Name: Driving the future of tech-powered collaboration
Uploaded: 2025-07-16T17:34:08.474Z
Duration: 1 h
Description: Driving the future of tech-powered collaboration

Transcript for "Driving the future of tech-powered collaboration": Hi, everyone. Welcome to Mural webinar, Driving the Future of Tech Powered Collaboration. We appreciate you joining us today, and we'll kick things off with today's guest shortly. I just need to cover a few housekeeping items first. First one's around technical issues. If you're experiencing any issues with viewing, please let us know by utilizing the q and a chat. And here's a tip, refresh your window. This is a web based platform, so that may help. As far as q and a goes, please use the q and a chat to ask our expert any questions that you may have. Feel free to submit the questions as they come up. And, yes, we are live and we're actually here, so take advantage of this. We will be recording this webcast, and you will receive an email with a link of the recording within forty eight hours after we conclude today. Now it's my pleasure to introduce our guest, Mickey McManus. For those in our audience who may not know, Mickey is a scholar at Tufts University, the senior advisor and leadership coach at BCG, he's a co author of Trillions, a book I highly recommend, and also research fellow emeritus at Autodesk, which is where we first connected. I like to say that Mickey actually lives in the future, but time travels back occasionally to share what he's finding out. So welcome, Mickey. Mark, thanks for having me. Yeah. I I don't I'm not sure. I I would say I'm not a futurist. Some people say that I am. And I I think it's because futurists are bad as good or useful useful as sort of a flipped coin. You know, you never quite know what will really happen. But I do I do spend a lot of time trying to build a model of what seems will really happen with human nature and with with technology and understanding what technology is, which I don't think a lot of people even that call themselves technologists understand the nature of technology, like, how fundamentally curious it is, and and then people are always curious. So super fun to be here. Thanks for having me. And and and I know we're gonna dive into what I appreciate most is your storytelling, which Mary's imagination and imagining what could be in the future with deeply steeped in in science and heuristics and and what you know about human behavior and and the technology. So, on that, I'd like to turn the clock back five years for everyone in the audience. Back to 2020. We had a global pandemic, massive, rapid, immediate change impacting everyone around the world. Mural was certainly caught up in that when we shifted from being a vitamin, like something be nice to your your remote employees and don't forget that the they're on the Polycom. Right? Overnight and within one month, we shifted to a painkiller, which is, you know, it's not a term I but it really felt like that. Suddenly, everyone needed it. Yeah. The the the pain was and inspired by one of our conversations, I started an internal program to raise MURAL's awareness of key thought leaders that I thought everyone should need to know. It was called Humongous Heroes Among Us sparked by one of our conversations about heroes and Aisha Vercel's work with heroes. And I pulled a visual. I wanna share this visual, here so everyone can can see. I'll give it a second to show up. Cool. Yeah. So, there's a there's the pool. Some of the people on the call might recognize themselves. So inspired by you, I gave everyone trading cards and kinda branded, were they facilitator, were they thought leader. And, and I thought I'd I'd share this artifact from from that period in our history. And a couple of things Hey, Mark. Out of curiosity, does everybody online have access access to this Mural board? I don't know. Or will they? But I'll ask Aaliyah maybe to to look at the opportunity to get people in here. Certainly, we can share any link. Things that you drop in there, we'll make sure get translated out to the audience. Yeah. So so, there there's a lot of interesting things here. And I think that this is kind of one of my reasons why I love so much working in and collaborating in visual ways is what you do is it covers a large surface area and the journey. There's an aspect of time, your background, you know, in in University of Illinois, and then Maya design, and and that Maya going to BCG, but also Autodesk and Ooma. There are a couple threads here. I'd love to some people may not be familiar even with the concept of of Maya. What does that what does it mean? Yeah. Well, yeah. I mean, I can explain that a bit. And I was just really honored and blessed to join Maya. I joined it not as a founder, but, as the CEO in 02/2001. And it had already been running since '89. It came out of Carnegie Mellon University, and it was three three folks, one person, doctor Peter Lucas, really understood cognition. And he worked with Herb Simon and was a post doc of JJ Gibson, who helps create the whole field of cognitive psychology and perception. And he was really interested in how complex how humans understood and dealt with an extended complex ideas. And then Jim Morris was at Xerox PARC and worked on the Alto and the Star, which, you know, folks, remember that probably because the Lisa and the Macintosh came out of that initiative. And Jim really understood how to make things. You know, how do you build stuff out of technology? And and then Joe Ballet built the design program at Carnegie Mellon over a few decades. Tom and Dave, who later went on to to do IDEO, were students of Joe's, a million years ago before that when they were at Carnegie Tech. Really, how do you give form and function? How do you how do you discover unmet needs? How do you build stuff to extend people's reach? And Maya was all about that. So they had started it and run it, for thirteen or so years. And then I joined, and and was blessed to be a part of it. And most advanced yet acceptable was the actual name of the company, was Maya. And that was something coined by Raymond Loewy back in the nineteen thirties, a French industrial designer who helped kind of design our way out of the depression with this new thing called industrial design. And he said, if you could find the most advanced technology yet acceptable for normal people, you had found the Maya zone. And that idea of the Maya zone, you can hear Brian Eno talk about it when he talks about music and when he pushes people as a producer to kind of go too far and suddenly nobody listens to your albums. They're like, what what did you do? And then come back a little bit. You've gotta find that space. And what I loved about Maya was it had a big mission, and the mission was taming complexity. And and and it was this notion that complexity is power, but you have to somehow balance that with the way people understand things so you can extend their reach. So it was really about how do we extend people's reach and give them superpowers, which is nice. It ties back to your superhero thing. But yeah. It joined in 02/2001, and then I ran it for until about 2014, and then shipped it to chairman and handed it off to Dutch. In the meanwhile, LUMA was started, which was pretty exciting, and that was Chris Pacioni and Pete Madder and Phil Lucas, who who grabbed a hold of some ideas about how do we teach other people, how to do create with people. And that was exciting. And I I was thrilled when when a friend, and a fellow Mayan said, you should meet Chris. He he might be a good person to take this crazy Maya Institute thing and and get it out get it out into the world. And he walked in, and I think an hour later at a whiteboard, we were like, well, we're hiring Chris. It's pretty sure. And that that led to LUMA. And what it what it It was a good journey. For for it's a simple vault complexity, Maya as a consulting firm has a business model and deep expertise, and you're pretty much approaching each engagement as a custom, but you're pulling from maybe some bits and pieces and stuff in that scaffolding. And I I love and have benefited as a certified lead instructor for many years now with LUMA. The the observation that there is a gap, that people need some type of scaffolding and then the work to consolidate that into it's like giving someone, oh, you got a new car? Here's a socket set. It's not everything. You're not gonna change your transmission. What you it'll keep you on the road and, and make it accessible, like bringing the bar down. Well, and it was really interesting when we you know, LUMA has also, started as an acronym like Maya. It was look, understand, make, advance. And it was this notion that we need to look at tools that help us discover look and un discover unmet, unvoice needs. That's the first thing. We gotta go listen to the people that are having the pain, and we've gotta do that. And then we've gotta understand. We've got abstraction ladder and step back and see what the model is. What's the information architecture of this? What are who are the personas? How do they relate? What's the root cause analysis of this challenge? And then make. How do you make something? How do you iterate? How do you prototype the idea? How do you design for, action? You know, you build prototypes to learn by breaking them. And then how do you advance the individual, the the community, etcetera. And so look, understand, make, advance was the kind of framing originally for LUMA. But what was amazing to me was Chris and Bill and Pete looked at over 1,200 mechanisms and methods that were developed over a hundred years, and they kinda boiled those all down into 36. And they weren't, like, the only mechanisms. We still had that giant list, but they built them into kind of a periodic periodic table. And that was the exciting thing, you know, that just like, Mendeleev, when he did the periodic table, he could predict elements that didn't we didn't discover for, like, a hundred years. And they filtered it by, you know, do every do real practitioners use this method every day to do hard things? Okay. That's number one. Is it used broadly? Is it used across whether it's architecture or product or building a system or policy, etcetera? And then, can you learn enough to be dangerous in a day or two during workshop so that you can actually start building your own muscle, like, building that bridge. And so we we still had others. They were sort of more esoteric, you know, only people that I do or only, you know, masters could do this after ten thousand hours. But we said, if we could teach people these, filter them down, and build some recipes and build some ways of doing things, then people could co create with the the folks outside, and they could kind of subvert themselves. I think design is about subverting yourself and and kind of taking on the needs of someone else and and be able to, like, bring that energy. So LUMA was, just a lovely one of one of six companies that kind of spun out of Maya because we were also a small kind of Pete, Joe, and I would invest in spinning out other kinds of companies. Well, that that resonates profoundly with me because at the moment when I encountered Ooma at Autodesk, I had been a designer. You know, I had built CD ROMs before the Internet. I had made pixel buttons to make it look like they're up and all that stuff, the the foundational things. But I felt like when it came to methods and really understanding how what is the problem we're trying to solve, I had a small I had some hard sorting. I had some strong things. I had some things I picked up from researchers along the way. It was in that Luma course where I felt like someone grabbed my binoculars and just turned them a little bit, and suddenly, the most profound thing was freedom from certainty. I didn't have to be the design expert you're paying to have the answer. I I have this I don't even have to know really what you're talking about. I just gotta hold space and be listening and sensing and drawing out the genius that's sitting there in the room. There was this wonderful booklet at Maya that was developed by Joe and the team, and Bill who later was one of the founders of of, LUMA. And it was how to be a Mayan. And and it was this kind of, road map for how to be a good facilitator and how to ask the dumbest question in the room, how to, like, measure the energy in the room, and, like, hey, I think it's time for a break. How to, like, push down the the the loud mouth, and how to bring the shrinking violet up so that you could might get diversity of thought, you know, and how how to get people to make together. And, I still have the PDF of that how to be a Mayan. And it's one of the most glorious. It's quite old and and but a new every new Mayan would get one. And it was, it was a lovely kind of thing about facilitation. And, and I do think that turned out to be one of the things that LUMA helped folks learn how to do as well. Like, you became a facilitator for for LUMA at Autodesk. And Autodesk was one of the ones that was very curious because I did not bring LUMA to Autodesk. Like, people would be like, oh, you're a fellow with Autodesk in the office of the CTO. You must have brought LUMA. I didn't have anything to do with that. I gave them two free tickets to a LUMA to a LUMA workshop at tech shop in San Francisco. Somebody went and they got excited. And then, like, a year or two later, I went to a Friday afternoon lunchtime thing, and there were, like, 800 people practicing LUMA and sharing. And I was like, oh, that's interesting. They're all calling themselves Luminauts. Yeah. So it was really they they grabbed a hold of it. Yeah. Yeah. The seeds fell on very fertile soil. I know the the two people that attended, and it was, yeah, it was a it was a it was a profound change, and it I think it's also because at the, leadership level, there was an understanding of how the intention behind where this can be applied. And the expectation that if you get a seat in one of these courses, you really have to carry some water and use these methods. Mhmm. And, I think internally, we we had brown bag lunches where we'd our our seeds to get people to attend where we'd say, well, we're gonna go into one of those methods you didn't get too deep on in the course and piece people would show up. But for me, it was always a conversation that happened in that casual discussion where Mhmm. I am not a mechanical engineer. I am not an architect, and yet those were the people who are building software at Autodesk. You know? And they would come in and say, I'm struggling. I tried this rose thorn bud, but I'm in an oil derrick in the North Seas. And these people are covered in grease up in their arms, and they're not gonna sit around with postcode. And someone I would I'd be stuck. I'm like, I have no idea how to help you. Someone else in the room would go, I was just with, you know, BP or Chevron. Here's what I tried. And that's when the like, it took off. Well and if you were to pull back a bit, I think that that this is actually gonna be a a bigger part of the topic today, which is that people actually love learning. Right? Autonomy, mastery, purpose. Right? We we learned this a long time ago. This is what motivates us. And and we had tapped into that somehow because it wasn't just a collection, a bag of tricks. It was actually it had an architecture to it. It said, look, there's this there's this collection and but this is how they relate. There's an architecture. Architecture, I think, is often misunderstood. But it's sort of how how how does this what makes a cup a cup? Like, what's the thing you can define that is always true? And stepping back and saying, this is not only it works, but as you pull the thread, it starts fitting together. You start to see how the bricks come together and build a house and how how that works. And I think that's important. I see that you're pulling back a little bit, Mark. Is that what you're I am I'm, like, suddenly just saying, like, I I know we we the conversations always We could go for a while there. I I wanted to nudge us because there are two threads that I wanted to pull on that I think are really relevant for for the topic. And one of them, both are, things that, again, stories that you shared in the past. And, I'll I'll look at them in an overview, and then we'll zoom into each one. So the first thread is this whole notion of being visual. And I remember, people may be familiar with, you know, Bernard's map of Napoleon's campaign. But you showed, in for those of you who haven't seen this before, if you were alive in the eighties and nineties and were in UX or a related field, Tuffy probably sent you a postcard with this on it. And what it is is a a probably one of the most rich multifaceted visuals of data showing Napoleon's march towards Russia and then his retreat, and there is layer upon layer of information encoded in there, the size of the troop, the temperature, ambient temperature, supplies, and etcetera. Well, the data is, this is the underlying data. And I think many of us in our roles in business, we're we're in roles really to make sure the data in this particular domain or that particular domain is good. But we really struggle. Like, having all of this data in front of you just is not the same thing as working visually. So that was a thread I wanted to pull forward into our conversations, the power of working visually. And the other one was the journey I I had the good fortune to observe at Autodesk, which really I look back now and go, oh, right. That was my introduction to machine learning. And there were these things we were doing with using the cloud to do massive iterations on designs, but humans were saying, here's a strut, here's what it's gotta do, here's the bolt pattern where things gonna line up, and then let the cloud generate thousands of thousands. Well, and I I'll I'll add I'll I'll just add one or two things to both of these. The Menard visualization came out of our research, Maya Viz, which was this notion of could we do information centric computing? Could you pick up the table of the latitudes and longitudes and just drop them on a map? Not document centric, but information centric. And could you see them show up on the map? Okay. Now this other stuff is not really a mappy thing. It's the he marched from this town to this town. He lost people because of dysentery. He lost that battle. Those are more like bar chart things. But then we're like, could you just pick up the bar charts and toss them on the bar chart thing? Okay. Cool. Could you steal the genetic code for bar charts and just teach the visualization that was a map new genetic code? Could you do recombination and paste it to the map? And suddenly, these bars would show up on a map, and it looked like Menard's visualization. And then the real secret was dynamic query because these things were uniquely identified in the universe, kind of it it helped resolve, the five blind men and the elephant. You know, you could dynamic query. You could slide a slider back and forth, and somebody could shake something on a map, and it would shake on the timeline, and it would shake everywhere. Because there was only one of them. It's just like me holding up this glass of water. You know I'm not talking about 22 glasses of water. I'm talking about this glass of water because in the real world, we get Pauli's exclusion principle. I can't be holding two glasses at one place and one time in the universe. But we don't get that with information. And we were saying, what if you did? What if you enforce Pauli's exclusion principle and what we would call interaction physics? And there are folks actually on this webinar in the chat that are far smarter about that than I am, I would say. But, having having been there. But that was this, which is that working visually and having physics so you can shake stuff and to see how it works allows us to see the underlying model and look for mechanism, look for causality, not just coincidence. Because one graph could fool you. But if the whole thing, like, just works, there must be some physics. And that later translates to Autodesk, which the huge change in the world in computer aided design was a shifting from flat files. Like, I've got a top view. I've got a side view. I've got an elevation and a plan. And if I change an elevation, it breaks the plan and vice versa to parametricism. I have a parameter that's tied tightly together. And my first project at Autodesk was this thing that you're showing here, which is Hackrod, and it was part of something called DreamCatcher. And it was the notion of, couldn't we define the goals, constraints, and obstacles about a problem, like designing a new chassis for a race car that works on the desert, but not define the chassis? Don't design the thing. Design the negative space and just set up the goal. So you drag around like an arrow that was a Newton and say, you've gotta support so many newtons of force under here because the wheels held up there and you've gotta you gotta do this and this. And then it would evolve. It would spawn this is called generative design. It would spawn thousands of agents. Some of them might be trying to do finite element analysis. Some of them might be trying to do fluid dynamics. Some of them might be, actual agents wandering around or acting as multiple cars trying to see if they can make it through a race and seeing if the forces could happen. And then you would come back with manufacturable results. And it was a very different way to think as a designer too, which it it actually hurt when we were working with designers. They would first say, oh, I'm using AI. Is it just gonna turn me into a curator? You know, am I just a curator? Because I didn't go to architecture school to become a curator. I didn't go to industrial design or whatever. And what would happen is and I remember seeing this with Felix and some of the early designers that were working on this hack rod, is they started saying, I'm out. It's too late for me. I'm too old. I'm never gonna use AI. And then they shifted to, oh, wait. This is like punk. This is like jazz. Like, I poke here and it pushes back. And I'm like, no. That's crazy. And then I poke here and it pushes back. And pretty soon, they were riffing together. And I think that was the moment I had was that the machine learning stuff could do a ton of things, but the machine learning doesn't have a mental model. It doesn't have what we would call wisdom. And so but, consequently, if you played together, it you know, it's sort of like, I'll never be the drummer in the jazz quartet, but I could be the best saxophonist ever. And if the drummer goes off on a solo, I'm gonna just run with it, and I'm gonna learn something. And that was what we started seeing with this with this work, which is a very different approach. And, of course, we saw some results that's now been rolled out of research. And and I love the so and the insights from that and this is where my introduction to machine learning was so nonthreatening. It was really in service of observing people in spaces making stuff. Right? This was an amazing project, and this was the idea. This was out of the Toronto team at Mars, in Autodesk. And what they did was they said, let's put let's put a watch so we can get the the movements. Let's put microphones out. Let's put goggles on that look at eye tracking. Let's put, you know, all sorts of telemetry on, and then let's bring in novice users, expert users, etcetera, and just record. And we'll put a point of video camera at it. And so we've got all this stuff. And what we did was we took that as training data, and we gave it to some experts to look at the videos and look at the data and go, can you guess how expert this person is? And then we gave it to a machine learning model, and we said, can you guess this? So we gave it some training data and then we held off some training data so that we could see if it would actually make predictions without having ever seen it. And it was pretty wild. We could see about 90% of the time we could predict the level of expertise of a person with a machine learning model, which is really useful if you're thinking about Autodesk is the sort of 800 pound gorilla for building most of the cities and structures and stadiums and things in the world. The most important job of a person on job site when you're actually building a skyscraper is the foreman. And the foreman only gets promoted when people can tell their expert at enough of the tasks that they can kind of keep other people busy or else you're losing money. And so consequently, this allowed us to, like, actually have a little system that could say, hey, they're right at desirable difficulty or productive struggle level. They could level up. And I think that's something we have to start investing much more in as we think about a learning company, sort of how do we help people level up. The ultimate outcome that you shared with me back then was this notion that the software can observe people as they're working and understand that so and so might be really good at the the physical appearance modeling and that kind of thing. Someone else is really good at maybe the fluid dynamics or something. But then as a business, you've got a case that comes in. And how do you, you know, how do you match people? And this whole notion of mentor see, this this was so nonthreatening to me. This was this was getting the right Zoom in even more on that for a second, Mark. I want I want everyone to be able to see kinda zoom super close in on that. Okay. So Tina wants to hone her generative skills. That's the one in blue. And she's got great skills. Mickey needs to learn these skills and there's a real project. This was for Volkswagen's electric bus. And so the system says, oh, there's a mentor match. The same way that when you're playing, like, a large scale multi user dungeon game, it goes, oh, you need muscles. You should find an orc. He's gotta have muscle. Oh, you need a wizard. So it would do the mentor matching because we did that other's process and it would suddenly say, oh, Mickey wants to learn this and he needs to learn how to be a follower because he's been called a lone gunman too much. He doesn't know how to play well with others. Tina wants to learn how to teach leadership. So it does a mentor match. And then as they're actually working on a real project, it's constantly tossing things back and forth to to help the person build their capacity, but staying within the proximal zone of learning. And again, we can do this. We know good science about how learning works, how you build neurons and build connections. It's just most of the systems aren't really designed to help you build capacity, You know, because most of the machine learning systems today weren't created by anybody who understands humans. They understand code. Now, I wouldn't. I mean, that's a gross generalization. Some people came out of the cognitive psychology world, but a lot of people are sort of high performing engineers who who who are passionate about this, but went into engineering because they didn't like humans, you know. So, you know, or maybe they didn't understand humans. And and I think we need both of those today. Absolutely. Well, the those two threads, being visual and kind of this this very different, perhaps, lesser known, if you're not you know, if you just grab the AI headlines, it's chatbots, it's images, it's deep fake videos and everything. And there's this other thread that I wanted to pull through about because everyone's confronted with, buying the companies, which five years ago, I'm just gonna lift a couple five years ago, you and Diana, joined us Yeah. And we had topics that we covered. In fact, we had too many topics to cover. So we we touched on three and then we had a spinner and we randomly touched on the ones until we ran out of time. But out of this universe from five years ago, I wanted to touch on and really reflect on, the gap between where what we were thinking about these things then to now, and then start looking forward. And the three topics that I wanted to draw out, the first one was the forced experiment. So that that disruption happened. That in many ways has has pressure tested systems, and maybe there's lessons we've learned as humans about dealing with VUCA, the volume going to 11 on VUCA. The second one was this, topic of the learning company. And I think it's meant something different than than now, especially as we're looking at incorporating AI into an agentic AI. And the last one was the Bionic company, where five years ago, we're saying and then I'll remember this phrase forever. Yeah. It what it means essentially is you're gonna be managing humans and algorithms, like in a reporting structure. And I'm gonna wait a minute. I go around. But here we are, five years later. Here we are. Yeah. And I think that, you know, obviously, I'm a I'm a child of the seventies. And so, and a lot of people who came up with this notion of the Bayonet company at at Boston Consulting Group are children of the seventies. And and I vividly remember Steve Austin, the bionic man. He's in a plane crash, tests you know, he's a test pilot, and he's got, like, an arm that's like a robot arm and legs and eye. And then Jamie Summers comes along, and she's the bionic woman. And it's like she's got super hearing. I think at some point, there might have been even been a dog. In any event, it was this notion that that, like, you couldn't tell where they ended and where the machine began, and they had these superpowers. And we said that's the way we have to think about how organizations, how firms work, is we need to think about how do we do the best that algorithms or the data big data could do, but how could we also do the best about how humans could do? Because I don't think you should write humans off just yet. Right. We've got, like, three and a half, four billion years of r and d behind us, and we outperform all those models in having any idea of causality. Of actually, like, big mechanism. We've got big data. We've got no big mechanism. And so we actually have to say, how do we invest in people? And that's where these two come together. The learning company back then, it was more like, yeah. How do you really change return on learning investment? The same way you've got, like, return on marketing investment. That's evolved now. There's something called BCGU, and I'm one of the faculty members there. And it's really about how do we scale, but have people do hands on experimenting and working to learn about this. And then how do we make sure that that the companies don't go off rails because they're not being ethical? They're not aligning with actual values. They're not, like, you know, tying it. Because it's very easy to fall for the kind of the the current Kool Aid about what AI can do. But it turns out that it's gonna take work, and it's forcing us to learn a lot more about how humans work. And not just humans individually, but humans collectively. And so there's been a lot more investment in collective intelligence in what's called, if you wanna be geeky, active inference, like how do cells, cephalopods, and children all work. And we're finding this crazy stuff in biology that's applicable to why humans are master collaborators. Why do we build societies? Why do we do this? And how do we could do that? And the the AI systems can't do that stuff, and it might take quite a while. But what we can do is look for ways to pull in what we're learning in nature and apply it to to sort of, organizations. Because organizations are kind of an organism too. It's just that they're, you know, they're sort of man made. Yeah. Absolutely. Well, I'd I'd love to focus the balance of our time on on, I think, these three questions that we've been kinda kicking back and forth. And the first one might be briefer than the other two, but, mirroring those three topics, you know, what lessons from the last five years are gonna help us in your opinion with the next five? There was a little word cloud I pulled out of, some research I was doing about. What what was the what were the issues over the last five years that we might encounter? Well, I think I think if I were saying what are some of the lessons for the last five years, you know, everyone thought after COVID that we'd all be virtual everywhere all the time. And and, you know, it has actually led to a lot more distributed teams and a lot more people going, wow. I could actually find brains anywhere in the world, and I could work with them to build to build capacity. That's actually pretty cool. I think because because humans have brilliance all over the world, and they often weren't tapped into because they weren't near that corporate office or something. So I do think it it forced everybody to see you could keep working, and you could actually find some pretty wild, amazing, smart people anywhere. But we also noticed that, like, meanwhile, you know, Apple built a giant Starship over there, you know, in Mountain View, And they actually are encouraging people to be there for a good reason. Humans are high dimension, and we actually are shaped by context. And a lot of the context, of the virtual digital world, I mean, Mural tries, but it's not the same. It's a very low bandwidth version of walking over to the coffee cooler and having a casual conversation. And, a few years ago, I went through this process of becoming certified as an executive coach as and and and, it was like a year and a half and I had to learn all these methods to take my hat off as a consultant and to take my hat off as an advisor and to actually just hold a mirror up to a leader and let them have a good conversation with themselves. And in this process, I learned all these really amazing methods. And I thought, wow, I wish we knew these when we were building LUMA. I wish we knew these when I was actually trying to advise Chris and Pete and the other folks when we were starting companies because I apologize now to what I thought I was doing was coaching. I was not coaching. And so what you learn though in all that is, two things that I think are really important. We're always moving towards safety, like this direction, we've gotta be safe, or towards thriving. Right? And and when we're moving towards safety, it's because there's something pretty scary going on. But when we're moving towards thriving, that's a chance for us to grow. But what's in our head is very different than what's we're doing with our hands and feet. And so what we're often looking for is, what can we do that that someone else could see with a camera, that you could watch in the room? Oh, Mark showed up differently today. Like, you did it with your hands and feet because action, it's this sort of, the phenotype versus the genotype shows up. And so so I do think in the last five years, we've actually gotten much more sensitized to, you know, broadly, this is about mental health and mental wealth. And and historically, mental health has been in the shadows. We can't talk about it. Or, only elite people, people with lots of money can get a coach. People with, like, a sports hero can get a coach or a therapist or things like that. So there has been a big investment in what you would call transformational tech that has to do with stuff inside our ears. The second thing I learned was context is decisive. So so this comes from that picture you showed way back over there, which was the Kiva, where where you put people in a room that was a giant whiteboard room. You got different things because people build in space and time. They it's called weird is is what it is and what it and and when it was is what it means. Context shapes our behavior. This is what Maya looked like. And and you would see people come up with things because they could build on what they knew, and they would use the space to remember. In Kiva's, you know, we were exploiting a little trick in cognitive psychology where humans can only remember seven plus or minus two things in short term memory, but every human can remember one, ten, a 100, a 10,000 things in physical space. The space remembers for us. It's because of these really cool cells in our brain called, place cells and grid cells. And so context shapes what we do. And if you look at Myers Briggs and all these other things that are sort of pseudoscience and not really based on how cognition works, none of them take into account context. And if you can't change the person, change the environment, and you'll get different things. And so a lot of the digital tools are trying to do that. I think we're learning more and we're investing in more. The other thing the huge thing that happened from five years ago to today is, of course, this weird paper came out that was that was, done by some folks at Google, and it was at NeurIPS, this big conference. And the paper was attention is all you need. And it was this really curious paper worth reading, and it was this notion of could we use these big GPUs, these big processors to pay attention not just to, like, the dog cross the road, like, how the relates to dog or whatever. But actually, open the context window and pay attention to all the ways these words work together and all the ways when you're decoding and encoding them. And this notion of building an attention mechanism suddenly unlock the ability for sixty or seventy million dollars to actually kind of rip the entire web and build a foundational model. So that was a big deal, you know, that happened in 2017 and ultimately, had GPT four come out in 2022 or so. And that unlocked a lot because it sort of democratized the world's knowledge in one way, but because it's purely statistical learning, it's only half a brain. It's missing, like, neurosymbolic. It's missing that mental model. But it's still very powerful. Well, pivoting to the second question, in what ways has the ambition of being a learning company changed? And and how do we how do we now leverage these new technologies in a in a thoughtful way? And how how should we be, thinking or approaching this? Yeah. I think it's changed a lot. I wonder if this would be a good time for me to show something. Uh-huh. If you're okay with it. I mean, I put up a few diagrams here that come from some of my research work. But why don't I why don't I share my screen if I can? Okay. Yeah. Let's do this. So I'll stop sharing mine. Alright. Now, by the way, are there questions out there? I don't see any people. Put in some queue questions in the q and a. Okay. I'm going to just do old fashioned share my whole screen. Oh my goodness. And look at this. Okay. And I think I'm gonna start with this, which which I think is useful. Can you see that, diagram, Mark? Turn of AI. So Yeah. So Kate Kate Compton, who's just this brilliant, game designer who got her PhD down at UC San Santa Cruz, she said to me one day, generative AI is somewhere between a hammer, an ocean, and a swarm of bees. And and I think that that's the problem we're grappling with. Like, what is generative AI? Is it a tool? Is it I mean, you can't marry the ocean. It's not quite that. And is it a swarm of bees? Is it is it this kind of broader thing? And I think the thing that helps help me explain this is I don't know if you remember. Do you remember AlphaGo when suddenly the best Go player in the world was beat by by a computer? Yeah. I do. I'm it's a novice Go player, but, yes, I follow it. Yeah. Well, in Go was considered, like, the most sophisticated sort of strategy game in the world played by for thousands of years. It's it's a much bigger game. It has more moves than atoms in the universe. So chess is a very fixed game. Go has a lot of moves and this Wired article actually frames it really nicely about what happened. So I just drew this picture. This is all the moves than atoms in the universe. Right? So so every human ever has only explored this little narrow path. And and all the precedents, all the books, all the discussions have always been down this path. So it's almost like we've only wandered into this galaxy and this solar system of how to play Go. And what happened is Lee Sedol, best player in the world, he's, like, playing the game and he wins the first game. Cool. And then he's playing, and then the AI wanders off to a different galaxy. And he's, like, he's holding his head. He walks outside. He smokes a cigarette. He's, like, he's never seen any human ever consider this because he's literally just been dragged to a different galaxy. And then it does it again, and it wanders off to some other galaxy. And so so this is what our opportunity is that we could be exploring very bigger, much bigger, different very bigger much bigger different spaces if we actually play with some of these machine systems. But I was kind of curious. Did everyone stop playing Go? Because this was, like, six or seven years ago. And and, curiously enough, there was a study, human learning from artificial intelligence. People did not stop playing Go. They actually started teaming up and doing Go as human, machine hybrids. And what they saw was that the the humans, the players actually showed a significant increase in decision quality and reasoning instead of the ones that were not using AI. And the reason was they were being they were being exposed to a much larger pool of ways of thinking. So that context was shaping their potential ability. And I think that's very powerful. And, you know, I'm gonna I'm gonna, jump to another thing that I think is really important. I'll I'll, I'll actually just jump to this picture because I think that's useful. I do this I teach a class that's like an eight week course, that we built with Kareem Lakani at at Harvard that's part of BCGU, and it's about data and AI platform thinking. And up at the top is kind of what we could do before 2017. Think of a factory. You get public data and private data and operational data and customer data, and you cleanse it, append it, and build some models. And then you've got a way of kicking out experiments and learning. And ultimately, out of a factory like cars, you get patterns, prediction, and process automation. But then this generative AI thing, I think of that as maybe that was the left brain. It's very rigorous. It's predictable. I I could tell you this is definitely gonna help automate this process. It's not gonna have hallucinations. It's kind of like your left brain or in some ways, you can think of this as thinking fast and slow or left brain, right brain thinking. Those are both just metaphors. But then generative AI came along and it's suddenly unlocking what I think of as the six c's. It's not that it could comprehend anything. But what it could do is you could say, please explain this to me like I'm a 10 year old, or please write this like a haiku, or, help me have an analogy for this. Or you could say, please help me, look at this like a BCG analyst or a McKinsey analyst and give me a two by two matrix. So suddenly, you it could help you and your team comprehend stuff. Could also create we've seen lots of things around deep fakes and generative, you know, audio and visuals and things like that. And that's kind of like, a web search for things that don't exist because it's mashing up these things in this, what they call a latent space. Then it can coach, and that's a weird thing. But I can say to an AI, hey, help me do prospective hindsight. It's a particular method that actually helps me run it forward, and then in hindsight, look at all the ways that things go wrong. You know, I could do sort of a a post mortem, pre mortem before I even start a project. And I can have it coach me, and that's kind of a gold standard. I could also do chatbots. No big surprise they're conversing. But then we saw NotebookLM come out, and it was a very different way of doing conversation. It was a podcast. And then we can command. We and this is the agentic space. You can say, go off and be a million Mickeys. Search out all these things. Come back and report to me. And then we're starting to see the ability to do casting, both backcasting and forecasting, where we can say play out a possible future with this, that, or the other thing. But what we're seeing in this whole space is that there's this iceberg. It's kind of the missing 70%. You start out caring about outcomes. Ultimately, it's affecting how you run a business, and that's where we're starting to see it. And we ended up doing a nice research study that if people point a camera at the at the QR code, they can see. But it was really looking at kinda how people are deploying, how they're reshaping, and how they're inventing entirely new approaches to things. And I think that's the fun space and how much, sort of future built companies are investing in the in the people side versus the stagnating companies. And I'll stop there because I think that that's that's kind of useful to, before we go much further, it's useful to kind of, to think about. And and I, you know, I wanna make sure we stay on track too. Yeah. No. Well, again, the I think the this is the scaffolding that helps us weed out the signal from the noise. Like, what it's, I think in the last year or two, it's been all too easy just to get, fearful. Right? If we're if we move towards thriving or fearful, plenty of signal to get us very afraid of this. And I think there's signal sharing here, which is like there's there is reason to be optimistic. It's just, perhaps not grabbing headlines as much or or, on one of the Well, and I see the question, ChromaJoy says, how can a product manager stay relevant even with the evolution of AI? And I think that this is actually a time for new science and a time for new experiments. And so, in some ways, I think as a product manager, you know, looking for how can I balance what's great with humans and what's great with machines? And I wanna cite two studies that I think are pretty provocative that I think, give you a sense of this. I'll go back to, to sharing my my screen here for a second, to answer that question. You know, if you think about a lot of people are like, AI is gonna make us stupider. Yeah. There was that MIT report. It was very small. I would wanna see it replicated. But certainly, if you use it the wrong way like anything powerful, you can cut your fingers off or you can actually build your brain. So but jobs require more than just tasks. They're not just automating tasks. They require metacognition. Right? We've gotta step back. We've gotta understand what a human's responsibilities are and the constraints and that wisdom that I I call the mental model. You know, so humans can do this crazy stuff. We can reason through analogy. We're constantly predicting things. We've got creativity. We love collaborating. We've got common sense. We've got passion. We definitely have biases, and it would be great if we had some more work by the folks like Joy Buolamwini and others who've been helping explore how to cancel biases or how to do things. Machine learners now can reason through analogy. You can ask it look at these two patents and find all the analogies. It could do prediction. It could do creativity. Doesn't have common sense that that half of the brain that it's missing. Doesn't have passion yet. It it's gotta be goal directed by somebody else. Definitely has biases. But the really interesting thing is here's this wonderful study where eleventh month old babies were challenged to try to predict the actions and goals of others. They had a little dots moving around like a jailer trying to keep somebody inside and things, and they outperform GPT four level models of AI. This is this incredible efficiency that humans have that, frankly, it's worth paying attention to because the the 11 old baby just needed some baby food, maybe some mother's milk. So I think we're entering this jazz age of AI, and the the environment will shape us. So how do we change our environment? If I'm product manager, what do I do differently to shape the environment? We did two studies that I think are worth looking at. One of them is called the Jagged Edge, and this was with, Ethan Balik and Kareem Lakani and the Boston Consulting Group and some folks at at both UPenn, MIT, and Harvard. And we took 700 of our own analysts and we baseline them, you know, from BCG. We said, right out of school, wet behind the ear, super rock star. We have baseline. And then we gave some of them access to GPT, just one model. Some of them, we gave an hour training to. Some of them, we didn't. And we gave them 17 different tasks. Some of them were creative tasks. Some of them were root cause analysis sort of business task. What we saw is they outperformed and created product innovation. It was just giving them access to that galaxy, but they underperformed with business problem solving because they didn't have a mental model. We also saw the bottom performers were raised up the most. So the people that really needed it, maybe what behind the air MBAs. And the top performers were raised least. And certainly, both of them fell prey to business problem solving actually kinda got broken. So it was a jagged edge. We think one of the reasons that the business problem solving went bad is that they were falling asleep at the wheel. And we've seen this happen over and over again when self, when autopilot came along to airplanes. There was a danger that the pilot lost track of what was happening, and suddenly the autopilot would say, stall and hand off, and you'd be you'd be trapped and the plane would crash. So because they sound so confident, and yet they're only statistical thinking, they do not have a causality, like deep causality model. They can sound really good when they're completely full of bogus. And so one of the things we saw is it certainly lifted up, individual performance for creativity, but we saw, oh, actually a depression in collective diversity of ideas. And we know that hybrid vigor matters, which means maybe we should have different kinds of models, and maybe we should have different approaches to things, not just use GPT for or or only get Geminis or only anthropic or etcetera. We might wanna have diversity of models along with diversity of people. But we did see that it leveled performance, which was pretty powerful. So we started thinking, could it be an accelerator or an escalator? Could it be the cyborg, like the bionic thing? Could it be a centaur? Could it be more like I do running really fast because I'm a horse, but my upper body does the shooting of the arrows? Could it be a swarm? Could we use the wisdom of the crowds and develop that? And that's where we're starting to see some really wonderful work in collective embodied intelligence. And then, a new paper just came out called the cybernetic teammate, and I would definitely worth reading. This was done with Procter and Gamble and, Harvard Business School. And they actually looked at this four different categories of things. A team without AI, an individual without AI, a team with AI, and an individual with AI. And what was interesting is they teamed up two different kinds of teams. One was an r and d team. One was a commercial team. So these were people that were in different parts of the world that actually had different jargons and different ways of thinking about things because they're different domains, but they had the team together. And I think what's really interesting is alone, you did pretty good. That's the baseline. The team with no AI did worse, and that's mostly because of Metcalfe's law. Every time you add another node to the network, it exponentially increases the number of people you gotta synchronize and stuff, you know. So that's not surprising maybe. But alone with AI, it actually outperformed really well and so did team. And of course, teaming was less because it takes it takes some to do coordination with different people, different mindsets. But we also saw, and this is interesting, if you look at the far right, this was, the quality of the solutions where people it was not their core job. So on the on the on the right, but but sort of inside the middle picture was people that was their core job, but the people on the right were were non core jobs. And it really helped them, which meant the commercial team could actually ask the model, like, what do they mean with that jargon? And they were using that during the teaming and actually understanding, oh, explain this to me. That was that coaching part or that was that kind of comprehending part. And I think that's very powerful. And, of course, we also saw very high quality ideas, and we saw that people were actually having fun when they were having this kind of project manager helping them or helping have access to it. Super early days. And the last paper I'll point to is they're actually starting to look at if you lead an AI poorly, are you a bad leader for humans? So could we actually send signals to leaders like, man, you are oppressing the hell out of me. You're asking or you're giving me bad direction. And could we use that to actually start detecting? Super early. But it'll be interesting to see what happens. And, I'll stop there. But I I think we could do science now, I guess, is the point. And we should be testing on ourselves, but we should also be learning more about what really works from science. And and it's hard sometimes to hear it through all the hype. Yeah. Absolutely. Well, I, I know we're coming up right in the end. We have, a couple of questions that I wanted to get to, but I thought, and we've touched on the other questions I had. We have we've covered the surface area, more so. Thank you for sharing. Enough time. Never enough time. But I was I was, just thinking, any kind of parting thoughts as people are kind of in the shoes that they have looking around saying, what what are No. I think you don't you don't grow neurons unless you play with things. You know, you have to play. And I think, I ask leaders, how often are you using this stuff to yourself? Not telling people to go use it, you know, because there's a lot of that going on where they're trying to, like, you know, lower the cost of things by getting rid of people and all that. Sound and useful leaders, and and I mean leaders at all levels, and kids, parents, etcetera. It's like play with it, but don't necessarily play with it the way that they tell you to. You know, play with it wrong. Like like, mess around. Mix it up. Poke on it, and and don't take it for for for truth. Think of it as kind of a jazz musician who's playing with you. Yeah. Serious play, Jim. That's exactly right. It's this notion of play to learn. And, you know, the the the idea of growth mindset is the opposite of fixed. You're either good or bad at math, or you're good or bad at things. Growth mindset says you don't know how to do it yet. So I would say yet is the most important word there. The struggle is when you're learning. And if I don't see people kinda struggling, myself included, I'm not probably pushing myself enough. And so so this is very powerful. It's still missing a bunch, but I would play. Well, it's it's interesting. We've in in the, the work that that I do with Jim Kolbacher, our chief evangelist, and Pete Maher, who you know well, we've we've adopted this this this language that says as leaders, you create the condition, almost like a gardener. You're creating the conditions where people can can play. And, you know, you lose a lot of IQ points. I mean, physically, you do not have access to a higher executive function when you're under threat. And, and it's so important to have those spaces to your point so that that play you can either initiate play before you're ready to as a way to kind of get in that direction, or you can actively create conditions where there is a psychological safety and trust and and a lot of, like, LUMA methods. Like, everyone understands the rules of what we're doing. All of those things are a way of of creating conditions where people can get access to that playful function. Yeah. Yeah. Well and I just I think play all over the place, you know. And, I was raised by mister Rogers. I don't know if you know who he is, but, it was lovely. I was on the board of, WQED where mister Rogers neighborhood was in the basement. So it was really fun. And he said, play is the work of childhood. And I think we need to be thinking about lifelong learning, and we're all sort of kids. And, you know, we have to we have to engage in that way, and that gives us that psychological safety. But we've gotta play all the time. And and I'm making physical things based on what the AI is saying. And it's horrible, but it's weird and it's fun, and it's it's, like, a little snarky and goofy. But that's part of it is sort of play. Yeah. And the and buddy system too. Play is more fun when you've got It's not yeah. This is collective. Yeah. And I think that's the interesting thing is we're getting smarter about kind of what works in collective intelligence. There's a wonderful group called CrowdSmart that's exploring, how to actually get more diversity of thought, more gender diversity, etcetera, through the act of anonymizing it through collective intelligence tools so that the best ideas come out, not the loudest person, and and making that scalable. And there's another group called Polyplexus that's actually helping find weirdos to solve big science problems, in new ways. And so there are some little glimmers of hope out there of new kinds of platforms. My own intern, Netta, is working on stuff around power dynamics and cognitive regeneration. And, you know, we don't we've got time management tools, clocks, watches, calendars, day timers. We don't have any cognitive management tools. We don't treat cognition as a human right. And and I think that's yeah. I posted something about silent spring. I think there's this danger that we don't actually focus on helping build capacity, and yet we could. Yeah. Well, I think that's a delightful point to land on and, and and a hopeful one. I wanna thank everyone for joining us. Reminder, this recording will be made available. Links and things we'll try and extract. I think we have QR codes and things. We'll consolidate all that and and make sure it's easy for folks to get to it. Mickey, thank you so much for taking the time. It's always It was always a pleasure, Mark. Yeah. And I always feel like this can go for a few days. We need to have them, you know, food and a little, you know, probation. Yeah. Have have a nice break, a little, you know, an Arnold Palmer, a little ice tea and lemonade, and then round two. Yeah. Yeah. I wasn't saying that. Thank you so much everyone for joining us, and thanks again. Thanks for everyone joining us. Absolutely.