The AI-Supercharged Dad

We talk about the moment an AI-generated goodbye song brought a room of parents to tears, and why that kind of emotional accuracy changes what “AI is capable of” for everyday people. We break down a simple AI fluency framework and share practical prompts, workflows, and safety habits that help us use tools like Claude without getting overwhelmed.
• using AI to create meaningful creative work that lands emotionally
• why “Grand Canyon moments” with AI are happening faster
• how different AI tools specialize and why that matters
• the 4D AI fluency framework: delegation, description, discernment, diligence
• real delegation examples that save time and ones that can backfire
• writing better prompts by adding context, constraints, and clarifying questions
• building trust through projects, review loops, and verification habits
• privacy, security, and the pause-before-you-paste diligence rule
• three tiers of AI fluency and the skills we share for each
• using certifications to prove learning agility as jobs shift
Listeners, if you hear this and that's something that resonates with you, please drop it into Discord, drop it into Insta, TikTok, wherever you're seeing this

Transcript

The AI Goodbye Song Moment

SPEAKER_02 0:12

Your 10-year-old daughter is more AI fluent than you are. And she didn't take a course. She just wasn't afraid to start playing, experimenting, and growing with AI. Now, my daughter and I had been going to this music together class for two years. This is a place where you go with your toddler, your infant, you learn music with them with other parents. So it was the same room, the same kids for two years, alongside this incredible teacher named Jenny. We saw first words together, first steps together. And when Kelly, my wife, and I decided that we were moving to Boston from Northern California, this meant saying goodbye to all of those people. So I sat down with an AI music tool called Suno and tried to write a goodbye song. I put the name of every kid in the class in there, every parent or teacher. I described what that room felt like on a Tuesday morning and what it was going to feel like to leave. After about 10 turns, Suno produced a song that was not just technically flawless, it was emotionally flawless. It captured something I couldn't have possibly written myself in a hundred years. And so I played it at our last class. The room was silent and then tears. Parents that I had been doing circles with, singing, dancing, running around for two years were crying because core memories were being unlocked in real time through a song that a dad and a machine wrote together on a Monday night. And that was the moment that I truly felt the uncanny valley being exited when AI understood the ones and zeros of my emotion, of that of the people around me, better than what 99.9% of people ever might all through song.

SPEAKER_01 1:57

Yeah, I mean, like this is before it was hot. Like, how did uh like what this is just me from a people perspective? Did the other parents did like they have any idea when you told them that AI created this, like that that was possible? I definitely didn't believe it.

John’s Path Into AI

SPEAKER_02 2:13

And it required some sharing of some links. And now you almost have to prove that you're human in some cases. I actually had to prove that this was AI, which was hilarious. But it definitely was me watching people have a true unlocking and a realization of what AI could be through a deeply personal moment.

SPEAKER_00 2:30

So you obviously were familiar with that AI tool at the time, and that was before most people were using AI on an everyday basis. But John, take us back a little bit. When did you first learn about AI and what got you so interested in using it?

SPEAKER_02 2:45

Well, I've been in tech for a long time, uh specifically in tech leadership. And back when AI and machine learning was just really called sort of advanced tech, uh, this was back in the day at Apple, it dawned on me that this was something that was really important to some really smart people. And so it should probably be really important to me. And so I've been excited about it. Also, the role that artificial intelligence has played in a very unrelated industry, video games, uh, something I've always enjoyed for a very long time. So I I really, though, got meaningfully invested in AI such that I meaningfully pivoted my career out of Amazon, out of big tech, into a pure AI play in 2021. And that was when I realized that there was something big on the horizon. Bezos, many others were talking about how machine learning, AI as a concept at the time wasn't really what it is now, anywhere near, obviously. I don't think anybody realized what Transformers were going to be. But that was when I started to realize that I didn't have any shot as a non-technical person at staying in a place of understanding with this stuff until I got to a place where I could play with it. And so uh I had some some great friends who were kind enough to talk to their non-technical schmuck buddy and help me learn how to play with AI earlier on in a sort of sandbox environment, which was which was really helpful. And that helped me see why I needed to change everything and get involved in this space.

Grand Canyon Moments And Claude Design

SPEAKER_01 4:04

And when like today, a lot of people, every day, people are waking up to the this being such a true game changer for everything. You know, again, it's not probably won't all be great, but I'm I'm optimistic that it'll be mostly great. Again, knock on wood. Like when did you have that aha moment that wow, this is really gonna change the world?

SPEAKER_02 4:23

You know, I call these Grand Canyon moments. It's one of those moments where you walk up on something and it's so vast, so seemingly infinite, that you stop and your brain reframes. And then when you start thinking again, you're doing so through a completely new lens. And those Grand Canyon moments, like what you're describing, Jeff, they used to be something that I would experience on a maybe yearly basis, like the first time you saw an iPhone or the first time that you interacted with the internet. Or at one point that was something that would happen every maybe decade, and then it was every few years. And now, because of Moore's Law, technology is getting so much smaller, faster, cheaper, and accessible, and AI is even further accelerating geometrically. So now those grand canon moments for me are happening nearly monthly and probably will be weekly here soon. I had one last night that we'll be talking about here in a minute that just similarly blew my mind so completely that it again made me stop and say, everything I've really known up until this point might just be obsolete, and it's time for me to stop and think again. Yeah. It's uh it's humbling and exciting.

SPEAKER_01 5:29

Well, I mean, like now I'm like on pens and needles going, I want to hear what you discovered last night because like this is my first interaction with you, John was you're always peeking around the corner. Like you always are kind of thinking through things that most people haven't thought through yet. Like a very proactive creative thinker. And so like when I hear you talk about this, like with AI, it's like, oh my gosh, not only are you peeking around the corner, you you kind of future tripped around the corner, like you know what's happening. And so, like I'm like, tell me. Like, like I want to hear what happened last night already. Varying the lead a little bit, right?

SPEAKER_02 6:05

For real. Um okay, well, I'll only vary it for a few more moments. What I the reason why I'm so excited to share, even just this somewhat small story, Jeff, it connects to this idea of future tripping and looking around the corners, which isn't just true of season one of Beast Games when we were constantly trying to figure out what even the color of someone's vest meant that was working on the crew. Trying to try to nostrà sit on every moment.

SPEAKER_00 6:31

We have way too much time on our hands so forth.

SPEAKER_02 6:34

Man, there was some real like beautiful mind mental math happening in Las Vegas right now.

SPEAKER_01 6:42

And even even on the bus, I think we were all on the same bus going into episode one when we were doing these models of I mean, the triangle, how many different, you know. I just it was that's I mean, that's when I realized this is who you are. Like, this isn't okay, John, research AI and talk to everybody about it. No, no, like you're always on the cutting at cutting edge of this stuff. Like, this is again, I know there's people out in the audience who are listening, kind of going, Oh, this is cool. I want to hear about it. I am sitting here in my chair going, I cannot wait to hear from you, John, because I know this is who you are. Well, you know what the funny thing, Jeff?

Making AI Less Overwhelming

SPEAKER_02 7:13

So Jeff Dir and I were part of like a little group and we wanted to get into our peak physical, mental, spiritual shape to compete. Obviously, Jeff did a little better at his preparation than anybody else, but we called ourselves something silly like Team Alpha. That was the first agent that I ever built. That was the first time that I ever stopped to take make.com back then. That and Zapier at the time were really the only automation tools that were publicly available, maybe a handful of other like small players. But I remember the lengths to which I had to go to get what was at the time Chat GPT 3.2 to actually connect to the internet to on a daily basis research trends and come back with training recommendations for mental, physical, overall mentality training that led to that recommendation of the triangle exercise, which is so funny how full circle some of this stuff starts to feel. Because that in many ways led to this conversation. Without further ado, the the Grand Canyon moment that I had last night, which is the second one I've had in the month of April, at the time of this episode's filming, this is just about 48 hours after Claude launched something they called Claude Design. And it is unbelievably capable at understanding an entire brand's design philosophy, helping you to adapt and iterate on it, and to ultimately create things that are so beautiful and so useful, far beyond what you could have possibly imagined when you threw in one prompt talking about how you wanted to design a new website for a tire store. All of a sudden, it's deeply interactive, it's connecting to socials in ways that you maybe wouldn't have ever thought of. But we're really getting to that point now where the agentic future that everybody talks about, you know, I think a lot of people they say agentic and they probably mean some form of doing it for me, and that, and rightly so. But I don't think we're gonna go from this light switch of us prompting everything and then inspecting everything. We'll we'll get to how we go about doing that in a moment. We're not just gonna go from that to all of a sudden we've got the robot from the Jetsons running our house. That is not an immediate light switch, it's not binary. I believe that we are gonna see an unbelievably fast transition into that lifestyle much more rapidly than we could have imagined. But on that very fast hockey stick ramp, we are going to consistently encounter these moments where machines are thinking and dreaming for us in ways that we didn't anticipate and helping us to discover things that we otherwise probably never would have seen. And that is for me the Grand Canyon, where you almost didn't even know it was coming, and all of a sudden you're just hiking, you turn a corner, and then your whole mind just gets shattered. You put it back together more usefully. But yeah.

SPEAKER_00 9:48

Yeah, there's so much to AI, John. It can be so complex and it can be overwhelming for people. So how can we break this down and say, hey, here's a couple things that anybody can start doing right now to incorporate AI into your life where it doesn't seem so overwhelming? I'm so glad you asked this question, Jordan.

SPEAKER_02 10:07

This this is the thing I am most passionate about in my world right now that isn't a human being that I get to love and see and create amazing things alongside. And that is the democratization and access to AI. And access means a lot of things. For me, it means the ability to meaningfully participate in something. And that means access to not just what AI can do, but what AI can earn. I think we're in this unprecedented time right now where AI is minting millionaires, even now, very recently, the first solo company that was valued over a billion dollars, founded by one person with a very, very small amount of seed capital. So we're in this era right now, and we won't always be, maybe for a few more years, where you don't even have to know how to surf. You just have to be willing to paddle out with a surfboard and mess around for a little while, and you're gonna end up on a wave so big that this the odds are better, I believe, that you end up doing something transformational than that you don't. So to that end, there's a there's a framework that I would love to encourage everybody to check out and and probably adopt. And this is not a new concept, this is actually anthropics. So that said, a little bit of frame. So before we break this down, um, and this is directly addressing our our listeners. So, first, I need you to forget everything that you currently hold as true about AI, just for a moment. Forget about the robot apocalypse, forget about the guy on LinkedIn who's constantly posting about how he's replaced his whole team with Chat GPT. That that's noise. If it's even kind of true, it's still just noise. What we're gonna talk about here today is an actual framework that was, again, built by researchers at anthropic that breaks AI fluency, and that's a term you're gonna hear a lot today, into four skills. Not for tech skills. This is not something you had to go to Stanford and get a CS PhD in. These are for human skills, things that you already do right now, and things that I'm going to encourage you to double-click into. And that, interestingly enough, is the thing that really, that really got me. You know, having been in tech leadership for so long, I mentioned maybe Apple and Amazon. When I went through this framework, I realized that I was actually only maybe strong in two of these four. And that was very humbling. Um, and I think it'll surprise you too, anyone who who might be listening to this right now. Uh so before we jump into the framework, Jeff Jordan, any thoughts or questions on anything so far? I'm ready to jump in.

SPEAKER_00 12:38

I'm excited to see this.

Which AI Tool Fits What

SPEAKER_01 12:39

I'm gonna have a thousand questions. I don't want to pull you off, but like uh I feel like everybody, and I hope hopefully you you chat about it, but like everybody talks about along a GPT this, but now you see more people using anthropic, then some people went with grok for real time. And then you also have perplexity. Like, are all AI, at least um, I don't want to call them search machines, but everyone the interfaces are they created equal? Like, are some dedicated more to certain different things?

SPEAKER_02 13:03

Well, you're absolutely starting to see specialization coming along. I think that in a field where often you see companies treating breadth as depth, meaning if we can do everything, then we will create enough of a flywheel of data capture to get even better at everything. And so everyone for a little while was trying to build the everything machine. OpenAI arguably could be seen as trying to build the everything machine. There's not really a lot that you could go to do right now that OpenAI can't, or at one very recently couldn't, do for you, from video to images to search to design to shopping to so on and so forth, replacing Google. So your question, Jeff, are all modern-day transformers created equal? Now the answer is certainly no. And part of the reason that now what seems like overnight, all of a sudden things change. Yeah, they really do, because the speed of advancement is coming so fast. Anthropic recently essentially announced that they're not going to do everything. And the reason why their announcement of Claude Code took trillions of dollars of market value and essentially repurposed it into more AI-focused projects is that they have said we're only really going after enterprise and code. We're not trying to be every single person in the world's assistant, like OpenAI is marketing themselves as. And some could see that as a really bad decision. Some could see that as like a three to four trillion dollar market that they're giving to their biggest competitor. Side note, you mentioned Grok, Jeff. Anybody that counts Elon Musk at anything is gonna regret it. That guy, you know, he's got this thing called Project Colossus that's been going on in the background, where he functionally will probably have access to more total compute than almost anybody else in the world, possibly even other hyperscalers. Um, and that might be a bunch of words that don't mean a lot to anybody on this call right now, but just go into your favorite LLM, large learning model, meaning ChatGPT, anthropic, whatever, especially rock, and uh ask, hey, you know, what is this project Colossus? What is Elon up to? What does it mean to have uh surplus GPUs? Everything is good at something. Most things are good at almost everything. But when it comes to things like coding, you're gonna see people really quickly go to Claude because it's a really good ecosystem. In some cases, Codex, which is an open AI product. The fact that Grok is directly connected to Twitter and can actually pull from that is wild. That is so powerful. And I find that very few people talk about it, which is really surprising to me. Meta, everybody's acquiring these smaller companies, some of which can do very specific, really exciting things. Meta is now releasing this Spark product or subproduct that essentially helps you put your Instagram on autopilot in a lot of ways, or at least the parts of Instagram that might not be what you as a human could do creatively. But at the risk of staying on the said soapbox, I hope that adds a meaningful color.

SPEAKER_01 15:58

It's good. I mean, and again, I feel like we could talk about this forever, but like I I do, you kind of teased us with what Anthropic talked about with kind of the the 4D. I'd love to hear more about that.

Delegation That Actually Works

SPEAKER_02 16:07

Absolutely. Okay, so diving right in. So again, one one really shameless plug. And you know, we're not yet uh sponsored by our friends over at Anthropic, Dario, anytime you want to anytime. Um while we don't have a direct relationship with them, I'm an unapologetic advocate for going and getting, and we'll come back to this again. I have a very, very specific recommendation for folks. And I don't want to bury the lead again here, but essentially, if there was one thing that I would do right now, if there's one thing everybody listening to this should be doing right now, it's going and becoming a certified Claude architect with the Claude Certified Architect Program, which Anthropic paid something like$100 million to co-develop with some professors. So I happen to have an MBA from USC. If you were to ask me right now to trade my MBA from USC, which was not free, for this anthropic education, which is free, I would right now absolutely trade those things. If I could only have one of them, I would categorically trade my MBA, no matter where it came from, for this education, which is free and which you get in about 100. I'm not expecting that. I know. Hot hot dads, hot takes is later, but I think there's uh there's my hot take. Coming back to the point though, let's talk about the four D's. You're gonna see these in the anthropic content if you go and you learn about it. And it is a really helpful, really meaningful frame through which to view all this stuff and it makes it understandable. And honestly, in my case, it helps me look as a human at human skills. So the first of those four D's is delegation. So delegation is the one that most people, when it comes to prompting, get wrong. So delegation is not about handing everything to AI. It is about knowing what to hand off and what to keep. And again, this is a skill. This is not just a thing you can read and then follow like a set of instructions. This is something you work on and you get better at. This is how you become that person in the office that's just always a light year ahead of everybody. The person who you may variable be looking at right now saying, again, Jim, again, with your awesome stuff you're putting up on screen that took me three hours and you didn't 30 seconds on lunch. Yeah, you should be Jim. This is how you become Jims. This is the same skill that you're using right now at work. You don't delegate the strategy meeting to the new intern. It's not fair to them. It's not fair to you. You delegate the calendar invites for the strategy meeting. AI is the same thing. So the question is, what's the calendar invite in your life, in your work, and what's the strategy meeting in that dichotomy? Uh, that's the first D, which is again, delegation. Any questions, Jordan and Jeff on delegation?

SPEAKER_00 18:46

Could you give us another real example of something that, you know, someone in the corporate world, for example, would would delegate? What's a good example of, you know, outside of the calendar invite and strategy meeting? What's another thing that anybody could delegate?

SPEAKER_02 19:01

Well, in a moment, I'm gonna actually flip it and ask you guys what you would delegate, but I would love to give a good and a bad example of what delegation could look like here. I think that writing a team schedule is probably a really good thing to delegate. It's very numerical. You're you're literally trying to make boxes fit together based upon demand. And this is true if you're running uh a retail store. This is true if you're running a delivery warehouse, if you're creating an on-call schedule for a set of medical professionals. I think that's a really great use of AI because it also injects this element of fairness where you're you're adding a layer of objectivities. Really powerful delegation. And I mention that because it's the same reason that you would delegate to someone else as a leader. Because you know, you know, Jeff, in your entrepreneurial backstory, right? Like you've got a lot of folks that probably rely upon scheduling to be where they need to be. So that I think it could probably be a great application. Bad application would be prioritizing the things in your email inbox right now. So you can get inference. I think that you can really teach AI quite well. Gemini is starting to do this well. There's all there's all kinds of third parties that that offer this service. And if it gets it wrong once, you could find yourself in some real hot water. And you can't really afford to miss the email from the CEO that happened to have one or two words that it automatically flagged a spam. Or worse, the email from a potential partner that was reaching out to see if Jordan, you wanted to come and uh live on a cruise ship for two years for several million dollars, right? But it flagged it thinking that it was like, you know, right. Those would be two examples of like delegation and how I might think about it. But let me throw it to you guys. What when you're using AI right now, what are you delegating and how's it going?

SPEAKER_01 20:45

So for me, your last example I got popped on because I was definitely because e email, if you want to get a hold of me, if you if you do not want to get a hold of me, email me. Email is not my friend. And so I'm like, I need to do a better job at this. So I kind of created some AI agents, but I I I lost some emails, and I still am trying to figure out how I did it. Um so like thought it would be easier to prioritize stuff, but yeah, it's uh some stuff got lost in the mail, so to speak. But um, with regard to like keeping on top of my everyday tasks, you know, with family, with business, with other stuff I'm doing is I kind of created an AI agent with anthropic, with another company, through a to-do list to just make sure that like things aren't slipping through the cracks for me. ADHD brain, I need kind of to tee things up. So just the little things that are weekly for me or that they're ad hoc versus like managing it through text messages with my wife or email with somebody else. I have it all go through this, and it's uh it's been helpful for me to ping me with reminders, to ping me with meeting prep. Um, say, hey, you're meeting with so and so. By the way, the last two emails, here's what it said. So it's been super helpful for me, and I know I'm just scratching the surface. Like I am in the top of the first inning. There's no outs. Like I have so much left to learn, but it's been it's been Really cool so far. So good.

SPEAKER_00 22:01

Yeah. And for me, there's two things that I've done that have really helped me out recently. Number one, being a content creator since 2021 is when I went full time on that. A lot of my income comes from working with other brands. And a lot of times that will take a lot of negotiation and a lot of time over email. So over the last five, six years, I've had a lot of emails come in and for whatever reason, you know, didn't get the brand deal. And so I had Claude go through and scrape my email from the last five, six years and pull out all of those conversations of deals that never got closed and pull out the contact for each of those. And hundreds of brands and contacts came up, and I was able to have Claude write a new email to re-engage in those conversations. And I mean, all of that took less than 30 minutes, and I had almost 300 new contacts that I could reach out to about starting those negotiations. So that was something that was really powerful. And I just did that. So I'll keep everybody updated on the results and how that plays out. And then I mean that's something that like a realtor could do.

SPEAKER_01 23:06

You know, you think about all the contacts you have. Like again, and this stuff is daunting. Like, do you have to go through your email the last six years and go through one by one? Okay, who is it? Okay, what should I write to them? Wait, what's their story again? Like the fact that Clod can do this in less than an hour is crazy.

SPEAKER_00 23:22

Aaron Powell Really, anybody that sells anything of any kind, this is extremely useful for, right? Going through back through all of the leads of your company, of your business of the last five years, six years, ten years, and pulling out all those closed cold leads that you never had a chance to close for whatever reason, re-engaging in conversations. I mean, the chances of getting at least one or two, three, ten, a hundred people to then sell them whatever product you're selling is very high. And it's something that takes very little amount of time to do.

Description And Better Prompting

SPEAKER_02 23:55

Little amount of time now, what could have been months of skilled human work until recently. And what a perfect segue, actually, to the second of our two Ds, because the way that we interact with these LLMs is the extent to which they are useful to us. The three of us could all have the same exact thing pulled up on the screen, and the three different prompts, we're going to get three drastically different outputs. And so for anyone who will in a moment be debuting a three-tiered way to kind of think about AI fluency, to bucket yourself, this is purely an academic exercise. And we'll be looking at some skills, which are basically a preset group of prompts that live in Claude and make it really easy for you to activate them. When we each put in those different prompts, we would get very, very different things. And so that comes back around to understanding what you really want AI to do for you. And that is why there's this incredible virtuous loop cycle between discovery, experimentation, redefinition, and play. Your willingness to get in there and just mess with it and just get it wrong and just see what it can do and be ridiculous. Half of the stuff I have learned is me just throwing a crazy thing in and seeing what it could do, 10% of which is stuff that I never would have imagined. And then I immediately adapt that and pull it into my workflow. Truly, truly great. To that end, the second D is description. And this has a lot to do with how we're talking to AI. So description. This is the skill of actually telling AI what you really want. And that's hard. I think it's it's hard for us to tell even our spouses what we really want sometimes. And so this sounds simple, but it's really not. This is probably the one that we all need to work on the most. But the better you describe your context, the better you get at giving the format, the audience that you have intended for whatever you're creating, to Jordan's point, especially if you're in a selling environment, hint we're all selling something. When we get to the career piece here in a second, you'll definitely talk about that. All of this, the sharper it is, the better your output's going to be. It's just like giving directions to someone. Go to the store just gets you lost. It doesn't mean anything.

SPEAKER_00 25:59

And if our wives asked us to do that, we would come back with all of the wrong items.

SPEAKER_02 26:05

Honey, you won't believe what was on sale at Home Depot. And just like that, AI will waste a lot of tokens if all you do is say go to the store. It'll come back with some stuff you don't didn't need. But again, the contrast is hey, take a left on main, park in the parking lot, grab the organic milk on the third shelf. If you can get to a level of description that allows for you to get exactly what you're asking for, you will get better outputs. And, and this is the wild part, I believe that we are learning how to speak a different language now. The parts of our brains that are engaged when we're learning new languages, music, new ways to express what we're trying to say, that's exactly what's activating when we're learning how to prompt more intelligently and more effectively. Because you're exploring, you're playing, you're recursively getting things wrong. And if you watch an AI tool learn, if you watch a neural network forming, it is identical to the way that human synapses form, which is no coincidence. We designed it to mirror our own functionality. So, description. This is the one that Anthropic actually describes as being the most underdeveloped skill for all of us. Most people type a few words in and then blame AI when the output is generic or has nothing to do with what they were imagining in their heads, but the output is only as good as the input. And description is the second of those two D's that make that real. And John, is this is what people have called prompt engineering or prompting. Yep. And that's all that there's people whose jobs, there are very well-compensated jobs out there right now that have the words prompt and engineering in them, usually right next to each other. And these are the kinds of jobs that you're starting to see. A real trend right now in the AI space is people getting weird jobs. These weird jobs that usually pay about half a million dollars a year are largely just a bunch of fancy words that that person put and connected to. I use AI to make people better. That's it. Prompt engineering is certainly one of those things where you're usually not just engineering prompts for yourself. You're usually empowering a group of people, especially a large one, uh, to do so.

SPEAKER_01 28:10

So, John, this is uh this is a time where you can tell me if I'm being an idiot or if it's a brilliant move. So, description prompting, I think I struggle with. So I will go into one AI, I'll go into Grok and say, hey, here's what's going on. Write me a Claude co-work prompt or Claude chat prompt for this. And then I just copy and paste it in the Claude. Is that, I mean, I'm having AI write my prompt or my description for me. For one, is there a what's the better way to do that? Is what I'm really asking. I love that approach, Jeff.

SPEAKER_02 28:40

First off, I would add a caveat, which you may already be doing, but you might not have thought this was valuable to mention, but it's probably the most valuable part of that. I do that too. I'll bounce and bounce back and forth between LLMs, especially if I know that, for instance, I believe open AI is still the best at truly abstract thinking. So if I'm driving somewhere for two hours, I'm probably going to turn on open AI, go to advanced voice mode, and talk to what is functionally an omnipotent being for two hours to help me unlimit my mind. Probably come up with some ideas, but but really I'm just trying to get it to help me stop thinking about finite world and start realizing how infinite things truly are now. And then I'll take whatever the output of that conversation is and I'll say, cool, package that up and put it in Claude, because I actually want to do something with some of this. And then I'll give it some prompts to do that functionally. But to go back to your point, Jeff, the one thing I would add to that is as you're doing cross-prompting, so from one LLM to another, I would also add, and then tell me why you did that in the way that you did it. Yes, yes. Okay, that's where I learned. Sweet. Because I would speculate, grok is like painfully honest. And if you grok yourself and ask it to tell you about yourself, you should be ready to hear some things, especially if you have any kind of presence that you haven't scrubbed. Uh, it'll tell you about the stuff you posted in spring break in 04. So cautious, cautious, y'all. Love it. But it'll tell you, likely, well, well, you said you wanted to go from grok to Claude. Grok really good at pulling current trends and identifying market behaviors, whereas anthropic is better at actually compounding using that reasoning to create things like digital artifacts. So that's why I phrased it this way. And then it'll probably follow up with is that right? Could I have done that better? And then when you respond to that, you're actually now engaging in a conversation where you're both learning. And there's nothing more powerful to do that.

SPEAKER_00 30:26

And one thing to add there, John, something I have loved doing lately is telling Claude or whichever LLM to ask me questions after as well. So I'll give it a prompt and say, here's what I'm thinking, here's what I want to do. Now ask me specific questions so that you know exactly, so you have a better idea of all of this. And then it will go through and say, Cool, I saw that you wanted to go for a steakhouse on Friday night, but there aren't any within 25 minutes. So are you willing to drive more than 30 minutes? Yes or no? Just stuff like that where it's asking me questions to get a better idea of all that. And I think that's been really powerful in my prompting lately.

Discernment And Checking AI Work

SPEAKER_02 31:05

You know what I love about that, Jordan, is you're taking advantage of a very rapidly widening context window that all LLMs are experiencing. For anyone who would fall into the first of our three AI fluency tiers, which would be AI Curious, which is a broad church. Context window, it basically means the breadth of consideration that a given statement or system will take into account when giving you advice or answering a question, or in some cases, even agentically giving you insights. So to Jordan's point, if you have a small context window and you were to ask something like, What could I be doing better here? What do you see as being improvable here in my world? If it can only see your email and it can only see your calendar, it may create an insight like that, like what Jordan mentioned. If it has a context window to zoom out far enough to have access to the internet and possibly even seeing what other user trends look like, it will probably say, Hey, so it looks like there's this recently opened steakhouse that's actually five more miles away, but it was just outside of your search radius, but it's like the hottest restaurant in town. How about that instead? And look at the quality difference there, right? It went from being just a simple response vending machine answer to what your best friend might ask you if you mentioned to them that you were going to that restaurant. And that's a step change. The next step beyond that, what comes next, this is all subjective, like people are going to start naming stuff, all kinds of different things here soon, is super intelligence. Superintelligence is anticipatory. It's able to take into account a massive context window, arguably omnipotence, to be able to take into account everything that a human could know and does know to give you the best possible recommendation based upon a massive context window. And uh we'll close down that curkling signature tech talk real quick. Jump over to the third D, which is discernment. So this is all about evaluation. So discernment is the one that I found I needed to work on the most. Now, AI will give you a confident answer every single time, but confident doesn't mean correct. I think we have plenty of examples of people who are confident, but not correct. Insert whichever politician or actor or influencer anyone's thinking about right now. There's a great example. Discernment is the skill of evaluating what comes back out of AI. So checking the facts, questioning the logic, knowing when the output looks right, but isn't. Anthropic, again, going back to that content that they they offer in their certified architect coursework, they use describe, evaluate, refine. So you don't just take the first output, you think about it, you push back, you iterate. This is exactly Jeff and Jordan, what you guys have both spoken to in the ways that you're cross-prompting, double prompting, asking questions about why it did what it did. And this is one that I would really make, I would recommend that everybody hearing this, make sure that this is where you're focusing, not necessarily first, but but certainly quickly. Because as you get good at AI, especially when it starts making pretty stuff, you could fall into a nasty bear trap of saying, well, it looks great. So I'm sure it's perfect. I'm sure that this report is ready to send to my board. And then it hallucinated one number, and even the best LLMs are still doing these things here and there. And that could be something that costs you a whole lot more than it ever could have made you. And it could cost you your job or worse, you know. So the discernment, the ability to still do your job, to still be a human with agency and command over the logic that got you where you are and will get you where you're going, is even more valuable now than ever, especially as we enter into this adolescence of AI fluency where so many people are getting into it, but few people are actually taking the time to use it well. It's a new tool, it's powerful, but it's like everybody's got a jackhammer and they're just like blowing up concrete everywhere instead of actually stopping to say, what should we build with this? So don't be the one that blows up the concrete everywhere.

SPEAKER_01 35:01

Well, and so John, I mean, with this one, I mean, I I've obviously we've all been a victim of hallucinations. Like, what is your process to kind of trust you'd verify?

SPEAKER_02 35:10

I'm a huge fan of working in projects. If I know I'm going to be working in a given space or category for more than perhaps one day, like let's say this is um part of a development project that I'm focusing on. I will create a project in Claude, and that's a that's an actual specific workflow. You go into Claude on the left in chats, you'll see projects. You'll also see this in cowork. You can make them collaborative, but you create a project and then you add instructions, you can add files, and it really gives you a specific set of definitions of your context window, so then you can rely upon that. I build a filter essentially so that I know that I'm telling it that I want it to double check its work every time. Not triple check. You're gonna blow tokens like crazy if it's doing the same work three times every time. So I'd be cautious with that. But I definitely do that because I want it to be more trustworthy. And then I'll usually break twice a day and re-examine all my work and to see if I've come off course anywhere. And I would say 25% of the time I find something that was either wrong or that needs to be better, that was misunderstood, probably because I didn't describe it correctly. And going back, I can usually see that root cause and I'll ask, like, hey, at what point did this become that? And it can usually come back and say, oh, you you essentially told me that. And I'll look and say, well, I can see how you would have heard that actually. Not what I meant at all, but I can see that. So two specific recommendations there, Jeff. One, if it's important, probably work within a project, make sure you're including files that you can include, adding instructions that seem right to put the level of diligence that you want to see consistently. Two is routinely building a routine around going in and actually doing that due diligence and saying, hey, what's what is the out quality of my output?

SPEAKER_00 36:57

So, John, along those lines, one thing I've noticed with Claude is that it's getting better and better at design. I use Claude to design client documents almost every single day, right? And they it can create a document so quickly that looks so professional and you know, beforehand would take hours and hours. And so over the last couple of weeks, I've wanted to take that a step further, and I've actually designed websites on Claude. One thing I've noticed when we're talking about discernment is even if Claude builds you a beautiful website or something that looks really, really nice, again, and this can be done in minutes, which is truly mind-blowing. A lot of times the code isn't exactly where it needs to be to actually publish that website or publish an artifact that's 100% ready to go. So where is that gap that you've been noticing? And what's something, what's a way that you know, someone like me or anybody listening can do to make sure that that is ready to go?

SPEAKER_02 37:54

Two thoughts come to mind right away. Um, because I had a similar experience even building the website for our podcast here, where the first two or three iterations were directionally wonderful, but had some pretty significantly broken links. The two ways I would recommend folks deal with that, and this would be another way to like look at this and frame this question would be hey, if I want to go from zero to a website that's selling stuff, my services, my products, my awareness of me as an individual, my brand, if I want to go from zero to that, what's the what's the fastest possible line between here and there with AI as the connector? I would recommend building into your workflow, especially if you want something particularly pretty. If you want something pretty basic and you're not trying to do a bunch of animations and you're not trying to do a ton of really complicated forms or connections to other tools outside, Claude Code can probably do Claude itself, even without the code enhancement, could probably do everything you need. You add design now, and it's going to go way above and beyond. But if you want to do something really beautiful, like these animation sites where you're scrolling and you've got something in the background that's following the cursor, like a koi fish that's swimming beautifully as you're scrolling, right? Like there's some stunning stuff people are making now. I would recommend utilizing cursor or lovable. Those would be tools that I would recommend in addition to Claude. And some of these are not super cheap, but let's say you're paying$100 a month for this stuff, even just a year ago, certainly two years ago, and five years ago, it's a whole different ballgame, you would be paying thousands of dollars on Fiverr. You would be paying tens, if not hundreds of thousands of dollars for this level of support if you were working with domestic engineers. So it's it's a is an unprecedented amount of quality output. And again, to answer your question, Jordan, to dial that back down, if you're trying to go from prompt to beautiful website as quickly as possible, the juice that you get for the squeeze of integrating cursor and or lovable into your workflow, workstream is uh 10 out of 10. Strongly recommend both. Uh, and it's not just because lovable is a uh a Boston company, not just up the street here. No, it's not just because of that either.

SPEAKER_00 40:01

I've actually used Lovable and they're fantastic. Again, not sponsored, but hey, we're open to it. Yet.

Diligence Privacy And Security Risks

SPEAKER_02 40:06

You never know by the time this airs, that could be that could be different. But that's right. The last D, and then let's just jump right over into uh the skills that we are so excited to share, which by the way, you will find all of our episodes will have an accompanying skill in quad that we will debut on our GitHub. So you can come on here. Whatever we're talking about, there will be a skill that you can pull in that will be able to teach you and actually execute some version of whatever you're hearing about on this show. And today we're actually going to drop three, one for each level of fluency that we'll be talking about. But before we get to that, let's talk about the fourth D. I thought there was only three D's on this podcast, John. No, there's actually four, and this one is diligence. So this is the ethical one, actually. So this is it, this is the one that's about safety. So diligence. This is where people are starting to get themselves caught up a little bit, particularly in large corporations. So you must ask yourself, and you've probably felt this and maybe even stopped yourself. You're you're about to drag and drop in like 2026 performance reviews. You're about to drop it into Chat GPT, and then you stop and ask yourself, who can see this? And then you drag it back off and you stop, and then you ask IT and they say, yeah, even though we have a ZDP agreement, we have like privacy with them, you still shouldn't add certain levels of privacy in terms of documents to an LLM. I'm not saying that's necessarily true of ChatGPT or of your company or your own practices, but that moment of stopping and saying, is this something that I should in a courtroom have to be able to say that I put online, you should be mindful of that. And this is not just true at work. This is true at home. Like what photos are you putting into whatever website you're going with? What are you saying on some LLM that you have no control over that's now going to be associated with you? Just be mindful. And I think the the last this diligence piece, it certainly starts at work, but it certainly extends out all the way out to you helping to make sure that your mom, your dad, your grandparents, everybody in your world is a little bit less willing or a little bit less, uh, I should say, likely to fall for bad actions, even if they're their own. So this is like the definitely the dad skill of the four, knowing that just because you can doesn't mean you should. Sometimes people are gonna have a teachable moment where they're gonna do something silly and kind of get burned on it. But yeah, that's the that's the fourth one. That's diligence. And uh yeah, any questions before we move into our our skills? That's like the downer.

SPEAKER_01 42:37

I'm going, gosh, what did I put in there that I shouldn't have?

SPEAKER_00 42:40

I'm like, oh this has to be the most overlooked of them all because I'm feeling the same way. I'm like, man, I have uploaded a lot of photos. I've uploaded stuff where now I'm gonna go back and review everything.

SPEAKER_01 42:53

I literally just went in and deleted a thread and I know that's not gonna do anything. Oh no, it's uh yeah, it could.

SPEAKER_02 42:59

It could. But yeah, it is definitely the so much about AI is all gas, no breaks, right? Like everyone's just going fast and building and deploying, which is why Anthropic's recent move. And calling it.

SPEAKER_00 43:11

By the way, John, that's my life motto. All gas, no breaks.

SPEAKER_02 43:14

Jordan, all gas, no breaks, Egbert. Hey! Well, the uh AI world is certainly going full YOLO, full send, like you, Jordan, which is which is why it was so atypical and almost this crazy departure and momentum shift when recently Anthropic they full they sort of dropped this new model called Mythos. And by now, if you're in the AI space, you've heard so much about this that you're probably tired of it. Largely speaking, what Anthropic did is they dropped this model called Mythos. It's so good at code that it essentially found decades old vulnerabilities in massively used systems. And so instantly they realized that this thing's so good that it could essentially obsolete a lot of the security that we have. And it could pose a global threat to humanity, not just like to some businesses. Humanity. Or at least that's what they're saying, right? And I am on the side of this being a more real than than just kind of a hype train moment, though there's plenty of people that would would maybe have that position. But I've seen what some of these systems can do. And if it can do even half of what they're saying it can do, it is something truly dangerous. Not quite Ultron, right? But but certainly way closer to Ultron than I think any of us want. So what they've done is they've done this thing called Project Glass Wing. They're working with a number of very large corporations to make sure that they have time with this tool so they can harden their own systems so that hopefully most of enterprise doesn't get broken when some version of this thing launches. We can only hope that OpenAI and their Model 5.5, also called SPUD, which could be dropping any second now. It could literally be dropping as we're filming this episode. We can only hope that they have a similar level of diligence in their way that they drop it. But yeah, who knows? Maybe all of our passwords are already obsolete and it doesn't matter.

SPEAKER_01 44:54

I mean, it's so funny because like going into this and become so excited and I'm like, oh my gosh. Like I'm just thinking about cybersecurity. What do I need to do? Like granted, I'm a tiny little fish, but I can only imagine what it must be like for a multi-billion or trillion dollar company to protect themselves or financial institutions. Like it's just like we are on the frontier, but we really don't know what's going to happen. Uh you just have to listen to John and hopefully he tells me the right things to do.

SPEAKER_00 45:20

Well, we went through the scary stuff. So now we should get into the fun stuff.

Three Levels Of AI Fluency

Skill One Rocket Fuel Prompts

SPEAKER_02 45:24

Seriously. Come on. Yeah, come on. Yeah, no, no more doom and gloom. Okay, so let's jump in. We've got three levels of fluency to talk about and skills to accelerate you no matter where you are. And you should probably use all three of these things. I would strongly recommend it. Before I bring this up on screen, the three levels we talked about them just briefly earlier. Number one, AI Curious. You are excited about AI, or you're just starting to be excited about AI. There's also level zero, which is someone saying, calling AI A1, which is super cool that government officials are still doing that. Remember who's doing that when you're voting, folks. Just saying out loud. But tier one is AI Curious. Tier two is AI power user. So if you're AI curious, you're excited to get in here, you're messing around, you probably have a subscription to one or more LLMs, you probably are at work hearing about, maybe even going to trainings on AI. If you are a power user, you're probably the person in the office who people are going to to ask, like, well, how do I even use it? What is a prompt? Like, what am I supposed to do with this thing? How did you crush it in that meeting so hard, uh, Jim? And that would be a good thing to be. So if you're already tier one, you're probably about 20 to 30 hours of messing around away from becoming tier two, just to say that really clearly. And if you're going to go down the Claude certified architect route or something like it, by the way, there's Gemini, Stanford, Harvard. There's tons of free resources. An unprecedented amount of free knowledge is out there right now for you to learn these things in ways that are actually certified, by the way. Like you get a certificate that you get to put on LinkedIn and say, I'm, I did this thing. I'm learning this stuff. Level three is what we're sort of jokingly calling the AI superpower dad. And that's the title of our episode here. It's also the title of the third tier, the top tier of the three. This is what represents some 5% approximately of users. This is the person who is probably doing a lot of stuff like what Jeff was describing, building agents that are actively looking at workflows, looking at your calendar, making recommendations, optimizing things, possibly even making financial decisions. You might be using something like OpenClaw to be automating investments. You're really in this thing. You are on the wave. You're surfing it. Um, and every single person here listening to this could easily be on this wave. Maybe easily is a bit of an oversell, but you could be on this wave. It's just a matter of playing. So let's take a look at the first of those three skills. This one is what we are calling uh AI superfluency rocket fuel. So this is the skill that we call AI fluency rocket fuel. Let's take a look here. So if you go to the GitHub and you load it up and drop this skill into your own Claude instance and use whichever wake word you choose to use for it, in this case, you could literally just say AI Fluency Rocket Fuel. It'll load this up and it'll take you through all of this. First, that's if you just download only this HTML artifact, you'll be able to do everything you see on screen here, including just linking directly to the GitHub. But there's a prompt transformer and a prompt library here. As we scroll down, you'll see it takes you through the 4D connection. We're really focusing on the two that are identified as the most underpowered in the market right now, which is we heard were delegation and description. These are the two of the D's, of course, from Anthropics Framework. This is all about mastering these two D's. So first, let's go to the prompt transformer. This is actually going to take you to a before and after rocket fuel. So you could type a vague prompt in here, like plan a date night. Let's see. And it's going to transform this prompt and it's going to say, nah, nah, nah, nah, nah, nah, nah. Here's how to make this great. And it will actually show you, and you can put whatever you want in here, and it will come back to you with a better version. And it'll give you a good example.

SPEAKER_00 49:03

This is cool, by the way.

SPEAKER_02 49:05

Well, uh, this is what I wish I had a year ago, maybe two years ago, because so much of this was guess and check and fumble around in the dark and wish that you knew this. But now, especially with modern frameworks, this is what will get you to a flawless output in one or two prompts.

SPEAKER_01 49:21

And looking like a hell of a husband in the in the in the process. I mean, like, this is like half the challenge with us as dads is time and energy. We want to take our wives out on dates, we want to have fun and make it special. But it's like, oh my gosh, what am I gonna do? This is awesome. Like, this is, and again, it's not obviously not just for a date with our wives, it's for anything.

SPEAKER_00 49:44

Like, any like let's do one more because this is exciting.

SPEAKER_02 49:47

Yeah, give me something else. Well, check this out. One more quick view of functionality. Let's say that you're in here and let's say you are actually like level tier zero, and you're like, this, even this seems like a lot. We've added this tool in here to make this super easy for you, no matter what. So you can actually hit the copy improved prompt. It is now on your clipboard, and you can paste this directly into your LLM of choice. I recommend Claude for this so that you can come in here and get as crazy as you want with this, right?

SPEAKER_00 50:17

And by the way, John, for for anybody on level zero, when we say input this into Claude, all we're saying is go to Claude.com, correct? Is that their website? Claude.ai. Claude.ai. Okay. Sorry. And copy and paste this prompt onto the very main screen there. That's all it is. Super simple. Bingo.

SPEAKER_02 50:37

You nailed it, man. It makes me think that we should launch a course called Dad Claude. Dario, again, please don't copyright that before we can.

SPEAKER_01 50:45

I mean, like, yeah, because I'm going, gosh, when do I use cowork? When do I use chat? We need we need a volume two and a volume three of this, John. Tell anyone. Hey, if we if the listeners love it, we can even do lives.

SPEAKER_02 50:56

We can even do live sessions together. That'd be sweet. Listeners, if you hear this and that's something that resonates with you, please drop it into Discord, drop it into Insta, TikTok, wherever you're seeing this. We would love to hear that feedback. And if this is something that would be helpful to you, we can do whatever we want with this. We could actually build an academy, or we could just do a couple of lives and party and have some fun. Here's another one. So help me with my resume. This is one of the more first thoughts that people have with AI. Great. So review my resume for you know, insert your resume. This is just obviously randomizing. Review my resume for insert your role here, for insert your kind of company there. I'm currently this, use my biggest strength, rewrite my top three bullet points. You can go way beyond this. This is like the very, very surface. But it's also going from like homo habiless here to Homo sapien here, and then maybe homo ultra habilist with the next version, which is where you actually get it to scrape the internet and figure out what the actual ATS AI is looking for and so on and so forth. But Jordan, you got all those definitions? Yeah. Well, here's some other suggestions where it like just is giving you ideas on where to start. Down here, this is where we've compiled some of what the uh what specifically Anthropic has volunteered as some of the most common themes of what people are asking about. And so we've got a suggestion. So here's for career, weekly status email, interview prep, a LinkedIn post, salary negotiation. I speculate that if you get good with these prompts, this could well be worth 10 figures for your family over the course of the next couple of generations. No big deal. It's just like$10 million. I'm sure nobody on this podcast has any familiarity with that kind of money. Family, date night plan, like we saw earlier, summer camp research. Here we go. By the way, any of these, you can go and again copy directly to your clipboard and then paste right onto Claude or ChatGPT or Grok or anything else, weekly meal plan, so on and so forth. So at the risk of being uh overly didactic, you can simply just go through here and and do these. I would like to highlight growth as a really cool thing where this one is specifically built to help you as a dad think about how you can accelerate yourself in ways that are meaningful to you personally. And this is one step towards building an agent to coach for yourself, which is a real good idea. Real, real, real good idea. Okay, delegation. This is the delegation diagnostic. So if you go through here, it's actually going to ask you a series of questions. You're gonna make some elections here. What eats most of your time outside of meetings? The goal here is to help you learn in the ways that's most meaningful to you so that you can get more excited about it, so you can use it more, so you can get more excited about it and use it more and ultimately become the master of this stuff. Let's just say email and message management eats my world up. Okay, uh, career development, I'm coasting, family planning, we wing everything. Let's say I feel the most behind in financial planning. What recurring tasks do you dread the most? Weekly status updates, meal planning, meal planning for sure. If you had an extra five hours this week, what would you do? Spend more quality time with my kids? Definitely that. And how comfortable are you with AI right now? Let's say, so basically, this is zero, one, two, and three. I'm gonna go with one. I've tried Jat GPT or Cloud a few times. Great. Cool. Well, here's your map. So now it's going to give you prompts that you can use for each one of the things that you've now identified that will help you start to walk through all of this. And this is based upon what you've indicated as being important to you. And lastly, and this is this is super, super cool. I encourage anyone listening to this to do this, even if you are in the AI superpower dad category already. This is a Claude system prompt. So what you do is you copy this into Claude as custom instructions. And you can literally ask it, what are custom instructions and where do they go? It'll guide you through it. It turns your Claude into a rocket fuel creator engine itself. So if you read through this, it essentially says you're a prompt engineering coach built by Dad Squad Cast for Ambitious Dads, who are just getting started with AI. This is just to give it a little bit more of a fun sort of energy in its voice to you. This is about teaching you how to use the framework. This is about how the tool behaves. This is about how it gives feedback. Here's this goes back, Jeff, to the way that you were asking, like, how do you do this in the in the area of discernment? This would be an example of always ask yourself before responding, do I have enough context to give a genuinely useful answer? Counterprompt. Always ask yourself is the feedback or the information I'm about to provide something that could be confidential? Just to like kind of bounce back and forth. Here's its personality, delegation coaching, and formatting. And there it is. Oh, what's this? Easy one-click follows for all of our socials.

unknown 55:34

Who put that on there?

SPEAKER_02 55:35

That's weird. I don't know. It's any questions on the AI fluency rocket fuel skill for our tier one dads.

SPEAKER_01 55:42

I want to go on it right now. You know, I mean, it's I mean, I obviously I know we're doing this, but this is like, and I I I thought I was maybe getting into level two, but now I realize like I am certainly AI curious. Like I am so early. And uh this this is just super helpful, even helpful for me. And it allows me to tell people in my network, because again, I live in Silicon Valley, so people think that I naturally know tech more than than other people. I don't. This will allow me to act like I do and point them in the right direction to get them started. But I think um I just think this is super helpful for people just to get their feet wet. And once you get your feet wet, you can slowly take yourself into the deeper end of the pool.

SPEAKER_00 56:20

And that's what it's all about, right? Is getting started. Because I know there is so much fear with AI, and you know, it's so complex and there's so much to it. And so this is the easy, simple framework, whether you're tier zero, one, or two, said here's the prompts. Go ahead and put these in and get started. And once you're started, you'll naturally think of other things that you want to automate in your life. And so, yeah, super powerful, super excited to use this myself. Awesome.

Skill Two Career Strategy Visualizer

SPEAKER_02 56:46

Well, then let's move on to our tier two dads, the power user. Let's talk about AI career strategy visualization. So, this one right here, this is probably the kind of thematically the thing I get asked about the most in terms of like helping people, friends, family members. This comes up in a lot of ways. One is like, how do I make sure that my career is less likely going to get obliterated because of AI? Like, what can I do right now to get ahead of that? This is actually our best shot at helping you think about your career to consider pivots and to do so in a way that is meaningful to you. So let's dive in here. This this could arguably be one of the most important things that you could be doing with AI. I'm not saying with this tool per se, we aspire to give you something that starts curiosity. That's it. And so to that end, let's check it out. So it's going to ask you about your career. And so you enter, and these are open text inputs, right? Head of something. Industry, choose your industry. You can also do other, of course, but let's just say some good old-fashioned software. And here's the top five tasks that fill your day. So as this thing progresses and you set your AT month career goal, you can also then add your level of AI comfort level. 10, of course, being what we would call tier three here, one being what exactly is AI? And then this is where things get really interesting. So risk tolerance. Let's say that you are feeling like, hey, I'm an accountant and I somehow still have a job. I am very afraid. I know that so much of what I'm doing is automated. I just put two prompts in and it produced a report that is better than what I could have done. I am mortified. You might want to go with, hey, I have an aggressive risk tolerance right now because I feel like the worst risk, the biggest risk, is not moving. And I volunteer everybody hearing this, and this is not hopefully like fear-mongering, but we should all be asking ourselves, how much faster could I be moving here? Even if you're part of this industry, and I am. And every single day I'm asking myself, how can I get more into this current and paddle with it? So let's say you're aggressive. Now, it's going to ask you how many hours per week for development do you want to spend? We maxed it out at 15 because let's just assume most folks are working full time and our dads. This is probably about the top end of realism, but you can also modify this in your own prompts. Now, as we go, this will do this inside of Claude for you. This does not have a Claude API connection to it, but it will actually generate your full career strategy for you. And all you have to do is have your own instance of Claude loaded up. And when you add this skill into your Claude, which is right here on our GitHub, it will do this for you with graphics as well. We don't have that API connected again right now, which is probably for the best because we're already filming our longest episode of all time. But this would be the number one thing we would do in a live setting if we were to actually have this.

SPEAKER_00 59:41

Could you give a 60-second overview of what that is gonna, what kind of report that will pop out and what that will look like? You can give the example of an accountant.

SPEAKER_02 59:49

Yeah, so it's gonna give you a bunch of recommended prompts to start playing with, depending upon where you said your fluency level is. Let's say you said it was 10 and you have a high risk tolerance. It may very well recommend that you go and get a wide set of certifications to pivot as quickly as humanly possible into automated or agentic versions of what you're doing now or possibly what you want to do exclusively. Let's say that you grew up in an agriculture setting and now you are working as an accountant, right? And you mention that in here and in conversation as you progress with the tool. It may ask you more questions about so you mentioned agriculture background. Tell me more about that. And you may very well discover that you see that there's a tremendous number of people who you might call farmers or ranchers who really need help with AI, but are at like negative five on this thing. And so it may guide you to take your skill in accounting to build baseline business plan tools for this group at an unbelievably high ROI for you and for them, right? To be that interface. That's one example of what I got this thing to spit out for me when I was doing test runs. And that's pretty exciting. And it also is probably pretty surprising. The good news is that if it does recommend something like that, it'll search the web and do this iteratively. Here's one set of steps that you could use to get there.

SPEAKER_00 1:01:06

So the more detail that you can give it about yourself, the better. Using discernment, of course.

Skill Three Meeting Mastermind

SPEAKER_02 1:01:11

Okay, let's jump to our tier three, our AI supercharged dads. This one is a fun one. This is the mastermind every meeting. So this is a very similar tool to something that I use myself right now. This is all about walking into every meeting as the most prepared person in the room. This is the answer to how to show up and be consistently seen as that person, which right now is about as valuable as it gets. So we're gonna go down here, we're gonna build a briefing. Six inputs. The more context you give, the sharper the briefing, to Jordan's point again. All four D's for Anthropics 4D framework are gonna work together here. So we're gonna talk about the meeting context. Let's say it's a one-to-one with your boss. Let's do something a little bit more public. Let's say it's team stand-up, all hands. Purpose. So just for fun, we're gonna add in some context here. You add in the names of the people that are here, their roles, and then hear your goal. What does winning look like here? How do you really want to show up? And then context. This is where you really get to add more and more in. Stakes, low, medium, high. This we added to just give a little bit more nuance so then you can play with it and get it to spit out something perhaps more detailed than it would have been otherwise. Like if it's low stakes, it's not gonna do that much research. It's gonna give you some pointers, probably some key bullet points to hit on, maybe some graphs, medium. It's gonna go a little bit deeper. High stakes, it's gonna go through and research every single person that you mentioned. It's gonna go through and research every topic that you put. It's gonna give you suggested versions of everything that you could be saying. And it will actually likely ask you if you want to walk through and coach some of those things individually. This tool right here, though, I will say, is probably the most primordial version of what any one person would want to build. Think of this as a template that gets you started on seeing what something like this could do. One immediate next step, especially if you are, in fact, a tier three user, would be building an agent that does these things for you automatically, using a connector to your Outlook, to your Google Calendar to automatically do research and brief you on the conversations that you have coming up. To Jeff's point, your tools should be giving you deep context to take mental load off of you when it comes to preparing. It should tell you you're meeting with Erica, you're meeting with Jenny, you're meeting with Bill, and you're meeting with Javier today. Here are the last two conversations you had with each of these people. Here are overdue projects that you have with them either way. And then here are some likely political problems that could come up here given the conversations that you've been part of. Also, if these are external clients, here's some research on them in the market, here's what their companies are doing, here's what they've said in the news lately. So you can go in ready for anything, knowing so much more than anybody else in the room until everybody else builds one of these, in which case then you have to make a whole new thing. This is uh one big step toward you having a Jarvis right here in your ear, telling you exactly what the perfect thing is to know at any given time.

SPEAKER_00 1:03:53

And I think the reason this is so powerful is because with AI moving so fast, the one thing that keeps us ahead of everybody else is knowledge and being prepared. And so this does both, right? Gets you to be the most prepared person in the room, which makes you valuable. And so that's that's what I love about this, John. Nice.

Hot Takes And Key Next Steps

SPEAKER_02 1:04:13

Well, those are our three skills that we have to share. Again, these are all going to be available for you on GitHub, as with every other episode that we post. And that's it. We're in our last couple minutes here. So how about we jump into everybody's favorite, a little hot dad's hot takes?

SPEAKER_01 1:04:34

Who wants to be our first hot take of the day? We gotta have the hottest.

SPEAKER_00 1:04:38

We gotta have Jojo lead us away today. All right, here's my hot take. AI made me a better dad this week. And here's how. I picked up my son from school this week, my five-year-old, and he said, Daddy, I really want to go to Chuck E. Cheese today. And I was like, We've been to Chuck E. Cheese a lot lately, and quite frankly, I'm getting tired of going to Chuck E. Cheese. And so I want something more fun to do, right? And so I opened up Claude, and I I've used a lot of bad prompts in my early stage of AI. And so I used a better prompt, and to the point of what we've been talking about today, I said, I want something fun to do with my five-year-old in this area of Las Vegas where we live. I want it to be unique, I want it to be something that we have never done before based on what you know about us, something that is active, where we can get energy out, and something that, you know, will create memories for the both of us. And it recommended this place about 20 minutes away that I had never heard of before called WALL-Y Wombats. And it was a children's paradise, ball pits, interactive toys, jungle gym, obstacle course. And I showed my five-year-old a video of this, and he just started saying, Wally wombats, wally wombats. I was like, This looks like the coolest place ever. And so we went, and for 19 bucks for all you can play, we spent About four and a half hours there, and he was just on cloud nine the entire time. And so I don't know if I would have found that place otherwise. So that was my AI dad win of the week. Love it. So good. So good.

SPEAKER_02 1:06:14

Wall-E wombatts. Wall-E wom.

SPEAKER_01 1:06:16

It's our new battle cryer for the pod. Uh hot dad, hot take. It's also a dad joke, too, which is AI is a-okay. I know AI can scare people and it can be freaky, but I think John spelled it out completely take your time and it can really help you. Even in the basic thing of planning a date with your wife, even if that's what you use it for, guess what? It's gonna make your life better, but it's only scratching the surface. So dive into it, teach yourself this stuff and teach it over time. You don't need to learn it all in one week or one day. Baby steps. So well said.

SPEAKER_02 1:06:48

My hot take is the most important thing, and I will go so far as to say the only thing that will matter about what you're doing professional five years from now is how you're demonstrating your learning agility and willingness to get in the AI current and start swimming, even if you feel like you're gonna drown. One way to do that, the way I'm doing that, and the way I recommend that most people consider doing that right now, is every single quarter to prioritize getting two to three certifications from known entities. Right now, I'm about to finish my cloud certified architect. And next I'm going to do the one that's offered by Stanford, that basically is like an all-star list of AI creators out there. After that, it'll be Google. After that, it'll be Open AI and so on and so forth. So that five years from now, no matter what happens with the industry, I can show receipts for the fact that I've been in this thing and learning from the people who are creating it. And that might be the only way that an increasingly agentic internet can see you as somebody who's actually relevant, no matter what you're doing, which by the way is hopefully scary and inspiring because whether you are listening to this and you're about to go clean the bathrooms at a sports club, or you're listening to this and you're about to go lead a board meeting and blow away your board as the CEO and founder of your company, both of you listeners that heard that are equally empowered by this and are equally responsible for doing this. So two to three certifications a quarter from known industry, known good industry actors. And that's uh that's it. Thank you all for uh for coming to our episode on AI. And again, if you found this interesting and you'd like more content like this, maybe even some lives where we get together and just like do some prompts and have some beers, we can do that.

SPEAKER_00 1:08:26

John, really appreciate the episode. This has been phenomenal and such easy, actionable steps to take to become better at AI.

SPEAKER_01 1:08:32

I know, and I'm afraid to ask more questions because I know we'll be here for another two hours. So I'm gonna save it for AI part two, but I want to know about how do we teach our kids with AI and how do we teach our kids how to use AI. So that's something down the road I would love for us to discuss. Maybe we'll kick off season two with that.

SPEAKER_02 1:08:49

But for now, see y'all in the next episode.