#519 From Bureaucracy to Agility: Jack Skeels on Leading in the AI Era

[00:00:00] 

Mehmet: Hello and welcome back to a new episode of the CTO Show with Mehmet today. I'm very pleased joining me, Jack Skeels, CEO of Agency Agile, and maybe you would guess from the company name more or less what we're gonna talk about. We're gonna talk about being agile, [00:01:00] but this time, you know, from a different perspective, as the audience knows, I don't like to take much of the.

Light of my, uh, guest. This is why I would leave it to you, Jack. Just, you know, tell us a more about you, your journey and what you're currently up to, and then we're gonna start the discussion from there. So the floor is yours. 

Jack: Wonderful. Well, thank you very much for having me on the show. I, I really appreciate it.

And, uh, this is my, my homeland audience, if you will, from the perspective of, uh, I, I came up in, uh, technology originally back, they called it programmer back then. Yeah. Programming and project and program management and the like, and, and built a technology. Consultancy in the 1990s and we won Inc. 500 awards and all that kind of thing.

Um, but I had, I dropped into, um, grad school after that to see how we did it and, and then, um, eventually ended up at Rand Corporation, the think tank in Santa Monica, California. And [00:02:00] my job was to study how knowledge worker organizations work. And I became a researcher and analyst on that, on that topic a few years later.

I ended up working at a complex project driven organization, commonly known as an agency, a marketing and consulting agency. And I noticed that people didn't really know how to run these organizations well, and that began a journey, basically a 20 year or so journey now, um, that resulted in agency. Where we've worked with almost 200 organizations, helping them run better, faster, and happier.

And uh, and more recently, in the last year or so, my book Unmanaged, which talks about some of the key principles that go into how you do something like that. And, uh, and so, and here I am, and I now we're in the age of ai and it turns out, by the way, a lot of the principles that we teach are things that you need to do to make AI work for teams.

Mehmet: Absolutely. And thank you again for being here with me [00:03:00] today. You know, it's, uh, an honor to have you jack with me on the podcast. So, uh, it's good. Like you, you explained to me, you know, the turning point, but I'm curious about, um, why these, you know, and I know like you describe it as, uh, you know. Some, some dysfunctions, right?

Why that happen? Why do you think clients face these challenges? Um, which you, you like to call hidden dysfunctions? Oh, that's very good. Yeah. They've done Why, why no one, why no one noticed them? Actually. 

Jack: Yeah. I mean, this is the, so. People think of managing and, and a lot of our idea of managing is, uh, sort of historically driven around military concepts, command and control hierarchy.

Um, the idea of leadership as being, um, telling people where to go, that kind of thing. And, um, and, and some of those techniques did work well in some situations they weren't even the best. But, and I go into this in the book, in great detail. [00:04:00] In the Industrial Revolution. And by the way, the, the idea of managing has only been around for about 140 years.

The word, the word used to mean to take care of animals like manger and manually, uh, dealing with animals, livestock, and that kind of thing. Um, and so it has sort of a, a strange history to it, but our idea of managing was quite flawed and it. Though it worked well when we had factories and and and things like that, and unskilled workers.

Today's workers are so in incredibly intelligent and. And the idea of what do, what does the worker really need from a manager is a, is a cornerstone question every manager should be thinking about. Now, the, the answer to this is workers need a lot less than you think they need for managers and they need some different things from managers.

And when you get that right. Teams, productivities go through the, through the roof. You know, we see boosts, [00:05:00] 50% boosts in productivity, um, dramatic reductions in rework, all that kind of thing. So a different way of managing is needed and, and part of it is that the work, to answer your question, maybe, sorry, coming a little bit circuitous here to answer your question more, more head on is the workplace today is so different than it was a hundred years ago.

Even 50 years ago. But the, with such skilled knowledge workers, what we really wanna be doing is unleashing them. Okay. Enabling them to go do, use their minds and do great work rather than sort of corralling them or bringing them in and constraining them. And that was the idea behind Agile. And I took some of those ideas, but we've added a lot more stuff.

And the idea is, at the end of the day. We basically unleash workers to their full potential, and that that unleashing unlocks a lot of business value, profits, [00:06:00] happiness, all that kind of thing. 

Mehmet: Like, you know, you, you walked us through a lot of things here, uh, Jack, right? So. One thing, which, you know, you mentioned, and this is something I didn't prepare to ask about, you mentioned about, you know, like we are using something which is kind of, has been with us for a long time and now, you know, the time change and you mentioned ai.

Why do you, you know, like why some of the management concepts in general are sticky? Like, why we don't change them faster? And I know, uh, at least, you know. From reading history, and I like read history about different topics, that it seemed that this happened multiple times during, you know, since, uh, let's say the beginning of the industrial age, right?

Mm-hmm. And you know, it used to take like 50 years, sometimes a hundred years under a change happens. What's the main reason for this? Uh, is it like, because we, we, we. We [00:07:00] refuse, uh, change. We don't like change. Or is it because, you know, management, it's, it's something different. It's, it's a different beast that we need to do.

Jack: Yeah. No one's ever asked the question that way. I think that's phenomenally asked. Um, so there are a couple pieces to it. One is we don't actually study and teach managing. Okay. I have an MBA. Okay. I can tell you, and in fact, one of the great researchers on the topic, a guy named Henry Mintzberg, wrote a whole book, um, which basically said Managers not MBAs, I think was the title.

And he pointed out that the modern MBA program doesn't teach. How to manage, how to manage people, how to be a great manager. What it, it's called Masters in Business administration. It teaches you to be an administrator, which is a very different thing than actually being a great manager. So partly we have an education system problem.

Right. Um, the other thing is that there are, there are two other things. One is I think, and you sort of hinted at it. [00:08:00] We tend, it's a good thing, by the way, as humans, we tend to repeat ourselves, right? If we see something working, we mimic that, you know, we copy it. Um. But the other thing is that the, the rise of the, what we call the meritocracy, right?

The, this is the idea that if I work really hard, I get a promotion and that gives me a title. And that title is a manager title. And that means I have people that I tell them what to, you know? And this idea of a hierarchy gets built. Around the idea of the better people are higher in the hierarchy and, and all kinds of sort of somewhat toxic things that actually make that.

What we call hierarchical behavior, which is not all that good a behavior that makes it much more prevalent. It reinforces the model. Um, when I get a manager, when I get to be a manager, then I get to tell people what to do. Um, and so the, all, all of our models are messed up and there's this resilience structure called the meritocracy or the hierarchy [00:09:00] that it seems to resonate for us a little bit.

Right. Maybe more than it should. Right. Um, but it is outdated for sure. 

Mehmet: It is outdated. Hey, this is, this is the world. Yeah. And, and we, we, we saw it, you know, not only in management, in, in multiple, I would say domains also as well. Uh oh. Yeah. Uh, so. From your experience, uh, are there like any kind of warning signs that, you know, that will give us that okay, we have this over layered thing that is coming.

You know, we are, we started to go wrong direction, so. We should get this, uh, awakened call that something have to be done. 

Jack: Yeah. So there, you know, one of the other things is we don't measure as managers, right? And the, there's, there are two problems. And this is a fairly complex topic and I'm, I'm gonna do my best in a, in a short window here.[00:10:00] 

One is, is. Our, our modern organizations are what we call multi-project organizations. That is any given person, a worker, a maker who does something, is usually allocated to multiple projects, which means they have multiple managers as well, and multiple other people that they need to interact with. And, and the, the challenge in this environment is that that's a lot of managing, a lot of managers, a lot of distractions, a lot of interruptions, that kind of thing.

In, in those environments. There are a couple things you can see. The first thing I noticed this is 15 years ago, is when the more multi-project, multi allocated your workplace is. Okay. The, the more managers you have. These man, the ratio between managers and workers can actually get quite severe. I'm, I'm working with a, um, 45 person agency, um, development shop, actually, um, a development shop that has, um, I think we counted 18 [00:11:00] managers out of 45 people.

So 18 managers and 27 workers. Okay. And that, that's one manager. What, how, what's that come out to? It's, it's, it's a two to three ratio. Okay. Two managers for every three people that are actually doing work. And That's crazy. It's a crazy number, but it's a result of how we overcompensate for the complexity of a multi-project organization by Overmanaging in the world.

And that's why my book, but to give a plug to that unmanaged is the idea, what if we manage less in this environment? And so that's, that's the real opportunity. And, and it turns out you can, if you're gonna do that, you need to measure. And to the real two measures that we use is one, is we ask people how productive are you during the day?

How many hours of truly productive time do you get during the day? And the thing we find is most multi-project, multi allocated organizations, people trying to do work, get three to four hours [00:12:00] of productive time a day. And that's a tragedy, you know? And it turns out you're not gonna get eight because there are meetings and there's stuff that happens.

But if you can go from, like we do with a lot of clients from three and a half hours to five hours, that's a huge percentage increase. In the number of hours that are available and the like, and you can literally boost productivity of a team and individuals by 20, 30, 40%. Um, another thing that happens with all this complex environment is people don't understand the work.

People get poorly briefed or they forget, and that's a whole other thing that you can actually deal with. If people understand the work better, they produce better work and we get better results.

Mehmet: Great, great, great insights, uh, again, Jack. So now let's talk about when you come in, right? So when you step in, um, what are like the first, the first thing that you do usually? 

Jack: [00:13:00] Well, you know. I think you alluded to this earlier, and I'm sorry I didn't follow up well on it.

It's a, it's a, even in my brain after all these years, it's a very complex topic, so I apologize for that. And your questions are great. The, the managers, lemme start that over again. Managers don't understand the situation. Okay. That is. I, I'll give you a real quick example of this is sort of like I can look at time cards, okay?

And I'm gonna see everybody's fully utilized right now. They're fully utilized. They're going to a lot of meetings and stuff like that, but it doesn't mean that they're productive, right? And, and so what we wanna do is actually help managers understand what's really happening. And that's why we use these measures.

Like how many productive hours are workers getting in the day? We don't ask managers how productive people are, what managers will say, by the way, they'll say, boy, they sure could be more productive if you talk to [00:14:00] workers. Why aren't you productive? And they're like. Well, 'cause it's, we're on the meetings, I'm constantly getting interrupted, both of which are managerial functions, right?

So what we wanna do is actually restore the first thing we do. We call it the leadership foundation workshop. We teach managers how to really look at the organization from a performance perspective, right? Mm-hmm. And the job of the manager is to actually help the team go faster, not to manage a lot, but help the team go faster, 

Mehmet: get the team, go faster now.

What about the friction that would happen because again, the same topic that I mentioned, so people, they don't like, you know, changes as I was saying. Right Jack? So, um, how much of I would say. I'm not sure if frisk management is the right word here, but you know, like putting into consideration that, uh, these changes will, will, will face some pushbacks.

Another scenario, which, you [00:15:00] know, from, I have watched a lot of companies happen, especially when they move from. Being like this nimble, you know, small company to becoming a more like, mature, like, uh, large company is the amount of bureaucracy that get gets in. Yeah. Um, right. And of course, like I, it's just like a coincidence.

I was speaking with someone, uh, yesterday or today before yesterday. He was telling me, look like we have perfect culture in our place, in our company. Uh, really people are nice. But you know, like you go talk to, to the one who's responsible, the decision maker, he says, you know, you know, like we, we found it like this way nothing can be done.

Or like, yeah, go talk to someone else. Once everyone agrees, I will come and agree. So these scenarios like are frictions usually, um, I know like there's, I don't like the word like magic wand or a secret sauce or something like this, [00:16:00] but how, how to deal with such situations. Also, Jack. 

Jack: Well, I, I think there are two pieces to, to what you're getting at.

One is, uh, in fact we only work with, with CEOs and their organizations if the CEO believes in the idea of, of change and, and empowering people. Right. Because if the, if the CEO doesn't believe in that, all of that trickles down into the rest of the organization and you just can't change it. Right? Uh, having a culture of, of ask and understand and measure, uh, requires patience and, and a dedication to trying to do better for everyone.

Right? And if the CEO's not organized that way, then. No one else will be. Right. Um, and and part of the process also is, like I said, you know, we do this, this very inexpensive quick workshop where we come in mm-hmm. And get everyone to understand, wow, here are all these ways we could be faster and more productive.

And, and in fact. [00:17:00] Being, having makers be faster and more productive isn't about making them work harder. It's about unleashing them and letting them be productive and get, get the work done during the day. Um, people don't realize, by the way, um, the hardest job in any one of these organizations is to be one of the people doing the work.

'cause you come in during the day, all you wanna do is just get your work done right and, and feel good about it and feel like you did a great job and the organization itself. Is sort of organized around interfering with that, keeping you from being as fully productive as you could. Right. So I, I think, you know, we, we do a lot around the awareness when we come in, helping people understand where the, where the organization could go to what it is that we could fix.

And that's the beginning of a whole change cycle that we use for, for transformation. 

Mehmet: Great. Now have you seen any situations where, you know. The changes started, you know, [00:18:00] in a very positive way. Right. But after a while it gets stuck in the middle, so, uh, it gets stalled. Right? So, so how have you seen what are like the one of some of the.

Things you advise, you know, your clients to do, to make sure that we don't have, and you know, usually we saw this in projects and you've worked in, in that space also, Jack. So at the beginning of the project, everyone is, you know, excited. Yeah, let's do this. We need to, to, to start kick things off. First couple of weeks, even months per everything, working perfectly.

And then we reach a stage where, I don't know. Is it the energy that went down? Is it like, again, the processes that stopped us, uh, what are the reasons for such things to happen and how you advise people to keep this, you know, you call it sticky change, right? We don't want this to be a like one time thing.

Jack: Yeah. You know, look it, it [00:19:00] happens in projects and it happens in transformations, right? Mm-hmm. Um, I'm gonna talk about the project example first, though, which is there tends to be, we are incredibly optimistic human beings, right? And, uh, optimistic creatures in that way. And, and we tend to think. Um, despite what's happened in the past, this, this time, it will be better, right?

And, and we, if you, if you map out as I have with hundreds and hundreds of projects over the years, you can see this cycle where we start with this, I call it the plateau of optimism. Okay? Everyone's really happy. Like you, exactly like you said, everyone's happy and excited. And what, what it usually is, is that we don't actually understand how hard this project is going to be.

And that's actually a very nice moment. Because we're ignorant in that way. Right? Right. And as we get into the project and learn how hard it is, learn more about the scope. We get better scope. We learn that we didn't know this, these things. And it's harder than it is, harder than it [00:20:00] seemed. We drop into this thing I call the chasm of despair, right?

It's the valley of despair. When we realize, oh my gosh, how are we ever going to do this? Um, this is a much harder project, et cetera. And, um, and the question is whether you've got what it takes to climb that, that mountain, to get out of that, out of that chasm outta that valley, right? Mm-hmm. And, and this is a very natural human cycle.

It happens with a lot of things. The, it happens and it takes, it takes a commitment to doing a couple things. One is you need to be thinking about how you wanna do things from the very start. And this includes good scoping. We teach great scoping techniques, how to do a great kickoff, how to get people aligned on things.

And likewise, when you're doing transformation, you get everyone aligned on what the problems are that we're gonna fix, and also how hard it is. How hard is going to be to fix those problems and, and sort of getting everyone connected that [00:21:00] way. Actually it makes the journey go a lot better and um, it removes some of the optimism from the project for sure.

Okay. But it adds realism, which is very, very welcome by the end of the project. 

Mehmet: Um, you know, like this is kind of, you know, opened my eyes on, on something else. I'm, I'm hoping also for the audience that will do the same because you mentioned, you know, crossing the chasm. Right. So, um, because it's the same life cycle of, of like how, how we talk about when a new startup, they try to find their product market fit and you know, they have the early adopters and then they have to cross the couch so they go to the majority.

So I think. Same when you mentioned this, like immediately it's like. How doing transformation or even a project, in my opinion, it's, it's run the same way as you run a startup. And then you need to, to make sure that you cross the chasm so you don't get stagnant there. And maybe you, you, you go to failure.

So this has made a [00:22:00] lot of sense to me now. 

Jack: Yeah. I'm gonna jump in real quick. So the equivalent, the equivalent in the, you know, in the, uh, Clayton Christensen and all that kind of, uh, chasm, that chasm model is really the. The amount of, I believe, and a lot of the research shows this, the amount of research you do on the broader market.

Mm-hmm. Beforehand. Okay. So this idea of completely lean startup, we will just find someone who buys it and, and that kind of thing has a, a fatal flaw, which is you still don't know how hard the big market is. Right. Right. And, and if, and if it's even there. And so that's, that's the equivalent of the project scope thing, which is, you know, finding a bunch of early adopters in a what will never be a big market is not a big win.

It's, it's a recipe for just spinning and, yeah. So, go ahead. Sorry about that. 

Mehmet: No, no, no. Yeah, I'm happy. I'm happy. You, you, you mentioned this also, Jack now. When you did the introduction, you know, and you start to talk, you mentioned [00:23:00] ai, right? So, and everyone talks now about the big topic, which is future of work.

So from what you are noticing now and working with a lot of clients, how the AI is changing the whole landscape of, of management and, you know, uh, the way that companies should be organized, uh. I don't want to make it a loaded questions, but because you know, I'm hearing sometimes we are heading toward like.

More agile, you know, quote unquote organizations maybe with less people. So we'll not have this problem. I hear some other experts saying, no, actually now it's, it's like more challenging because not only the AI part, and now we are talking about. You know, uh, something that comes into the mix, like hybrid working and you know mm-hmm.

People who can be sitting anywhere. Um, you know, culture, uh, might, uh, become like a little bit a challenge because people are not seeing each other's, [00:24:00] uh, on daily basis. We're not coming to the office. What are you seeing? What are the trends there, Jack? How, how things are moving. 

Jack: So I, I mean, you know, it's a huge topic.

Okay. We could spend hours on this. Um, I'm, I am gonna refer your listeners to a, uh, piece of research done, um, in 2023. You can go onto Google Scholar, a scholar@google.com. Yes. And it's called GPTs are GPTs. And the authors talk about the, the workforce implications to gpt. Okay. And the, the idea of the paper is that gpt are general purpose technologies.

And that's, that phrase is a special word in the world of industrial economics. And a general purpose Technology is something like electricity or the internet, right? In other words, you can't say. How, what's electricity gonna do for jobs? Right. Okay. You can't because it's, it's a general purpose [00:25:00] technology and it really depends on the job, what electricity will do for it.

Right. And the same thing is true for ai. So, um, anyone who gives you broad prognostications, um, is just making shit up to some degree. Right. Okay. Now. I, I can tell you, and I, uh, you did the intro on Agile and there are a lot of people talking about agile and, and ai and it in the sense it's justified. I mean, all of our work that we do is sort of agile inspired, but advanced beyond the basic agile stuff and the like.

Um, and we recently did a, a very, very cool, um. And it's not even a pilot 'cause it's real, it's a production, it's real life project with a big agency and their big automotive client in the like. And um, it, there was some very interesting lessons out of it. And the, the idea was this. And they had a great toolkit, a great AI toolkit in the like, and the whole agency had been using the toolkit, but they [00:26:00] hadn't been really getting much for results as is the case for a lot of people.

They were enjoying using the tool, but. No big business impact. And it turns out, and we, what we did is we set up a pod and we gave that pod the assignment. And this is the people who were gonna do the work, but we trained them in the tools, but we gave them the assignment and the help in re-engineering the way they were gonna work with the client.

And we got the client on board with it as well. And together the team and the client re-engineered how to work faster and, and better and all that kind of thing. And, and the reserve, the results were quite miraculous. And for as an example, and this was just their first six months of the work, they cut the span time for their typical project in half.

Okay. And a lot of the span time was reviews and comments and revisions and all that kind of thing. So they cut the span time in half and they cut the labor by 35%. Okay. So still a lot of labor in there, [00:27:00] but much more efficient labor. And, and the client change the way the client operates changed as well.

And so this, I think this underscores a couple important ideas. One is it will restructure how individuals do work, and certainly even how a team works together and what they work on. Um, but it'll also restructure the value chain if you do it right. Okay. And this is where the AI thing comes in, because if we had just, and the rest of the agency, this is an AB test, it's kind of a cool test.

The rest of the agency was just letting their teams work with their clients and use the tool if they want to, not getting any big results. Okay. We got amazing results because we focused the team on. How do you implement ai? How do you make your process better? We said, forget the process you have. You come up with a new process and that enabled ai, and there's a lot of research that supports this.

It is the undoing of organizational structures that enables AI to have big impacts, [00:28:00] okay? Mm-hmm. Is that if you don't undo your organizational structures and process, then those, those structures and process constrain ai. Okay. A AI can only optimize within that. And what you really wanna do is, uh, release it all.

Hey, where do we put, let's put electricity everywhere we can. What do you know? That kind of thing. So, 

Mehmet: uh, you know, it, it totally makes sense. And this is very similar to, you know, when people started to talk about, uh, digital transformation and, you know, shifting to cloud. 

Jack: Yeah. 

Mehmet: And they start, say, hey. Like we didn't see benefits.

Yeah, of course you didn't see benefits because what you are doing wrong or not wrong? I mean, in Unoptimized way on pen a paper, you just took it and digitized it in the same way. Yeah, yeah, exactly. Yeah. 

Jack: Yeah. 

Mehmet: So, so, and the same thing with ai because people, and this, this is something, maybe I'm being philosophical a [00:29:00] little bit.

People looks for shortcuts and for magic wands, I'm repeating myself because they think, oh, this AI will serve fast. I'm not sure if you have seen this. Uh, you know, like funny, uh, kind of a beam or, uh, you know, sketch where they put like every CCEO is asking for ai. When do you want it now? Yeah. Right. So it's, it's like everyone want to be on the ai, but actually they don't know that they need to do some work before bringing the ai.

Now still talking about 

Jack: ai. Yeah, I think, yeah, I think there's a reason for that, which is that in general technological, and that's why the GPT or GPT paper I think is really cool. Right. In general, technologies come as individual tools. So, hey, we've got a new programming language, or we've got a new programming platform.

How do we implement that platform to replace the platform we had before is actually a fairly straightforward kind of thing to do. Right? But when it's, but it's in general purpose technology. Like while we have [00:30:00] electricity all of a sudden, yeah, okay. It's just what are you replacing? You know? And, and so if you treat it like a little tool, all you'll get is little tool changes to it.

Yeah. 

Mehmet: Uh, I know like it's kind of fictional, uh, question and maybe science fiction question, but are you seeing AI really being placed in managerial positions, take decisions? 

Jack: Yeah, I, I think, you know, I've been publishing a little bit of stuff on this. I have more to come. Uh, I think there's a trust problem with ai.

Okay. And people talk about this a lot. I've, if you. Look, look me up on LinkedIn. You can find, um, I put together an AI trust matrix, um, a trust gradient. Um, there are some things that AI can provide quality, high, highly trustable input on. And, um, and the, then there are a lot of things that can't. Right? And, and the problem is the AI can't, it doesn't do a good job of telling you where it is on [00:31:00] that gradient.

Right. And I think that, yeah. A lot of people have written on this, but the idea of judgment, um, quality assessment, things like that, that are hopefully managerial functions, right? Yeah. Those things are not easily replaced by AI right now. Um, the, the funny thing is that AI can make great observ. It's like AI can watch a manager in action and, and critique a manager in terms of here's some skills you need to get better at, and your conversational, like, Jack, you're talking too much.

That kind of thing. They can ab absolutely do that observational stuff, but the intuitive of what it really means and how, how I might coach Jack that's. That's really more of a human thing that humans can outperform AI in that all day long AI outperform humans in the observation task, for example. 

Mehmet: Yeah.

And automation and, you know, the, these kinds of tasks also as well, you know, I was just checking, [00:32:00] so I, I don't, uh, you know, replace any, any word with the wrong word, so, uh, on the trust. And I think many people, maybe they have seen this. There was a, uh, a training manual from IBM dated back 1979 saying A computer can never be held accountable, therefore a computer must never make a management decision.

Jack: Oh, that's it. That's awesome. I'd love for you to send that to me. That's great. I love That's fantastic. Yeah. 

Mehmet: I saw, I saw it multiple times. Like, um, you know, and now, because you mentioned this about the trust. Um, yeah, because like even myself coming from technology perspective. I heard things, I'm not sure about it here and there, like how these LLMs are programmed, uh, how they are trained to behave.

Uh, and I spoke with some really, you know, experts in this domain. They told me. Little bit scary, not scary, I would say, like things that I, I should and everyone should be, uh, [00:33:00] worried about, about. I don't know if manipulation is the right word. 

Jack: Absolutely is the right word. Yeah. No, no, no. I'm, I'm like 

Mehmet: I, that, I don't like to be harsh.

Right. So, but yeah, it's a manipulation. Yeah. So actually, and this is a fact, this is where everyone talks about ethical ai. Every single company that developed an LLM, all of them without, without any, you know, exclusion, uh, they can manipulate the way. The chatbot can reply to you. Uh, well, in 

Jack: fact, yeah. So let, lemme jump in on this.

This is, I find this is a fascinating area. In fact, I'm doing some work with my old alma Mater Rand corporation around this, uh, recently. So if you take, and I forget the exact numbers, but if the, the actual LLM model is not all that big. Okay. Mm-hmm. The actual, the ve, what's called the vector, the token vector space is I think five gigabytes or something like that, which is a frighteningly small number.

Given that it came from like 30 or [00:34:00] 90 terabytes worth of information, right. Can, there's not a lot of knowledge in it. What there is inside of that, that token vector space, that's the LLM, the large language model is. There are relationships between words. Okay. And, and, and that, that can be very valuable if the, if the relationships between those words are part of a factual truth, it can pull it out like just like magic, right?

And it pulls out those tokens, but then it's gotta turn it into something we can understand. So 95% of all the data that A GPT uses, okay, 95% is what's called the transformer layer, which takes the output of the vector space math and converts it to English that we understand. Okay? So that is, it takes this, these, these word token combinations and it figures out how to tell it to us in the most persuasive way.

Now if it, [00:35:00] if in fact the tokens that it got out of the LLM are really weak, like they don't tell a truth, they tell an ambiguity, they tell whatever. It will still make it sound like a convincing truth. Okay? And this is a scary thing. They're the LL, the gpt are designed to be highly persuasive. And even if it doesn't know something, it will try to persuade you and give you the best.

And this is the deceptive thing about, uh, about GPTs is the GPT doesn't measure itself by whether it's correct. It measures itself by whether it gives you an answer that you like. Okay. And if I said there's a person, me, if I said, do you have a friend who, who answers you, not truthfully, but in ways that you love hearing, even though it's not the truth, what kind of person is that?

Okay. I mean, I think of that as a sociopath, right? That's sociopathic that, right? And I'm, I'm gonna tell you whatever I need to tell you to make you [00:36:00] happy, regardless of whether it's the truth. 

Mehmet: Uh, yeah. Yeah. It's, it's scary. It is scary, you know, and this is why I, um, I, you know, the more I go deep in this, I also started to dislike the word intelligence in this.

Oh, yeah. 

Jack: It's not intelligence in it. Yeah. Yeah. 

Mehmet: It's not intelligence. It's, it's, it's just a soft, it's a computer software. It's a com It's a piece of code. Yeah. I'm not saying it's simple. Don't get me wrong. I'm not like underestimating the amount and you know, of, of time and money and. You know, uh, that the developers have put to train these models, but it's not intelligence.

It's like driven. And I was having like also another discussion about how we were like hooked for these tools and we are able to use them. So it's absolutely, you know, it's, we need, I know we need hours to talk about this. Jack, uh, Jack, tell me a little bit about the book. Like what are, like the major things you discussed there?

Like what people can, you know? [00:37:00] Distill from the book. So just let us give them a, a, a, some teaser so they can go and check it out. 

Jack: Yeah. Thanks. I appreciate it. Uh, you know, well the book was written largely. Um, I'm just gonna say a little bit about the beginning journey and then I'll, I'll come back to your question.

Um, I had had a moment back, uh, almost 20 years ago now, where, um. I'd been laid off during the global economic downturn, but my manager said I wasn't, hadn't been a very good manager. I was running a, um, 105 person, uh, Los Angeles office of a co, a big consultancy, and I thought, I haven't been, wasn't a good manager.

You gotta be kidding. I'm like, you know, I'm MBA and I studied management sciences at Rand Corporation and blah, blah, blah. They started to write the book and, and I did all the research and I realized, well, I didn't actually know how to be a good manager, so I went out and did it. And that was the, the genesis of the company that I have now, agency Agile.

Um, but the book essentially is that, is [00:38:00] that whole story. In fact, it starts with that story. Um, and the story of me trying the techniques out that I'd done the research on and them succeeding and saving a big project, um, basically there are a bunch of principles that you as a manager need to understand and we teach these in our transformation trainings, projects we do with, with our clients.

Um, but you can use it as a book to sort of guide some of your decision making as a manager and certainly to raise your awareness as a manager. Uh, among the many things we, we talk about are the. How important it is to have people understand the why. We have a model called Why, what Go Grow, and having people understand the why dramatically improves project success rates.

And then the what, which is the scope. We, we, there's a whole set of principles around that. And then the go, the execution, and then how to grow people. So a very wide ranging book, 380 [00:39:00] pages, we've won five awards for it. I have no idea why I can't believe anybody's ever read it. 'cause I, I, after having worked on it for a year and a half, I can't even read it.

But, uh, people say nice things about it and I, I think it's generally useful as a man for a manager. 

Mehmet: Uh, yeah, uh, definitely. Uh, Jack, like if, if we want to, and this is part of the thing that I, I try to do it with my guests. Uh, we have a new generation coming up now and, you know, they're ready to take on live so.

If you want to give them a piece of advice to, to do, you know, this whole organization, uh, management in a proper way from day one, and they avoid maybe the mistakes that us, I mean, I mean, not all of us, but I mean the, the, the previous, uh, generation have done, what do you advise them to do? Like, how, how do, how do you, how do you.

Push, you know, the [00:40:00] younger generation to, to be the driver for the positive change. 

Jack: You know, I, I think certainly if they are starting their own organization, it's a lot easier than to be the internal change agent as a, as a, a new hire or something like that. I, I've learned a lot from, I, like I've worked with.

Probably 300 executives over the years and, and I learned a lot from them. And one of them, very successful guy, Ben Kirchner, uh, had an agency called Elite, SEM, and. And Ben had built this amazing org, like it was like a hundred people and it had eight managers or something like that. Right. And I was like, wow.

How did you do that, Ben? Well, what was the secret? Because he hadn't, he hired me after he'd done that. I mean, and. I said, how did you get there? He said, you know what? I worked in a traditional setting for a while and I hated it, and when I started my company, I said, what I'm going to do is every time there's a decision I need to make, I will think about what they would've done in my old [00:41:00] company.

I'm gonna do the opposite. And basically he, and this is, I've, I've, this has come up as a theme multiple times with clients of mine over the, over the years. Is this idea really think different? Go back to the old Apple theme, you know? Yes. Do the opposite and see what happens. Okay. Because the orthodoxy is not a powerful orthodoxy.

It's, it's a very fragile one. And you can go very far like he did, just saying Maybe we don't need to do that. Okay. Or what if we did the opposite? So, uh, ask, think, and ask questions is, I guess my, my broader answer, 

Mehmet: uh, if, if you allow me to add one thing, I, we know anyone that comes today to be, and they ask me, like, not necessarily they want to start a company about anything.

Maybe they are still, maybe they want to go to a normal career. The first thing I tell them to do. Uh, don't be shy of asking why anything is done the way it's done today. Absolutely. 

Jack: [00:42:00] Yep. 

Mehmet: And, and don't be shy to ask can it be done better? Right. So if you have this mindset. You know, I, I, because you know, like, maybe I was like this.

Yeah. You would be a little bit hated out. Just joking. Yeah. Like people don't, like, don't like change. No one likes change, so you would be questioned, like they will consider you the weirdo guy, you know, over there. But what I. I'm happy about and I'm optimistic about, I see the newer generation I know, like from my own daughter for example.

Uh, they question a lot and, and they ask like why it's done this way. Can we do, can we have done it in a different way? And exactly the the thing, different thing you mentioned, Jack is a hundred percent, uh, you know, relevant. To to, to the whole thing we were, uh, discussing today. 

Jack: Well, let me, let me throw something in real quick.

So I, this is in the book a more detailed version of the story. Um, but at Rand Corporation, I had an opportunity to have some of the top researchers there explain to me, [00:43:00] and they did it through a very fun little lesson if, if you read the whole vignette in the, in the book. But basically it came down to this is, um, I, the, they look for people who ask questions.

This is. Top flight, think tank in the world, and the only people they hire are not people who know things, but people who ask things. And they explained it as, as this, having the answers is the end of all learning. Okay? We don't hire people who aren't learning. Okay? Having the questions means that you're still learning, you're still growing.

And especially the question like you said is, is this the right way? Or if it's not the right way, what could be the right way? Those questions are where you grow as a human being and where we move forward as human beings. So your, your comment is spot on. I think it's a, a very, a noble way to live is to actually question everything.

And it's not negative. It [00:44:00] certainly is weird, right? But it's not negative. It's very positive. You'll live a very good life if you do that, by the way. 

Mehmet: Yeah. And, uh, you know, you'll not only be benefiting yourself to your point, noble thing because you, you might be benefiting a, you know, bunch of people, maybe a whole generation by questioning, you know?

Yep. What I like to call it the sit, the status quo. Right. So, absolutely. Uh, as we are coming to, to the end, uh, Jack, uh. Where people can get in touch and, uh, you know, if they want to, uh, leverage your knowledge and maybe, you know, work with you. So where they can find out more. 

Jack: I, I think the best way to get ahold of me is on LinkedIn Jack Skeels

Uh, well, there's only one. So it should be pretty easy to find. Um, we're, uh, we're actually rebranding the company shortly and that'll be obvious on there. So better to use me as an anchor that way. The book you can find on Amazon we have in other places. But no, I can tell you empirically, nobody buys my book anywhere other than Amazon.

I don't know why I don't think anybody buys [00:45:00] books elsewhere, uh, these days. But, uh, it's on Amazon, uh, five stars and the, like, if you're, if you're curious. And, uh, I think the ebook is like especially cheap right now. So, um, if. 

Mehmet: Great. So, so for sure. Yeah, they don't need to go and search. So everything, uh, you know, all the links will be in the show notes of, uh, anyone is listening, uh, on, you know, your favorite podcasting app.

You'll find them there. If you're watching this on YouTube, you'll find that also as well. Well, Jack, really, I appreciate very much, you know, the time you gave us today, 

Jack: I feel like we could go for. Yeah, another two hours. Um, 

Mehmet: AB absolutely. Abso no, no. You know, like, uh, I feel like we need to do part two, part three maybe.

Absolutely. Let 

Jack: me know. I'd love to. Yeah. You ask wonderful questions and you have a great mind, which is, uh, you know, 

Mehmet: thank you. 

Jack: A benefit for your audience. I'm sure. I'm sure that's why they won your show so much. Thank 

Mehmet: you. Very, thank you very much, Jack. And uh, this is how I end my episodes. This is for the audience.

Uh, if you just discovered this [00:46:00] podcast, thank you for passing by. I hope you enjoyed it. If you did, so give me a small favor. As you seeing and listening, we are trying to do an impact, so. Subscribe, share it with your friends and colleagues. Maybe someone will benefit out of it. And if you are one of the people who keeps coming again and again, thank you for the support.

Thank you for making the show trending on the Apple Podcast, uh, in multiple countries in the entrepreneurship, uh, category. This cannot happen without you. And again, uh, I'm very happy that everyone encouraged, you know, uh, all the effort that I'm doing. And as you know, uh, the, my. Book also is now out, uh, on, on Amazon.

Uh, it came out on 10th of September. Um, so you can get it. Uh, thank you. Thank you, Jack. So you can get it from Amazon and like Jack, I don't think someone buys any book under the Amazon, so it's the only place where you can find it. Nowhere to next. Uh, and it's the lessons that I learned from my personal experience and talking to people like Jack on the podcast.

And until we meet next [00:47:00] time, thank you very much for tuning in. Thank you. Bye-bye.