Conversations on Applied AI

David Quimby - Systematic Innovation and the Art of AI:

• Justin Grammens • Season 5 • Episode 7

Share episode

The conversation this week is with David Quimby. David is a principal at Innovation Radiation, specializing in systematic innovation, experimental design, and technology forecasting. He is a patented inventor in web architecture and user experience, and co-founder of the Minnesota Change Management Network. David holds a BA in Mathematical and Developmental Economics from UCLA and a Master’s in Organizational Behavior and Sociotechnical Systems from UC Berkeley. With extensive experience in technology analysis, consulting, and product discovery, David is passionate about bridging technical and social systems to drive innovation. He has been an active member of the applied AI community, sharing his expertise at events and workshops. He is dedicated to advancing the field of AI through both practical application and thought leadership. 

If you are interested in learning about how AI is being applied across multiple industries, be sure to join us at a future AppliedAI Monthly meetup and help support us so we can make future Emerging Technologies North non-profit events!

Resources and Topics Mentioned in this Episode

David: [00:00:00] One of my favorite quotes of Albert Einstein is we can't solve our problems with the same. Kind of thinking that created our problems, and that's really kind of my impetus towards thinking more systematically versus kind of some of our conventional ad hoc methods. Albert Zen Yogi, he won the Nobel Prize in physiology and sometime in the thirties, and his quote is, discovery is seeing what everyone else has seen.

And thinking what nobody else has thought. Part of this fabric as well is how can we see things that we're missing? A lot of things are hidden in plain sight. Both, both problems and solutions. How can we identify them better? 

AI Announcer: Welcome to the Conversations on Applied AI podcast, where Justin Grahams and the team at Emerging Technologies North talk with experts in the fields of artificial intelligence and deep learning.

In each episode, we cut through the hype and dive into how these technologies are being applied to real world problems today. We hope that you find this episode educational and applicable [00:01:00] to your industry, and connect with us to learn more about our organization at applied ai mn Enjoy. 

Justin: Welcome everyone to the Conversations on Applied AI podcast.

Today we're talking with David Quimbee. David is a principal at Innovation Radiation, where he practices systematic innovation, experimental design, and technology forecasting. He's a patented inventor in web architecture and user experience and co-founder of the Minnesota Change Management Network. He's earned a bachelor's degree in both mathematical and developmental economics at UCLA and a Master's degree in organizational behavior.

And sociotechnical systems at uc, Berkeley. I've known David for years, but I actually found out some little known facts about him recently. He's climbed Mount Whitney, done trekking in Thailand, photographic safaris in Kenya, sea kayaking in Alaska, and has traversed the US and Canada by motorcycle. So that's amazing, David.

All of this adventure I know will lead to some rich conversations as exploration and innovation go hand in hand. So thank you, David for being on the podcast today. 

David: Thanks [00:02:00] Justin. Uh, really glad to be here. Well, hopefully we'll connect those dots around exploration and breaking through barriers. And thanks for doing the podcast, Justin.

I really appreciate the way you build community in the Twin Cities. I like to say technical innovation is 80% social, so you really are a boost in that regard. 

Justin: Thank you. Yeah, no, excited, excited. I always learn something from everyone that's on here, so it's one of these things that I just enjoy getting people on the show and sort of talking about artificial intelligence and then.

You know, AI is such a broad topic, so it spans out into so many different areas. So systematic innovation, I know we'll end up connecting the dots to that. So I talked a little bit about, you know, sort of where you're at today and the innovation, the systematic innovation work that you're doing. But maybe you can sort of lead us a little bit on a journey with regards to how you got to where you are.

David: Sure, thanks. And, um, since we're gonna be talking systematic innovation, I'll probably take a little bit of a systematic approach. I'm not sure I can do anything unsystematically, quite honestly. Probably a little bit compulsive in that regard, but I think of some of these disciplines as building blocks or it is, there is a kind of a method to the madness.

I like to say life [00:03:00] is lived, looking forward, not looking backward, but when I look backward, I kind of see. Oh, I was like, this is what I was doing. I was kind of scheming. I was kind of like putting together some building blocks to enable me to get to some destination, which I'm not really sure. I knew the destination.

That's kind of this notion of intentional wandering, which will come back to systematic innovation, but I'll be maybe a little bit systematic about the building blocks that I see as forming this direction. I like to think that I've, to some degree, bridged the web and AI areas to some degree, and I think we've noticed there's a few, few members of our.

Community that have had that same type of journey. So as I look backward, I almost see this kind of metacognitive orientation that I was headed forward in a direction and I was kind of building some building blocks. I call it an interdisciplinary approach or intellectual diversity, or call it what you will.

So all I was ever gonna be for the longest time was. An architect. I took architectural design, you know, architectural drafting [00:04:00] courses in high school and that sort of thing. I went to the University of Minnesota. I was very fortunate to go through the pre architecture program. I got into the architecture school.

I kind of thought, wow, I've arrived. I'm an architecture school, I'm gonna be an architect. And then I started realizing. This is not ringing my chimes. And if I reflect on it, I think it wasn't, I was not getting kind of the theory. It seemed like ad hoc to me and it just seemed like, oh, this is good design.

That's good design. The history was fascinating. Studying architectural history was absolutely fascinating. It was a great program. It just, it wasn't really ringing my chimes from a theoretical perspective and I said, well, I think I need to. Make an adjustment here. And so I actually had this kind of vague notion of I'd hitch up the wagon's head west.

I was gonna go to architecture school at uc, Berkeley was my next sort of step in the process. And there was a guy there, his name's Christopher Alexander. He was doing something called design patterns, which will come to later, and you may have heard of him in the software engineering realm. He originated this whole concept of design patterns, which [00:05:00] were not.

Getting much traction in architectural design. He was a professor of environmental design, uc, Berkeley. But it just, there was a lot of resistance, there was a lot of pushback because, oh, architectural design should be like ad hoc. It shouldn't be like systematic. But that was kind of in my genes at that time.

And due to various circumstances, I ended up making a pivot after heading west and I was gonna work in California for a year, get residency, tuition was a lot cheaper, that sort of thing. And I ended up. Pivoting down south and going to UCLA studying, switching my, my major studying mathematical economics.

So that was sort of another kind of node on the journey. Yeah. Not switching up too much. You're not doing any zigzags are we? But I really, I'm very analytical. I'm very quantitative, but I also had an interest in the social sciences, which we'll see come out later in terms of like organizational behavior, things like that.

My first job outta college was, oh, I think I'll become a systems analyst. And, uh, I'd had a. A programming class or two things like PL [00:06:00] one and Fortran and uh, I ended up being very fortunate, pivoting into Bank of America where it was actually the largest. Very large scale environment, largest data center in the world at that time up in San Francisco.

And I started moving more in a systematic direction in that regard. They were looking for, interestingly enough for broader background than purely, you know, computer science and talk about a confluence of events and there you go, metacognitive orientation. How did I know and how did these events converge?

But it was opportunistic and I started getting more technical after working as a systems analyst for a while, which. I guess these days we'd probably call it more like a business analyst or business architect. Then I pivoted into even more technical, getting into software engineering, developed some pretty large systems at Bank of America, helped develop, you know, getting into large scale system development.

Then my next pivot, getting, uh, itchy for the next move. I pivoted over to Deloitte Consulting and what I got to do there, based on my background was. Really mostly focused [00:07:00] on what I'd call technical and economic feasibility analysis, mostly for the US Air Force, but looking at feasibility studies for very large scale development efforts and kind of adding that longer term sort of orientation perspective, that sort of thing.

And by this time, after, you know, couple years doing that, I was starting to see this tension between. The technical side and the organization's organizational side, or the social and cultural side. And I thought, where do the twain meet? Are they meshing, are they coming together? And that sort of like became my focus.

And so pretty soon it's time for another pivot. Right? And I decided I wanted to, uh, go study, uh. Organizational behavior in what they called sociotechnical systems. And I was fortunate to get into uc, Berkeley, and do that program. Technically it was an MBA program, but I, I like to think of it as a master's program in organizational behavior.

And so added that node we're kind of building the building blocks. And again, I'm not sure you know exactly where things are headed, but I'm kind of having this vague notion. On the horizon and kind of pulling in building [00:08:00] blocks to move in that direction. And after uc, Berkeley, my first job was over at Stanford Research Institute as a, uh, technology analyst doing technology forecasting.

And my focus was in emerging technologies, really oriented towards a. Artificial intelligence, believe it or not, of all things. So that would've been like early nineties. And my orientation was what we'd call knowledge-based systems or expert systems, inference engines, whatever you might call 'em. We were also getting early in neural networks at that time, and then speech recognition and natural language.

And as you may know, SRI is one of the origins of. Technologies like sir, it was really a real leader in, uh, speech recognition and some of that stuff at the time. I also had the good fortune to collaborate with a guy named Doug Engelbart at SRI and there the focus was social computing and what was then called groupware and eventually became social media.

And, um, he's the inventor of the graphical user interface and real orientation towards what I'd call human-centric design. You know, I almost think of him as the [00:09:00] really, one of the. Early pioneers in that whole notion, and he had led something at SRI called the Augmentation Research Center. He was real interested in augmenting human intellect, was kind of his focus.

So all these nodes, hopefully the breadcrumb trail is leading closer to where we stand today. The next pivot was having an opportunity to go back to work at Bank of America with some previous colleagues. And working in what was kind of an autonomous business unit, sort of an exotic business function, large multi-entity transactions that were like really complex and needed a little more advanced technology.

And I had the opportunity to bring some advanced technology in there and serve as technical and architectural lead to introduce things like object oriented programming systems and client server architecture. That sort of begins my segue. Toward the web. The orientation towards client server architecture was kind of like a natural pivot into web architecture.

And then combining that with human-centric design is what led me to thinking about problems on the web and [00:10:00] what was next. And at that point, I pretty much, that's really the last. Job I've really had, I've done some contract work, you know, at Best Buy in emerging technologies and around product discovery, that sort of thing.

But I really set off on this journey, unknown to me, but, uh, I call myself an accidental inventor. I invented a web presentation technology or, or a tool for web presentation, and it kind of combined the backgrounding in client server architecture with the whole notion of architectural design and moved us into this era of.

Inventing and realizing, oh, there was like a method to the invention, and maybe I'll take a pause there. Hopefully that wasn't too long-ish, but it gave us a little sense of the building blocks, and that'll kind of bring us up to this whole notion of what is systematic innovation and how does it apply to ai, if that makes sense.

Yeah, 

Justin: no, it does. It does a lot of vocabulary, a lot of terms, I think we have to unpack here as we're sort of talking through this, we've got things around [00:11:00] metacognitive orientation, intellectual diversity. But as you're getting to systematic innovation, I think that is a great place for us to start because that seems to be sort of at the core, the things that you're focused on today.

I think probably a lot of our listeners, maybe they hear the words systematic and innovation, and they try to put those two together, and maybe you can help us understand what that means when we're talking about systematic innovation. 

David: Absolutely well pointed. So the first thing I want to do maybe is offer a couple of quotes just for perspective, because they kind of set the stage.

One of my favorite quotes of Albert Einstein is we can't solve our problems with the same kind of thinking that created our problems. And that's really kind of my impetus towards thinking more systematically versus kind of some of our. Conventional, you know, ad hoc methods. And then I'll mention a quote from an inventor who's much more obscure.

His name is Albert Zen Georgi, if I'm getting that right. I think he was Hungarian, he won the Nobel Prize in physiology and sometime [00:12:00] in the thirties. And his quote is, discovery is seeing what everyone else has seen and. Thinking what nobody else has thought, and that's part of this fabric as well, is how can we see things that we're missing?

A lot of things are hidden in plain sight. Both problems and solutions are hidden in plain sight. So how can we identify them better, articulate them better, if that makes sense? It's kind of a segue. 

Justin: Yeah. And so there's an approach here that you're saying there's like a systematic approach for us to solve our problems with a different kind of thinking.

David: Absolutely. That's kind of, yeah, setting the stage to get a little more particular, I'd say systematic innovation is a more structured approach to generating and implementing. Ultimately, I'd say product and service value doesn't mean it doesn't apply to non-industrial domains. We'll get into that realm, but maybe for starters, it's a structured approach for generating product and service value that may be unseen.

I like to say unlike a spontaneous creativity or ad hoc brainstorming, which they can be effective, you know, there's nothing wrong with them, it's just are we being as [00:13:00] effective as you know as possible, systematic innovation uses more deliberate. Principles, methodologies, and frameworks to be more consistent about producing meaningful and impactful value.

If that helps. And I'll, I'll introduce a couple terms here and then I'll try to simplify them. You'll hear me using, and they're built into my method using a lot of terms, and they end up tying back to classical philosophy. I think there's more need for philosophy in ai. In innovation, in tech in general.

I think STEM needs more philosophy and it, and I didn't realize, I wasn't deliberately incorporating these capabilities into my method. I kind of realized like retroactively that, oh, I was using these, like these capabilities. I'd somehow learned them somewhere along the way and I'd kind of built them into my way of thinking.

But I wasn't consciously building a method that used these capabilities. I'll call the two terms I'll mention here are ontology. Morphology, and I'm gonna say they come from, you know, ontology comes from Aristotle and he was of course a student of Socrates. And [00:14:00] morphology comes from Plato, who was a student of Aristotle.

And ontology simply means, in the simplest term, we could use it. I mean, there's more nature, more texture to it, but we can think of it simply as categories and how do we look at information through categories and existence, through categories. Morphology, in its simplest terms. Just means forms, and I find these to be useful levers for thinking systematically.

Another thing I'd mentioned about systematic innovation is it's a stepwise process versus, say an event like an explosive event of like idea occurs. Here it's, no, it's very deliberate. It's a stepwise process. We might say, you know, even kind of a journey and pick your stages and they may vary depending on the method.

But it's a deliberate stepwise process. And I'd also say it's relatively problem focused, and I, that may sound like, well conventional, but I think it's differentiated relative to conventional problem solving. I think we tend to sometimes. Slide over the problem [00:15:00] definition and it prevents us from really solving the problem or even solving the right problem.

It's very expensive to solve the wrong problem, and I think systematic innovation is better about being very problem focused and not proceeding until we know we have a problem definition. And what I would like to say even further, unambiguous problem definition, and I could give a couple of examples of methods that are out there, if that might help.

Justin: Yeah. Yeah, for sure. I think you said something key there, like, it sounds like in its crux we need to make sure that we're Well, a, that there is a problem there, I guess, right? Is what you're saying. I mean, I talked to a lot of companies. That are just building AI for the sake of building AI or trying to implement a technology solution when actually the old tried and true method in some ways continues to work just fine.

And sometimes you're trying to fit a square peg into a round hole. And what can be difficult is not only that maybe the technology is there and it's available, but there is no return on investment, right? There is no clear path forward of like doing [00:16:00] this is actually gonna save the company more amount of money.

And so it, you end up kind of pushing a rock up a hill. And everything ends up kind of going sideways on these projects. But yeah, that's, you know, that's kinda what you're saying is think more of a systematic approach. There was some examples that you were gonna share. 

David: Well, and I think, again, you're very pointed in your perspective, you know, Justin, in terms of, and almost sounds like hammer and nail.

As technologists, we love to use technology. We love to point that technology in any direction that may work versus almost like working backward. You know, start from the problem. And for some technologists it's a real. Tough thing to do. It takes a lot of discipline to say we aren't moving off the starting line until we have a problem.

And until we all agree on the problem and the problem definition, and then we'll, you know, then we'll work backwards from there toward a solution. But we have to agree that we have a problem first. So anyway, just that concept right there. I wanna point out kind of as we go, that's a systematic thought that you just expressed and that we just kind of agreed on.

'cause we're looking at the large system, a principle [00:17:00] of. Systems engineering. And of course there's a lot of systems engineering that comes into systematic innovation. Every system exists in a super system. Systems are composed of subsystems, they operate within a super system. And in this case we're talking about a systematic approach to the holistic situation, the problem and the solution, the social or organizational context, and the technical approach that we're gonna take to solving it.

And that's very much systematic thinking. They're just saying we're gonna start with having a problem before we even move forward. Okay. So couple examples that I'll mention just in case they're kind of like landing points for the audience to explore. An author named R Van Wick had a method that I was kind of exploring in this realm, in the genre, and he wrote a book on the core theory of technology.

He was looking at like, what is technology broadly what cuts across all technologies? And he stumbled across a presentation I was giving in one of the technical communities where I'd done. This view, I guess I called it a palette. I didn't call it the Quimbee palette. He called it the Quimbee Pal [00:18:00] in his book.

Interestingly enough, and I did, I kind of mentioned and connected the dots among some of these genres, and I said, I think it'd be actually a systematic innovation. Endeavor to think about these things systematically and connect the dots among some of them. But one of them, I've already mentioned design patterns by Christopher Alexander, where he's looking at, you know, kind of the combination between modular construction or modular approach to things.

And yet we can still generate a profusion of complexity with. A modular approach, but we still have kind of an underlying backbone, that sort of thing. Maybe the essential idea under design patterns. Fascinating field. Christopher Alexander, fascinating guy, might be interesting to the audience. Another one that comes to mind is lateral thinking.

Edward de Bono, kind of one of those creative. Thinking techniques where you look at, well, what are the adjacent ideas? What's to the right, what's to the left, what's ahead, what's behind? Again, fairly deliberate, fairly specific. There's a, a technique called Riz or [00:19:00] TRIZ. It's a Russian acronym, stands for Theory of Inventive Problem Solving.

It was originated by a Russian, imagine that probably in the sixties, maybe even seventies, he looked at the Russian patent database, took all the patents and distilled from them what he called the inventive principles. Other one last one I'll mention 'cause it lands pretty close to my method, is called a morphological analysis developed or originated by a guy named Fritz Wiki.

Interestingly enough, he's a Swiss astronomer or was a Swiss astronomer, and he was looking at kind of like the way I defined morphology, earlier forms. He's looking at like, how do we look at forms? How do we identify things in terms of their forms? How do we pivot off? Different forms to produce new forms, things like that.

And so those are maybe a few little more practical views on types of systematic innovation that are currently in use. 

Justin: Yeah, no, this is great. We'll be sure we put all these links to these books here in the liner notes, so if anyone that's listening or whatever, I'll get [00:20:00] these there so people can go off and explore on their own.

David: It's a little bit of a deluge, I suppose. Like I said, we got a long way to go, a short time to get there. So, and maybe a couple things I'll mention at this point, because the audience might be interested in, well, like, so what, what is the benefit? Or what's the advantage or the differentiation? And I think I've maybe intuited to some degree.

By the way, another thing I might mention. For helpful clarification is we kind of know what innovation means, or we have ideas of what is sys, uh, what does innovation mean, and we have some ideas around like, what does systematic mean? And to a large degree, we are just combining those terms and we're saying, okay, we're doing innovation, but we're doing it.

Systematically, and that may sound like a paradox or even like oxymoron. I'd like to think of it as a synthesis of two different types of things that we're kind of meshing together. And I'll be getting to synthesis when I talk about my method of systematic innovation. I'll also mention that a lot of conventional innovation.

Might actually be subconsciously systematic when we take things apart. [00:21:00] I've had the experience of taking apart, like for example, I looked at the graphical user interface and I said, well, if Doug Engelbart was such a freaking genius, what was the systematic innovation or the systematic. Method and in, in my approach, I'll be talking about opposing forces, and in retrospect, I was able to look at the graphical user interface and see that it solved a contradiction or it identified and solved a set of opposing forces.

Now, I'm not so for a minute saying, and I, I, I knew I had the good fortune to know Doug and collaborate with him, and I'm not saying he thought about the graphical user interface systematically. The way I'm gonna talk about it anyway doesn't mean that he wasn't subconsciously systematic. And that maybe gets into this, what I'm really talking about when I talk about metacognitive is I'm really talking about beyond the conscious, maybe not sub, but super conscious.

And, um, anyway, so a lot of conventional innovation may actually, in retrospect, be systematic. Sometimes when I talk to entrepreneurs, I look at their idea and I can actually decode it [00:22:00] into. The contradiction that they solved the way they're actually thinking systematically, even though they're not conscious about doing it.

And I guess my segue would be. It doesn't mean we can't become more systematic in our way of thinking, and what I think that enables us to do is become more continuous innovation. It enables us to become more distributed. I think innovation can go much more broadly across the organization than it does.

I think those who are tasked with innovation can be more continuous or more predictable to some degree with innovation. I think systematic innovation potentially produces more disruptive innovation or more radical. Divergent innovation. And I also think it enables us to be more collaborative because we have a more kind of codified concept of what is the problem and how are we approaching the solution.

Justin: So, quick question here that, 'cause Yeah, we are gonna start talking a little bit more about some of the AI concepts, right? We're gonna talk about alignment and re and stuff like that. But 

David: Yep. 

Justin: You got me thinking. Are there any cases where you don't wanna use this? [00:23:00] 

David: Wow, that's interesting. I've been asked the question like for example.

What if it takes too long? And my experience has been to a large degree, maybe it's kind of a platitude, but there's never time to do it, right? There's always time to do it over or do it again. And you know, just like, as you know, it's very expensive to solve the wrong problem. And so I guess that would be a possible.

Objection or a possible resistance as well. It's too formal. It takes too long. I'm skeptical about that. Possible Objection, and you got me thinking about, well, what other reasons? That's the only one that jumps to mind as a possible concern is well, do we really have time to be formal or systematic? 

Justin: Yeah, I guess that's what I was just kind of getting at.

You're saying this should be used everywhere, and are there particular cases where. Uh, this idea of just a random walk that you, you would just kind of experiment and try around with other stuff. There have to be some sort of use cases out there where potentially it would be fine. Right. And some of it is like [00:24:00] if the company has unlimited budget, they don't have any goals really set.

Like it's just kind of. Exploration for the sake of exploration. Maybe that's what I'm saying. 

David: Ah, well that, yeah, that kind of gets me into this hybrid that I call intentional wandering and how the two come together there, it comes in waves, it unfolds in layers. One thing I should be careful to articulate and, and some of this will come out as I talk a little bit more about this specific method that I've kind of developed, is it's not only systematic.

I'm actually combining elements of randomness. And how do you do that? That's almost like a third level. It's almost like, well, we don't wanna be. Purely random. We don't wanna be purely systematic either. We also want to introduce elements of randomness or non determination or, or that sort of thing. And I'm gonna say at a metal level.

Mm-hmm. We got Russian dolls here. One within another. I'm gonna say at a metal level, the concept of integrating. Systematic and random is a systematic process. And, and I'd like to think I've done that. It's kind of a layered [00:25:00] architecture. Yeah, sure. But, but there is a random element in it, so. 

Justin: Yeah. Yeah.

Well it's uh, that song by rush. It's like, uh, even if you haven't made a choice, you've still made a choice. I think it's free will, I think it's a song by rush. You know, so it's like interesting. Even if you, even if you haven't made a choice, you still made a choice. So that's what you're saying is like basically combining these together, you've actually created a new systematic approach.

So let's, let's talk about your systematic approach. 

David: I guess I'll be a little bit systematic in explaining a bit, hopefully fairly compressed. I saw the contradiction, I like to say early in the 21st century and. I saw, I was looking at the web and I'd had some background, as I mentioned with, with Doug Engelbart, and I was thinking, you know, human-centric design, user-centric design, and I'm going, what are the problems in the web?

And search was being solved at that time. And I saw a problem with navigation or presentation and the underlying forces that I saw being kind of intention, where the notion of web content is very granular. It's very, very fluid, very modular. And yet, how do [00:26:00] we view content in streams? It's this clunky kind of point and click, how do we put these objects together?

And on the other hand, we could, you know, we could view animation and video and movies in streaming format, but it wasn't very modular. And I said, well, I want some combination of both. I think that would be like the next generation of being able to do content on the web. And of course everybody told me, well, that's impossible.

I said, yeah, you're probably right, but it's worth trying. And I just kind of imposed my architectural will on the situation. And that's where architectural, you know, design and architectural thinking kind of came in. And lo and behold, I was able to solve the problem and I was able to reduce it to practice.

And I did get some patents on it. I, I have four, four patents and web presentation of architecture and it's being used in, uh, large scale e-commerce to this day. I have some licenses out there. And it comes down to this notion of these opposing forces. But all I was trying to do is solve the problem. I was not trying to invent an invention method, but sometime after that, upon reflection, again, this [00:27:00] looking backward kind of this retroactive perspective, I realized.

There's a method here. And then after, uh, you know, for some time after like realizing, oh, there was a method to this invention, I realized, well, it applies to, has other applications in the technical realm. And then after a while I realized, oh, it not only has other applications in the technical realm. It can apply to domains like the organizational domain, behavioral problems, social problems, that sort of thing.

I actually recently had a chapter on applying this method in the realm of social change, and as you know, I've got that interest and background in sociotechnical systems and change management and some of that kind of stuff. At some point, again, as it's kind of gradually being unraveled unfolded, it unfolds in layers, it comes in waves.

It's kind of this journey meme. I'm gradually uncovering these layers. At some point I realize, wait a minute. This is like one of those methods of systematic innovation, which I was like, you know, had been brought to my attention and I realized it fit in that genre thing, which led me to the quote unquote Quimbee [00:28:00] palette and seeing, oh, this is one of these methods.

It fits in this realm of systematic innovation. And then the last elbow joint I'll mention is I realized retroactively. Kind of metacognitive, I guess, that I had tied together these aspects of classical philosophy I got from Socrates, the dialectic or opposing forces, thesis, antithesis synthesis. That's a deep subject.

We probably won't have time to go into it real deep, but just the notion of two opposing ideas and then coming up with an idea that combines the best of both of those ideas. It's not a compromise, it's not an optimization. It's a actual resolution of the. Yep. Yep, exactly. Net new and best of both worlds kind of thing.

And then I'd also realized I'd been incorporating Aristotle's ontology or categories and Platos and morphology or forms. And so again, accidentally discovering that this method is rooted in ancient classical philosophy, and it ties to this modern genre of systematic innovation methods, and it does.

Trigger and [00:29:00] leverage this metacognitive orientation or unleash the subconscious, not just purely conscious and also combining linear thinking and non-linear thinking. Also combining a systematic aspect and a and a random aspect. 

Justin: Yeah. That's cool. I like this approach that you're thinking about. There's a lot of different ways to go with this, but I wanna try and bring it back to AI with regards to how you can tie systematic innovation.

To ai. So I mean, I'm just thinking I about a more practical use, right? So I'm building an AI system, for example, a couple different like layers here. One is like, how would I use systematic innovation in what I build, but also like maybe what are some core philosophies or core principles that I could be bringing from this?

I guess from this philosophy into the systems that you build in ai, whether it be you and I talked, you know, before we got on the show here, but this alignment problem, right? For example, these things hallucinate. Are there certain techniques or practices that you have seen that you're like, aha, like if we apply systematic innovation to this particular [00:30:00] challenge we have with ai, this will help us solve it?

David: Beautiful again, right spot on Justin. Yeah. Where do these two things come together? And it, it's almost like an ideal fusion in a way. And again, I approach these things independently. I didn't set out to develop a method that could like leverage AI or anything. It's all a matter of like, you know, what do you see when you get there?

And again, I've got some background in ai ancient times, and that is AI has like coming, been coming to the surface, of course. Can't resist, like starting to dabble with it. And at some point I realized, oh yeah, there's this whole alignment thing. They hallucinate. They're not totally accurate. By the way, wouldn't it be nice to use them for innovation and for creative thinking and those sorts of things, but it's very non-linear by the way.

As a side point, not exactly in this direction. I did think about applying. Systematic innovation to the contradictions in ai, and I had a, had an article published in a journal and operations research, looking at it from that perspective, that's one way of looking at the intersection of systematic innovation in [00:31:00] ai, and particularly from.

The perspective of how can systematic innovation contribute to ai. In this sense, I'm thinking more about, yes, this alignment problem and they hallucinate, and how do we discipline them? How do we make them like more useful in that regard? And we've sort of come across these. Approaches like retrieval, augmented generation, and I think of, of course, with open ai, we have the custom GPTs and also part of the equation is the human in the loop and the prompt engineering.

How do we prompt these things and we get better and better about prompting them and chain of thought and some of these things to keep them between the guardrails and the irony is it's a contradiction, and this is one of the contradictions I wrote about the contradiction between generativity and what I call fidelity.

Keeping the thing between the guardrails, and we don't have any way inherent, at least that I know of, or maybe you know of, but that we can inherently, because they're neural networks, they're neural network architectures. They do what they do. They're trillions, billions of variables and we, you know, at this point, they don't really show us their work and we don't have any internal mechanisms [00:32:00] for keeping them.

Between the guardrails, but we're using some of these things like better prompt engineering, chain of thought, state breaking things down, stepwise, a retrieval, augmented generation using external knowledge basis, that sort of thing. And so at a certain point it occurred to me, well, gee, could I use systematic innovation as that component?

And here's where I'm gonna revert to the quote by Albert Einstein. I think there's two parts to this approach. One is, okay, we've gotta use things like retrieval, augmented generation, better prompt engineering. But I think we also need to think differently. I think conventional human thinking isn't gonna.

Be the best answer in terms of our approach to keeping the AI between the guardrails. I want to take a swing at applying a different way of thinking, applying systematic innovation to this challenge, and so I, I just, I trained a custom GPT using my method and kudos. Shout out to my colleague Dan Olson, who I know from Best Buy and.

We're [00:33:00] co-founders on the Minnesota Change Management Network, et cetera, and shared the vision with me, and he's a real builder and doer. He can like put these things together like seamlessly and really glad to have him as a longtime colleague and had the same vision. So we built a custom GPT that operates the method.

Justin: That's fabulous. That's fabulous. And it's, and so it, you're satisfied with what you ended up getting out of this? 

David: Justin. That's the part that's really interesting. This, especially when we're experiencing all the hallucination we can tend to experience. I was, to me it was profound how well I was able to train it in my method.

And the interesting thing that Dan Olson says is he likes to say, now I have a muse. The AI. Knows my method and I can interact with my custom GPT on my method. I now have a muse. We interact and I'm gonna say it's even like to some degree suggested, maybe extended it maybe filled some gaps in it. It we're actually seeing, you know, through the process of interacting with it, where it can only apply the [00:34:00] method and it comes up because it's connected of course, to a broader LLM, it can operate on.

Domains that aren't built into the training I gave it. That's a pretty interesting, you know, aspect as well is it's got the whole world at its disposal and it applies systematic innovation to it because it knows the way to do it and it's, it's not limited to the information that I used in the training, like example contradictions that I made might have given it, that sort of thing.

So that's kind of the interesting combination there and I've been very surprised and pleased with how well it seems to. Be able to operate the method. 

Justin: Wow. So that's truly what you're talking about with regards to just the different approaches. And I think you and I have probably seen this taking approaches from a different industry and applying it to something else where people get in their blinders, they just think, this is the way this has always been done.

And sometimes you can come in and say, well actually they do it this way in banking and finance. Have you thought about doing it this? You know? So kind of bringing those approaches. This idea that you have this framework, [00:35:00] this higher level metacognitive framework that you can apply, and now you have essentially trained in AI on that.

That's super powerful to be able to point that at other things. And it kind of also goes back to, you know, I was meeting with a potential customer the other day, and a lot of them are, they're thinking of these same ideas. Like, okay, I want to throw a bunch of my documents into a retrieval augmented generation system, or whatever.

And then I wanna be able to have people come in and um, you know, ask it questions. And what kind of dawned on me was, okay, that's level one. And that's where everyone's at. But you know, the thing that I think is more interesting is ask the LLM, how do I solve this problem? And not to say that you're gonna get the right answer, but I'm continually trying to push it and say, use it for creativity, right?

Because it oftentimes will, it may not be the first, second, third, or fourth answer, but there might be a fifth answer down there as I keep iterating and prompting with it and saying, well, what do you think about this? What do you think about this? It has the opportunity for us to unlock these different paths.

We wouldn't go down if we were just sitting there isolated [00:36:00] and maybe even kind of getting metaphysical. Now you throw a bunch of humans together. I always say we're a bunch of, you know, carbon based people, but now there's this silicon based thing that is ultimately in its core, gonna take a different approach and might actually end up.

Sending you off on a different path that you even wouldn't have gotten if you were brainstorming ideas with another human. I don't know, just sort of throwing that out there. 

David: No, very, very pertinent. All across the board. Justin. I think a lot of things goes in many directions. This whole notion of how are we using ai, we're only at the beginning, and Andrew Eklund had a real interesting presentation last week on the notion of.

We're using AI to do things like note taking and find the action items, you know, in this transcript and things like that. I think the meme or the cartoon was the user asking Albert Einstein to take some notes for him and Albert Einstein's over here going, well, yeah, but how about I solve like quantum computing or do something like pertinent or important And, and I think if you look at that spectrum, I'm trying to nudge forward on that [00:37:00] spectrum and how do we use AI more powerfully for things like.

Creative thinking and product discovery and some of that territory. And by the way, combining it with the notion of, I think we need, you know, at the front end, more powerful human methods that we're complimenting using to compliment the ai. So, Ooh, 

Justin: there you go. So you can basically, yeah, it's one thing just to throw it out there and say, do some stuff, but you're actually taking the best of both worlds is kind of basically what you were saying, taking these two approaches and now you have a new way to do it.

David: You got it. Yeah. That's kind of the idea. And it's a fusion at some level. It's a synthesis at some level. And how do we take the best of both these things? I, I, I guess I would say, well, obviously when I kind of saw the light and said, oh gee, I should build a custom GPTI should, like, I'm just dabbling with, and I can't tell you what exactly, you know, lit the light bulb, but at, at a certain, because it wasn't, it wasn't where I had it, it was, I didn't start out to apply.

AI to my systematic innovation method. I was just dabbling in the AI stuff and [00:38:00] at some point I kind of go, wait a minute, you know, could the two of these pieces come together? And certainly that's kind of when the light went on and I said, well, I could actually extend and empower the SI method by attaching it to ai.

And at the same time, AI has this alignment problem. And so it's really like best of both worlds SI can align the AI. AI can extend and empower the si and I think that combination, it, it's obviously an artistic event to combine them, but that fusion or that combination is better than either one alone. Si combined with AI is better than AI without it or random human techniques si with AI on the backend is much more powerful than me doing SI and my old like manual methods.

I've kind of gotten bored with the old manual methods. If that makes any sense. 

Justin: No, it does, it does. I think AI makes things fun again in a lot of ways, and it take, allows you to be able to experiment around with it. I, I took a note earlier that you're an accidental inventor, which I think that everything you just said right there is, it basically [00:39:00] sums up there because you were looking at building something or looking at a certain way to do it, and you actually invented kind of a new method to bring these things together.

You know, as we sent out, kinda wind down the conversation, you had mentioned something about sociotechnical systems and I think I did want to touch on that because coming up through engineering and building software for 25 years, what I've learned, and I didn't know this at the beginning, right, and I think many software engineers don't understand, and even architects that are just in their own technical world, it is so much more than just fingers on keyboards and writing code right there.

It's the whole human interaction with. The thing that you're building and not even within your team. Yeah, that's just one piece, right? You've got product managers and product designers and all this type of stuff. There's this whole socio piece there, but then even out to the customers, right? So you know the, this idea of, I had never really put a term to it, but I, I feel like the, is that what you're talking about when you talk about these sociotechnical systems, kind of the idea of bringing technology, and I dunno if socio, I means sociology or society, but [00:40:00] is that kinda what you're getting at with some of these things?

David: Very, very pertinent. Sorry, we've covered some of these kinds of things fast and, and, and thin, but yeah, very, very pertinent. And what I'm saying is when I talk about a sociotechnical system is it's got both social and technical elements and there is no system that's purely technical. It's just not, these things don't emerge from vacuums, so they're always built.

With, I like to say technical innovation is a sociotechnical system. It's like all the social elements, organizing the humans, getting them together on the same page. It's all a system. If the system is gonna be a whole product, you know it's gonna emerge from a holistic sociotechnical system where those parts are better integrated and then they go in enact.

Back to Doug Engelbart and human-centric design and that sort of thing, that social or human aspect of any particular application has those aspects to it. And in a way it maps to, again, it's a dichotomy. We were back to dualistic philosophy and the dialectic or thesis antithesis, and it [00:41:00] kind of maps to this whole notion of systematic and random, and I, I think I've conveyed, you've been careful to draw out in me this aspect of systematic innovation isn't.

Only systematic. It's not like formulaic. It's not like just deterministic. It's really this combination, this dance or fusion or integration of the systematic and random aspects. And likewise, a sociotechnical system is kind of a combination of those sort of aspects that we'd typically consider to be maybe linear and non-linear.

The social is the more non-linear part and the more amorphous part and the linear being, the more well-formed part I guess. 

Justin: For sure. Yeah. No, you can't have one without the other. They basically are intertwined with each other. And the other thing to sort of kind of wrap it up too, that was in my head as you were talking, so you're talking about design patterns, and it's one of those things that I kind of backed into as well.

I wrote a lot of software and then over time I realized, and then I kind of opened up the book, you know, this Gang of four book that, you know, yeah, there 

David: you go. 

Justin: And that's like, oh, I was doing this. [00:42:00] Oh, and there's another pattern there, there's another pattern there. And then you start to realize the philosophy behind it and why it was done and all that type of stuff.

And, and like the power in it. And I think it's fascinating that it actually, it sounds to me like, you know, you mentioned, you know Christopher Alexander, Alexander, right? I wasn't exactly sure. Now I think the guys who were kind of wrote, there's like Grady Boch and some of these other guys who sort of like did that.

But it, it makes a lot of sense to me that it actually came out of, you said it's came outta more of the metaphysical or the human world. Not so much from a technical programming, engineering standpoint. 

David: Yeah, architectural design and by the way, gang of four, that that crowd got theirs. That that was a progression from the Christopher Alexander stuff.

He was at the root of it. He was kind of at the origin of it, trying to apply it to architectural design, but at a certain point he pivoted, oh, pivot, pivot. You know? 

Justin: Yeah, yeah. 

David: You know, that's the nature of the beast being paying attention to, well, where are we getting traction? Where are we not getting traction?

So yeah, there is that breadcrumb. And by the way, I also wanna point out to you, Justin, you were doing it. Then you [00:43:00] found out there was the method. Yeah. You know, and, and so again, I say a lot of times we're maybe operating in this subconscious way systematically. Well what if we were more conscious about it?

Just like I say, well I was kind of subconsciously, you know, using some of these classical philosophy disciplines and, and some of that kind of stuff. And I think a lot of times we are. Kind of under the covers. Yeah, thinking systematically. And I'm just nudging us towards, let's be more deliberate about it.

And then by the way, can we also be deliberate about. Artistic combinations of the random and the systematic. Of course, we don't wanna be totally deterministic. We want deterministic elements. That's the nature of innovation, is producing something new and can we produce systematic catalysts for producing artistic combinations of the systematic and the random elements.

Justin: Yeah, that's fabulous. Heavy 

David: stuff, huh? 

Justin: Yeah. Well, back to what I said at the beginning. I always learn something with these talks that I, when I talk to people for, you know, 50 minutes to an hour or so. To wrap this up, David, how can people get ahold of you? What, what's the best way for them to find you? [00:44:00] 

David: My website is innovation radiation.com.

Justin: Okay. 

David: And there's plenty of propaganda and rhetoric out there. Just because it's rhetoric doesn't mean it's not valid. And then also if they wanna take a swing at Matrix Mentor is the custom GPT that I built with Dan Olson, and they can find it at, at bitly. It's right now at at at bitly bitly slash matrix mentor.

Nice. Okay. 

Justin: I'll have these links in the liner notes and in the transcripts and everything like that. So. It's been fabulous. David, I appreciate the time. We gotta have you back 'cause yes, one hour is not enough for sure for us to talk through this stuff. We just scratched the surface. I think there's a lot of different ways we could go, but I appreciate you taking the time and coming on the program today and sharing your knowledge about systematic innovation and how you're applying it to AI today.

It's been great conversation. 

David: Thank you Justin. Thanks for building community. Really appreciate it. 

AI Announcer: You've listened to another episode of the Conversations on Applied AI podcast. We hope you are eager to learn more about applying artificial intelligence and deep learning within your organization. You can visit us [00:45:00] at applied AI Dotn to keep up to date on our events and connect with our amazing community.

Please don't hesitate to reach out to Justin at Applied AI if you are interested in participating in a future episode. Thank you for listening.



Head Shot

Justin Grammens

Host