonate the system. Science did a little bit more. But a few years ago I started working with a group an oral history of all the original Dartmouth 1956 AI conference people and she worked with them and knew them on a personal level. So they opened up in ways that I don't think they would have for just someone who was just an interviewer and she put together an awesome book. So in looking at that book and she passed away in 2021, unfortunately and she donated all of her archives to CMU. So I started going to CMU, diving into the archives paper archives, some that haven't been digitized yet, and I'm seeing all these things.
Sean Manion:And for the non-Pittsburgh natives like Sean and myself.
Mark McGrath:Carnegie Mellon University.
Mark McGrath:Carnegie Mellon University, not Central Michigan, which used to be Carnegie Institute of Technology until the early 60s and that's where Herb Simon and Alan Newell that actually did the first and only working AI program at that 1956 Dartmouth meeting. That's where they were based and, as Herb Simon who won the Nobel Prize and the Turing Award and was the first person to be an AI person to win it in the 70s, everyone forgets about that with the recent Nobel Prize for AI. They were the center point, cmu, along with MIT, along with Stanford and AI back in the day. But everyone talks about 56 and the Dartmouth meeting as being the kickoff of AI, sometimes the 1950 paper by Turing. But what I started seeing was there were precursors to that and there's always precursors to everything. Simon in one of the interviews and I listened to the interview she did with him he's like he's like, yeah, we were doing AI before 1956. We just called it operations research. So you look into it with new light, you start to look at what's going on. And there was a guy by the name of Norbert Wiener who who coined the term cybernetics. But even before he coined that term there was work he was doing developing essentially a predictive targeting for airplanes and for gunners to hit airplanes using the pilot and the airplane as a combined entity, using the gunner and the gun as a combined entity, and he set up a program that did that in 1941 and 42. That was classified. It didn't make it out but there was a lot of that work that went into. Shannon was one of the students so he kind of helped inform the whole idea of entropy as information theory.
Mark McGrath:But this time in the 40s and into the 50s there was this tremendous amount of excitement and activity almost to the level we have now with AI, about cybernetics. There was a parallel group in Britain with Turing involved called the Ratio Club, where they got together they drank beer, they ate food and they presented on all of these things and they would bring over the head of the defense intelligence from the US because they knew them from the war and these people were all having conversations about thinking computers. And this is in the late 50s, in the early 50s, prior to any AI talk. And so all of this stuff with cybernetics had this rich, very distinct history and for a variety of reasons some was untimely, deaths of a few individuals involved and some was very political the US decided to pursue AI and thought that the intelligence in that, you know, connected with intelligence. We used it and set the Central Intelligence Agency.
Mark McGrath:Meanwhile the Soviets dove deep into cybernetics and at the same time cybernetics in the US and in the West in general kind of took on a very social and antisocial component of the counterculture drug use, a lot of art, a lot of different things and so it took on a different tone in the US than what it had in other places. And a lot of this culminated in there was a guy named Stafford Beer who was a sort of second generation physicist and there was a second order cybernetics, sort of thinking about the cybernetics of cybernetics, and all of this just sounds Boydian when you start to dive into it. And he advised the Chilean in what was a coup there. But they had a group called Cybersyn or an approach called Cybersyn, which was a cybernetic approach to centralized running of the government. Now the feedback was solid in thinking of the loops, but the information gathering and the time lag were too slow so they weren't able to do it that effectively, probably because of the centralization, but it was really the death knell in the West of that being an acceptable area of government expenditure and things that helped keep that funding going. So, as you had some of the early proprietors, like Warren McCullough who worked with Wiener and did one of the first artificial neural network designs in 1943, he passed away. The other people who were involved ran out of funding and it just kind of died a slow death through the 60s into the early 70s in the US.
Mark McGrath:There were some threads of it that didn't go away entirely but it was very much just kind of viewed differently and viewed as this dormant old thing that nobody touched anymore and that to me was fascinating. So I've been spending time getting deeper into it. I've met the people who are involved in the American Society for Cybernetics, again based at CMU, the head of that. Right now he's starting up. There was a series of meetings. I'm going through here's the minutes of the 12 meetings that they had in late 40s and early 50s. It's absolutely fascinating read. I would have loved to have seen Boyd get a hold of this information. He definitely got into cybernetics but if you look at what Ocinga says he was looking at some of the most popular end stage pieces. He did start into some of Wiener's like on human use of human beings, but I think by and large he hadn't gotten to delve all the way back yet, which would have been amazing if he could.
Brian "Ponch" Rivera:Sean, there's some connections to like stochastic process or stochastic math and the Brownian motion. I'm not sure if you've been diving into that at all and the reason I bring that up is the connection to economics. Boyd had a background in economics and Moose has a background in economics, as do I, but stochastics are being used and I recently one of our guests, actually Enos Hippolito, brought out a point that large language models are stochastic parrots, right, so there's some interesting connections again that connect back to things like Brownian motion and cybernetics. Can you speak to any of those connections that you've come across between I forgot Brown's first name, who came up with Brownian motion, maybe Mandelbrot who looked at that and stochastic processes? Is there a connection in there that you've come across?
Mark McGrath:I don't know the history of those things exactly, but where it does start to intersect with cybernetics is with Norbert Wiener, the man who coined the term, who set up some of these initial cybernetics groups and the focus on it. He had actually had success as a mathematician before World War Two. He was, he was working at MIT and Brownian motion was the area that he had focused on the most at that point, and it wasn't until he was tasked with essentially figuring out the time series activity involved in targeting a plane, where the time to get to the plane was seconds to a minute and in that time the pilot could see you firing and make adjustments. And so it was some of that randomization, that trying to break down what seems random but how we can project what random things could move forward.
Mark McGrath:I think Wiener's work with Brownian motion, which Einstein had done as well that was an area Back in the late 40s. He was almost as famous as Einstein for being a scientist, but then he pissed off the Soviets, he pissed off the FBI, he pissed off his colleagues at MIT and he didn't play well with others. So I think he kind of tarnished his own image and didn't get remembered for as much as he should have, but he definitely was approaching some of those areas way back before AI was even up. Twinkle in John McCarthy's eye.
Sean Manion:I just put up John Boyd's collection in Quantico. You see, norbert Wiener's the Human Use of Human Beings, cybernetics and Society was one of the books that he had.
Mark McGrath:Yeah, that was when Wiener came out with cybernetics in 48. In 49 he published the time series um book that was actually published in 43 but classified, and then 50 he started getting upset with. He was upset with a number of things. He stopped taking government funding. He didn't like the military use of of some of his ideas, um, and he published on human use of human beings which was really a very prescient tale of these. Automations are going to take over jobs, they're going to ruin human lives. And most interesting and it was interesting to me that Boyd had that 67 version because the original 1951st edition had a chapter in it called Voices of Rigidity. It was the last chapter where he warned that the Soviets, the Catholic Church, the FBI could use technology to survey and to abuse their power with individuals. They made him take that out of the second edition. At that time he was under investigation by the FBI and his publisher was like we're going to take that chapter out. So I don't know if Boyd ever saw that chapter.
Sean Manion:Wow. So this is the Anderson US Anderson Success Cybernetics.
Mark McGrath:That one, and then I think Psycho Cybernetics was listed.
Sean Manion:Maxwell Maltz. Yeah, this one. Fh George, cybernetics.
Mark McGrath:FH George and I think he was at Bell Labs and then Ashby or two people he read Ashby was one of the original British cyberneticians. Those would have been very much the first order cybernetics. So he was starting to get some exposure. But I don't think he got as deep. He didn't get back to wiener's original 1948 cybernetics. So I feel like he only scratched the surface on the first order. The, the psycho cybernetics, the success success cybernetics in the late 60s and 70s were starting to be that sort of pop psychology, you know, sort of feedback to improve yourself and how to improve your life. Valuable in in some ways for some people more than others. But I don't know that that gave him as much depth as I would have really loved to have seen what he would have done with if he would have gotten into some of the meat of it.
Sean Manion:Is SAP Arena cybernetics within us?
Brian "Ponch" Rivera:I think there's so many connections inside the archives you just passed over, mandelbrot.
Sean Manion:Yeah, I saw Mandelbrot and.
Brian "Ponch" Rivera:Mandelbrot. So we have a connection here. We have Wiener with a connection to Schadklanen, with information theory. Schadklanen was a Weiner student. What's the correct German pronunciation? Weiner or Weiner or Wiener.
Mark McGrath:I say Wiener, and I haven't been connected it might be Weiner, I don't know.
Brian "Ponch" Rivera:We'll ask one of our german-speaking friends on that. So a lot of connections in here. But can we unpack what cybernetics is? So we got a lot of business leaders that listen to this, that they're probably going. What the hell are you talking about?
Mark McGrath:yeah, they're thinking. They're thinking it's like I was a few years ago. It just means like something with computers. Oh, so what it was referred to as is the science of control and communication in animals and machines animals including humans, and it was a look at it was really also a science of feedback. It stemmed out of a lot of control theory from the late 1800s and early 1900s and then thinking how that interacted with individuals and entities and those entities were computers that were starting then and they were thinking about how computers could think like humans and how you could model humans and computers together. It did set us into a little bit of a tight loop and that I think has been a little challenging to get out of humans as computers and computers as humans, but not any way to think of them differently to really expand upon that.
Mark McGrath:But it's that science of communication and control in animals, humans and machines, and how feedback can give you fine-tuned complexity out of very simple devices. And I was teaching just the other day and using a textbook I teach brain behavior and cognition and there it is. There's a, there's a set of feedback cells and I'm demonstrating it for the undergraduates and it it really is textbook, what, what, what Wiener was thinking about, what all these others were thinking about. And the thing about Boyd is he was coming at all of this without the same background that they had, but finding the same conclusions. He was moving in the same direction. And it's you know, it's you know. We all have a certain amount of time and it's a shame he didn't have more, because I think what he probably would have done was start to absorb some of that the way he had, you know, different military strategists, and started to change his thinking based on all this new information he was absorbing. But that that science of control and communication some people think it did sort of work its way into some of the managerial systems.
Mark McGrath:Stafford Beer wrote you know, a brain of the firm, how to design a firm. You know, design an organization around, how a brain works. A lot of robotics. The whole British cyberneticists were very heavy into robotics. They had essentially, like you know, little robots that could go around since when their battery was low and then find their home to recharge. It was a Roomba without the vacuum to it they were playing with like it was a game in the late 1940s and 1950s, and so the science itself is very broad. It did spin into systems theory, complexity theory, chaos theory. All of these really grew out of cybernetics, so there are elements of it still around. It's just some of the core thinking of it just got left behind in history, I think.
Brian "Ponch" Rivera:And the concept of Stearman. That's what it means in Greek. Is that correct? The term they actually that's what it means in Greek. Is that correct the meaning?
Mark McGrath:The term they actually. There's some interesting. John von Neumann and Norbert Wiener and Howard Akin hosted a conference at Princeton in early 1945, I think January 1945. They called it at that time, they called it teleology and some of the early 1943 papers.
Mark McGrath:Wiener's paper called it teleology, but that had that and that means sort of a purposeful action, but it had a religious connotation. So they decided they needed a new name and so he went with and he kind of developed this on his own and got credit for it with his 1948 book. He called it cybernetics, which means steersman, comes from the Greek, same root as governance, so it basically means governance. People think it was a nod to Maxwell, a physicist who developed control theory and governance theory. So it had been used once in 1800s France to mean control and governance of a population, of a government, but he didn't know about that, so he thought he was kind of giving it a new life.
Mark McGrath:But it's a very useful name in that respect because it really does mean to steer and to direct something. And so how do you steer something? You steer something not only with the control of it but with the feedback. You need to make the right choices with that and it really gets into what Boyd's general thinking and Odo Oda was also arriving at different language sometimes, but really similar concepts that just aligned in a lot of different ways and in ways we probably still can.
Brian "Ponch" Rivera:I can't remember if cybernetics has the internal model of the external world, or what's that called. Does it call for having an internal model that reflects the outside?
Mark McGrath:world. There are like just a representation model. There's threads of it that do. I don't remember that as being a core element, though Maybe in some of the second order.
Brian "Ponch" Rivera:Okay, and the feedback loops. So again looking at it, when you're sailing a boat, now, I'm not a, I don't sail. What do you call sail people? That sail boat? I'm in the Navy, I can't tell you Sailors.
Brian "Ponch" Rivera:Sailors, sailors, that's right. Yeah. Targets, yeah, well, yeah, you got trim tabs on all there. You got feedback, yeah, shoes Driving a boat. You get feedback from the ocean. You have currents, you have wind, you have all these things. You just can't go straight, right. You got to adjust to the external environment and I think that's what the feedback loops kind of reflect. There is you take the feedback from the external environment and update two things, right, or maybe one thing your perception of the external world. That's what you're trying to do. Or take actions, shift the boat, turn the boat to counteract the forces in nature or the features of the world. Right, right, Moose.
Sean Manion:Yeah, yeah, I mean that's. I was just looking over Cybersyn, the history of that, and it's interesting how, as you pointed out, when the coup happened, they destroyed it Like they destroyed all the off-centers. Now I imagine too, there's probably a dystopian view of that too, that, cybersyn, if we lived in an algocracy which is a word I just learned democracy or government by algorithm, do you think so? I mean, maybe that's a discussion point Do you remember when the doctor got thrown off of a United flight and they beat the shit out of him? I mean, they grabbed him and they dragged his ass off the plane?
Sean Manion:Yeah, and it was just basically, though and Rob wrote about this that's what the algorithm said to do. It said to free up the space so this pilot could jump the plane to get to his next location because of whatever, and they had to pick a passenger, and they picked this guy who's a doctor, and they not got his teeth knocked out or whatever, but it was saying that no one was able to intercede above the algorithm. The algorithm said to do this. No one injected any humanity in it to realize that we should not be beating the shit out of passengers and tossing them off of planes.
Mark McGrath:I think there's an important aspect that's missing from our current use of AI technology and technology in general, and that's trust and technology in general, and that's that's trust. We tend to trust systems and then that trust is somewhat mislaid for algorithms that don't do the right thing or situations where we're used to someone having the ability to intervene if something doesn't make sense One of the things that you know I mentioned. I kind of come back to Boyd again and again. I went to grad school trying to understand the world just at its basic level. I was applying it when I was playing soccer there, just to try and figure out how it worked, and it worked wonderfully, because I was a terrible soccer player and I needed some help, and Boyd's loops helped me in that respect. But then it went away for a little while came back to me when I was in an administrative policy position in DC a lot of bureaucracy. I called it bureau science.
Mark McGrath:Understanding those systems was very interesting to me, and so I revisited it again, but then, when I left government in 2017, got very interested in blockchain technology and not cryptocurrency.
Mark McGrath:Per se, that's one application and a popular one, although a troubled one.
Mark McGrath:Per se, that's one application and a popular one, although a troubled one but what you can do with a distributed system of information, of knowledge and an automated governance across that, and so that automated governance and at that time I had no idea cybernetics kind of tied in with governance as well, and I wish it did, but I'm starting to get all the dots together.
Mark McGrath:I got very interested in how Boyd would have perceived or how, the sort of net centricity that he implied as something you were able to do if you had the right trust in a group that everyone's operating on the same principles and so can go off and do their own thing without having to have tight command and control on those units. And I started looking at that with respect to different governance structures and control on those units. And I started looking at that with respect to different governance structures and blockchain different applications of blockchain in health and life sciences, health and human services. I had a white paper contest in 2016 that I was studying and trying to understand how it would work, and I actually reached out to Chet Richards and a great guy I'm sure I don't know if you know him personally, but you know who he is and he's retired now and I hope he's still with us. I haven't talked to him.
Sean Manion:Yeah, he just wrote the foreword to our transcript of Conceptual Spiral and he's been on two or three times with us.
Mark McGrath:That's right. I see his blog once in a while, Slightly East of New.
Sean Manion:But we had if you go to slightly east of new. Right now we're the featured article at the time. Really Nice.
Mark McGrath:That's that's all coming together but he was very gracious to like listen to me I mean by email. We had a brief diet listen to my questions here about this weird technology he didn't really know anything about and and it was. Then I got busy in the startup world and I wasn't able to. You know, once I had dreams of like kind of putting some papers together or something on it, but I really see one of the gaps in trust of AI, of institutions. You know, it's not just the algorithm we don't trust all the time, as being perhaps augmented by this automated trust. If you know what governance went into the system, if you know what things are going to be done, you have a better sense of whether you can trust an algorithm, a group, it's when, it's when an or it's you know something he had purchased a ticket for and he was able to go sit down and then that didn't happen. That's. That's completely outside of not just the algorithm but what the organization was doing.
Mark McGrath:So how to bring trust back? What blockchain and cryptocurrency came out of was the financial crisis, and how do you trust that if you say you have a unit of currency? And you say you have a unit of currency and I say I have it that I don't try and tell you. I'm going to send you one and send you the same one. Well, we trust banks, we trust institutions, but what if there was a process, an automated process, that allowed each of us to have that ledger of what was going on? That simplistic audit trail for any data system, that transparent governance, as well as agreed upon governance, can allow a lot of things to happen with data and a lot of things that happen with data going into algorithms, and so there's a lot of interest now in how we can keep data private, but also trust that data that's going into an algorithm is appropriate for the decision-making level that we give it. You can't take a sample of 10 college kids and then take that health data and apply it to a geriatric population. It just doesn't work, and yet we don't know what's going into these algorithms, and so having a transparent audit layer, which blockchain or distributed ledger technology allows us, is a way to bring that together.
Mark McGrath:So I think, going back to your question, that trust issue is missing, and finding ways of bringing trust to a system gives you the value that Boyd saw and discussed, whereas if you have trust, you can operate much more independently across a network. Where I think things like Cybersyn failed was the authoritarian need to control it in the center and not to make them my example. But I was a DHA employee and a little bit antiquated and maybe the wrong direction. There were some good people working on it, but it hasn't played out the way it was supposed to. And does decentralization of those systems, or re-decentralization of those systems, provide better mobility, better agility and a better sense of independent operations? And I think if you look at what Void was going into not simply in the OODA loop it does point to that being more successful Centralization, and we have the ability to automate some of that trust and governance now, which I think is valuable.
Brian "Ponch" Rivera:So explainable AI is critical now. And then there's a connection to net centric warfare that I think a lot of people like the military you know we were taught net centric warfare was about systems like computer systems, and I know that's not what the intent was behind it. So the distributed approach to decision-making is something that we put into the flow system, right? We call it distributed leadership and I think in one of your articles you kind of highlight the difference between a decentralized approach and a distributed approach where you can kind of graph it out and see that what it looks like in a network, right? So the example of blockchain connected to explainable AI, which I don't think we have right now. We have these black box approaches and you gave a good example of that that you don't know what you're going to get. They're hallucinating at the moment.
Brian "Ponch" Rivera:My view is, if you take some of the things we talked about with cybernetics and neuroscience and you move that forward and you borrow some of those ideas, because you can't go one for one and copy a human, as far as I know you can take a lot of lessons from natural science and apply them to AI. We get closer and closer to explainable AI, something that we can trust? That's where I see the path right now. Is the current status of LLMs? Is they're that closed, linear OODA loop that we push back against, right? We can't tell you what's on the inside. We want an open system approach, a lower energy approach, one that adapts, one that learns, one that perceives just like we do. Can you talk a little bit about the connections to network-centric warfare and this path to explainable AI and what you've seen?
Mark McGrath:I can. And there's probably another important element to put in there, and this goes back to the stuff that I was doing in my prior company, equidium Health, and now we've spun off. I am the co-founder and chief scientific officer for a nonprofit called AI Mind Systems, where we're trying to apply blockchain technology for that audit layer as well as an incentivization layer. In more advanced cases, AI, but in this case, federated AI. We are obsessed with this. Bring all the data to a center and put it through the paces there. One, it makes data less valuable because you're copying, pasting it. Two, you're getting further away from the source of the data, where people who know how it was collected which in health matters I mean, if I take my blood pressure at home versus it gets taken in the clinic, versus someone who's very well trained and being very careful with the best equipment is taking it in a hospital setting for a research study. Those are three similar data points that look the same to a data scientist, but you need to know which is which, depending upon which questions you're going to ask of it. So keeping that data decentralized, keeping data at rest and sending compute to the data is another technology advantage we have right now that's entirely possible. So I think being able to look at net-centric warfare health business as how do we operate without the time lag of needing all the data to come to a central repository speeds up decision-making while still enabling all the data to be available and that trust. And there was a wonderful example of this in 2017 or 2018.
Mark McGrath:The first blockchain system with authority to operate in the federal government a friend of mine, jose Arrieta, who went on to be the chief information officer for health and human services. He had put together a program called HHS Accelerate and it took. They had 20,000 buyers. So if you're an HHS, you can buy stuff for some unit or entity you have. You've got the credit card to buy. If you are making a, you know whether it's software or whether it's nitrile gloves. If you are making a purchase above a certain amount, you need to shop for the best.
Mark McGrath:The only way for them to do that strategic buy was to go to a centralized office that was crunching all the numbers from 20,000. It would take them four months to tell you where to get the best purchase. What he did was set up a system that allowed all of those individuals to have access in almost real time to all the data across all those other individuals. It was a closed system, so it was open within there, but it wasn't open to the public and so, instead of four months, they were able to have the best value and the target of where they're going to purchase in seconds, and it saved the government 25 million in the first year they had it. He got promoted and they sundowned it because no one wanted to maintain a system that you know undercut their sundown because no one wanted to maintain a system that undercut their budget. The government is, unfortunately, sometimes incentivized to not cut money, but it was an amazing demonstration of what you can do with a simple thing like purchasing.
Mark McGrath:Walmart used similar technology to figure out how they could track through their entire supply chain. They did one test in China and one test in the Americas pork in China and one test in the Americas pork in China and mangoes in the Americas. It used to take them days to figure out if there was mango recall where all their mangoes came from. So they just have to pull them all off the shelf, throw them all away, because they didn't want to get sued for making somebody sick. But with having all that data of all the supply chain in the hands of the guy with the handheld computer inside one of the Walmarts, they could know in seconds if a batch came from the bad place, where there was a recall or somewhere else and they could leave it on the shelves. So this ability to look at data stretched across a network without having to centralize that data speeds things up remarkably.
Brian "Ponch" Rivera:So this is interesting. I don't know if you know the answer to this, and I'm curious. It sounds to me like the brain kind of operates this way already, right?
Mark McGrath:The brain operates not only like this, but levels and levels above that that we barely understand. I mean, I'm a neuroscientist and I'm the first to admit we know little about the brain and we know amazing amounts, but we keep getting these little incremental bits that we haven't integrated. There are governance structures at the subcellular, cellular and systems level of the brain that we haven't even studied. How it's, with light bulb level energy doing everything. My heart still beats, my body temperature is being maintained while I'm talking to you, while everything else is going on. It's doing all of that in multiple different systems operating and where they need to be integrating, and we we've barely scratched the surface. So I think, yes, the brain is an example of this, but at the same time, you know when, when you ask a neuroscientist, you ask anybody where is the memory of you know, Central Catholic, my graduation, where does that sit in my brain?
Brian "Ponch" Rivera:We have no idea.
Mark McGrath:We simply don't know. We can, you know, we can poke holes in it until I don't remember it anymore, but that's not very effective or useful, especially if it differs across people. So what the brain is doing probably has some similarities to it, but probably goes well beyond in ways we don't even know. Which is why I say that the most successful computer or tech companies of the future are going to start employing neuroscientists and doing their own research, rather than just waiting for it to come out of NIH.
Brian "Ponch" Rivera:Right. So the naturalistic approach to engineering. I hate to call it engineering, but we should mimic what's in nature. We should learn from nature anthropology, biology, neuroscience and try to figure out why. Why does that work that way, and see if we could scale it. When you read about the Toyota production system, to me it kind of mirrors everything we just talked about, right, what Shingo and Ono were looking at and what John Boyd looked at before he got into the neuroscience and complex adaptive systems. I think he looked at cybernetics before he looked at TPS.
Sean Manion:Yeah, because cybernetics was one of the that was for destruction and creation. It was one of the sources.
Brian "Ponch" Rivera:Yeah, but just taking a look around and seeing what's actually working in nature, what actually scales, is important. And to me there's a fractal approach behind all this too, because if you could take the OODA loop and say this is how a cell works and for the most part this is how a neuron works and a person works and the team works, then you just reduce the energy required to understand and again, it's not a full understanding of how the world works. It gives you a good abstraction, but that's the way Moose and I are looking at this is. It's a fractal approach to understanding how natural things persist through time, right, and if you understand it, how do we fend off the second law of thermodynamics? Right, that's what we're trying to do here and I think, by the way, is there a connection between cybernetics and the second law, by any chance that you know of?
Mark McGrath:I feel like we're looking at it from an information perspective. Cybernetics definitely kind of grew out of the same information theory component that both Wiener and Shannon. Shannon got most of the credit, but Shannon actually referenced Wiener's still, you know, not public work in his work. So there's a there's an information theory aspect and there's an entropy aspect to information theory and to some extent, figuring out what we're doing with AI and the internet right now. Are we, are we facing an information entropy death of the internet by allowing it to continue to feed garbage, garbage, garbage and then learn off of the diluted garbage?
Mark McGrath:This model collapse that could be some, you know, parallel of the heat death of the universe is the information death of the internet. So I think there's definitely things to be learned from that and I think you know what Boyd did in Creation and Destruction is conceptually start thinking about these things Now, applying that concept to different areas. You know he was applying it in a general respect, but thinking about it at that time mostly in a combat situation or in a war situation. But if you think about how that same lack of information, how that same incompleteness of a system and how that same uncertainty in the system come together. That can be applied in a lot of different places.
Brian "Ponch" Rivera:I'm going to throw this at Moose. Everything Sean just said to me connects right back to Boyd's Trinity. Take it from there, moose.
Sean Manion:Yeah, entropy, incompletenesseness and uncertainty by by fusing second law with uh girdle's incompleteness theorem and heisenberg's uncertainty principle, thus required us to. I mean, that's really where oodaloo sketch comes from, is like, because of these three, three, because of these three things. How do we create mental concepts in our minds to help us engage that, as time unfolds, as the circumstances in our environment?
Brian "Ponch" Rivera:Yeah, and the internet bringing that trash together, collapsing on itself, is a great analogy of Boyd's strategy, which is you cannot change the system from within, right, you have to go outside the system. Yeah, and that's why you can't determine the nature, you have to go outside the system.
Sean Manion:Yeah, you can't determine the nature of a character of a system within itself. Yeah, and the more that you attempt to do so, will you increase confusion and disorder there you go.
Brian "Ponch" Rivera:That's what Sean, I think what Sean just said.
Mark McGrath:Yeah, that's what Boyd covered very, very succinctly.
Sean Manion:I still read that paper and I'm just amazed at how just just short and to the point.
Mark McGrath:It is um that's, that's hard it's still dense though it's not. It is. It is, but you know it's not a hundred pages it's. You know what like have you?
Sean Manion:ever. Have you ever looked at evolutionary epistemology by chuck spinney? So, chuck spinney, I I call it the codex of destruction creation. You know my my history slash economist mind had a hard time really grasping destruction and creation, other than I could see the quote unquote, stonewall building. But anyway, chuck Spinney, who was Boyd's closest collaborator that helped him illustrate this OODA loop sketch, he created this evolutionary epistemology which you can get on Chet Richard's site in the article section, and one of the cool things that it does is it draws a boundary between the internal and the external role. You know Ponch talks about this a lot, that you have to understand what's inside your cognition and what's in the external world. I also think that it does a pretty good job of helping us understand the difference between perception and perspective. You know.
Mark McGrath:I haven't looked at that. I definitely need to. I've, you know, slowly working my way through, like all the different threads, but it just amazes me. I need to revisit the same things each time, because the depth that they, they, they got to, and even the things that they arrived at independently and I'm starting to say they, is it not just boy, but you know the lack of a better term the acolytes that he had but pierce sprayay with, I don't know. Have you guys ever checked out what Maple Shade Records did, what Spray did after he retired? He was a record producer. I knew that.
Mark McGrath:He set up a completely analog record studio in Pennsylvania in an old house. He actually had a studio in Baltimore. I walk by there sometimes and no one was ever in there. I would have loved to have met him. But I walk by there sometimes and no one was ever in there. I would have loved to have met him.
Mark McGrath:But he does completely analog recording and it's beautiful, and I know people who know music a lot better than I do and there are things on analog records and now they brought records back, but a lot of times they're just recording the digital recording onto them, which loses everything, and so he's got this and at the same time I'm looking at it and I'm talking with people. There's areas of neuromorphic computing and analog computing that are starting to come into play because we're hitting caps on energy for digital and if you can do analog at the edge, if you can have certain types of analog processing happening, you get more exacting this to it. It wasn't feasible technically years ago, but now they're starting to be analog computing companies that are coming into the AI discussion and it might be the future and I'm like they were moving in directions and sensing things that I don't even think they realized have parallels to what the rest of the world is catching up on.
Brian "Ponch" Rivera:Well, that makes an interesting connection to fine temperament tuning. You can't digitally shift the frequencies. It doesn't work that way. When you digitally do it, it has to be done in situ. When you tune an instrument, is it 440 or 432? See that as a connection to something else we're talking about quite a bit on the show, which connects to aliens. So some other time.
Sean Manion:Well, it's interesting, interesting. We've talked a lot about. We've talked a lot about jazz.
Mark McGrath:Maple shade is a jazz label primarily yeah yeah I think it's got some gospel and some other things, but it's primarily jazz, huh dang, do you know what your cousin and I used to listen to?
Sean Manion:a lot that comes up, that's come up on the podcast. Quite a bit was grateful dead we were, and there's a lot, there's a lot of those concepts in there.
Mark McGrath:Yeah, it takes you into a different dimensionality, whether that's aliens or other dimensional beings. And there's a great, there's a neuroscience slash, psychedelic book called Reality Switch Technologies Love it.
Brian "Ponch" Rivera:We've been trying to get them on the show forever. That thing blows my mind getting into DMT and then to get into affordances and attractor states and fitness landscapes and complex adaptive systems. And I read it. I'm like everybody needs to read this book and of course people are like what the hell are you talking about?
Mark McGrath:I mean we could go in this direction as well. I actually my niece, was an undergrad and she and I and another colleague presented on the medical evidence that's out there or lacking that's out there on ayahuasca at the last Society for Neuroscience meeting and I got to dive deep into some of the literature on ayahuasca and you know what's going on. I think it's the Bronx, va is heading up the psychedelic research area. Rachel Yehuda, who was a PTSD researcher that I cited heavily in my thesis work for the government, is now, you know, on on VA side doing psychedelic research. I was in Baltimore, hopkins had a big community there and so I think I think seeing not only what it can do for people with trauma but also how some of the impacts on the brain parallel different approaches to looking at other technologies elsewhere.
Brian "Ponch" Rivera:Yeah, and I think that reality.
Mark McGrath:switch technology gave me that sense that there's again kind of like Boyd arriving at some of the same cybernetics conclusions without fully diving into that whole library. There's different approaches to the same questions and answers and you can't dismiss the stuff even though it's got a negative aspect to it which we could go into too. A lot of weird things.
Brian "Ponch" Rivera:We can. We've done that quite a bit. So I want to connect a few things to the Alien Switch Technologies book. In there he talks about how we perceive reality, some connections to Anil Seth and the free energy principle.
Brian "Ponch" Rivera:Going back to the psychedelic piece, when I drew the entropic excuse me, the entropic OODA loop, it was inspired by the entropic excuse me, the entropic OODA loop. It was inspired by the entropic brain hypothesis through Robin Carhart-Harris and Carl Friston, right? So that's that neuroscience connection there. And then the way we explain it, when we say we're trying to minimize, surprise, we're trying to get rid of free energy, not get rid of free energy. But the OODA loop exists for one reason and that's to say, let me just say, minimize free energy. That's where we get that from is from the neuroscience of the free energy principle and active inference, right. So my view is that Boyd's OODA loop was on that path and had he been an academic, other academics would have picked it up and they would have made the connection to Bayesian updating. Active inference perception is top-down, inside-out. It's a controlled hallucination.
Brian "Ponch" Rivera:With the OODA loop, the way we draw it, with the boundary and the entropic boundary in it, I can explain ego, death default mode network and how you press that down and gain access to novelty using things like counterfactuals, using the pathway that moves from act back to observe as a pathway. A counterfactual pathway that helps you update your internal orientation so, again, we've been looking at a lot of these things on here actual pathway that helps you update your internal reorient, or your, your orientation. So, um, again, we've been looking at a lot of these things on here and, by the way, that's all connected to cybernetics, complex adaptive systems as well. Um, and and then, on the market side of the house, moose and I and some others are looking at, um, understanding Brownian motion stochastics, using sacred, sacred geometry, which again is connected back to a lot of psychedelic therapy. So, yeah, we're, we're loving that space right now. I think a lot of so much, so much potential.
Sean Manion:I mean, this is this is where, like you hope, your competitors stop at the linear oodle loop. They looked at boyd he's a fighter pilot, he has this little tactical thing and that's where they stop and they don't realize this whole other world that he opened us up to.
Brian "Ponch" Rivera:Hey Sean, in your updated orientation I looked at this and I was trying to understand what memetics is. Am I getting that right?
Mark McGrath:Memetics I was using an older term to probably cover two broad things and just for your audience back in 2017, I kind of got it in my head that you know. Another thing I wish Boyd would have had more time with is to dive deeper into neuroscience. I think he got deep, but you know he was coming from a different place, whereas I was coming from the neuroscience place and saw all these connections, and so I attempted to update some of the small orientation sub-details with some things that may give it a better framing for how it fits into what we know about neuroscience.
Brian "Ponch" Rivera:Let's build on this. This is great. So in there you have biases. Sometimes you call them heuristics. So let's paint the picture. The current orientation, excluding analysis and synthesis and new information, is comprised of three components. It has three things genetic heritage, cultural traditions and previous experience. Help us build out what you would put underneath genetics, to kind of genetic heritage, to kind of try to build this out.
Mark McGrath:For genetics. I think you know that there's a major component of genetics in there, but I think the thing that he you know it wasn't really popular until several years after his death and it's still something we're just scratching the surface on is epigenetics. And I think the epigenetics which, for people who don't know you know your genetics is a, is a, you know, a ACGT code that that is in all your cells and is relatively the same. Epigenetics is changes in methylization spots or areas that, and methylization spots or areas that so your DNA translates to, transcribes to RNA which goes to your proteins that are created. How much of any stretch of DNA gets turned into proteins is dependent upon not only how many copies there might be but also how actively that is accessed, based on different chemical structures in the DNA and how it sits can change during someone's lifetime based on experience.
Mark McGrath:Exposures stress a lot of things, but some of that epigenetic change which isn't in the code can also be passed along through generations.
Mark McGrath:So you hear things like generational trauma. There's been studies where you know you've got you've got stressed rats who have babies and then stress unstressed rats who have babies and they switch the babies with the mother and it's, it's DNA based impacts on how you behave and how you react and how much your body makes certain proteins, and it's it's an area that we I mean there's there's a million methylation sites and that's just one type of epigenetic change on your DNA. So every cell has a million different options of ways things could be different based on things that aren't DNA related Intelligence, how the brain works, what's going on in the brain, is probably far more heavily influenced by that than we realize. So I think that genetic component is. It was smart, it was what he knew at the time, but it's just the surface. So I think, adding to that, I think it was right on the experience, although the experience starts to cross over physically with memory in that respect and we want to be conscious of that.
Brian "Ponch" Rivera:Let's go back to the psychedelic piece here real fast. So epigenetics you brought up trauma can be inherited from other generations. I've learned recently that there's something known as a multiple hit hypothesis, which connects to genetics, culture and experiences, right, so those three things kind of make up how we see the world. It kind of clouds how we see the world from, like a PTSD or even TBI. But the point I'm trying to make here is inheriting the trauma of your ancestors is not proven yet. We're starting to understand that can be carried through the DNA and that could actually cause trauma or the way people view the world. But that slide I shared with you a few months ago has epigenetics as a way on how we explain genetic heritage, right, and some people push back on that and say hey, it's not proven yet. So can you talk a little bit more about where we are on the I mean epigenetics is not fantastical, it's a reality.
Mark McGrath:What and where can be epigenetically translated? That's something that we don't have the full edges of. So sometimes people will take a known or something that's been demonstrated or has got evidential support, like epigenetics, and then translate it to a much broader thing, and that's where we maybe have extrapolated it too far. We don't know how many generations it can pass through, what types of changes occur, what types of things get carried down. The challenge with science is it's very exacting, and then the communication of science is much more of a narrative and those two don't always line up, which is, I think, what they're bringing up. But it's kind of like the psychedelic piece where it's you know, you've got these things that, for whatever reason, people are afraid to touch or afraid to admit, and on the one hand they sometimes are overblown. On the other hand, we're not lacking in some evidence that there is something involved. It takes a balance to be able to say we want to look at this intently and we want to look at this scientifically, but we're not going to dismiss it because there's cultural, dismissive aspects of oh no, it's the hippies taking the drugs again. That I think is happening a little bit.
Mark McGrath:People are very reactionary to having their narrative changed and so when that comes into play, or if they've had somebody push a broader narrative with, well, there's, you know, you know, I'm of Irish heritage and I could say, well, you know, I might, my ancestors haven't been in Ireland for five generations.
Mark McGrath:But I could say, well, you know what the British, you know, there I still have, am suffering because of the what the British did to my ancestors five generations ago. There might be some validity to that. But if I push that too hard, should policy be changed with how we interact with Britain because of that? Well, you know, we need to not take it and extrapolate it too far, but if you're used to me doing that, then anything I say about epigenetics might become well, you know, that dude hates the British and that's all there is. And so we don't want to get too caught up in the broader narrative when we're trying to talk about the small specifics, and we don't want to get too caught up in the specifics when we're talking about the broader narrative. But again back to Boyd he talked about, you know, general to specific and specific to general, with the analysis and with the synthesis, and I think people get tripped up being on either one side or the other of that and it's hard for them to flow back and forth.
Brian "Ponch" Rivera:I want to talk about what's behind you and make the connection back to where it would go.
Sean Manion:in orientation I was going to say something about it till you. Beat me to it.
Brian "Ponch" Rivera:You got the biases. It's the great graphics.
Mark McGrath:That was a Valentine's Day present from my wife. That is he knows me well.
Brian "Ponch" Rivera:Oh man, yeah, where were they? Are they a combo of culture? Is it genetics? Where do those biases fit in? How do they?
Mark McGrath:manifest themselves inside of orientation. It's one of those situations where it's a term and I'm gearing up to bring this to my class and I've touched upon it a little bit. The term bias is used very broadly and sometimes we talk about a social or a cultural bias and sometimes we talk about a cognitive bias, and these are not completely discrete things. But when we talk about a social or cultural bias, it's generally a macro bias against a group of people or against certain things, and it oftentimes it oftentimes is a negative and it almost always has a negative connotation.
Mark McGrath:When you talk about cognitive biases, a lot of times they're just shortcuts.
Mark McGrath:Our brain works on a very low power and does a lot of stuff because we don't have to analyze everything constantly, and that makes us jump to conclusions. And now there are some things, there are some biases you've got it right there. There are some biases that are extremely evolutionarily beneficial, and if we didn't have them we wouldn't be sitting here but at the same time, they can be exploited, they can be exacerbated, and when they are, sometimes they show up as those cultural biases, and so we need to be clear where we're thinking about it. Are we thinking about it at the neuronal level? Are we thinking about it at the brain systems level for how we make quick decisions with limited information, or are we thinking at the policy level of whether or not we've got this bias impacting that policy decision? I think that's one of the reasons we often and in science we do this we assume we've got bias from certain things, and so we look for a consensus process to actually refine the information before we move forward with it and you said sometimes these are good.
Brian "Ponch" Rivera:We talked to a lot of neuroscientists who say that 2% of our body weight burns 20 to 25% of our energy, that humans are naturally lazy. It just saves energy to have these things and that's the way I kind of look at it is when you get into implicit guides to control those habits, those skills, that type zero, type one cognition, that system one thinking, whatever you want to call it, muscle memory, that is a low energy approach to dealing with the external environment. So athletes, dribbling a basketball, hitting a baseball I don't know, hitting a baseball is a good one. But those repeatable things that are built up, those skills that are built up over time, to me they kind of parallel those biases. You know they're formed in our orientation, but they really emerge and manifest themselves through that rapid pathway that takes us through implicit guidance, control and then out to the external world.
Mark McGrath:And potentially, if you change the term muscle memory to muscle bias, it wouldn't be that inaccurate but it would take on a whole different connotation.
Brian "Ponch" Rivera:Good or bad, I don't know. Help me out. I want to make sure we get this right, yeah.
Mark McGrath:I don't know. Did you have something you wanted to say there?
Sean Manion:No, I mean like the implicit guidance and control aspect of that is huge. I think that that's probably the most misunderstood thing about Boyd, generally speaking.
Brian "Ponch" Rivera:I never heard muscle bias, I just want to make sure.
Mark McGrath:I just made that up.
Brian "Ponch" Rivera:Okay.
Mark McGrath:Because the way you were describing it, I was thinking of the. You know we often call that muscle memory. But if we called it muscle bias, because it was just a, you know, efficient way we move, it would take the same connotation as the cognitive bias in many respects.
Brian "Ponch" Rivera:There you go, it's an efficiency.
Mark McGrath:And yet it wouldn't have a term that you know, if we called it muscle memory. If we call these cognitive, you know memories gets confusing because we do do memory in the brain. But if we call them cognitive repetitions or cognitive shortcuts, people would get less up in arms about them in their own right because they wouldn't confuse them with the cultural, the social bias. That is usually what the problem is.
Mark McGrath:Yeah but they can lead to that. You know, I think you've got muscle, you know, think about someone with PTSD. Maybe while they're sleeping, they're having a nightmare and they, you know, reach out and they hit their spouse. I mean, that's that's. That's a muscle bias, that's a muscle memory. That's going on. There's no intentionality to do harm, but in that case some harm can be done because of what that individual is trained to do when they're stressed, which is to respond in a physical way. And so I think there are examples where, like a cognitive bias translating to a negative societal bias, you could have muscle memory translating into this negative connotation, but we just don't call it bias.
Brian "Ponch" Rivera:I'm curious about this. Back to the genetic side, genetic heritage. Is there something in our DNA or bodies that prevents us or filters information from the outside world and I might be talking about the reticular activating system or something like that but don't we have a filter that keeps information from flowing in?
Mark McGrath:Well, the thalamus is sort of the gatekeeper of information. Well, four of our five senses, all the information that comes in, goes through the thalamus. It's a grand central station, as I refer to it as olfactory. Our smell does put stuff through there, but a parallel path goes, like directly to the amygdala, which is why you can smell something that's a threat and humans don't have a threat response to smells as quickly as some other animals which must much more robust smells. But you can have an immediate fear of a fight or flight response to a smell because it doesn't even have to go through the thalamus. You can also have you know we get nostalgia with that it can kick off emotions without going through that thalamus, that area that's doing the initial processing.
Sean Manion:Isn't smell the most? Isn't that the most memory-inducing or the most triggering?
Mark McGrath:That's why you could put it as it's not being filtered, it's going direct to the reactive mechanisms of some of our emotional memory. But at the same time you do have some initial we'll call it initial screening of information flowing in from all of your other senses before it starts being processed. What that screening does and I said the term screening, that's probably inappropriate or incorrect in a lot of different ways, but it's going through the thalamus and it's the first step of processing and so that processing we don't have a full understanding. I mean, what would happen if we didn't have a thalamus? Well, I don't know that. We've got those case studies to understand that.
Mark McGrath:Could we start to computationally look at that when we have fine-tuned enough imaging or neurophysiology testing? You probably could. Maybe it's out there and I just haven't come across it. There's a lot of new work being done every day, but what happens if you stop screening that out? What if you change your screening mechanism to mine or vice versa, because I have a different epigenetic profile than you do? Or we could be twins and then be raised separately and I have a different epigenetic profile, and does that change how sounds are processed in my thalamus before they even make it to my auditory cortex. I don't know the answer to that. It seems like it's possible, but that's a hypothesis. I don't know whether it's true or not.
Brian "Ponch" Rivera:Let's continue the going around orientation with genetic heritage, cultural traditions and previous experience. Anything else you want to add that you think Boyd could have really put into the orientation? By the way, I don't want to say, well, let's pack orientation with everything. It's just what else would influence Boyd to have put genetic heritage, cultural traditions and previous experience in there?
Mark McGrath:I did use the term in almost a double entendre with using mimetics with an MI. Mimetics refers to sort of like representation in art and those things versus memetics, which is another. I've seen those misspelled or deliberately used similarly sometimes, and I think I was doing the same it was 2017, because I was interested in memetics, what the neural underpinning of memes and how. Not means necessarily in the modern computer, on the internet sense, though that captures an element of them, but in the in the Richard Dawkins who, who first coined the term as being. You know, if I watch you do a task or create a tool and I remember how to do that. That's a meme. That's a transfer of information. Now, what we see in these pictorial memes is like an image meme that translates to some, often mocking or some assessment of a situation in a way that then I can quickly translate to you. But those all, that whole collective, what are the things that we transfer to each other? Both, you know, religion is a meme. Tools is a meme. Cultural, you know, ideas, is a meme.
Mark McGrath:If you look at the original use of the term meme, and even including the newer use of the term meme, which is probably taken over for it. There are components of what's been put in my head that could be called experience, but I think needs to be thought of distinctly. So thinking of the genetic and epigenetic thinking of experiential aspects and then thinking of these, call them complex memory, or sometimes in memory research is called engrams, like the E-N-G-R-A-M, like the memory engrams of where a memory lies and we don't know. It's a pattern, it's something that moves about in the brain. There are complex engrams that are included in that memetics idea, that are different than the contextual aspects of an experience or the specific details of an experience or that more general genetic and epigenetic aspect. So that's kind of what I was going for. It wasn't so much adding to what Boyd had done, so much as starting to detail and articulate it.
Brian "Ponch" Rivera:Yeah, I like that, I like that type of thinking, so using a triad. So when you're talking about the let's see memetics the one with the eye that to me it kind of sits between genetics and culture and previous experience, right, it doesn't, it doesn't fit in one box completely, and that's what I. That's why I put the triad out there to kind of explain things that you know. Epigenetics are not just one thing. It's influenced by the external environment, those previous experience. How many hours are you putting on your iPad a day? That type of thing, right? So there's a nice triad approach to kind of get people to go okay, it's not one thing, it's not just this, it's connected to maybe two or three, any other things you would put within previous experience, cultural traditions or genetic heritage.
Mark McGrath:There, no, but I think the you know, I kind of you know, did that just kind of. There's the layer of the input coming in, there's a layer of the output going out, the analysis and synthesis.
Mark McGrath:I think there's definitely almost a repeat of some of those components that are going to be an either. Actually, your question before, how information is processed and can you know generational trauma change how that initial input is processed? I mean it might be, but also, you know, do we tend towards analysis? How are we doing analysis? Do we tend towards synthesis? How are we doing synthesis? Can we jump back and forth those things together?
Mark McGrath:I think that there's complexity in each of those side layers that mirrors some of the central complexity too. So you have sort of a center pool of, or center filter of. It's going to go through these three filters, but the conveyor belt on either side is going to work in a certain way that's also impacted by the filters, or the same way the filters are impacted Makes sense, yeah, and so I think there's nuance and details that can be teased out there, both at the psychological level and at the neurological and neuroscience level. And this is this is getting back to some of the discussion about ai we've kind of jumped the gun with ai and talking about intelligence. We don't know what intelligence is. We don't have a good way of doing it. We go with the I everyone's oh. The iq is garbage, oh, but we only have one way of doing it. There's like what the waste for?
Sean Manion:that's one. It's not one of your titles of your articles and says stupid or something and I.
Mark McGrath:I drew that from a book that I reference in there, where there's a french philosopher.
Sean Manion:What's the?
Mark McGrath:proper title, that's 8.1 of the brains are complex. Intelligence is stupid, yeah okay and the next one coming out is actually called ai ain't intelligent, because it simply isn't. And we talk about the term idiot savant, where my calculator is smarter than me and putting math together. But I'm not going to be like oh my calculator is intelligent.
Sean Manion:No, it's not.
Mark McGrath:It does one task way, way better than me, and maybe some of these things do a couple of tasks. But here's the thing Humans are really good at fooling themselves. We are. I mean, if you look at any of like you know, persuasion and how, like you know, they used to run seances and trick crowds with mediums like we're doing that to ourselves with these LLMs to some degree, and these are really smart people, but they're really narrowly focused in their intelligence and I'm talking about some of the groups that are not just hyping it up but are like, oh my goodness, this thing is rational. Well, I mean, in the 60s, carl Sagan.
Mark McGrath:Now Carl Sagan is famous for saying you know amazing, you know, statements require amazing evidence, or forget the exact term.
Mark McGrath:He was a scientist, he was evidence based and he looked at ELISA, which Eliza was like a psychotherapy, not an LLM, but effectively a psychotherapy AI tool.
Mark McGrath:That had people wowed and even people who knew it was a tool started revealing all their secrets to it and he said this is going to change everything with psychotherapy, there's not going to be psychiatrists or psychologists anymore, there's just going to be these computers doing it, because it's that good Novelty, especially, especially, really gets us.
Mark McGrath:Our brains really get tricked by novelty, and then we want to believe that there is something on the other side of that box. And I think a lot of what we've done is take a very effective writing tool for really middling writing and decided oh my goodness, it's got these amazing abilities to it in the same way that people used to get. You know, there's a sucker born every minute, and the circus used to make a bunch of money off of them. And I'm not dismissing the importance of what we're creating, but I'm saying that we haven't even gotten close to what is intelligent, because we don't know what it is, and so we're just targeting one system and you had said earlier the modeling after the nature, before we build something. So a lot of people say, oh, but planes don't fly. You know, airplanes don't flap their wings like birds, so therefore an LLM must be super intelligent.
Brian "Ponch" Rivera:They sure look like whales, though, don't they?
Mark McGrath:I like that, but we also understood flight a lot better, and there's like four components to flight that need to be matched. With regard to intelligence, we don't even know what that is. So, you know, I could, I could, you know. Oh, this pen is intelligent because birds flap their wings and airplanes, don't that? That doesn't ask for a logical argument.
Sean Manion:Yeah, it's kind of like, yeah, you'd said something. I was just trying to think we're trying to avoid myopia, like we're trying to. Who's going to decide what comes out of an LLM? Who's the ultimate arbiter? If something's not open source, or if something's not at some point, someone's going to have control over it.
Mark McGrath:And I mean we could going back to the discussion of AI and having data, being able to track the data that goes into an AI, the model training as well as testing, as well as decision making, what the weighting is for all of that. We spent a lot of time with that in my previous company and you can do that. It's going to take some time and some structure, but it also includes the individual, and the individual is self-sovereign, being able to make a decision or be part of the general consensus decision that is being made. If the model is making a decision, I want to know what it's making that decision based on, when it was last updated and who set up the governance for it. And that's what you're getting to. Most is who's setting it. And in the brain there's lots of governance that includes a variety of cells or a variety of structures involved in how the patterns are passed around to make an action or make a memory. We have very good for us historically, governance we build into our laws and things like that, but it can be so much more fine tuned once we start to connect things and, like it or not, sooner or later you've got not just through a keyboard but through neurotechnology we're able to.
Mark McGrath:You know, let's say we all live on a block with 20 people and we have to make a collective decision what time in the morning someone can start up their lawnmower. Well, right now it's like don't start up the lawnmower before nine o'clock. But if we get to a point where, like during the summer, when it gets hot out really fast, as long as we've got 17 of us that ping on, you can start up your automatic lawnmower. That goes on over there, we can start to have decision making collectively in a way we haven't had before. And I don't mean from a, you know, a socialist standpoint. People get political.
Mark McGrath:When you say collective, I mean from a group consensus standpoint, based on a governance we decide upon. And you could say 17 out of 20 is enough to start start up the you know the lawnmowers, but it takes 20 out of 20 to make a more important decision. You can grade decisions. You can have a lot more nuance to decision making instead of again relying on a third party with a delay in incentivization. Like a politician, you could start to set up you know group algorithms to be a representative. There's even a movement out there called like AI in 2030, or something like that, where replacing, you know, replacing some representatives with AI decision making that's just fed in by its constituencies. Could you have that in real time? Could you have that with an appropriate level of feedback? Would it be less corruptible? Would it be more reliable?
Sean Manion:Hey Sean.
Brian "Ponch" Rivera:I have to drop in a minute.
Sean Manion:Yeah, we're going to pause round one. We're going to send people to the unjournaling.
Brian "Ponch" Rivera:Actually, I want to try something again. One question to you. Given your knowledge of neuroscience and your understanding of the current landscape in AI, can you give us some insights? What are the good technologies? What are the emerging things? What are those things on the edge that people don't know about that? May that excite you, I guess.
Mark McGrath:I mean, it's not brand new, but it's less appreciated than it should be is the area of neuromorphic computing. Neuromorphic computing is trying to learn from how the brain works and we're really at a nascent stage with it but meme resistors and things that can change their state rather than just be on off of current transistor status. So the neuromorphic computing world is amazing. The brain is a digital analog hybrid in ways that we, you know, have known for years, and yet we're relying fully on digital. I think the hybridization between digital and analog computing is something else that is coming. And finally, I would go with.
Mark McGrath:You know, we've got all of this connectionism with the machine learning that has come back into vogue but was really set on the chew gum at the same time. It's not going to be one or the other. We've got several dozen systems and multiple subsystems in our brain. One algorithm isn't going to replace that very effectively. So there's going to be a variety of integration of different types of systems and I think you're going to see connectionist, symbolic and other. We don't know yet that'll probably come into play. So that's where the sort of three things that I would say definitely have the cutting edge along with you know a federated or distributed data approach, the blockchain, auditing and incentivization layers all of these things come together.
Mark McGrath:So you know the future is a five lane highway. If you're just looking at one lane, you're going to get run over.
Sean Manion:We'll do this again, for sure, sean, and we're going to send everybody to the unjournaling.
Mark McGrath:Thanks for coming on and we'll send people your way to both the unjournaling and to your ex. I appreciate that and I thank you for having me on and actually kind of bringing me back into the next phase of me diving into Boyd, because I threw a question out on X and it was, you know, because I looked at Boyd in cybernetics and you were like yeah, it's in a Senga and I was like I gave my. I have a bad habit I give books away when I really like them, and so I didn't have a physical copy anymore and I went into the electronic copy. I was like, oh yeah, there it is. So good stuff.
