784: Exploring AI & Productivity, with Michael Hyatt
01:27:26 Share

Hello and welcome back to MacPower Users. My name is Stephen Hackett. I'm joined as always by my friend and yours, the birthday boy, Mr. David Sparks. Thanks, Stephen. Hi, everybody. I'm a birthday boy. Yeah. Well, a few days ago. A few days ago. Okay, it still counts. I just want to jump right in, David, because we have an awesome guest this week. Yeah. Welcome to the show, Michael Hyatt. Thanks, guys. So fun to be on with you again.

Yeah, we always like having you on, Michael. Your shows always are fun. They go fast. You always bring a lot of energy and good stuff to the show. And we started corresponding a few months ago about artificial intelligence, and I was interested in your take on it. And you said, I'd love to talk about artificial intelligence from your vantage point. So here we are, and we're going to do that today. For folks who don't know you, Michael, tell us a little bit about you.

Well, I'm the founder and the chairman of Full Focus. And interestingly, especially since we're talking about technology and AI, we publish an analog planner called the Full Focus Planner.

We also do tons of coaching of all kinds for business owners and basically high achievers who are really struggling to get some semblance of work-life balance. That's our primary audience. They feel overwhelmed. They're not sure it can ever change. And so we have a process that we take them through to help them get what we call the double win, which is winning at work and succeeding at life. We don't want anybody to have to sacrifice their health or the most important relationships

and kind of the altar of their own ambition and get to the end of their life and have serious regrets. Yeah. Isn't that what it's about really? I mean, I'm, I have a toe into productivity with some of the stuff I've been doing and I feel like the big motivation for me is, is this idea of momentum, Mori, right? The weird, none of us are getting out of this alive, so you better make it count. Yep.

the, uh, I had personally some friends passed away in the last few years and, and, uh, seen one who lived life right. And one who lived life wrong and seeing what that does to you at the end, boy, look out, but, but people did not log in for that today. So we will real heavy four minutes in guys. Sorry guys. Sorry guys. I, I, I got myself turned. Michael is one of my favorite people. So every time I talk to him, he gets it all out of me. Um,

But what have you been up to since the last time you've been on the show, Michael? Well, my priorities have shifted a little bit since I turned over the company to my eldest daughter, Megan, who's doing a bang-up job as the CEO, which allows me to really do kind of the things that I love.

which are coaching. So I have several one-on-one clients that I coach. And then I do group coaching. We've got a group of about 500 plus certified pros that have been trained in our IP and our methodology. I lead a couple of masterminds. So that's fun. I'm still doing the podcast with Megan, but we rebranded it as the Double Wind Show and we changed the format. And part of that is I just get bored.

You know, you're amazing. You guys have been just like doing the same thing forever. And it's always great. The shows are outstanding, but I kind of get bored. So we have to fuss with the format of the branding, but now we're doing interviews and oh my gosh, you know, one of my top five strengths is learner. And I feel like every time we have a guest on, you know, I just get to pull out the yellow pad and take notes and really go to school. And so that kind of feeds that insatiable desire to keep learning. Um,

What else? I think one of the big things with AI is we've made a renewed commitment to analog, which kind of sounds crazy given the fact that we're talking about AI today and we're talking about technology. But I think they go really hand in hand. And what we're concerned about is that we don't lose the human touch, physical presence,

Face-to-face interaction, tangible experiences. You know, we want to champion those in addition to AI efficiency and what AI can do for you. Yeah, I've been on my own journey with that, Michael, because I knew in my bones that

that I wanted an analog component to what I do on the daily and also in kind of my review process. I wanted to be more mindful about that stuff, but I could never quite get the tools right. And me being a nerd, that's a big deal. But just in the last year, I have found my, my happy place where I do most of my planning and reflection on paper and

but I also use all these tools and I can tell you it does matter. And I think one of the things as nerds need to get over is the efficiency idea. Like I get it. You can be more efficient with a digital tool, but maybe there should be a list of things in life that you don't want to be efficient at. And, you know, figuring out what's going on between your ears is one of them. You know, it's so true. And I think sometimes when you've got the reality of,

you know, virtual reality. And like I was watching the Superbowl and

the ads for the, you know, the Ray-Ban meta glasses. And I thought, oh, I'd love a pair of those, but do I really want a pair of those? I mean, I just, I want some spaces in my life that are unaffected by technology. And I think that sometimes we don't do a good job as humans of thinking of the implications of technology, particularly like what the phone has done to us and what social

social media has done to us as a culture and as individuals. And it's changing our brains. And this concerns me. You know, I just want to be more thoughtful and less of an enthusiast because I tend, you know, I'm exploring tons of apps. I buy everything new that Apple comes out with practically. And, you know, there's a place for that, I suppose. But I just want to be more thoughtful.

Well, I think there's like a bigger story here. I mean, to me, this is kind of my motivation as Max Barkey is that like when I was a kid, you know, there's a, you and I are of a certain age and the computers were kind of on the rise. The personal computer was a new thing.

And the promise of it was, wow, this is going to make life so much easier. You know, we're going to use these, these devices to do our work faster and have more time to do things like fly fishing. I know that you love that. Um,

But what happened was somewhere along the line, the story got discombobulated and we turned into the product and the people making the devices realized that they can make a bunch of money advertising to us. And like the whole thing got turned on its head. But I do feel like that original dream can still exist. We can still use this technology to be more productive and pursue things that are important to us.

But we have to be the captain of the ship. We can't rely on Silicon Valley to do that for us. Absolutely, because our interests are not aligned. And I really believe in what Apple's doing. And for the most part, they don't use advertising in their stuff. But still, they're pushing a model that tends to amplify usage, even though they have screen time and all that kind of stuff.

but you can get sucked down that rabbit hole and not have a life. And I was just with one of my granddaughters who's now an adult, and she brought her boyfriend home for Christmas. And he wouldn't look me in the eye.

And he could barely have a conversation. His head was constantly in his device. And maybe that makes me sound like old-fashioned or whatever. But I even read a report last week where it's changing our physiology, like this whole, I don't know what they call it. I think there's a name for it. But where your head's kind of stooped over because you're looking at a screen all the time. And it's got some implications because it's happening incrementally so that nobody really notices until, like, you know, the frog dies.

Boiled in the kettle. It's too late. We wake up and we're cooked. Yeah. It's, you know, I like you. I feel like, um,

Every generation as they get older says that the young people don't know what they're doing. I think that goes back. It's a universal rule. And so I'm aware of that as I get older. But at the same time, like you, I've witnessed people that have serious addictions to the device in their pocket.

And, and I struggle with it. Like we, it comes up on this show. We talk about it. And also on the focus podcast, like, especially with children because their brains aren't developed there. They're at such a risk. My, my daughter, who's a high school teacher, the very first day she taught high school after getting all her degrees and getting her first job,

I said, so how was your first day? She's like, oh, it was good, Dad, but these phones are a real problem. Wow. And it's like, after I told her, they grew up with me preaching this

But, you know, you know, these kids, they, uh, they're so addicted, they can't really help themselves. And, and then that turns into an adult who can't really help themselves. It's, it's a problem. I had a moment just this, this past weekend, uh, it was my wife's birthday and we all went out to a basketball game together and we're going through security, you know, to get into the arena and you got to take your phone and, you know, keys and stuff out of your pocket. And, uh, we have a 10 year old, our youngest is 10 and the, uh,

the woman who was kind of checking people and was like, oh, hey, like, do you have like, take your phone out of your pocket, sweetie, like talking to my 10 year old. I was like, I just kind of bored. I was like, he didn't have a phone. Like, like that's crazy. But then I was like, oh no, like that's maybe not even all that. Maybe we're in the minority, even, even at 10 or it's different with my two high schoolers, right. Who are out and about and out with friends and stuff. But like my fourth grader, like, I don't feel like it's crossed that, that threshold yet, but it really is.

it really kind of stopped me in my tracks. Like, like, Oh gosh, like this, this is just a question that was asked for a reason. Right. And, and that's one thing I love about y'all's analog tools and analog tools in general is that as more and more of our world becomes, or is digital, especially at work,

I find it really useful to have that context, that mode switch. Like, I am going to write something down. I'm going to read a book and take notes in the margins. Like, it is important for us to break away from those things. And so I think now more than ever is a great time to lean into the analog because I think for a lot of us, it gives us an opportunity to think and act differently than we would in front of a keyboard. Yes.

Well, I couldn't agree more. You know, I was highly influenced recently by Cal Newport's book, Digital Minimalism. And then there's a great blog called the Digital Minimalist.com. And I don't even know who the guy is. His ex-handle is HeyEslo, E-A-S-L-O. And one of the things I was inspired from him to do was to put everything on my phone in grayscale. And you can do that through the accessibility settings.

and the color filters. And so I've got just an Apple shortcut that I've created that I can turn it back on to color if I want to look at photos, for example. But otherwise, it's grayscale, which makes it less tempting. So I was writing a newsletter article about this and started doing some research about the use of color in terms of encouraging engagement on phones. And there's actually been some studies done on this. And it's dramatic. It's like 2x if you have the color on how tempted you are to spend more time on the phone.

And then I installed an app that I think I first installed this 10 years ago, but it's come a long way. Freedom, which is an app and web blocker. And I've just created these automatic schedules. I like to say I have the discipline.

But you're kind of up against the algorithms and all the technology and all the research that people who build these devices and these apps are doing to get you to stay on them 24-7. So the thing I like about Freedom is it allows me, like, I have a schedule that goes off at 5 in the morning.

And blocks me from all social media and everything about except the things I need to get ready for the day. And then at five o'clock, because I would be tempted to just keep working through the evening, you know, it shuts all that stuff down. So I don't even bring my laptop, for example, into the house. I'm in a building out back, but I don't even bring it in because there's nothing I can really access. I have to do crazy stuff like talk to my wife.

Or engage with my kids. Or read a book together. Or do a puzzle. Yeah, you know, I was going to give you my holier-than-thou speech, how I really am not a person that gets very tempted by social media and stuff. But the second half of that part, where you intentionally brick your iPad every night, I could probably benefit from something like that. It is kind of surprising. And Freedom even has this setting where you can lock it down.

And it'll give you the opportunity if you lock it down. And you got to do this when you're really, you know, got a good vision for what you want and you're sober about it. But you can actually end a session once a week, but that's it. Once the session initiates, you're stuck. That's fascinating.

Yeah. I mean, all of those are tools that I think can be helpful to people, but also I think we all need to work on the basic discipline of this stuff and, and also just get an appreciation for what's actually going on. You know, all these things that are here stealing our attention are not trying to make our lives better. They're just trying to make, you know, somebody that already has too much money, get more money. I think it's not enough to just remove stuff from your life.

You have to replace it with something meaningful. Yeah. So, and I think, you know, for me, kind of scaling back on the digital side has given me the opportunity to dive deeper and be more present in my most important relationships, to be outside and enjoy nature. I mean, that's something that, you know, the youngest generations now don't get a lot of.

just fresh air. And, you know, when I was a kid, my parents would turn me loose in the morning and they'd say, look, don't come back till we call you for dinner. And so we had to figure out stuff to do. We had to invent play. We had to make things, you know, and all that was great. And we've got a whole generation that's never known that experience. Their whole experience, most of their experience with the world is mediated through a device they can hold in their hand.

It's tough. I think it's tough for parents. There's a book out now, if you're interested in this topic, called The Anxious Generation, which is making a lot of waves. And it's got all the science about how this is bad for kids. Like I said, I think kind of the...

The ironic part is nobody's talking about what it does to adults. It's just as bad. But I think we're starting to see like there's legislative changes. Like California, I believe, just recently passed a law that's going to ban phones in schools, I believe, in a year or two. If I had a time machine, I would go back and build a company that can make

phone lockers for schools because i feel like there's going to be a lot of business for those in the next few years which is all good so i think we're more aware of the problem and i told my kids i think it's going to be like smoking where in the future they're going to look back and say you let 10 year olds have phones you know like like that is crazy and

I think you're right. But I don't know. I don't know. That book, The Anxious Generation, is fantastic. Yeah, it's really good. Even as someone in his late 30s, I read it and I was like, oh, this book isn't just about my kids. This book is also about me. I distinctly remember the entrance of the smartphone into my life. It was in college and how everything was different before and after.

But it's interesting to me because like the Mac power users audience is, is very alert to these issues. Like when I talk to listeners, I mean, these are people that like technology, but also understand that it's limitations. But I think we're a minority. I think a lot of people like give kids phones and don't think twice about it and don't realize, you know, the hand grenade they just threw into the house. I've been watching, I have five daughters and my fourth one,

has my youngest grandson. He's two and a half, but he literally has had no screen time. And so many parents use an iPad, you know, as a way to basically babysit the kid and get them engrossed in something so they don't have to be. And that's, you know, it's a commitment. It's a big deal, but she's committed to that. I mean, he plays outside every day and I can see the impact on him. He's very social.

really good communicator and so inquisitive. And it's just been fun to be with him. He looks you right in the eye when he's talking to you. You know, he's just, I'm hoping he can be the exception that's not conditioned by all this digital technology. Well, now that we've explained all the sins of technology, let's talk about AI. This episode of the Mac Power Users is brought to you by 1Password.

Go to 1password.com slash MPU for 20% off your plan and use 1Password, the app that protects your most sensitive information.

The internet is a dangerous place. There are lots of people spending lots of time and effort trying to get your details. With 1Password, you can protect yourself with a secure password manager where you can store sensitive information and access it anywhere. Passwords, credit cards, and sensitive information of any stripe. I've got my medical information in 1Password as an example. You can keep it all in one place and easily access it on any device with the 1Password manager.

1Password doesn't just protect your information, it also helps you share it securely. You can create shared vaults where everything in that specific vault is shared with the people you share it with, or create expiring links for short-term password sharing with anyone, even people who don't use 1Password.

And I think my favorite feature is Watchtower. Knowing you can keep track of all the data breaches in the news, Watchtower does the hard work for you by flagging compromised passwords and other security issues related to your saved items. It looks at your saved items and compares it to a list of compromised sites and lets you know if there's trouble. You can then update the affected account with the new password before an attacker can do any damage.

It's like having your very own security team. It works on a bunch of different platforms, including all of the Apple hardware. You can have a cloud system, so if all your devices go up in smoke, you can still get access to your data. 1Password is used by small individuals like myself and big multinational corporations.

1Password is a big deal, but I was there at the beginning and I saw how the people that started this company were obsessive about security and that runs through the DNA of this company. If you want a secure password manager and a bunch more features, check out 1Password. Go to 1Password.com slash MPU today to get 20% off your plan. That URL one more time is 1Password.com slash MPU. Get yourself set up with a great password manager and let them know you heard about it here on the Mac Power Users.

Thanks, 1Password, for all of your support of the Mac Power users. Okay, so Michael, when I heard you wanted to talk about this topic, I was very eager to discuss it with you because my feelings, like I love a lot of the stuff that AI is doing. We've done several shows on it, and I'm constantly exploring how I can use it.

But the one place for me that AI has kind of fallen down is I always get emails from people and listeners saying, well, don't you want AI to like organize your calendar and your tasks for you? And there's, there's products on the market now where you feed it all that stuff and it tells you what you should do every day. And I find that stuff kind of repelling. Like I don't want an AI to tell me what to do. And I don't know what's your, what's your thought on that kind of as an overview going into this discussion? Yeah.

I'm 100% with you. And I think it's important to understand what AI is good at and what it's not good at. But for me, I'm kind of like you. I just rebel at the idea of somebody, even my assistant, who schedules everything for me. You know, I will oftentimes deviate from that. And I'm talking about like time blocking for tasks, just because in the moment, I feel inspired to do something else. And I magically get it all done. But yeah, I have not...

really used any calendaring AI or task AI? Well, I've tried them because people tell me that they're great. And they just, you know, this is much more of a, I don't know, is it emotional or spiritual, but at some level there's, there's a part of me that knows when I'm best at doing this thing or that thing. And the AI doesn't have sufficient knowledge of me to tell me. And I think if it did tell me, I would probably be a contrarian and not do it anyway. So yeah,

Yeah, that part of it doesn't really land for me. But then there's another piece of it. Well, how can AI help you do the work better? And I feel like there's a lot of opportunity there. I do too. And that's like even an email. You know, I use Missive and I love it because I can communicate with my team inside of it. We can talk about emails and we don't have to forward them and all that. But the other thing too, it's got AI built in. Now,

The thing I like about that is, you know, there's sometimes when somebody's written a really long response and I just want a draft of some way to reply to them that captures everything. And I want an AI that's going to be less curt and warmer than I would be normally. And this is particularly true. I use this actually in social media when somebody, you know, rags on me on social media or has kind of a nasty reply.

Left to myself, and this is why I'm not in customer service, left to myself, I would give a really terse, negative, probably semi-insulting reply. But I built a custom GPT for that very thing, and I call it my conflict resolution assistant.

And I uploaded some books like Nonviolent Communication and Crucial Conversations to it. And I'll just paste in the message. And I'll say, here's what I want to say. I want to be direct. But I also want to make sure that it's kind and empathetic. And I've had people literally write me back after that and apologize for their initial post. Hmm.

Yeah, the sort of tone shifting in a lot of these writing tools is really interesting. We just did a big review of Apple Intelligence, and that's something that I feel like Apple's AI kind of cranks that to 11. Like if you tell it you want something friendly, it's like,

bubbly and over the top and lots of exclamation points like okay like simmer down uh but it it is an interesting type of tool where it can take maybe your initial message or something generated and and shift the tone uh do you how do you think about that from like an authenticity standpoint do you feel like there's friction there i do but i think that most of the friction comes when you don't train the eye ai to writer to talk like you

And I have a custom GPT. In fact, I've given the link to you guys so you can share it with your audience. But it's called Write Like Me. And it basically, you upload some samples of your writing. It deconstructs it and gives you a voice instruction that is pretty comprehensive that you can upload to various AIs to train it on your voice. So I would say it gets about, for my voice, and I've trained it on a ton of content, but it's probably about 90% accurate.

Now, I never think of AI as something I'm going to delegate the writing responsibility to 100%. I use it more like a co-pilot or a thinking partner or a collaborator that at least gets me started so I don't have to just look at the dreaded cursor on a blank page. Sure. But I think that most people who complain that the AI doesn't sound like them just haven't taken the time to train the AI so that it does sound like them.

One of the challenges for me, and I guess it really depends on context, like responding to a nasty email is one thing, but writing toast for your

for your anniversary is a different thing, right? It's like, there are some parts of writing that I feel like I need to do it from scratch myself. I need to go through the process of collecting my thoughts. And I don't even want an ugly first draft from AI. But there are areas where I do use it for that. So I just never want it writing for me. I just want the...

you know, that was the initial thought, but I have found context where it makes sense. And what you've done as I, and I want you to explain this. It looks to me like you're building custom GPTs and chat GPT with a paid account, right? I am. Okay. So explain that for folks who don't know what that is. How do you do that? Okay. So chat GPT among other platforms has a way to build custom special purpose assistance. That's the way I think about it. And so let me give you an example. I think

This will make me clarify, but I have one called Book Summary Assistant, which I've also given the link to you guys to share with the audience. But basically, it says to you, it starts, initiates a conversation, and I write out the instructions in a custom GPT. And this is way easier than it sounds, although your audience is pretty advanced. They'll figure it out immediately.

But so I've said to it, OK, the first thing you need to do when this GPT is initiated is ask the user to upload a copy of the book they want to analyze. And I've got a whole procedure for converting a Kindle book to a PDF. Now, we have to respect authors' intellectual property rights. As an author, I'm very concerned about this. So I insist, even with my team, that if you're going to use a Kindle book, you've got to buy it.

And you just can't, I don't feed these books that I upload to chat GPT or custom GPTs. I don't make those publicly available because to me, that would be theft of the author's intellectual property. But for your own use, and if you bought it, it's incredible. So in the book summary example, I basically give it the instructions of the output that I want.

So I may, I want a book summary. I want a biography of the author. I want an outline of the book that's annotated and gives me the summary of every chapter. I want the 10 most remarkable quotes in the book. And I want it to, I say to it, based on what you know about me, what are the biggest takeaways from the book? What should I be paying attention to? Yeah. And then I have it come up with a reading plan that if I want to read the book in seven days, how many pages or

you know, whatever, I have to read a day to do that. Or alternatively, it'll tell me if you want this read, you know, like a chapter a week, how long is it going to take you to read this? There's a lot of different ways you can configure that. So that would be an example. That's another example, though. I have a medical research assistant.

Now, this one I don't share publicly, and here's why. I have uploaded it, all the functional doctors that I've read and follow, people like Mark Hyman, people like Peter Attia, Andrew Huberman, you know, and others. And so I can go up there and upload, for example, a PDF of a blood lab report that I get, and then I've given it the instructions. I say, analyze these blood labs,

Tell me what's out of range, what's concerning, what I need to give attention to on a priority basis. And if the problem can be solved with a supplement, give me a link to buy the best reviewed supplement. And I put that in a, in a, in inside the, I actually do that one inside of a chat GPT project so that every conversation that I have with this project, it retains what I've given to it before.

So it knows what my blood labs were six months ago. And I've got like 10 years worth of data that I've put into that, which is helpful. Yeah, yeah. Another one. I've got one that has all of our full focus content. So this has all 12 books that I've written. Again, I just bought the Kindle editions and converted those to PDFs. It includes every course we've ever recorded, the transcript, every live event that we've ever done, the transcript,

Every podcast we've ever done, the transcript. And when I'm creating newsletter content, there'll be times when I say, oh, there was a story that I told about going to Maui and having a snorkeling experience where I was caught in a riptide and almost died. But I can't remember where I wrote that. What book did I write that in? Feed it back to me verbatim. Yeah.

I think that sort of use is really fascinating. I know a lot of people are doing that with Notebook LM from Google where you can upload a bunch of content. And I think for people like us who make stuff and have made stuff for a long time, I got 16 years of blog posts, right? I got 12 years of podcasts, like,

And even just NPU, like I've been transcribing them and putting them in Notion just so we can like, okay, I know we have talked, I know we have touched on this, exactly what we said. Like, what did we say in the past? Or maybe I'm just looking for it as a reference point of like, hey, we talked about this topic eight months ago. If you want to go hear more, you're kind of being able to direct the audience. And I think it's a great use if you're in a creative profession, just like,

Give me insights into my own body of work is powerful because if you're anything like me, after it's written or after I've said it on a show, it's in my mind for a while. But then generally over time, those things fade, right? It's like, have I shared this story? Have we talked about this? It can be really useful. Well, it's kind of embarrassing if you're a keynote speaker.

And you tell the same story to the same audience. And, you know, some people want to hear it. You know, it's kind of like, I guess, if you go hear the Eagles, you want to hear the greatest hits. But I'm not quite at the Eagles level. So I think people just think, man, you know, he forgot that he told us that before. So it's helpful for that. And sometimes I want to reuse that same story in a different context. And that's good, too. Now, another thing I've done for stories, I actually have another custom GPT that I built and trained called Storytelling Assistant.

but I uploaded to it about 14 years worth of day one entries for my journal. And I said, okay, I need a story about this, whatever it is, to illustrate a point. Go back into my journals and find where I've had that experience or have a story about that. And the first time I did this with my journals, I said, go through all my journals and find out what stories would be useful as anecdotes in my writing or speaking.

And it like very quickly came back with 200 different entries. Do you worry about privacy though? Like putting day one into chat GPT, that gives me the willies. You know, this is an interesting thing and most people don't know this, but I wrote a book back in 2001 called invasion of privacy. And I was the champion for privacy.

And I talked about all these methods, you know, clear back then how you could protect your privacy. The book was published a month before 9-11 and sank without a ripple. Okay. Because nobody cared about privacy. My attitude towards it now, and listen, I don't want my identity stolen. I get all that. But I kind of operate with the assumption that all that's going to be made public sometime.

Even though we try to protect it the best we can, you know, I've written an article on ChatGPT or OpenAI's privacy policies. And, you know, I think, you know, there's ways to lock it down. But, of course, that's only as good as they're willing to live by their terms of service. Yeah. But in ChatGPT, usually when I'm doing a training on AI, that's the first thing we do. You know, we go into settings.

And then you can go down to data controls and then you can go, you know, basically train the model and make sure that that's set off, set to off so that it doesn't train AI's models. Now, whether it does or not, I don't know. I probably would load my personal financial stuff into it. Or if I did, I would make it sound like it was from somebody else. But there is a lot of financial analysis that you can do that's very helpful with AI.

Well, I mean, you talked about medical. I've done the same thing recently. I've started uploading data to it because I just wanted to get analysis of some medical things that me and my wife are dealing with.

And it gave me really good answers, you know, better answers, frankly, than the doctor gave me. And I'm not saying I'm going to replace my doctor, but it is nice having a little buddy that's pretty smart about these things when I ask questions. But I do struggle with the privacy element of it. There was a recent study done, and I can't cite it here, but it was basically a comparison of doctors who were AI-assisted versus doctors that weren't.

And the AI, particularly on the diagnosis side, was more accurate than the doctors. Now, on the remedy side, it's a little bit more nuanced. And so that's where a doctor coupled with AI was a better solution. But you think about it, you know, the doctor's got to hold all this stuff in their minds or go back to their office and, you know, scour their medical books or whatever. But when AI has access to all that real-time information, it can do a pretty amazing job.

Yeah. Well, I mean, you know, you even compare that to what we were talking about us as creators. I've done similar things to you where I've uploaded like on certain topics, everything I've said about it. And then when I work on new content around that topic, I ping it.

And I've been using Notebook LM. I frankly hadn't really thought about doing it with ChatGPT, but it sounds to me like that's your weapon of choice. It is. But it is fantastic to be able to kind of pull my own thoughts because, you know, there's a lot of them and I don't always have immediate access to them or sometimes the AI sees patterns that I didn't notice. That's right.

And it gives you a jumping off point. And that's where I'm not going to have it right for me, but I'm going to certainly have it help me surface ideas as a jumping off point. I'll tell you where I do use it in writing for me is that I'll write the first draft or I'll just have a brain dump of ideas. But when it comes back to me or when I feed to it a paragraph or two paragraphs or whatever, and I want to say, make this better. One of my favorite AI prompts, make this better.

And it does what a really superb copy editor at a publishing house or a journal would do. And it just helps me be more clear and more to the point than sometimes I can be, depending upon my state of mind when I'm writing. Yeah, I mean, and you worked in publishing, so that says a lot from you. Yeah, I mean, I spent 35 years of my career in the book publishing industry where I was an editor of various stripes and acquiring authors.

And dealing with all that. And so I kind of, yeah, I kind of know where you can use this stuff and where you can't. I think editing is a big, big thing. If I were an editor at a publishing house, you know, I'd be making sure I was up to speed on AI because, you know, it's like somebody said, you're not going to be replaced by AI. You're going to be replaced by somebody who knows how to use AI.

This episode of MPU is brought to you by Squarespace, the all-in-one website platform for entrepreneurs to stand out and succeed online. So whether you're just starting out or maybe you're managing a growing brand, you can stand out with a beautiful website, engage with your audience, and sell anything, products, services, and even content. Because Squarespace has everything you need all in one place, all on your terms.

With Squarespace, you can make the most of design intelligence. Not everyone has an eye for design, but Squarespace's tools make it easy to build something beautiful. Design intelligence combines two decades of industry-leading design expertise with cutting-edge AI technology to unlock your strongest creative potential.

It empowers anyone to build a beautiful, personalized website tailored to their unique needs and craft a bespoke digital identity across one's entire online presence.

Those design elements go out with things like newsletters or invoices that you can send with Squarespace. And everything is really integrated with their SEO tools. Every Squarespace website is optimized to be indexed with meta descriptions and an auto-generated site map. But if you're doing things like newsletters or sales, those analytics follow you around. So all of your Squarespace stuff...

You can see where people are coming from. You can see what's performing, what's not performing. And when you combine things like design tools, SEO tools, and analytics, you can make informed decisions about your brand. I love building on Squarespace because all of these tools are made to work together. You don't have to be an SEO expert or a design expert or an analytics expert to take advantage of what Squarespace has to offer.

So head on over to squarespace.com and sign up for a free trial. And when you're ready to launch, go to squarespace.com slash MPU to save 10% off your first purchase of a website or a domain name. Once again, that's squarespace.com slash MPU for 10% off your first purchase and to show your support for the show.

So, Michael, we were talking in the prior segment that you're building a lot of these out in chat GPT. You know, kind of this notebook LM has been very popular with folks like us lately as kind of a resource for looking at your own materials or building your own AI assistant. Why chat GPT? You know, when I first began this journey over a year ago, I decided I was going to do a deep dive into AI. And I went on this buying spree.

where I subscribed to every shiny new thing that showed up in my feed. And I tried dozens and dozens of different AI. And first of all, the experimentation was a blast, but I found myself continuing to come back to ChatGPT. And part of that is, at some point, just experimenting can be extremely disruptive to your workflow.

You know, you start building something in one AI, you get a workflow nailed, and then something comes out that has the promise of doing it better. And so you jump from what you're currently using and whatever productivity gains you'd hope to get are lost just in the experimentation and in the onboarding and leveling up to use that new tool. Yeah.

So the other thing I would say is that there are very few tools out there, and Notebook LM is one of them, ChatGPT is one, where you can upload lots of knowledge. Like the requirement in ChatGPT, you can upload up to 20 files, and you can upload up to 500 megabytes of data. And as it turns out with PDFs, that's a lot of data. And 20 files, like, for example, I took all my podcasts, and I put those into a single file.

Yeah, that's the trick. I did the same thing in notebook LM. Yeah. Like when I did the productivity update, I took all of the webinar transcripts and made it one file instead of, you know, 14. Yeah. Yeah. I think, I think that's one of the things that keeps me coming back to chat GPT is the ability to do that. I've uploaded my voice instructions to chat GPT so that it's, it's always writes like me. It always talks like me. The other thing that,

I like about ChatGPT, the paid version, and we're on the Teams version because everybody in our company uses it. But the thing I like about it is it doesn't, I never hit the limits. Like I love Claude. Claude is great at writing. I don't think it's that much better than ChatGPT at the moment, but I hear they're on the cusp of releasing a new model. But the problem is you quickly run out of tokens.

And I've been in the middle of a project or almost finished with a project, and it says, oh, you have to check back in several hours or check back tomorrow because you're out of capacity. I mean, I'm paying for this. And I much prefer the model where you hook up your own API connection, and then, like, I'd be willing to pay lots of money to finish the project, but it just cuts me off. Yeah.

And that's frustrating to me. Yeah, a lot of these companies are facing real capacity challenges. And that's one thing that for all of its concerning elements, something like DeepSeek or other newer models that can be smaller and lighter, require less hardware, require less capacity, require less electricity, all these are good things. I think Claude is kind of the front of the list right now where like,

They are just running headlong into, we don't have enough computers to make this work. And so they're passing those limitations to their users. Well, that's the other thing about ChatGPT. We know it's well-funded. Yeah. Those guys are raising all the money. I mean, seriously. And so it's tough to compete. Now, I don't use it for everything. You know, I've gotten really into multi-agent creation. And there's better tools in ChatGPT for that. In fact, you can't do that in ChatGPT.

Period. There are some plugins, Chrome extensions that allow you to do it. One of them is called Ultra, and it'll show you what's possible with the Chrome extension, which is mind-blowing. But you can just Google Ultra for ChatGPT. But for multi-agent kind of work, and this is where I think the future is, and by the future, I mean like the next month.

But for the future, you hand off specific tasks to specific agents that may be employing different models. So the one I use the most frequently is one called MindPal. And so I can do a newsletter, for example, that'll ask for human input. It'll say, okay, what's the problem you want to solve in this article? What's your solution?

And then it'll pass it off, you know, pass off that input. We'll generate some stuff. Let me choose or let me polish and then pass it back to me so I can continue to iterate with it. So instead of having one giant prompt that does the whole, tries to do the whole newsletter, which I've tried to do before and get unpredictable results, I'll break it down into the component pieces. So if I'm working on a headline for an article, then, you know, that's one prompt. Now it's in a

It's in a multi-prompt system, which is different than multi-agent, but I can do multi-prompts. And it works much better if I can keep those prompts focused on a specific task, complete it, and then move on to another prompt. And agents can do that too. They just make all that work programmatically.

Yeah, I want to talk about agents. But before we move on, are there any other custom-built chat GPTs that you've made that you rely on that we missed? Yes. One is Show Prep Assistant. And this is one where, in our case, we're going to interview a guest. And so we take their most recent book and upload it. And this used to take one of the writers on our team about three hours a week. And now we can do it like in two minutes.

And it's no substitute for reading the book, to be clear. But it's enormously helpful because it will spit out the biographical information about that author, including things that aren't revealed in the book because it goes out on the web and searches that stuff. It will also give us a breakdown of the book. And this is all part of the briefing process. But then it will give us the 10 best questions that we could ask based on the book's content. And it also gives us like the top criticisms of the book, just so that we're aware of that kind of thing.

But it makes it very serious. It's kind of like I'm a big fan of templates. So if I ever am going to create something that I think I'm going to reuse, I just go ahead and take the extra time and create a template and try to optimize it from the very beginning and then improve that over time. I do the same thing with these chat GPTs, these custom ones. It's not an agent. I don't know what we call it, a GPT, I guess. Yeah, I think it's GPTs.

I have another one for show notes so that we can take the transcript from the podcast and then we can feed it back and it gives it, it outputs in a very specific format. So we're able to specify in multi-markdown language exactly how we want it back. There's a title that's an H1, you know, then there's all the different parts that are H3s. And so we can do all that, specify that in the output format.

for the custom GPT. And I even use these in my coaching sessions because every one of my coaching sessions with private clients are recorded. And then so we upload the transcript of that and get it to produce a set of session notes for that particular conversation. And that saved, my executive assistant was doing that and it was taking him probably a day and a half a week. And now it literally takes him

a minute or a two minute. The hardest part of it is going and grabbing the transcript and then putting it into the GPT to generate that. Yeah. Another one, and I think this is going to be the future for a lot of people like me who create courses online or who have group coaching programs, and that is to help our clients get the results faster, easier, and cheaper. So in my book, Your Best Year Ever, I teach a very specific goal-setting methodology.

And I have templates for these and it takes people forever to get it. You know, I think I'm clear. I've written it clearly. I've taught it clearly, but inevitably they struggle to come up with it. Or I help them identify their key motivations as to why that goal is important, because I know that they're going to get stuck in the messy middle where they, you know, they're too far into quit, but they're not sure they got the resources to complete. And so I created a custom GPT that's like having a coach that's

right by their side and walks them step by step through the process. And every time I demo that to a group of clients, their heads explode because it makes it so easy. It gives them output that they can write directly into their full focus planner and

And I see this kind of as the future of course creation. I'm doing one now for our course, Life Focus, that'll help people construct a personal mission statement, a personal set of core values, and a personal vision statement for each of their nine life domains. And the fact is, most people aren't writers, and they struggle with this stuff. And if you can have a GPT that makes that easier and faster,

All of a sudden it puts them back in the game and now they can participate in a way that they might've not done in the past because they got stuck. Yeah. It's just so much of this stuff is, is helping with that. The friction points of this work. Yes.

Yeah. I could tell stories, but I think that you're, you're right on track with this, but you know, I am interested, like when you talk about your reliance on chat GPT, it sounds to me like part of it is that you have developed expertise with chat GPT. So why would you move on? You know, it knows you better and you know it better. Right. So, and I think there's something to that. You've got to pick a partner in this thing. If you're going to like really maximize your benefit from it. It, you know, it's really true, David. I,

Like I'm committed to ChatGPT, but I wouldn't exactly describe it as a marriage. You know, I mean, I still look at other stuff. You know, I came across this phenomenal multi-agent program yesterday. I'm not recommending it because I haven't tested it yet, but I watched some of the videos and my head exploded. It's called Mind Studio. And it looks like an amazing piece of software. So I want to experiment, but I keep coming back to ChatGPT.

I know how it works. I know, like any favorite digital tool, you kind of know where to find that thing that makes it do a certain thing. And I've really learned how to do the prompting. And one thing I use ChatGPT for is sometimes I'll write a prompt, oftentimes I'll write a prompt, and then I'll go to ChatGPT and say, refine this prompt. And that's the future, I think, is using AI to create the AI.

Yeah, well, I kind of laugh when people tell me they want to pursue prompt engineering because I feel like if you want to talk about a job that's going to get knocked out by artificial intelligence, it's prompt engineering. Absolutely. Because there's nothing better at making a prompt than the AI itself. I think there's some value in knowing it now. Just like, you know, it's helpful if you're going to use a calculator to learn math the old-fashioned way. But, you know, do we even do that anymore? Do we teach kids? I mean...

Like I don't even try to add anything anymore. You know, I'm just automatically using a calculator on it. But I think that there may be some value in learning the process. This episode of the Mac Power Users is brought to you by NetSuite. Go to netsuite.com slash MPU for the leading integrated cloud business software suite.

Nobody knows for sure what the future holds for business. If you ask nine experts, you'll get 10 answers. Some will say rates will rise or fall. Others may say inflation is up or down. The only way business owners can know for sure is if someone invents a crystal ball.

Until then, over 40,000 businesses have future-proofed their business with NetSuite by Oracle, the number one closed ERP, bringing accounting, financial management, inventory, HR, all into one fluid platform. With one unified business management suite, there's one source of truth, giving you the visibility and control you need to make quick decisions.

With real-time insights and forecasting, you're peering into the future with actionable data. When you're closing the books in days, not weeks, you're spending less time looking backwards and more time on what's next.

The business world is competitive, and the only way to survive is to have that single source of truth with forecasts and actionable data for your business. I saw it as a business lawyer for years. The companies that get the information and know what's going on are the ones that survive.

NetSuite is there to help you with exactly that. Whether your company is earning millions or even hundreds of millions, NetSuite can help you respond to immediate challenges and seize your biggest opportunities. Speaking of opportunity, download the CFO's Guide to AI and Machine Learning at netsuite.com. The guide is free to you at netsuite.com.

That URL one more time, netsuite.com slash impu to get that free guide to AI and machine learning. Check it out. Some really great information in there. And our thanks to NetSuite for their support of the Mac Power users and Relay. Okay, so you started talking about agents earlier. I shut you down because I wanted to finish ChatGPT, but I am so excited about this idea of AI agents. And you made the joke that it's a month away, but I don't think you were really joking, were you?

I really wasn't. You know, I feel like every time I state a limitation of AI, I have to add the phrase, not yet. Because the models have come so far in the last few months. And I mean, it's continuing to increase exponentially. And every time you turn around, there's a new tool that does something the other tools don't do.

And I think this year is going to be kind of the year, and I'm not, this isn't original with me, but I do think the focus is going to be on agents. I mean, even looking at much smaller use cases, things like what's happening on a phone, right? We're just coming out of Samsung's recent event. They've got kind of this mashup of their technology and Google's technology to like do things on your phone. And yeah,

I think on the Verge cast, David Pierce was talking about like, hey, I would, you know, find me five popular YouTube videos about this subject, pull the URLs into this note, right? And it's hit or miss, but that's sort of the idea of a lot of these things. It's help me get a task done. And that can be within the bounds of chat GPT, but you're also seeing like what Apple has promised with Apple intelligence that it's going to work across the apps and your personal data on your personal device. And, you know,

I think there is something to that. We want our devices and our software to be smarter and to help us get things done and be able to pull information from across different sources. And whether that's going to pan out in something like Apple Intelligence or the new Samsung phones, I think is to be determined, but it certainly is interesting. Are you guys hopeful for Apple Intelligence?

In the long term, I think it's going to be good. I just don't think it's there yet. You know, I run the betas software on my iPhone and on my iPad. And so far, I'm not feeling it. Yeah. Part of it is that Apple has branded their AI foray into the mind of most people via Siri. And Siri is such an unbelievably bad experience.

You know, I find it's almost never right. I'm constantly having to correct it, or it just tells me it can't get the information. And, I mean, if I were Tim Cook, that'd be the first thing I'd fix, you know, try to build back, you know, that they have competency in this kind of thing. Now, you know, on my phone, for example, I have the action button program to immediately go to advanced voice mode with ChatGPT. Yeah.

And the answers there are almost always over the top. In fact, I sat with a friend of mine, a new friend of mine from Honduras. He didn't speak hardly any English. I didn't speak hardly any Spanish. And I told Chad GPT, you're going to act as a translator in real time as we have this conversation with each other.

And I don't speak any Spanish. I speak a few words, but I don't really speak any Spanish. And I asked my friend, I said, how is the translation? And it was happening just like a translator. It wasn't a long, awkward pause. It was just happening. And he said, impeccable. Wow. And that's for years, that's been the sort of go-to example from tech companies of like, you have this thing and you sit on the counter between you and the other person, like you can converse in your own languages and like,

It's a compelling example because like that unlocks all sorts of potential, right? Like you can have a conversation with somebody you couldn't have a conversation with otherwise. But at the same time, like if it gets it wrong, then, you know, instead of asking for the bathroom, you're asking something inappropriate, right? The stakes can be high there, but that's encouraging to hear that they've got that passable at least.

You know, one, one, two quick hacks on advanced voice mode. First of all, I named the chat GPT Claire. So I feel like I'm having a more personal experience and it really feels like a conversation. Sometimes I have to remind myself that I'm talking to a machine. But the other thing I've had to tell it was I said, when I pause, don't interrupt me. Wait at least three to five seconds to make sure that I'm finished. Because sometimes I'm just thinking, right? And I said, say that to your memory.

Yeah. And these things do evolve. Like you can give it additional instructions. I have one, one of my chat GPTs is an optical character recognition machine. We talked earlier about analog. And so I'm writing a lot of things down. I write an architect block text and in my, my lifetime, I've never found a program that can OCR my text.

But I started talking to Chachi PT and I told him the problem. I write like an architect. Everything is a capital letter, but not everything is meant to be a capital letter. And like, I started kind of telling it what I was doing. And now it flawlessly conducts optical character recognition on any picture I sent it, you know?

And like, that is a thing that evolved over time. And once I got it modified, then it's good. You know, in fact, I've even given specific instructions when I copy it, what I want you to copy. You know, I don't want you to copy the part where you told me that you did the work for me. I just want the stuff that you translated. Exactly. You know, and things like that. So it's like,

It is like a conversation. When you said earlier, this is not highly technical. It's not. You have to get your head out of the idea that this is programming. This is a conversation with an assistant. You know, it's a digital one. But if you explain in plain English, quite often they get it right. Yeah, it's kind of amazing. By the way, one of the things I didn't say about ChatGPT that I like is I love the Canvas feature where you can create an editable document, you know, much like the Cloud's Artifact feature. Yeah. But I like it even better.

So that's another reason for ChatGPT. So I think you almost have to differentiate between multitask and multi-agent because those terms often get confused. So, for example, MindPal refers to itself as a multi-agent platform, but it's really a multitask platform. It's not going outside of the MindPal ecosphere. I mean, it may search the web, but all of them do that.

But it's not actually acting on your calendar or scheduling things. I mean, you can do that with webhooks and connecting to make and all that kind of stuff. But it's not natively doing that. I think multi-agent kind of things are going to be the things where they actually can interact with other apps like ChatGPT can on some apps now. And I think that's going to be the future too, where it can really be your whole ecosystem is interactive and AI can talk to it and read from it. So explain to people...

because we talk about the word agents a lot, but give an example of how you see one of these future agents being helpful to you or the ones you're currently using, I guess. Well, and this is where that kind of gets fuzzy because, you know, if you think of it, and this is kind of the metaphor that I tend to use. If you think of, if you're a business owner or you lead a business or you lead a department, the future is going to be that you're going to have human assistance or special purpose assistance that are humans. Right.

and you're going to have those that are digital. And so you're going to have a hybrid workforce. And there'll be things that you can delegate to an AI agent, for example, that can even further delegate to other softwares, to other apps, websites, combine it all together. And probably the best example is something like Make, where you can have it go through, for example, you could have it

create a series of social media posts, give you the opportunity to approve it, and then go out and actually post those on the appropriate social media platform. That's already possible today. I think that's going to get more complex. You know, for example, when I'm writing my newsletter, I know that there's about 17 discrete steps in just writing a newsletter. You know, it's everything from identifying the big idea, you know, the problem, the big idea,

coming up with an outline at the highest level, doing the research on each point. These are each discrete tasks and in an agent workflow, and a lot of these like MindPaddle now have this and MindStudio has this, where you're laying out much like you used to do with Infusionsoft, where you would map out kind of the flow

Even with branching, you know, if you encounter this and you go to this branch, but you're mapping out the workflow. And then once you get it mapped out, then it's automatic. And with both MindPal and from the demos I watched on MindStudio, you can specify the appropriate tool for each part of the process. So even like in writing my newsletter, I try to come up with an image that, you know, captures the essence of the article. And so in that case, I'm

you know, flipping over to a image creation tool like Flux. And you can do that. You can tell it you want it to use that image or idea. Ideogram is another one that I like to use. But you may specify different ones for different kinds of projects. So instead of just being locked into one large language model for the entire process, like you would be in ChatGVT, you could say in a multi-agent model, you could specify the appropriate and the best

agent to use. Yeah. And it goes beyond mere text at that point, you know, where it can go like on Safari and do a research and go on the web for you. It can download software and, you know, build a spreadsheet, not, not only give you the ideas for the spreadsheet, but actually open Excel and create the spreadsheet for you. And it's,

This stuff is already exists. It's just not public yet. Now at the highest tier chat GPD has the ability to go search the web and do work on the web for you. You know, I, I was joking with somebody that the agent test is, can it order a pizza and get it delivered to my front door? You know, and you know that, but that's a thing that by the end of the year,

this will be accessible. I believe to everybody listening to this, where it can go on the internet and do stuff for you. I did too. Like I, I ordered some shirts and they were the wrong size. Like I was thinking this morning as I spent 10 minutes going through that process, is this something that I will have to do in a year or will I be able to tell an AI return, get my shirts returned, you know, and then I show up and there's a label on my printer and I got to just go to the post office, you know, that kind of thing. I,

I don't think we're that far off from this stuff. And it's something to be thinking about as you're working. You know, the opportunity for us to scale ourselves, you know, the constraint has always been time. You know, this is a lawyer. You know, there's only so many billable hours in a week. And the thing that AI is going to enable us to do is to scale in ways that we haven't thought of before. Like, I've been experimenting a lot with 11 labs, and I've cloned my voice there.

And it sounds just like me. You know, I can envision a future where like one of the most tedious things to do as an author is to read the audio book in the studio because it's going to take a day and a half of your time. It's very tedious. It's mindless in a way. And I don't really enjoy it. But the AI can do it with my voice.

Which, by the way, raises some important ethical considerations. The technology is always going to move faster than humans' ability to process the ethical implications of it. But here would be another one. You know, if you use something like HeyGen, which is a text-to-video thing, and you clone yourself in that. Like, I have this friend, Julia McCoy, who's the head of a company called First Movers. And she's amazing.

very much a, I don't know, optimist when it comes to AI. But one of the things she did, she's got a YouTube channel that's worth checking out. And so one day I'm just watching her video. I don't think anything about it. It's great. You know, I'm getting some great content from her. And then she discloses that that entire video was generated with HeyGen and 11 Labs. It wasn't her at all. She wrote the script. Then a couple of weeks later,

She continued to do this. Now she's disclosing. So like a lower third comes up under her explaining who she is. And so it's like Julia McCoy, digital clone. So she's disclosing this. But what really gets interesting is that she goes into the back end of YouTube and shows you that the engagement that she gets on the videos that are AI generated is better than the one she does herself. I don't know how I feel about that.

Yikes. That's how I feel. Yeah, I mean, it's complicated. It's so complicated. It's fascinating from a tech perspective. But there's also a part of this where you just go, I don't want to be wondering every time somebody's pitching something to me or teaching or talking, is that really them? Yeah. And that's where I think there's going to be a huge opportunity for people that are willing to be very transparent about what they're doing, what's AI and not. Like, my daughter just wrote a manifesto

about how we're going to use AI at full focus. So one of the things we want is our audience to have confidence. And so we just said, you know what? We're no longer going to sign our name to anything like emails unless we actually wrote it. Because in the past, we have copywriters, you know, that would do it. Humans, but still copywriters. And we said, it even made us question that practice.

It's tough. And these questions are just getting raised and, and the potential for fraud when you can reliably duplicate a face and a voice goodness, you know, that's going to be, that's going to be a whole nother problem. It is. It's coming. It's coming. I mean, there's no question. It's,

And my take on it has always been, as we're technology people, is how do you find a way to take advantage of this for you to help you do the hard stuff faster, but also give you more time for what's more important to you? I mean, there is an angle to this. And I don't think the fact that it can be abused for fraud means that we shouldn't be looking at this.

I agree. And in fact, this is not a new problem. You know, every time there's a new technology going back to the printing press, you know, there's the opportunity for abuse. Same thing with the Internet. This episode of Mac Power Users is brought to you by Google Gemini. I used Google Gemini the other day, and the most impressive thing to me is just talking to it. You go live with it, and it's just like you're having a conversation. You can talk about your day or have it explain something to you or start brainstorming ideas.

I'll give you an example. I pretended I had a job interview coming up and asked it to help me prep for the interview. It's been a long time since I've had one. It immediately started suggesting common questions I might get asked. And then I started talking through my answers out loud and it would give me feedback. And this is all happening in real time, like you're talking to a career coach. And that's just what I tried first, but you can talk to it about anything. That's the magic of it. You can have this back and forth and it's all seamless.

If you haven't tried Google Gemini yet, it's definitely worth checking out and you'll see what I mean. Our thanks to Google Gemini for the support of the show and all of Relay. So Michael, we've talked about a lot of different tools so far. There's a bunch of stuff in the show notes people can go check out.

But I want to get down kind of into the weeds a little bit. A while back, you mentioned Missive as an AI tool. And you said something that caught my eye, so I kind of flagged it to come back later about this being a tool not only just for you, but for your team. And y'all are implementing this.

a lot of this technology across the organization with different people. I'd love to hear a little bit more about, particular how you're doing email in missive with a team, but sort of more broadly also where these things apply to people you're working with and sort of how those interactions have gone. - You know, how we started this, Steven, was my daughter, I feel like I work for her,

She's the CEO. And so I, she asked me if I would do an AI training, knowing that I had fallen down the rabbit hole and was pretty deep into it. If I would do a training for the team. So she blocked two, four hour sessions for us to do that. And it was very workshoppy. So I kind of gave a basic orientation and I taught him how to do some basic prompt engineering. Cause that's where we're at today. And then I gave him some assignments and then we came back a week later. And so that got everybody acclimated to it. And I think if you're a leader, right?

it's your company in any capacity. I think it's important that you lead by example. And I think that that's one of the requirements of leadership, you know, is that you go first. If you want to build this into your culture, you know, they're going to be, people are going to be looking at what the leadership is doing. And so, you know, we've really adopted it first and foremost as a leadership team and full well knowing the limitations of it. So we signed up for a team account on ChatGPT. We've got a few other

team accounts on other AIs as well, but that's the principal one that we use. And the great thing about these custom GBTs, like the one that I call all full focus content, which is all my books and podcasts and all that, is that we can actually publish those for the entire team to use. So that anytime any of our writers or even our designers are creating a product or creating new content,

They can draw on the existing body of work simply by selecting that from the explore GPTs because it will give you the team GPTs up near the top. Some of those I use just for myself. They're personal GPT. Many of them are team GPTs. But yeah, you've got the opportunity to share, you know, keep it private to yourself, share it with just your team.

everybody with the, you could give the public link to. And by the way, inside your team, when you share it, you could also give them the ability to sort of explore the administrative backend so they can actually modify the GPT, which I almost, by the way, never do. You know, I say, if you want something changed in it, tell me, cause I want to agree. So I don't just suddenly find it's not working or not working the way I expected. Did I answer your question? Yeah, I think so. I mean, it, it's,

In one way, these technologies are very personal. And so I think applying them to a team setting is really interesting. And I know just from my own time leading Relay over the last 10 years, like sometimes a tool that I've built for myself, when I hand that over to my team, it's like, oh, they actually don't think about this problem the way that I've thought about it, or they want something different from it. And there's always kind of back and forth when you've got a group of people thinking

But what's different about

that sort of problem solving in the AI era is that you can iterate much more quickly, I think, than in the old days. You're like, okay, this is our process for doing this within the business, right? When you're talking about an AI tool, you can iterate and change. And like you said earlier, these tools are getting better at a shocking speed. And so you're also seeing your processes and your tools getting better on their own, which is something that,

that hasn't always been true in the course of civilization, right? Like, you know, the first stone-carved axe didn't get better overnight, but, you know, now the AI agent can. Yeah, it's pretty crazy. And, of course, you've got AI purportedly in the not-too-distant future iteratively coding itself or improving itself. And once that gets unleashed, who knows what's going to be possible. Yeah. Yeah.

Do you, I think some people hear that and kind of have a chill, a chill go down their spine. Do you have, I mean, I know these things are complicated, but like, are there lines that you can see that maybe these companies or these tools shouldn't cross? Man, it's complicated because I want exploration. I want development. I think we need it to be competitive, you know, globally because, you know,

whatever guardrails we put up for ourselves, others won't install those same guardrails. And so that raises the question, should we do it anyway, just because it's ethically appropriate, or do we have to remain competitive? And I think there's going to be some lines that we're just going to have to say, as a society, we're not willing to cross this. And I don't know what those are, because again, I think

The technology changes faster, and this has been true in the entirety of the modern era. The technology changes faster than our ability to process it and come to ethical conclusions. And I remember back when I was a philosophy student, we were talking about this, and this was, you know, 40 years ago. So it's even more important now. And my biggest fear when it comes to AI, and I'm generally optimistic, but my greatest fear is

is that we will be solving ethical problems that no longer exist because the AI has moved so much past that. Do you have an example that comes to mind there? No, I don't. I do think, because I'm optimistic, I think a lot of the problems that come up can be solved. There may be an intersection, for example, of blockchain with privacy where we could guarantee it.

But again, I don't know. And I don't know enough about blockchain to really know if there's an application there. You were asking about Apple intelligence earlier. I feel like that is like, there's an opportunity here as the AI progresses to have a somewhat inferior local model. I mean, like even with deep seek, they, they open sourced it. So people are already running it locally and that's a nice privacy solution. And I think that actually helps Apple, uh,

if they get to a point where these models get small enough and powerful enough that you can just install it locally. And then that solves a lot of that problem for people. If you can run it locally. Yeah, I think it would too. I think that's a real possibility, but it's going to have to get a lot more user friendly because maybe for us geeks, we can install it locally and get it all running. But for the average user,

And if anybody could do it, Apple, I mean, if they, if they're good at one thing, it's simplifying complex stuff and make it accessible to the masses. Yeah. I, the other thing that concerns me about the future of AI is once you wed AI with robotics. Yeah. You know, I mean, that's, that's where I go to the Terminator, right? Yeah. And, you know, I, I, I do understand the possibility of having, you know, people that could help you with chores around the house domestically or in your company, but

or just do physical things that you can't do with just AI by itself. But there's also the potential for enormous abuse and

Well, another fear, I think a legitimate fear is, is disruption. I mean, you're talking about your assistant taking a job that took them a day that gets done in a minute. Well, how much longer before that assistant isn't necessary or not as many are people are going to, I mean, disruption always comes with technology, technological change, but I do feel like I'm not sure society as a whole realizes how big of a change AI is going to represent and,

And I think that's a surety. I mean, I guess there's a possibility we'll create the Terminator world, but I think it's certain that we are going to have massive disruption in what people do for a living and what jobs are relevant. Yeah, and there's lots of people proposing all kinds of solutions from universal basic income to mass training. But yeah, I think the problem is really not

new, as you stated, but it's the scale at which the disruption will occur that makes it of concern. It's complicated. We keep coming back to that, but it really is. I mean, and what I'm struck by in talking through this with y'all is that some of these concerns are

old concerns that just come back, right? Like there were people sitting around, probably not on podcasts, but, you know, sitting around in the, as steam, you know, began to power industry, right? It's like, okay, this, you know, we're in this industrial revolution, right?

People are going to be replaced by machines. And that happened to a large degree and it didn't happen to a large degree. It's like some of these, these concerns are, are not new. They keep coming back around. Uh, but some of them are new because of the speed and the, the power behind this stuff. And, and,

I think it's up to each of us as we explore these tools and how they interact with our workflows and our lives is to understand that, to understand that this is something where not only will people disagree with each other on where the lines are and what's good and what's bad, but that the lines themselves are probably going to move when we're not looking and to be aware of that. Agreed. And I think...

This has so much uncertainty as we look to the future. And if there's one thing, one of the things that humans don't do well with, it's uncertainty. It creates anxiety. We've already got an anxiety problem. But I think just for me, I just remind myself, you know, yes, there's a non-zero chance that we could end up with the Terminator scenario. But I don't lay awake thinking about that.

I'm thinking about all the possibilities. And I really believe in entrepreneurialism. And I believe that any problem that gets created, whatever it is, that humans are going to figure out a way to solve it because we are wired for survival.

I think we've got some great questions. I do think anybody out there who's shying away from AI should be exploring and experimenting. No matter what, you know, this turns into, it's happening. So the ones that figure it out are the ones that are going to be in the best position to deal with it.

Michael, we always like to finish up the show with a few apps and services that are bringing you joy and delight. I think we've covered a lot of them during the show, but I see on your list here, there's a couple that we've never talked about in the Mac Power Users. I thought maybe you'd want to point a few of them out. Yeah, one of the big ones that I'm enjoying right now is MindSara. I've been a longtime day one journalist. I've been journaling for, I don't know, 14 or 15 years.

Probably on average, three times a week. But I've gone through stints where I do it seven days a week. And I just found an AI assistant. There's actually two of them. Mine's Sarah and one called Rosebud that take all your journal entries and are able to analyze them and analyze your current journal entry based on what you've written about previously. So it can recognize patterns, emotional states,

And it's like having a therapist sitting with you as your journal. And I just am very excited about this and for the insight that it's given me into my own life. Gamma AI is an AI that I use for producing documents. Like, for example, I teach a master class to our Double Wind coaching program on Monday night.

And so I've been experimenting with taking the transcript from that masterclass and turning it into a beautiful ebook. So I have to clean up the prose, you know, make sure that it reads well, because obviously I'm giving it a live setting, but then I can pour that into gamma and it will create a beautifully designed ebook. And it's kind of best known for creating slide decks. And I don't, I really like working in keynote and I've got my workflows and all that. So I haven't used it for that.

But I think some of these design things are amazing. But we got to know the difference. Like I was going into Flux Ultra for most of the images I wanted to create. But I was like endlessly tweaking the images to get them because they weren't quite right. And I had to do some more thinking. And there's a certain level of randomness that it gives you. And so, you know, six fingers and all that kind of stuff on one hand. And so then I discovered, you know, I just need to go back to stock photography because it's 10 times faster.

So there's some stuff that AI has the promise of doing better at, but it's actually a giant time suck. And I think you got to know the difference. I've got one for you on your images. Okay. Napkin AI. Yes. Okay. Well, see, you're ahead of me. No, no, no. That's great. I actually love that one for diagrams. I'm talking about images like a photograph. Yeah. Stock photography. Yeah.

I'm very curious to see how all this develops. We're going to have to have you back in the future as we are all in our AI worlds. But you've given me food for thought. I'm going to try and do some comparisons between chat GPT, GPTs and notebook LMs because I have gone down the notebook LM route pretty far at this point.

But I like the idea of tuning up a chat GPT and I'm not so hung up on privacy with it in certain categories, you know, but I think we all got to start somewhere. We should all be familiar with this stuff because it is coming and it is going to affect a lot of change. Any, any big wishes you have, like what we talked a little about agency and the agent models, like what in one year from now, what do you see yourself doing with effective agents that you're not doing now?

Well, I think it's going to get easier and easier to do the tasks that we want to do. And I think that when AI can write the prompts, and MindPal has a version of this now where you go in and you just describe a workflow and let it build it out. And it will literally populate a canvas with that. And I think that's going to even get better and easier going forward. But yeah, I don't know. I just...

Every time I see a new AI, my head kind of opens up, my brain opens up to new possibilities. And it's coming faster than, frankly, I can process. I don't know what's going to be possible a year from now. But I do think that we're going to look back on 2025 as a pivotal year. Thank you, Michael, for coming on the show. It's always fun talking to you and hearing what you're up to. For folks that want to learn more about what you're doing, where should they go? Probably just our main website, fullfocus.co.

All right. We are the Mac power users. You can find us at relay.fm slash MPU. Thank you to our sponsors this week. That's our friends at one password, Squarespace, NetSuite, and Google Gemini. If you would like to join more power users, the ad for extended version of the show, you can sign up for at relay.fm slash MPU, and we'll see you next time.