Mental Models with AI: Think Better, Decide Faster

Meg Smith (00:12)
Hello and welcome to the AI Advantage. I'm Meg Smith here with Mark Smith, my co-host and life partner and one great love and biggest champion. I'm feeling a little bit mushy today because I got some really cool news on the weekend. was actually, it was Saturday night and I flipped open my emails, my personal emails, and I don't usually, I turned my notifications on my emails off a couple of months ago for peace of mind and I love it, highly recommend it.
 
If you can. Anyway, so I popped it open and I saw the subject line and it blew me away. It said that I'd been accepted into or invited to join the Microsoft Most Valued Professional Program. ⁓ And that is a huge thanks to you, Mark. So I've put you on the spot here. I didn't tell you I was going to do that, but I wanted to thank you because there's no way that I would have got there without you. So, and also you listening. 

Mark Smith (00:53)
Woohoo! Awesome. He did all the work.

Meg Smith (01:08)
Thank you. It was lot of work and it is a long time. there's, you know, I've been adjacent to the MVP program for a long while because you've been part of it for 14 years now. And I've talked to a lot of people about their involvement and the work that it takes. And, you know, I've talked to people who have chosen to not pursue renewal because of the time that it takes. And I think that's super valid as well.
 
I'm really excited to get involved and see what it looks like from the inside having observed from the outside for so long. But today.

Mark Smith (01:39)
I think it is awesome that you've achieved this. It's crazy because I looked at my MVP profile on the weekend and I've been 15 years in the program and I was just so buzzed when Meg got it. It took me back to my first time receiving it and the excitement that I had. I think the program is still amazing.

So proud of you achieving it. So, so good. So well done.

Meg Smith (02:08)
Thank you. And today we are, I'm really excited about today's topic. It's all about how can we use AI to help us solve problems. Problem solving, complex problem solving. This is a big one. And we are going to talk a little bit about some of the ways that AI has helped us. This is a fun exercise to do at home if you feel like it. One of the things.
 
⁓ we each prompted our various AI tools with this morning was looking back at our conversations. What problems have you helped me solve? And I was a little bit surprised by what came up. but before we do that, there is a wonderful ex Googler called Mo, I'm going to butcher his last name, Mo Gordart, I think is how you say it. he wrote a book that Mike and I both read, ⁓ at a time in our lives when we were living through a lot of grief. He wrote a book called Solve for Happy.

⁓ and I highly recommend you going and reading that book. If you are just at a time when you're looking for a little bit of direction or a little bit of grounding. I found it really helpful personally. So I follow his, his, ⁓ his work with interest. He was part of, what became Alphabet or spun off as the moonshot factory in Google. So these were the, this was the part of Google that was looking to solve big problems.

that would have global impact. Things like here in New Zealand, they launched Lunar Balloon, which was the balloons that went up into space and were able to broadcast signal and internet sort of to places that wouldn't otherwise be able to be connected to the internet. Obviously now, probably 15 years later, we've got things like Starlink that do that at scale and have commercialized that. But...

I follow his commentary on AI with some interest and he said something today. I just wanted to share it as a bit of a grounding for our conversation. So he says, disruption that is about to happen is so significant that when anyone tells you that the existential threat of the future comes from AI, they don't. The threats come from bad actors, bad humans who use AI to their benefit at the expense of all of us. We can all win.

It's truly no longer a technological challenge to create an amazing world for everyone. It's an ethical challenge, a moral challenge, and that's something you can influence by teaching AI to stand for us and by speaking up to those in power.

Mark Smith (04:28)
I love that as in next week we're going to talk on safety and everything around that kind of area of ethical decision making and safety with AI. you know I was at MVP Summit last year in Redmond with Microsoft and I remember

There was, I asked the comment, know, how do we, you know, is AI going to take our jobs? And I asked it in a internal session and the response was brilliant. The response was AI is not gonna take our jobs, but greedy capitalist will.

Right. Very, very insightful there. And just like Mo was saying here, and what Meg didn't tell you is that she's had a personal back and forth with Mo when she authored her first book, following us reading that book when we were traveling through Scotland, I think it was at the time, must have been 2017. But he's come out with a book since Mo and it's called Scary Smart. And it's of course, it's about AI and

Meg Smith (05:18)
Yeah, it was.

Mark Smith (05:30)
In more recent times I felt Moe has taken a dystopian view, a potential dystopian view of the world. And so I'm not necessarily fully agreeing with what Moe says, but it's very interesting what he says here. And I think that is gonna be one of our biggest challenges as we move forward is that AI being used against us, not in a, it's very obvious or blatant.

but I think it'll be very subtle, it'll be manipulative, and that of course creates risk.

Meg Smith (06:03)
Just before we jump into today's, I just wanted to say one last thing on what you just said there. It's a really good example of this critical thinking assessment that we need to train ourselves to do because we are living in this world that is trying to polarize us all the time. These algorithms are very, you know, evolved and clever, but ultimately they're trying to get you to be A or B, have a really strong reaction.

Mark Smith (06:05)
Yeah. Yeah.

Meg Smith (06:26)
⁓ for, for example, in this scenario, for Mo, so you read his content and you're like, yes, I love this. This is everything. So you engage and you comment and you share it. And maybe you talk about it on a podcast or B I hate this. This is terrible. I'm going to throw everything out about this person and, that we are not even realizing or not even cognizant a lot of the time.

⁓ We just feel the reaction and then make a choice and kind of move on with those fixed things that become part of our mindset and how we see the world. It's something I've really noticed in myself the more that we've been doing this work. And so I'm trying to build in, you know, without overloading myself in my mind, but trying to build in the checks and go, okay, can I think that something that someone has said is good without agreeing with everything that they say?

I kind of think you can.

Mark Smith (07:15)
Totally agreed. You know, when you sit down to a good meal, you don't have to eat everything. You can eat the bits that are tasty and great and are going to be beneficial to you and you can spit out the bones. You don't have to have it all. So I like that. One of the things that, you know, when Meg said, hey, why don't we do this prompt and see how AI has helped us with problem solving, complex problem solving. I ran it across

three main AI tools that I use and each of them had different lenses, you know, on things that I'd done. And so I thought we'd split today into first of all, some things that I've used for business, and then things that I've used in my personal life and problems and challenges that I've used ⁓ AI to help me with. And what I'm really keen to though is to hear how you were using it.

And so as you listen to this Perhaps you've come up with a unique way to solve complex problems and would love to hear What the problem was how you solved it or how you used AI to assist you in solving that particular problem, but to kick off The first one I have was I use AI a lot when I'm wanting to develop mental models

And I like the concept of mental models because it's knowledge in that it is here's the answer to something, it's here's how you should think about things. And therefore, based on the situation you're dealing with, you will have different mental, yes, sorry, you'll have different outputs from running those mental models. And so I've been using it a lot around mental models for AI adoption. How do I think

about adopting AI into my life. What kind of structure should I put in place? And so when I typically will kick off a sequence like this with AI, I'll say, hey, listen, I want to get clarity on a mental model around whatever the topic is. And I will ask the AI to ask me questions until it's fully understood what's going on in my mind around the nuance of this topic.

And only then do I ask, because what I'm challenging it to do is ask me challenging questions so it can get insight into it rather than just a one-shot prompt, so to speak. And out of that, I find I get really deep and it helps me unpack my own thinking and then also brings to light inconsistencies in my thinking.

And even as I say that, I'll give you an example of an inconsistency in my thinking. I like to be open-minded, I like to be broad-minded, and even just as I said, as it came to mind. I was brought up deeply religious, so for 35 years of my life, I was massively trapped in religion and Pentecostal Christianity, all that type of stuff. My parents were, and that took me on a life journey.

Until I really confronted that and really unpacked it. And I suppose I did most of that unpacking actually while I was living in the United Kingdom and really put to bed this kind of concept in my life. The concept of religion as a construct. And so anyhow, I saying, hey, I'm open minded, blah, blah, blah. This is a conversation I was having with AI. And it came back and said, you know, I opened up a different religious belief. And I was like, fuck that, no way. I don't want anything about religion and you know, that I'm open.

challenged me and said, well, why do you say that? You said you're open-minded yet. And I was like, ⁓ very good, very good. Bought out a, you know, a blind spot for me.

Meg Smith (10:45)
love that. And I think I remember when you had that interaction and you really took it very well. Like I was trying to imagine like if you had heard that challenge from a person, you might've been less receptive. You would've been more likely to dig in, I think. Because when we're talking with people, you have that emotional connection, which I think is a good thing. We don't want to lose that. But in some topics, we are so emotionally connected to...

Mark Smith (10:59)
Yeah.

Meg Smith (11:11)
our beliefs or ⁓ our interactions with them in the past maybe, you know, we, we bring to it all this extra stuff. Whereas in that scenario, I think you were more objective in replying to the challenge because AI was like, didn't have the emotions in it, for example. That's funny. I, ⁓

Mark Smith (11:27)
Yeah, I like it.

Meg Smith (11:30)
I hadn't really thought of the phrase mental models ever until you started explaining it or using it. And then I think one of the examples we included in our book, which we might've talked on the podcast about before is Edward de Bono's six thinking hats. And so when I think of mental models, I literally think of changing hats ⁓ and being able to tap into

Mark Smith (11:48)
Yeah.

Meg Smith (11:53)
a different way of thinking than is your natural way of thinking. My natural way of thinking is quite optimistic and it can be really helpful to have a challenge. So I think that's black hat thinking to be able to ask AI to play the role or put on the thinking hat and have that interaction back and forth can avoid me falling into I'm definitely guilty of confirmation bias. only look at

Mark Smith (12:10)
Yeah.

Meg Smith (12:17)
I only look for the facts that support my ideas or the decisions I've already made. Or I can be guilty of making connections too quickly and jumping to conclusions. That can be a strength, but it can also be a weakness because I just jump in and go, yeah, yeah, I've got all the thinking done. And it's really helpful to have another perspective so that you are really doing that deep thinking time on decisions that are you know, a really important when I did that prompt with my Copilot, my work AI, the thing that came out was strategic thinking. That's the thing that I'm often asking Copilot to help me workshop or brainstorm is in a particular scenario, what's the strategy? And sometimes that strategy or that scenario is what I kind of think of as like templated and that it's an area of domain that as a ne entrepreneur, new CEO, I've never done before. I've never ⁓ had to do some of the types of planning or risk management or, you know, the kind of crude scenario is I've got this much resource and how can I multiply it? You know, that kind of strategic thinking. I'm really finding that's what I'm turning to Copilot a lot for ⁓ in our work, which was cool to see reflected back in that prompt this morning.

Mark Smith (13:34)
Yeah, another one I use it for is, it's been a thought partner on whatever business challenges that I'm up against at the moment. And I will say, listen, you are my thought partner, here's the challenge I have, and I mean, a recent one was around branding and using brand consistency across business, so I was able to give it a whole bunch of data.

around a particular brand and it was able to help me look at it from these different lenses around consistency and around did my brand reflect and I'm as in not my brand but the brand I was working on, was it reflective of what we were trying to say, do that type of thing and so it, and so rather than give me answers, it asks me questions and

And the beauty is the questions unlock my thinking on subjects. So there's often times I'm not using it to solve a complex problem, but I'm using it to help work my way through a complex problem so I can get to a conclusion with that problem.

Meg Smith (14:43)
It's a starting point for me that so often has become where I want to go with AI. Because if I was going to you or another colleague, maybe when I was working in different environments, I would want to have my answer already in a pretty good place because, you know, coming back to the emotional thing, like that's how I like to think of myself as someone who has the answers.

⁓ it's about, you know, how do I show up and add value? I bring to it all of this expectation for myself. but with AI it's kind of like, well, I don't mind if it thinks I'm dumb. I'm going to ask my dumbest question. I'm going to be, you know, I'm not looking to impress. Copilot or chat GPT. I'm looking to help use it to help me get to an outcome.

So I'm with you on that one. It's like, can be something that I would have put on my to-do list to go, you know, that would keep going to the next day. I'm just going to start, I'm just going to ask this question here and start to get my head around this little bit. And then it kicks in the kind of traditional processes of going and talking to a friend or colleague or doing some more research, maybe doing some learning or a course. But that catalyst for me role in solving the problem is one I do a lot.

Mark Smith (15:57)
Yeah, another one I use it for is code reviewing, giving it a piece of code and saying, hey, what is the problem with this? Or sometimes explain it. I'll say, can you explain what's happening with this piece of code to me? In fact, just last Friday, I think it was, you and I had the scenario, we're doing an update on the 90 Day Mentoring Challenge site, and it has a management system behind that called Zenla.

for whatever reason, it was creating this massive white space on the page and we couldn't see where the error was, how it was, and we're just able to pass the URL into an AI and say, hey, what's causing this big white gap on it? It said, here it is, here's exactly where it was in the code. And we'll then just to jump into a code editor, identify it, remove it, problem solved.

That was after trying to fiddle maybe 20, 30 minutes to work out why it was doing this or what we were missing. And that in that situation, AI, it just sees everything, right? And it can solve so quickly a very binary type problem like that.

Meg Smith (17:05)
Yeah, and in the past we would have needed to, well probably spent, you know, another however long couple of hours trying to work it out. Rebuilt the page. Yeah, that was what I was going to do. And then, we would have gone to someone, you know, who could have helped us and then we would have paid them to do it. So this is an example of that. And it wasn't, it's not about either of those things really. It was that this thing was a blocker to the thing that we were trying to do. And so.

Mark Smith (17:12)
rebuild you just rebuild the page from scratch without you know yeah yeah

Meg Smith (17:34)
If you are finding yourself in situations where you're blocked on the thing that you're trying to work on, working a muscle to go, AI might not always be able to solve your problem, but what if it can just get you to the next step?

Mark Smith (17:46)
A couple other ones I just go through them fast here. I use it for fact checking. So if I get something that I'm just getting a gut check on, I'll use it. I'll give it the data and say, hey, can you reference this for me? Find it for me. I'll use it a lot for researching complex research and really takes me into areas on subjects that I'm not familiar with. And then we unpack those and learn them. Productivity and workflow.

I gave it like, here's how my day is, how can we optimize? And it'll say, hey, we're gonna create a bunch of power automate workflows to address this piece we're gonna do. So it's not just tying into one AI tool either, it's saying, hey, we'll use these range of tools. And I definitely have become super efficient around email handling and things like that. And I've always operated on a model of zero inbox, but it always took a cognitive load. Now, you know, I'm waking up.

and within three minutes of inbox, it's empty again. That type of a scenario really can, because of the automations and stuff. And then I suppose the other big one I wanted to highlight, that was handling challenging conversations. I have used it where I knew I was gonna have a conversation that was not going, that's a tough conversation, and I've used it to play kind of the devil's advocate. I'm gonna have this conversation with this person, here's the background.

what mindset might they be coming from with this? And then how could I, if I said blah, know, and I have this kind of role play of a conversation that might be coming up or in writing or that I'll be dealing with. yeah, those are a couple of my last business examples. Meg, do you have any other business examples before we get into some personal?

Meg Smith (19:23)
I was just going to say one thing that I found really interesting when I compare and contrast how we use AI. I'm more likely to go to use AI for problems when I'm already in the problem. And I'm like, I just want to move through this problem and solve it as a one-off. But because you're a systems thinker, I've noticed that more often you're likely to go, I want to solve this problem. I want a system to solve this problem. So I never have to face it again. And,

Yeah, I just thought these were really interesting, sorry.

Mark Smith (19:52)
Yeah, no, good. One of the things I've been using it for in my personal life is that, we are heavily invested in ETF funds and particularly S &P 500. And, you know, when you look at what's happening in the world, the way the world is running, is, things are changing, right? And so, as in when we look at world powers and dominations and markets and things like that, and

China is massively growing in the space and so I did a whole bunch of experimentation around identifying Chinese based ETFs that I could buy through the Hong Kong Stock Exchange and that diversified my US and European interests also with the China market and of course that allowed me to do some deep research on equivalent type ETFs but that are in China.

and that I can buy through my stock account in China. So probably we'll put that into personal investment, but that's something where I've used it to really, I suppose that was more research than problem solving, but the problem I had is that I wasn't moving forward with an initiative in the space, because I didn't know there's so much that I didn't know. And so it really helped me get clarity on that.

Meg Smith (21:09)
Yeah, I think that ⁓ I was, had a similar one example in our work as well, but it's the problem is diversification, right? When you have a lot of income coming from one customer or in the scenario we had across our different share portfolios, significantly more invested in the US market than anywhere else. And so then the problem we have to say is, Hey, we want to

we want to mitigate our risks or exposure to risk in these scenarios. What are some ways that we've never thought about? And it's that whole answering the question of we don't know what we don't know. Help me, help me see some other things. I'm definitely not going for decision-making to AI. I'm not letting it make the decision on my behalf. I'm wanting to understand for myself, how can I be a better steward of whatever resources I have and what are some ideas I've never thought of and I might

Mark Smith (21:47)
Yeah.

Meg Smith (22:05)
have never met anybody who's taken that idea.

Mark Smith (22:07)
Recently I've got into beekeeping as a hobby and so I bee hive and I suppose I got into it less than 12 months ago and in the first season I lost my entire hive of bees from wasps came in and robbed the hive in winter. Got a massive increasing wasp problem here in New Zealand. Anyhow I

an early spring which is very unusual very early in spring almost just the end of winter I was able to purchase a new nuke what's called a nucleus which is five frames with a queen bee and some honey brood and and start off my hive and so I started the hive and I was feeding it a lot of sugar water to bring on or encourage the queen to lay a heap of eggs.

And was doing that and the hive was buzzing and we're buzzing like literally, heaps more bees out collecting pollen, going crazy, that type of thing. Meg and I walked past the hive on the way to the garden one morning last week or the week before and everything was fine. By the time we got to our vegetable garden, there was this massive buzz in the air. We look around the sky, it was pretty black as all our bees had left the hive and I was like, ⁓ shit.

I assume this is a swarm not being experienced with it before. Sure enough, our hive was swarming because they felt that they were growing at such a velocity they're gonna run out of space. And I hadn't put the next box on which allows them to grow into because I was just giving it two more weeks as I was doing some mite protection varroa mite. Anyhow, the swarm happens. I've never dealt with swarms. I've never caught a swarm.

I'm allergic to bee stings so I don't want to be too crazy with them. So anyhow, I open up ChatGPT in voice mode and put the camera on and what's going on? They said, it like the start of a swarm from that information. I said, what do do? They said, well, you just gotta wait till that swarm of bees settles down. It'll settle down somewhere close by. About five minutes later, sure enough, found a tree, created this big droopy bunch of bees hanging on it.

And so once again, I'm like, what do do now? And it goes, well, you need to capture that swarm of bees. I've never done this. Meg's going, you should call somebody and everything. like, yeah, I got AI. I don't need to call anybody. I'm like, I got this. Sure enough, AI steps me through the process of what I need to do. I put a box under it and I go to shake them in and Meg was on a call and I was like, I'll wait and I'll show her. She can watch, they'll observe this happen. And.

When I came back from seeing how far away she was, the bees were all gone off the branch and they all gone into my box. And like textbook, just, it all worked. And then I was like, well, how do I reintroduce the bees back into the hive? Because I don't want to lose all these bees because we're just about to have the honey flow start and I want a maxed out hive. And so I found out that there's this newspaper method where you, how you introduce it and how you need to remove the

the virgin queens that are in the hive and let the queen come back and all that complex, never done before, all in real time, all done with the assistance of AI.

Meg Smith (25:15)
I do. I didn't remember saying like, Hey, you know, we've got people you can call who you can help. And I just loved the confidence with which you were like, actually, no, like I've done my research. I'm, this isn't just like, ⁓ you had one conversation with chat. GPT at the time, like you've been learning about bees for more than a year. Reading books and going, you know, watching videos. And so it was like, it was just really cool to see that you were able to feel so confident that I think I'm as

Like, and also there was a time, there was some urgency around the timing. You were like, I don't want to spend time doing that when I'm pretty sure I know what to do and I'm just going to go and give it a go. And then it did work. And, you know, nervously, we checked the hive every couple of days and it seems good. We seem good. Hopefully.

Mark Smith (25:49)
Yeah. Yeah.

Meg Smith (26:06)
I just wanted to thank, ⁓ Alex and our community cause he, he did the homework, not that you need homework, but I asked in the last episode, ⁓ anyone if they had had experiences with AI chatbots or with chatbots generally, and how are they finding that experience? And his response, ⁓ he shared a little bit about, you know, just the frustrations.

He didn't know if it was an AI chatbot, but the chatbot experience with his local water company was worse than an old school phone system where you could choose your menu options, right? And it kind of made me think, if the solutions that we're putting in place to solve problems for our customers are creating more problems, those aren't the right solutions.

So thanks Alex for sharing. And I know Mark, you kind of gave a question or gave a challenge to our listeners at the start of this episode that ⁓ if you've been using AI to solve problems, particularly complex problems, let us know how it's working for you and what your approaches are. We're always wanting to learn. And what your problems are. Awesome.

Mark Smith (27:05)
and what your problems were that you solved. Yeah, just to wrap up, if you have been listening to this, it has a video version available on Spotify and YouTube. Of course, if you want to engage in the little community we have around this podcast series, you can join the WhatsApp group for the link up here on screen. It's in the show notes. As always, we love to hear your feedback and your ideas and topics you'd like us to discuss.

The next episode, as I said, will be on ethical decision making and safety with AI and really the responsibilities on all of us. I've just, as I said, recently read George Orwell's 1984 and I'm really hoping that we're not heading in that direction and it's on us to really control our use and how we work with AI and to set the future and not just be

Meg Smith (27:47)
84.

Mark Smith (27:59)
ingested into somebody else's vision of it for us.

Meg Smith (28:03)
you went to quite a dark place after reading 1984. I remember we were talking about it and you know, I think it's understandable to go to dark places at the moment, right? Because sometimes it feels like everywhere we look, the news isn't good. But I also think we can make a...

sort of a focused effort to find the good news stories. And there's one that I, as you were talking about 1984, I remember we talked about, there's an amazing recipe developer in Australia. She's responsible for the site Recipetineats, Her name's Nagi and I've got a friend that works for her. And she's got this great like website that has just lots of recipes, right? But the coolest thing about what they do in their business is she has a

a team in her kitchen in Sydney in Australia that every single day they make 400 meals that are given away into community programs. And I remember finding that out and thinking, that's amazing. How cool. I wonder how she's funded. She must have some government funding to do that, or that must be something. No, she just takes a portion of her profits and decides that that's important. Yep, that's what she does. And I just love that as a challenge to go, hey, whatever your sphere of influence.

you can choose with the resources that you've got to do something that helps people, you know, and that's a choice we all have. So I absolutely love that. And I'm trying to train my brain as well to go when it feels like the world is really dark and bad, like let's look and shine our light on the things that are great and the choices that people are making to help other people.

Mark Smith (29:43)
See you on the next one.