# Artificial Organizations: Human Judgment and AI Infrastructure

**Podcast:** Product Momentum Podcast
**Published:** 2026-08-25

## Transcript

What's really fascinating and what is changing probably because this technology has really come to the fore is that you can now start thinking about a company that really is compounding every piece of information that it's gathering and using it to inform better judgment systems and decision making across a company.
It's this idea of a judgment system and judgment infrastructure that supports.
Teams, leaders, companies that want to make better decisions.
First of all, there's human judgment systems, like how we make decisions.
Most people really never evaluate the way that they make decisions.
They're not really aware of this internal algorithm that they're using to make these decisions.
And then at the same time, inside companies, the ability to make decisions is getting good insights, good information moving around the company.
That could be with humans, it can be with machines.
But having great infrastructure in place means people can make better decisions when they have better quality of information.
So pairing those two things together, if I ask many product leaders, what percentage of your time do you spend on complex, hard problem solving, strategic work?
Like when I ask people that, it normally goes in this sort of 80-20 mode where they feel like 20% of their time they're actually doing the creative problem solving.
and 80% of their time is stuck in the admin, the necessary admin of updating documents or capturing tickets or moving statuses.
What these tools can offer is that they can start to take away some of that administrative work and create more capacity actually for the problem-solving work.
And I think most people are happiest when they're actually working on tough problems.
Every call I'm on, I have a transcription on it.
It captures every conversation and treats it like a data asset.
I take those calls and I put them basically into a database that's constantly churning, saying, what are the problems that keep coming up on these calls?
What are the words people are using to describe why they're struggling with AI adoption or they're struggling to understand how to use these tools to make their businesses more effective?
How do we kind of deprogram the executive tier that's getting?
AI access and starting to build their own things the first time and joining the builder workforce.
One of the things we learned a lot in our venture studio is it's very easy to prototype.
It's quite hard to productionize software.
I can't tell you how many things I've seen or been privy to building over a couple of hours.
I'm getting super excited because it's a nice UI and people can understand the concept.
But once more than 30 or 40 people start using it, it just collapses under the strain of being non-performant, essentially, right?
Because the code is messy.
It's not optimized.
There's certainly very limited security there.
And if you've any way, you know, a large volume of people dealing with it, the machine blows up.
Dan, we just got done recording with Barry O'Reilly.
He came to us from the other side of the world.
Right.
We spent a lot of time talking about the organizational application of AI and the value of human judgment.
Yeah, really wonderful, insightful conversation talking about the nature of decision making in the AI world and how AI can strengthen human decision making.
We talked about his process for writing a book, how to unlearn and improve your practices as a product person when using machines.
Really, really interesting stuff.
So go in and take a lesson.
Let's do it.
All right.
Joining us on the pod today is Barry O'Reilly.
Barry is an entrepreneur, executive advisor, and bestselling author.
He's been helping senior leaders redesign how organizations perform, make decisions, and innovate at scale.
He is a three-time author, started lean enterprises, unlearn, and artificial organizations, and advises boards and executive teams worldwide on enterprise transformation, especially for now in this dawning age of AI.
Bauer, welcome to the pod.
Yeah, no, it's great to be with you guys.
Thanks for having me on.
So, right, why don't you bring in artificial organizations?
You know, we hear a lot about companies that are being AI first and AI native and AI this.
I'd love to hear how are you describing an artificial organization?
How's it different from those other types?
Yeah, so the way I tend to think about it is actually the components of the company.
You know, often a lot of companies previously have got a really, I guess, sort of linear view of how they do projects or work.
You know, someone has an idea, they bring a group of people together, they execute, they get stuff done.
But what's really fascinating and what is changing probably because this technology has really come to the fore is that you can now start thinking about a company that really is compounding.
Every piece of information that it's gathering and using it to inform better judgment systems and decision making across a company.
So a couple of the points I really talk about a lot are in the book.
It's this idea of a judgment system and judgment infrastructure that supports teams, leaders, companies that want to make better decisions.
And even if you just think and take those two bits apart.
First of all, there's human judgment systems, like how we make decisions.
Most people really never evaluate the way that they make decisions.
If I asked you, you know, what strategy should we do, A, B, or C, people often just reflex and come up with an answer.
But they're not really aware of this internal algorithm that they're using to make these decisions.
And then at the same time, inside companies, the ability to make decisions is getting good insights, good information moving around the company.
That can be with humans, it can be with machines.
But having great infrastructure in place means people can make better decisions when they have better quality of information.
So pairing those two things together, and this is one of the reasons in the book I called it building better judgment, speed and results with human and machine intelligence, is that you need great human.
judgment systems about how people look at decisions, look at options, look at assumptions, test them.
And you need great judgment infrastructure, i.e.
telemetry, analytics, access to data, so people can do those things together.
And really, when people talk about AI, because it's such a massive word for so many different technologies, machines are really good at some aspects of judgment, right?
They're really great at capturing information.
They're really good at synthesizing it.
They're terrible at making decisions.
That's what humans are really, really good at.
So that's what really inspired me from just the work I've been doing over the years.
I started an AI venture studio six years ago.
Nobody Studios, we've been building that.
The mission is to do 100 companies in five years.
And everything I learned from trying to build those companies, from build the studio itself, from working with entrepreneurs, about helping them be more productive, that really was the culmination of what the book is about.
Yeah, it's been super fun as well as coaching a bunch of executives.
Again, all case studies in the book from American Airlines, Slack, Spotify, Skyscanner, all these amazing companies I've worked with over the last four or five years.
So yeah, hopefully there's lots in there for people who are curious about how they can use these technologies, but ultimately for better judgment, decision-making and operating sort of model.
Yeah, we bought it.
The Silk Wormby.
ways back and right he said you know one of the core jobs as a product manager and a product leader is to have opinions and make decisions yes i'm curious you know for you know the product managers and product leaders out there who you know maybe maybe the current decision model is more uh vibes and feeling based how do they get started going down the path that you describe yeah so most people do have a like you say they have an instinct and and for many years in work We sort of rewarded people with having a good gut, you know, or they've had tenure in a company.
They worked in a domain for 25 years.
They've, you know, they've relied on instinct and that's great.
But we're in a world now where, sure, people can have great instincts, but you're up against people who have great insights.
These tools now where, you know, I used to just today, I'll give you a very simple example.
I had.
four NDA contracts I had to read from four different companies.
I had to build four statements of work, each for individual assignment.
Typically for me to do that, that would have been maybe two or three days work.
One going through these NDAs that some of them are three pages long, some of them are 15 pages long, line by line.
Then crafting a statement of work for different types of businesses.
Each has its own objectives, outcomes, uniqueness.
That literally would have taken me three days to do all that work.
Today, I got it done literally in four hours because I've built systems now where I'm like, well, I know the type of work I'm doing.
So when I look at an NDA, there's a certain sets of characteristics that need to be included for me to do my work.
Things like intellectual property, confidentiality, usage of materials during event and after.
And do we want to do co-marketing of events?
All these stuff that I've just learned up and built up this sort of infrastructure, if you will, around me, because I'm not a lawyer, but because I've done the thing so many times, I've built up this really great best practice.
Now, I sent one to a global financial institution today, like a huge one, when they sent me their NDA.
And they replied back going, this is great.
We're going to take some of this and fold it into our normal NDA that we send out to people.
And all I had done is use repetition, use tools to sort of go back and forth with building an NDA that's going to work for me.
And then, you know, it's fair on the person I'm working with and improve it.
Similarly with statements at work.
Now I just have a 30 minute discovery call with a client.
or we talk through the problems they have, the challenges they're facing.
I take transcripts from that call that they're aware of.
And so I capture everything.
I transcribe it to a markdown language.
And then I synthesize it in one of these tools, whether it's a ChatGVT, Gemini, Catalog, whatever you want.
Again, I have built a system that allows me to generate statements of work that are really robust, vision, objectives, the strategy.
details of how the program's going to work, all the way through it.
And from a 30-minute conversation, I can be generating a 15, 16, 18-page SW, like a statement of work, and sending it back to the people I'm working with.
And they're blown away.
They're literally like going, well, this is exactly what we talked about.
This is brilliant.
I'm ready to sign this document.
And again, that would have taken me days before of sitting there sort of like trying to grind out.
What were the four points, you know, that Daniel said and did he say it this way or that way?
So there's aspects about how, so my decision-making process in all of those examples is it's a review.
I never have to necessarily do a massive creation.
I can create output really fast, but the judgment.
is mine at the end of the day right i've created frameworks templates tools for how i want the machine to respond to the inputs i give it but i still have to double check the work and go okay actually that that's not right no they they didn't say they wanted to do it in august they said they wanted to do it in april i'll just fix that right and i get very small things but the judgment still lies with me But I've all this support infrastructure around me, whether it's a meeting co-pilot or an LLM tool.
But my decision systems or judgment systems are all codified now because I've made routines or templates for best practices that I've always had.
And the machine, again, has helped me improve those.
Like my first ever Esto statement of work I wrote, I thought it was a good one.
And then I asked the machine, how could I make this better?
And got into this sort of using it as a thinking partner.
And I think these are some of the most fascinating use cases for me.
It's not that the machine is making these for me, but it's become the ultimate sort of coworker where it can execute really fast.
I can coach it.
We can debate things.
It can improve how I think, and it can capture and synthesize my work.
And that has sort of just, you know, radically changed the way, certainly where I spend my time.
on a day-to-day basis.
I don't want to go into the new market of gambling podcasts and talk about the chances that people are reading the full 15 or 16-page document that's produced in this new age of lack of attention span that we're going through.
But I really like how you are defining what an artificial organization is versus...
versus that concept of like AI first or AI native, because I've never really understood what those meant.
And they felt more like marketing terms to me.
They are.
Whereas an artificial organization is purpose driven and you can't have judgment without purpose.
So I really thank you for that break.
It makes those terms even less relevant for me now.
So I'm going to move forward with this thought process around artificial organization.
You've talked about...
You start talking about the ways that your working process has changed, right?
What is something that used to be really common in the product practice or like a necessary thing that needed to be done when executing a product practice that or was like a strength of an organization that now we think is like actively holding product managers or product management organizations back?
So I don't tend to think of them like eradicating.
good disciplines, I tend to think of it almost like a speedometer or ratio, right?
Like if I ask many product leaders, what percentage of your time do you spend on complex, hard problem solving strategic work that you would like to be doing versus what percentages of your time is on this sort of administrative, like necessary, but tedious work like updating jira tickets or whatever that might be right like when i ask people that it normally goes in this sort of 80 20 mode where they feel like 20 of their time they're actually doing the creative problem solving and 80 of their time is stuck in the admin the necessary admin of updating documents or capturing tickets or moving statuses you know and look that work is necessary because it's discipline but it takes time What these tools can offer is that they can start to take away some of that administrative work and create more capacity actually for the problem solving work.
And I think most people are happiest when they're actually working on tough problems.
And that's what people enjoy, figuring it out.
But we get so little mental capacity for that because we're so busy coordinating work.
trying to drive decisions, going to meetings, preparing for meetings, following up on actions, tracking statuses.
And those are really necessary administrative tasks.
But there's ways now that they could be done more efficiently or they could have less of a burden on your time.
And for me, that's how I think of it more.
It's not like we suddenly need to get rid of doing customer interviews or synthesizing, you know, customer complaint calls to try and find maybe new product ideas.
No, you still need to do that.
But there's the way you can actually focus and build a system around you.
So my little example was, you know, every call I'm on is I have a transcription on it.
It captures every conversation and treats it like a data asset.
So every time I'm sitting on calls, say with customers, I take those calls and I put them.
basically into a database that's constantly churning, saying, what are the problems that keep coming up on these calls?
What are the words people are using to describe why they're struggling with AI adoption or they're struggling to understand how to use these tools to make their businesses more effective?
And then literally, like at the end of my week, I sort of get a synthesis back from all of these calls that are customer discovery calls, if you want to call them that.
That says here's like the top five problems you heard this week.
Here's some quotes that exemplify that problem.
And to me that like it's a really healthy reminder of, okay, that's what people might be struggling with.
So, you know, I'll write a blog about that maybe at the end of the week as a way of sort of driving that feedback loop, if it makes sense.
And I don't have to spend all the time.
you know, going through my notes and pulling things together or relying on gut instinct of what was that conversation I had with the guys?
They said something that was really, really encapsulated the problem they were experiencing.
Well, I've so many mechanisms where I could just jump to my meeting co-pilot and go to the meeting and look for the exact words.
I have a machine that's trying to synthesize those types of sentiments.
And then again, I have a reminder at the end of the week to tell me like, these are the things you heard from calls that you did this week.
So this whole infrastructure is around me and it's making me better.
You know, and you know, the famous saying, you don't rise to your ambitions, you fall to the level of your systems.
And that's one of the things where I feel like I've been able to build a lot of these systems around me to support.
how I do my best work.
And it's not going to be the same for everyone, but it's how I do my best work.
And that I think is just giving me leverage to move that needle from 80-20 to maybe 50-50 to maybe 60-40.
You know, but those are massive improvements when you think if I only had 20% of my time on creative problem solving and now I'm at 40%, that's like a, that's a 100% improvement.
And the kind of work, like I genuinely feel like I'm doing the right work with my time, where before I would often be frustrated because I felt like I was being held back from doing the work I wanted to do because I'd had this like necessary disciplines, like follow-up work to do, if that makes sense, or admin tasks.
And that's been a fascinating experience.
Yeah, I think I have had a similar experience.
And yeah, I found like in my AI journey, the transcript piece too has been like a game changer, right?
Like I had the experience you have, you have that awesome meeting, right?
And there's a lot of good stuff comes out of it.
You kind of jot down some notes and then 14 other things happen.
And then you go back to synthesize and you're like, I don't remember what this means.
Um, guess we're gonna have to do all over and now right now you have the transcript and now you can kind of cross those conversations.
So yeah, thank you for bringing that up.
Um, because I think that is a very tangible thing that you have product folks that start doing today.
Like, Hey, we have the recording.
You can analyze this and actually have the facts.
One thing I, I'm curious your thoughts on, yeah, especially as product people, or we get busier, right.
You have that, that looming 80% of kind of the stuff that you have to do.
When we look at.
decision making and using ai as a co-worker how do we avoid falling into a trap where ai starts making yeah starts making decisions or ai offers that suggestion like yeah that seems good enough all right that task's done off the list right and we're kind of giving up that yeah the decision right yeah and and um it happens very quickly you know and i think that's probably one of the traps i think you everybody has to be aware of is um Like judgment is like a muscle.
If you stop using it, it starts to erode, right?
It goes into atrophy.
And it's very easy as you're describing there, right?
Like when something looks good enough and you're busy, I'll just go with that.
You know, that's good enough, right?
And the way I, again, I sort of think about it again, it's like every time I don't actively participate in...
a judgment call.
I'm giving up my agency.
I'm deferring my responsibility to a certain extent, right?
Like that's what we're paid for, especially as product leaders.
You're paid to make decisions.
Like that's the job.
That's really what it comes down to.
So this moment that you start to give up on making the decisions is it's a road to perdition, right?
And again, I get it.
Like these machines are...
intelligently programmed to make you like their answers.
You know, like there's psychologists like working on the other side of these products, behavioral, you know, economics, everything in place where it's designed for you to keep interacting with the system.
And it's going to tell you the things that you want to hear.
And this is, again, one of the most important.
components I've found, and I talk a lot about it in the book, when you're using AI as a thinking partner, you almost have to use it as a disconfirming, brutal co-worker and demand that it tells you the ugly truth, you know, and continuously challenge it to keep telling you that, to ask for blind spots, to ask disconfirming questions.
Because if you say, I want to build a ideal customer profile for, you know, a retail product that I'm, I'm creating for book, bookstores, you know, like it's going to create something really nice.
And you're like, yep, that's, that's 90% there.
Okay.
That's good enough.
Now onto my next of my a hundred tasks, you know, and it's, it's alluring, right?
Like that's what's, that's, what's fascinating.
So many people about these products, like we've never had products to give their first response is so sort of enticing to you.
And yet the first response is the worst, most limited probabilistic response that it can come up with until you start getting into a dialogue, right?
And pressure testing both your thinking and the output.
So again, it's just, it's fascinating how amazing these technologies are, but you just always have to remind yourself, like one of the things I have here, say when I'm like, talking to people or telling stories or posting anything, I have a post-it note on my desk and it says, keep telling authentic stories.
Because again, in a world where everybody is generating probabilistic output, the one thing people can only do is create their own viewpoint, their own words.
And that's actually quite powerful in a world where everybody else is playing the probabilities of of some output to might sound kind of unique, but it's not, it's the lowest common denominator, you know?
So you have to remember, and this is so important to have human and machine in the title for me, because if we lose one or the other and we're not getting the best of both.
Yeah.
It's, it's interesting.
I'm saying, uh, at ITX, you know, our core values is integrity and it feels like, you know, maybe it's, maybe I'm wording it.
kind of strongly right but like when you are abdicating decision to the eai right you're kind of out of integrity in owning your responsibility absolutely i'm also reminded i saw a silly meme on the internet right it was like you know uh chat you could be like oh yeah you know i'm sorry yeah 125 of your max heart great for this to work out is too much right now it's somebody lying dead in the track it should have been 85 right like that that that same idea of like hey that first output maybe just poke at that a smidge right yeah absolutely you know and like these are again like it it is confidently wrong a lot of the time as you're describing in your meme right and and and then it's easy for us when we're under pressure to produce When a tool is producing a polished output, it's very easy to say, oh yeah, that's good enough and move on.
Like you've still got to do the work, you know, but I just am really conscious because this is one of the other, again, sort of second order effects that are starting to happen.
And I think you sort of alluded a little bit to this, you know, a couple of minutes ago, Sean, is that the cost of production has gone to near zero.
So again, the cost of my SOW production or stay-wind-to-work production has gone to like near zero for me.
The cost is time.
I'm waiting for it to be done and a quick sanity check of a 20-page document.
And then I push it to the person who has to process all that information manually, right?
So you could imagine in a company, and this is one thing we've seen a lot in the studio, like executives are, overwhelmed with people sending them 20 page documents and asking for feedback on it and you're sort of literally going well it costs you like three minutes to make that and now you've put it into my an executive's queue which means they in order like when they give feedback they have to really look at things so that's a 20 minute piece of work you've created for them that cost you three minutes to create so you're The bottleneck is just like being pushed onto the next person.
Again, which to your point around integrity, Daniel, like that's not how I want my colleagues to treat me and my time, right?
I want people showing up to say, look, I've done the document.
I looked at four different options.
I pressure tested three.
I think we should do C.
Here's why.
And I need your input on this very specific piece because the decision I want to make is pick option three and go.
Like that's what these interactions should be like because you're doing the work of pressure testing your thinking and you're asking your colleagues for expertise that they might have in a, in very specific niches or it may be their job that they've assigned to you to look at the challenge or the initiative and you make a decision on their behalf, right?
They didn't ask you to generate a 20 page document and ask them what, what they should do.
that's just putting the work back onto them.
Like what's the value there?
So I think this is sort of, again, some of the emergent behavior that we're starting to see with these tools when people, again, like they're using them, but they haven't figured out like their responsibility for using them.
So I've experienced that.
That is like what I call like you validate, you do the human validation for me.
Like I pressed the button and gave it a template and it created this thing.
Tell me what you think.
Tell me what you think is actually your job.
And then you're supposed to iterate on that and give me a yes-no decision to make.
That's really what I'm looking for as a decision maker.
We also have been kind of saying like treating, like talking about AI as like a co-worker.
One of the things that I'm starting to subscribe to is treating AI more like it is a machine.
Remember that it's a machine.
You don't necessarily, you should.
be nice in your daily interactions with anything that you're working with but you have to remember that it is it is just a machine it is not a co-worker it is something that is working with you it's something to pay for right and to your point earlier right it's trying to get you right there are marketing teams behind you behind it trying to get you to continue to pay for your license right so it is going to give you friendly responses but treating it more like a machine that is your thought partner versus like this Treat giving it like anthropomorphizing it.
I can't pronounce that word directly.
Giving it a name, right?
Calling it a coworker.
It's not necessarily what it is.
It's a machine, right?
So I think we could go, we could go pretty far down that route.
One of the things that I'm, that I've been experiencing working with multiple clients is something where I, we essentially have to do some like executive deprogramming, right?
So they've seen what the tools can do, right?
They understand the decisions that it can help with.
But what they see is that what we were talking about a little earlier was like, hey, instead of like a 20 page document, I produced a working prototype, right?
This front end looks great.
I just got Claude access two days ago.
I built this thing with a couple of prompts.
Why is it going to take us six months to deliver a new capability into market?
How do we like kind of deprogram the executive tier that's getting?
AI access and starting to like build their own things for the first time and like joining the builder workforce.
What's the approach that we should have in terms of actually this is what we're going to need to do to make this happen?
Yeah, I think, you know, like one of the things we learned a lot in our venture studio is it's very easy to prototype.
It's quite hard to productionize software, right?
I can't tell you how many things.
I've seen or been privy to building over a couple of hours.
I'm getting super excited because, you know, it's a nice UI and, you know, people can understand the concept.
But once more than 30 or 40 people start using it, it just collapses under the strain of being non-performant, essentially, right?
Because the code is messy.
It's not optimized.
There's certainly very limited security there.
And if you've any way, you know, a large volume of people dealing with it, the machine blows up.
Right.
And, and so, you know, the, the promise is still probably an over promise from our friends at these frontier model companies with their massive marketing budgets, you know, telling us this, you know, we don't, we, all our jobs are going away and machines are going to build and run software forever.
And we don't need coders anymore.
You know, like, and we've gone through this, like the radiologists were meant to have disappeared years ago.
And yet we need, we're a shortage of radiologists and, you know, we need more of them to look at x-rays.
So it's, it's one of these things where I think the business of software or the craft of software will always persist.
You know, architecture, thinking deeply about how to create robust systems, secure systems, like that is a practice that people need to learn as an engineer.
You don't suddenly just never have to understand that again.
But the upside is that people who want to express their ideas now have a really amazing forum and capability to turn what they're thinking into a prototype.
But it's just helping them remember the word prototype is a reason that it's called that.
these things are not necessarily production ready, that you can't just drop into highly regulated, you know, enterprise level environments and expect everything to be okay.
So I think it's just, again, it's a part of the process of, like, we've had no code and low code and, you know, middle code or whatever was before that, you know, like, so, but there's still a lot, a lot of way to go.
And we've certainly learned that in our venture studio is.
Once a company gets any real traction, you immediately have to get off a lot of these prototyping products and start looking at something that is much more robust because they're not there yet.
I'm positive of that from our experience in the studio.
Right.
We don't put passengers on the prototype plane.
Certainly not 400 and tell them we're flying for 12 hours.
Exactly.
Yeah, I want to go back to something you and Sean were talking about.
Because I think as humans, right, the first way we learn how to use something is kind of the way it sticks, right?
So if you learn how to use AI and you start relying on it to make decisions or you're crafting prototypes and you're really excited about that now that you think that the prototype is the way to go, how...
Yeah.
And you wrote a whole book about I'm learning.
Yeah.
How as product people, you know, if that first way you learn the thing maybe isn't the best way, how do you start the process of breaking those habits and kind of building the new way to work?
Yeah, well, this is a, especially I would say with the latest technologies is one of the biggest reframes.
You know, most people rarely evaluate or redesign how they work, as you're sort of describing, right?
Like.
I remember when I was writing Unlearned, I thought writers sat at a desk beside a roaring fire with bottles of wine and velvet jackets and they just sit there and I turn out pages as they're typing.
So I sat by the fire and drank lots of wine and tried to type, but I didn't really get very far.
But one of the things that was sort of the reframe for me was...
You know, I had been conditioned that that's how writers write.
And I forgot actually what I was really trying to do.
My goal was not to type.
It was actually to create content.
And there's so many different ways to create content.
Typing is just one mechanism to do it.
And how I do a lot of my best work is actually talking.
It's having debates at a board with a team member or...
You know, conversations going back and forth with people.
That's how I generate my best information.
So when I was writing on Learn, I realized actually the best trait for me was talking.
And the task was content creation.
And the tool was actually, well, why don't I just hire a journalist?
And they can interview me and we'll transcribe the conversation.
And so I would bullet out points of my chapter and the journalists would jump on the call like this and they'd ask me a bunch of questions and we would talk through the chapter and transcribe it using an AI tool.
And this is like in 2016.
So I went from sort of a blank page to like 10,000 words almost instantly.
And the journalist would take the transcript and copy edit it a bit and send it to me as like a minimum viable chapter.
So I was going literally from blank page to like a first draft of a chapter in like two and a half hours because I was using these tools to accelerate the way I do my best work.
Like sitting there and typing was actually slowing me down.
My brain couldn't put into my physical fingers the speed of the things I wanted to say and do and put together.
Right.
And talking gave me speed.
AI gave me a capture tool, turning it into data.
The journalist did synthesis.
It's like playing the role of an LLM does today for a lot of people.
And then suddenly I had this output that I was like, okay, actually I forgot to put this story here and I'd move that around there.
So I was way more in an iteration mode within hours rather than waiting like a week.
of struggling and drinking all that wine before I could get a first chapter together, you know?
And again, for me, that sort of exemplified this sort of unlearning, if you will, for me, where I was like, I actually have to really start with myself and how I do my best work and then think of the tasks I need to do.
And then if I know how I do my best work and the tasks that I create value, then I find the tool that will help me.
And that's, again, one of the formative models in the book called the 3T model.
And that is what I have found has helped people, not starting with the tools.
The tools are, again, someone's locked-in opinion, to your point earlier, Daniel, of how something should be done, right?
The locked-in opinion that a book must be typed.
So, you know, Claude has its own opinion on how you should interact with the tool for certain tasks.
that someone's programmed that and it hasn't, it's not you.
It's not optimized for you.
So you have to remember that.
And again, start with yourself and then think of the tasks and then pick the tools to help you.
This has been awesome.
Thank you so much for being with us today.
I want to do a couple of takeaways, but before I do takeaways, tell us about your most recent book.
Yeah, well, first of all, thanks for having me on, guys.
And it's always fun to share.
Yeah, look, the new book, Artificial Organizations Build Better Judgment, Speed and Results with Human and Machine Intelligence.
This one is self-published.
It's available on Amazon.
The audio book actually just came out last week and I did it myself.
So if you like listening to me and go on on this fantastic podcast, you can listen to me for three hours on the audio book as well if you want.
So, yeah.
And then Barry O'Reilly dot com and artificial organizations dot com.
You can you can get access if you if you don't go on Amazon to get it.
I think I've got my next audio book lined up.
So thank you.
Thank you for that.
Also, I'm glad that you'd like I want to learn more about like the self-publishing process, what it's like recording an audio book.
But there's probably another episode somewhere in there.
Right.
Yeah.
And it's painful for both of those things.
So, yes, happy, happy to do it at any point.
Feels like we should go back to like the like the roaring fire and the wine instead of.
Yeah, that's way more fun.
Yeah.
That feels more productive.
Right.
All right.
So here's some of my top takeaways.
Right.
Early on, you led with machines are bad at decisions and humans are good at decisions.
And I think that that's if we step back from everything that we're talking about, like that is a main theme.
Right.
Get to the point where you are the one making the decision.
Right.
A lot of the some of the examples that we talked about today really involved.
recording meetings, like record all of your meetings, right?
That should be a general practice right now.
I think that I assume that every meeting that I'm on is recorded, which often involves me forgetting to hit the record button.
But I always assume that everything's recorded and that everybody can see my screen at one time.
We actually had a guest almost two years ago now named Dan Chu Parkoff.
We were at the industry conference and he said that this was going to be one of the first big like corporate applications of AI that was going to return a ton of value.
It's recording meetings, getting the decisions that were made and not having to remake decisions over and over again.
We're learning a little bit more about just like the context we can get from meeting recordings, but like record all of your meetings, right?
You said that one of the best parts about, you know, being in product is still like the time spent solving complex problems, not necessarily.
All of like the tactical work that we have to do, but finding a way to maximize your time where you're like solving complex problems is really a goal of what we should be doing with AI.
Judgment is a muscle.
I really liked that explanation.
That was great.
And then where I was, where I asked the question about kind of like deprogramming, like executives who have like recently gotten like AI access and think that like things can be built, you know, in a few days.
One of the things I think the way that you answered that question was like, hey, in the pursuit of that reality, don't discount the opportunity to express your idea through prototypes.
Right.
So don't it's not necessarily something you want to shut down, but you have to have the right conversations afterwards.
But don't discount that opportunity that people have been given that they've never had before.
Right.
especially with the roles, you know, with the different roles that we work with on a daily basis.
So I really liked that answer because that kind of flipped the way that I was thinking about it on its head.
So this has been excellent.
Barry, thank you so much for being with us today.
Yeah.
And I love enjoy the other side of the world.
Yeah, no.
And I love this sort of little takeaway recap at the end.
It's really, it's really nice to hear what you're taking away from the conversation.
So thanks very much for having me on.
And I really enjoyed it.