All Episodes
Sep 25, 2025S1E1500:36:51

15. One More Breakdown: Season One Finale

Music from Uppbeat: “Make It Happen” by All Good Folks. License code: 6BVFZAHAWVMGPFRE.

Or listen on
Pete BishopHost

Pete Bishop

View bio

Carrie RobinsonGuest

Carrie Robinson

Co-host (Season 1)

In their season one finale, the hosts of Artificial Breakdown reflect on their favorite episodes and key learnings about AI. They discuss the importance of 'play' for adoption, the necessity of human 'taste' to evaluate AI output, and the future impact of AI agents on e-commerce. The hosts also share practical tips for using tools like ChatGPT and preview topics for their upcoming second season.

What Were the Most Impactful Lessons from Season One?

The most impactful lessons centered on the importance of play for AI adoption, the necessity of human taste to guide AI, and the concept of unconscious incompetence. These themes highlight that human expertise remains crucial when leveraging AI tools.

The hosts identified the episode with Mike and Sammy on 'AI and play' as a recurring favorite. The idea that people need to experiment and get their hands dirty to truly learn and adopt AI has been a constant theme. This approach is evergreen because it is the most effective way to understand how these tools can be applied to individual needs, rather than just following a generic course. Another key concept came from the episode with Tailwind: unconscious incompetence. This describes a state where a person lacks the industry knowledge to even ask the right questions to get useful AI outputs. It underscores the value of domain experts, like experienced marketers, who can query an AI with the necessary context to generate meaningful results, such as a brand strategy, instead of just a generic logo.

How Does Human 'Taste' Complement AI's Capabilities?

Human 'taste' is the critical, often intangible, expertise that allows a person to judge the quality and appropriateness of AI-generated outputs. While an AI can generate numerous options, it lacks the nuanced understanding to determine which is strategically sound, original, or well-executed.

The hosts discussed a concept from Scott Irwin, which posits that 'taste' is the key differentiator. An AI like ChatGPT can produce five campaign ideas, but it needs a person with experience to discern which one is actually good. Without this expert lens, a user might mistakenly assume all AI-generated answers are valid simply because they were produced by the AI. This is a significant red flag, especially in creative and strategic fields. The AI's function is to provide an answer, and it will execute a command if given one. It requires a human with taste and experience to apply the brakes, provide direction, and ensure the output aligns with strategic goals. This is particularly true when building foundational assets like a brand's tone of voice, which should be developed by an expert before being fed into a custom GPT.

What Practical Tips Can Improve Your Experience with ChatGPT?

Two key tips can significantly improve ChatGPT usage: leveraging custom instructions to globally refine outputs and creating Custom GPTs for specific, recurring tasks. These methods help make the AI's responses more consistent and useful for your individual needs.

A helpful trick is to modify the global custom instructions for your ChatGPT account. One host recommended adding instructions to remove em dashes, which are a common tell that text is AI-generated. Another powerful instruction is to ask the AI to provide a percentage score of its confidence in the accuracy of its answer, which helps you gauge the reliability of the information. For more specific workflows, Custom GPTs are extremely valuable. They allow you to build a version of the model with persistent memory and knowledge for a particular topic, such as a client's brand or a specific project. This avoids the need to re-explain context in every new chat, making it much more efficient for businesses or individuals managing multiple brands or complex tasks.

How Might AI Agents Reshape E-commerce and Marketing?

AI agents are predicted to fundamentally reshape e-commerce by gaining the autonomy to make purchases on behalf of users. This creates a new audience for marketers to target: the AI agents themselves, leading to a shift from business-to-consumer (B2C) to business-to-AI (B2A).

The hosts explored the future potential of giving AI agents a budget to spend. An agent could learn a user's habits, such as their tendency to order pizza on Thursdays, and then autonomously order and pay for it. This level of automation represents a massive shift in consumer behavior and will have huge implications for marketers. The target audience will no longer be limited to people, but will expand to include the AI agents making decisions for them. This gives rise to new marketing models, like 'B2A' or even 'AI-to-AI.' As a result, optimizing content for discovery by AI agents, a practice some call Generative Engine Optimization, will become an essential marketing discipline.

How is AI Being Integrated into the Education System?

AI in education is best treated as a tool to be used after a student has grasped foundational knowledge, much like a calculator in math class. Schools are beginning to teach AI literacy while ensuring students still master core curriculums, sometimes by using AI-free final exams.

A discussion from the episode with Muhammad provided a balanced approach to AI in schools. You would not give a student a calculator before they learn basic arithmetic, and the same principle applies to AI. Students should first learn foundational skills like writing without AI assistance. It is also crucial to arm them with AI skills for the future, as they will be expected to use these tools in the workforce. The hosts also emphasized the importance of AI literacy, a topic from the episode with Stephanie from Amy. Since children will inevitably learn about and use these tools, it is better for them to learn from trusted teachers in a structured environment. This can help them avoid unmoderated and potentially unsafe chatbots. Some universities are finding a middle ground by allowing students to use AI for learning but requiring that final exams are completed without AI assistance.
“Just because AI gave you an answer doesn't mean it's good.”Emi
“If you don't understand the industry or the vertical well enough, then you can't even ask the questions you need to ask to get AI working for you.”Peter
“This tool will do what you want it to do. Ask it to do the stuff you want it to do, so that you can focus on the stuff that you want to focus on as a human.”Emi
“You can create custom GPTs, but you can also just put in the instructions for your own overall GPT.”Peter
“Our audiences as marketers is people and agents.”Peter
“I think there are still these people who are like, well, it has no place in welding, you know? It's like, well, it does have a place probably in the business process.”Emi

Key Takeaways

  • Adopting new AI tools effectively requires hands-on experimentation and a sense of play.
  • Human expertise and 'taste' are essential for judging the quality and strategic value of AI-generated content.
  • The concept of 'unconscious incompetence' highlights that you need domain knowledge to ask AI the right questions.
  • AI can automate background processes in any industry, including CRM, operations, and accounting.
  • AI literacy is one of the most important skills for everyone to develop in the current landscape.
  • AI agents with purchasing power may soon become a new target audience for marketers.
  • Custom instructions and Custom GPTs are powerful features for personalizing and improving ChatGPT's output for specific needs.
  • AI should be used to augment human tasks and free up time for connection, not replace human experience entirely.

Frequently Asked Questions

What was a standout episode from season one of Artificial Breakdown?

The episode with guests Mike and Sammy discussing AI and play was a consistent favorite. The hosts found that the idea of using 'play' as a method for learning and AI adoption is a powerful and recurring theme that resonates with many listeners.

What is 'unconscious incompetence' in the context of artificial intelligence?

Unconscious incompetence is the idea that without sufficient knowledge of a subject, a person cannot even ask the right questions to get effective results from AI. This concept, from the episode with Tailwind, emphasizes the continued need for industry experts who can prompt AI with the necessary context and nuance.

Why is human 'taste' important when using AI?

Taste refers to the human expertise and judgment needed to evaluate AI outputs. An AI can generate many answers, but a person with experience is required to determine if an answer is good, creative, or strategically sound, as the AI itself lacks this nuanced understanding.

How is AI impacting the music industry?

According to the episode with James O'Brien, AI's role in music is not just about generating songs. It is also about creating tools that preserve the human element of creation, collaboration, and live performance, ensuring people can still make music together.

What are some practical tips for using ChatGPT more effectively?

You can add custom instructions to your ChatGPT account to refine its answers, such as telling it to avoid em dashes or to provide a confidence percentage for each response. Creating Custom GPTs for specific tasks is also highly effective, as it allows the AI to retain context without needing to be retrained for every chat.

How should AI be used in education?

AI should be introduced in schools after students have mastered foundational concepts, similar to how calculators are used after learning basic math. While students can use AI to learn, some institutions ensure mastery by holding AI-free final exams, preparing students for a workforce where AI skills are expected.

What are AI agents and how might they change e-commerce?

AI agents are systems that can perform tasks autonomously, such as making purchases based on a user's habits. This could fundamentally change e-commerce by shifting the marketing focus from targeting people (B2C) to targeting agents (B2A), as agents will make buying decisions.

What topics will be covered in season two of the podcast?

Season two plans include revisiting past guests like Scott Irwin and Dave Nagy, and exploring new topics such as AI agents, early childhood education, and AI in government. The hosts also want to interview more artists and businesses about how they are incorporating AI into their work.

Topics: Generative AI, Podcasting, AI Adoption, AI and Play, Unconscious Incompetence, AI and Marketing, AI and Taste, AI and Music, AI in Education, AI Literacy, AI Agents, E-commerce, Prompt Engineering, Prompt Optimization, Custom GPTs, AI Safety, SaaS, AI Ethics, AI and Government, FinTech

Tools mentioned: Riverside, ChatGPT, Claude, Co-pilot, Custom GPTs, Alexa, iPhone, Samsung, VIC-20, Commodore 64, Oregon Trail, Stickler

Artificial Breakdown (00:03.086) Okay, our levels are hidden in yellow, so that's looking pretty good. Yeah, this is kind of an experiment because we learned kind of the hard way at Inventures. What happens when you take like three really sensitive mics and put them really close together? the answer is chaos. Yeah, in our minds for some reason the answer was this will work perfectly. Yeah, I you know these some of these tools like artificial Like Riverside is supposed to be kind of one button all kind of automated ready to go Which is great until it's not until you try to do something tricky and then it's like, okay So we're trying it with one mic now, which is I think it'll work fingers crossed Well Listen, I'm having a lot of deja vu because it's been about a year since our first episode. Yeah, that's crazy. Which is crazy. It speaks to a very slow editing process. When we get a new team, we're going to fire the old team. Yeah, I can't wait to get a new producer. Yeah, and... I remember when we first started doing this, we recorded like, think, two full episodes that just never got published because the mic wasn't on or something wasn't working. Yeah. Look how much has changed. That's something I can probably get. We'll never stop experimenting and doing dumb shit and hopefully figuring it out by the end. Yeah. I think it's gotten better. Anyway, we'll see. But it'll be nice to get back to a regular pace, I think. know, Inventures was super fun. Super fun. That's what's very Incredible people to talk to. I just find it was, we're a little out of our element with tech, but that's okay. We're figuring it out. And I think if we do it next year, we will be 10 times better. Yeah. I think we're two people who are both fake it till you make it. And when you put two of those people in a room, it's, we'll see what happens. Yeah. What's free. I love that phrase building the plane as you fly it. I don't think I've ever heard that. Oh, really? That's funny. That's so me. Yeah. Artificial Breakdown (02:24.056) that is 100 % my life. There's a gif of a dog from an old cartoon and he's on a train and he's building the track as it's going and the train's going. I think it's from an old Bugs Bunny cartoon or something. But yeah, this is our final season one episode. Yeah. What do you think when you, this might have been a question you had, but when you look back at episode one to episode now, what are we on, 15? 16? I think Sabah was 15. Yeah, 16. Sabah. Sabah. We'll be cutting that out. Yes, we're on episode 16. Yeah, I actually think it's 15 because we cut one. I remember we cut one out. Yeah, but I thought he was gonna be 16. I guess we can check. It's not like we don't have the technology. Yeah, we should probably know this. you're right. You're right. You're right. I hate being wrong. Especially caught on video. Yeah. when you about, So when you think about, from episode one to episode now, which is 15, big changes, do you feel like, man, we didn't know what we're talking about? Or do you feel like things just change really quickly? What are you thinking? Well it's kind of funny, I was listening to some clips of episode one and it actually and I thought I was gonna go in there and be like oh my god we sound so stupid but I think there there was a lot of good nuggets in there about like where this is all going and I think we're pretty open about the fact that but this was all going to change pretty quickly and we had no idea where it was going to end up in a year. But yeah, think the thing I feel them, I feel so much more comfortable in front of a mic and sitting here talking for like, you know, I was like, Oh God must know what I'm talking about. And I'm like, no, I don't. What about you? When do you find the biggest difference? Yeah, I don't know. Like I think going into it, I really had no idea. That was the whole point of this for me was to learn. Artificial Breakdown (04:46.333) So yeah, like I still feel like there's so much of it. I don't know but there's It's been good like just having like listening to it from a bunch of different angles just makes you realize how it can be everywhere like the Fashion episode where we talked about it. I was like how would AI really be in this but you know It may not be in the actual crafting of materials or the service items yet But it's in the background as far as the CRM and the operating software and the accounting and all that kind of stuff can be automated. So it can kind of go anywhere. Totally. Which is cool. That's so true. Yeah, we talk to a really wide range of people. Sorry, I have like the squeakiest chair. Every time I move in it, it's just like, sounds like an old haunted house door. Ooh, creepy. I'm like... Yes, Halloween episode. Right. Thank you. my God, whoa. If you were going to dress up as AI for Halloween, what would that look like? Oh, interesting. I think I could see that being a costume this year. Yeah. Yeah. I just talk a lot and be super positive and my name would be Claude or something. Just a little badge. That's the best idea I've ever heard. To say that to everybody. Wow. You've really thought this through. Yeah, this is very polished, very well written. Yeah. Or You can wear aviator glasses and like a little scarf and go as a co-pilot. That's so good. And you freeze up and break down every once in while. I cannot complete your request. that's great. Yeah, there's got to be costumes already in development. Yeah, I never even thought about that. That's Costumes in development. You can find it on our store at artificialbreakdown.com. Jokes, jokes. Okay, what about, let's talk about favorite episodes, Emi, stand out for you. I mean, my favorite and the one that comes up the most when I ask people what their favorite episode was of the two people that I know who've listened to every single episode that we have is the one with Mike and Sammy where we talked about AI and play. I think that really struck a chord with people, including myself. Artificial Breakdown (07:09.549) I think about it all the time now, especially when it comes to like teaching and learning. And obviously we're still doing AI adoption internally here. And that's something we're going to be offering our clients and all that kind of stuff. And it's like, okay, how do we do that and make sure it actually works and it sticks. And it's just this idea of play just keeps coming up, keeps coming up. And I think they really captured it. Yeah. It's a little evergreen, I think, cause that's going to be a probably, I would imagine a theme forever. As far as how do you get involved while you gotta play with it? You can't just wait for everything to be sorted out because I don't think, it's such early days, I don't think there's any definitive, like here's your go-to to learn everything. It's just everyone's creating little mini courses all over the planet. Totally. And it's still hard to even figure out unless you actually get your hands dirty. Well and then also it's like when you think of if you're using this tool for whatever you're using it for. course might be teaching you about something completely different like for a business or for a marketing or you know whatever and you're just trying to use it to I can't think of a single example now for some reason but you need to play with it to see how it works for your purposes. Yeah yeah exactly I wonder how the mic's gonna pick up the screaming sirens. Yeah interesting. Reminds me of my dog instantly because second sirens are in he's howling. Howling. Just like Snoopy. That's so cute. I think what I read was is that they are trying to tell whoever the sirens are that they're there. It's like a way to connect with every dog. That's like that scene in 101 Dalmatians where they all start. Scotland Yard. Have you seen the Instagram where this lady's got like four huskies and they... Yes! go fucking berserk. they are howling. It's almost like screaming. Like four people screaming. It's like children screaming. Yeah. Yeah. It's terrifying. Oh, I'm like, how do you know how to monetize the crap out of this? I mean, I tell that to my cats all the time. How come you guys are making me any money? Start making bang. Like be cuter. Be cuter. Why? I need to start my TikTok account, obviously. Yeah. You're way behind. Oh, fuck. Artificial Breakdown (09:35.15) Let me see. I know you didn't ask what I'm going to ask anyway. Please. What was my favourite episode? Yeah. Oh, you're not interviewing me on this Oh, Yeah, Peter, what was your favourite Good question. Let me see. Let me think. Yeah, there's a few. There's so many that kind of changed how I thought about stuff. Like Nagy with his... Dave Nagy. That's another top one. Yeah, with the environmental impacts, it's just a factor having considered it. Like you just, when you think of electronic stuff, I just never go to like environmental impact, but the implications are incredible. So that was cool. I really liked Tailwind when they talked about unconscious incompetence. I've been using that a lot. Yes. Which I love that phrase. like, and it's me in so many ways, but the idea of like, Where there's some saving grace, I think, with some of these things is that if you don't understand the industry or the vertical well enough, then you can't even ask the questions you need to ask to get kind of AI working for you. So that's the unconscious incompetence. And then once you kind of do figure that out, then you can be more effective. But it still speaks to having industry experts or in our case, like marketers who could actually talk to AI on a and understand the questions they would need to figure out in order to undercover something like a brand or something like that. Cause the average person is going to go directly to make me a logo, you know, give me a color palette and you're going to skip most of that brand building and you're going to end up in a kind of probably the wrong place for a company. Right. So I think there's that, that gives me some comfort around that go, you know, that stuff. Yeah, I totally agree. It's, It's like if somebody was going to build a custom GPT, I think about this all the time, if somebody is going to build their own custom GPT for their tone of voice, it's like, well, do you have a tone of voice? Do you have a tone of voice doc that somebody's actually built that makes sense for you and your company? Or are you just like right in the tone of voice of this company? But yeah, it's like you said, you have to be able to ask the right questions and ask the right questions, but also be able to look at the output from that. Artificial Breakdown (11:58.772) Expert lens. Yeah. Yeah, you know speaking of I've chatted with Scott who we almost had on in season one, which we should get back but he Scott Irwin Scott Irwin. Yeah. Yes, and he was talking to me. Have you heard this concept of taste? Like the concept of taste I'm sure you've heard of the concept taste. As opposed to AI this one guy was talking about us. But he's saying that is the difference is taste. So when you Like, ChatGPT be like, okay, you know, I've come up with five different campaigns. What do think? Like, which one should I run with? And you're like, whoa, whoa, like you need someone to be like, this is good. This is not good. Like this is, there's something off with this or this is, this is, I've seen this before or whatever. It's just this, it's almost as intangible of like someone who is really, really good at their craft being able to be like, this looks like shit. This looks good. Like, I think this is good. whatever that is, is taste. Right? And I think that's what's lacking a little bit right now. Cause you look at it, it's like, I would run with any of these. It's like a race car just like revving its engine, waiting to go. And you're like, no, no, no, wait, wait. Like these aren't good. Right. But it would go if you just let it. Yeah, totally. Because we don't totally know how AI works still and it's evolving faster than we can even comprehend. But there is something we know, which is its purpose is to give you an answer. So if you say, give me five campaign ideas and it gives you five campaign ideas, you need that taste level or else you're just people who aren't maybe, you know, trained in marketing or trained in creating campaigns. Are you going to look at that and be like, these are five great ideas. I like this one. And clearly this is a good answer because AI gave it to me. And it's like, that to me is such a big red flag. It's like, just because AI gave you an answer doesn't mean it's good. And especially when it comes to stuff like that. That you do you need that taste you need the experience you need everything behind it. Yeah. No, I totally agree the You the other episodes that I liked was I really loved the the music one. my god. Yeah Ben O'Brien. Yeah Ben O'Brien. No James James. my god. I know another James O'Brien and his cousin's name is Ben O'Brien. That's where that came from Artificial Breakdown (14:25.165) Yes, James O'Brien was, that one blew my mind. That made me want to move to LA and work in AI music. Oh yeah. It was such a cool guy and that was such a great topic. Just AI and music, there's so many areas that you're like, wow, yeah, of course this has got a massive implication to it. Well, and it's another topic where you assume, it's like, okay, AI music means somebody's making AI music. Boo. But it's like then you talk to him and it's like no it's about so much more than that. Yeah it's about ensuring that you can still sit down and make music on your own as a human and jam and hang out and preserve the humanness of music. Yeah and I like the thought of out of that one what I liked was just like yeah you know perhaps maybe the silver lining of this is like it's going to force us to do more things together like we've been being separated ever since phones came out. And now maybe this is the impetus to start to come together more because that's our safe space, right? Like doing live performances or joining together for drinks or a campfire or something like that. Those are safe spots for us. Yeah. Yeah. Well, I think there's this craving. This was my other deja vu. When we first, our first episode, I was talking about going on the solo camping trip that I do every year without any technology, et cetera. And I'm about to go on that. tomorrow. And I just, I have such a craving to be away from technology. Not all the time, but like you need that break. And I feel like more and more people are, you you're sitting around a campfire, leave your phones in the house. Like, how nice is that? It's just like, you can sense it in people too. So it's like, even when you rebooted your computer earlier, you sat back, you took a breath. It's like, finally. That one minute of freedom. back up, you're like, shit. I was just checking my watch. But yeah, that was a good reminder from him. Like, yeah, don't use it for... Well, and this is a phrase I've been having, I've been trying to say a lot, especially at the office and just with people in general who are kind of freaking out about it. It's like, this tool will do what you want it to do. If you ask it to do stuff, then you're like, wow, I can't believe he's doing this for me. It's like, well... Artificial Breakdown (16:48.631) Don't ask it to. Ask it to do the stuff you want it to do. So that you can focus on the stuff that you want to focus on as a human. Yeah, it's kind of interesting when you start to bake it into your daily life and it does really make things better. Like we're trying to find a used car at the moment and it is so great for that. Yeah, hey. Just cutting through all the bullshit out there because most of the stuff out there is written from perspective of a seller. Right. So you don't know what to believe, right? You can't really trust, you know, the garage is out there and you, cause everyone's got an agenda and there's all biased and it's not like AI is not biased too, right. But it's still at the moment of the objective view on like what you could use cars. Cause it's not just like cars. Cause that's factual. It's like, here's the stats. Here's what it's worth. Yeah. Here's the things that maybe take the price up or down. Exactly. Yeah. So it is handy. Like it's just. Again, it just depends on how quickly it gets involved in your life. But this is going to happen fairly quick. imagine. I know that they're coming out with the, the Alexis, where you can just use them as chats. I don't think I'll ever get one of those. Yeah. Would you? I don't know. Probably. We use one. Well, there you go. So, yes. I really don't like the, I know this is the listening part. I don't like that it's listening. And I, for some reason, don't like... Yeah, I think that that is it. It just feels like a spy. I don't know. I'm still in that. And everyone's like, well, your phone's a spy and your computer's a spy and everyone's a spy. like, well, I don't want one more spy in my house. You know, Jill's talking about it, which at first I was like, okay, you're being completely paranoid. But now the more I think about it, I'm like, oh, this may be an idea. Where the whole family takes all of their electronics and leaves them in the house. We go out to like Fish Creek and we come up with a safe word for the family. Artificial Breakdown (18:50.559) And then in the future, if we'd have to keep doing it if we ever used it, but then if you get like that spoof call or something, you can be like, what's the word? A hundred percent. And then we know that it's real or not. Cause I really think we're going to get to a point where we have no idea. Well, and these scams are getting better and better. Man, if AI has done anything for any industry, it's the scam artist industry. Yeah, totally. The emails, the texts, everything that's coming out. Yeah. Yeah, and so like I've noticed now it's like if I'm expecting a shipment I get fake shipment. I get a lot of those now too. So it knows what else you're doing and can tie it in well and that's just gonna get harder and harder to figure out. There was, I really liked Muhammad, the one about the education. that was really cool. It was good to hear around A couple things. One, I love the notion around, you when you're learning math, you don't use a calculator. But when you're going out to the world, obviously you do, right? The same with AI. When you're trying to learn English or, you know, some of these basics, you shouldn't be using AI. But then at some point in schools, you should be armed with AI. Because when you get out of school, you're expected, you're going to be expected to know this. So that, like, when to use it, when not to use it was making sense to me. Totally. Well, especially when you think of like kids this day and age, they're going to learn how to use it. They're going to figure it out. Would you rather they learn that in from a trusted teacher or would you rather they learn that from each other and then they end up on all these super scary chat bot sites specifically made for kids that are not guardrailed at all. Right. It's a frigging disaster. Yeah. Yeah. In my personal humble opinion. Well, and I liked his thought around, um, Hey, use it as much as you want, know that the final exams are going to be AI free. you got to know it. You can use it however you need to do to learn. I think that's more university level stuff, but that's where my kids are now, right? Which is like, okay, yeah. So it's cool to see that schools are completely shying away from it. No, and it's interesting because I think there are a lot of people or a lot of teachers in a lot of places like a... Artificial Breakdown (21:07.743) Education districts and stuff that like they're leaning into it, but some of them are still they don't know what they're leaning into so But that's why because that's another one of my favorite episodes was the one we had with The episode we had with Stephanie from Amy because then you get these companies like Amy who are all about AI literacy and You know, they're about making sure that yeah Everybody gets a chance to learn how to use these tools to help themselves Yeah. Which I think is just the most important thing in the world right now is AI literacy for everybody. Yeah. Yeah, exactly. It feels like in education when, you know, my mom can't use text. It's like, man, I wish she'd kept up a little bit. And that's where, like, I think that this could easily be a huge gap if we don't stay up to date. I'm just speaking as an old parent, but this could be the gap where it's like, I can't keep up to my kids because I'm not latching onto this technology now that feels, you know, scary and strange and wild and wonderful and I, you know, that's like easily slip out of my fingers. Totally. Yeah. Yeah. And then it's kind of like where, because how we use it is going to be so much different than how the youths use it. Um, and it's like, but how the youth uses it and how say kids like my nephews who are, you know, seven and nine, um, how they use it is how The world's going to use it moving forward and they're going to think of ways to use it that we never would. And it's like, gotta keep talking with them. What are you doing over there? Can you show Auntie Carrie what you're doing? She's gotta talk about this on her podcast. The new version of the iPad for the oldies that just like simplifies AI just for the gray hairs out there. Yeah, I know it was really good. Like I found the whole thing was like, again, there were so many guests that were, you was great to see how everyone was leveraging it and using it their own space. But yeah, I'm hoping that we have another round next year and just keep growing. As far as my biggest revelation, think was probably with, it started with Carl. I don't think I really understood it, but agents was a big deal for me. Wants to figure out. Yeah, you're on quite the agent Quite the path. Yeah. I'm, yeah, I find that. Artificial Breakdown (23:33.122) That made a lot of sense. Once I kind of unpacked what agents actually do, suddenly things opened up a fair bit for me. So I think that's been helpful. And that was the spark of it, was Carl starting to kind of unpack it a little bit. Yeah, and I think we both met with individually after that, and it's fascinating what these agents could potentially be doing. Yeah. Yeah, exactly. It's just essentially the chaining of all this stuff together is the agent part of it. Um, but that was, that was crazy. Um, any tips, tricks like that you've learned over the year that you think that anyone using the stuff should try? I've got one that I was just like, a buddy of mine told me about this is I just stuck in just yesterday actually, but you know, it's chat, GPT, you have You can create custom GPTs, but you can also just put in the instructions for your own overall GPT. And on my personal one, I stuck in a little thing, two things. One, get rid of dashes, because it's such a giveaway. God, it's brutal. You know where I see it all the time? Dating apps. my God. So many people, it's such a dead giveaway, somebody's answer will have this huge dash in it. I'm like, there's no way, that's not on your keyboard, on your phone. Like, what? I like long walks, dash. It'll be like... by the fireplace. Because you know they have the little prompts in there and it's like dating me is like dating me is like a long walk on the beach, dash. Fun and... what? Sandy. Like I don't know it's so dumb. It's like you don't even care. You don't even care. You can't even write your own answers in this friggin thing. The other thing is... So I get it to like at the end of every answer give me a percentage of how confident you are that this is the right answer. Oh, that's cool. Super good because you see it combine 90 % like that's all it does is at the end of everything it's just like this is 90 % but if you see a like 60 % or 70 % you're starting to be like, okay, so does it give you a lower number sometimes? Yeah, sometimes. Interesting. Yeah, it's really neat. Okay, that's really good. Yeah, I like that one. So it was a idea. was a friendly line. But it was was brilliant. Artificial Breakdown (25:54.381) That's interesting. Yeah, my personal one is full of recipes. Like I'll be sitting at home being like, I want a chocolate cookie, but I want it to be healthy and I don't have any oats. I'll like put in all these ingredients and it will give me a little cookie recipe. Oh yeah. What is the one you did the other day with you swatched? Oh, I had, uh, I made a grapefruit cardamom sugar cookie and I had originally asked it. a while back to it because I did an orange cardamom cookie. And then I was like, okay, what about if I put grapefruit instead? And it's like, well, grapefruit's more potent. So I would use less, zest and less juice. And they turned out freaking amazing. That's not, sound good. I mean, they were, I, they were amazing. Next time I'll bring you some. Too bad you didn't bring any, know, whatever. It's fine. took them to my other job. that and, and I, I am, I'm a huge custom GPT. I like it because you can build it and you decide what goes into it. And it's just, especially, you know, our work one, we have the memory off for a lot of different reasons for safety and whatever else. within that chat GPT, there are the custom GPT, sorry, you can maintain all of that stuff and it learns and you can go back in and be like, Hey, you remember that thing you did? Can you do that again? And I just find it. It's so much better than training a chat every single time. So I am, I'm a huge fan of the custom GPTs. Yeah, it makes so much sense if you're a business or if you work for multiple clients or you're touching different brands to pack all that stuff in so you don't have to keep iterating, remember this, remember this, remember this. Totally. And like I said, we have clients where we've written them a tone of voice and a brand doc, that needs that expert, needs that taste. And then you put those into a custom GPT, you're laughing. But it's all based on stuff that I know is solid. And it's making a lot less stuff up. And that's something I've started putting into, we got a really good prompt, what was it called, a prompt? Artificial Breakdown (28:17.876) a prompt uplifter. I'm a writer. That sounds amazing. But basically that's what it does is it's a prompt that helps you. optimizer. A prompt optimizer. Thank you. Which is such a fuchsia. love optimizer. I like uplifter. Yeah. So this prompt uplifter just makes you feel good. No, it helps optimize your prompt and it helps reduce hallucinations and... Yeah, I quite like it. Cool. Does that work for image prompts? Yeah, for sure. I'll send it to you. It's funny. I'll send it to all of you. Oh, Thanks. On behalf of everybody. I've noticed that especially with video how specific you need to get. Oh, yeah. How's your video exploration going? Good. Yeah, it's a little lacking at the moment, but it was good. I had a prompt. It's like a page long of what I wanted it to do. And it's funny, it's still, there's a lot of variation. what I haven't found yet, which you can do with video or images, but you can't quite do it with video yet, is put in reference. Like, use this can, use this shirt. It can't really accept that stuff yet, whereas image stuff you can, right? You can be like, use this can. So I'm waiting for that little Yeah, it's kind of funny, hey? It's like, you've seen how it's evolved with text from start to now. and how it's evolved with images from start to now. And now you're seeing that video. Okay. It's kind of where images were last year. Right. And then pretty soon it'll be like, yeah, make sure she's holding the stickler can. Shout out to stickler. Yeah. Your best winter beverage. Yeah. And it's, so when you think about the future, this is the stuff that gets me excited slash a little nervous, but the, can, it feels like, like the agent building feels like early, early days. It's very kind of techy. chaotic, but as you, as you can imagine these things growing, they talk about hey personal software where it's like, okay, I don't need, I don't need an outlook to manage my email. I can just visualize my software anytime I want. So I can have Peter's version of outlook and I can customize it any way I feel like, and it's going to show up just for me. No one else is going to have the software. It's just my as needed software. And I can change it on the fly and be like, I want this to Artificial Breakdown (30:47.948) move like this and that and that. it's when they talk about like the threat to SaaS software, it's a bit of that, right? Where it's like, okay, we don't necessarily need to have a defined piece of software that has to be worked for everybody. You can just generate stuff that works just for you on the fly, which is exciting from a customer point of view. Well, depending on the type of customer you are, yeah. So I used to work for Telus and you know, kind of in the iPhone boom. And I remember people would come in and this was always kind of our sales pitch. They're like, well, I don't know which smartphone I want. Do I want an iPhone or do I want a Samsung? Because those were kind of the two options at the time. And this was kind of the sales pitch we would always say is like, well, iPhone is intuitive and it gives you what you need, but you can't really adjust anything. And Samsung is a lot more customizable. You can change the colors, you can change the whatevers. And then, depending on who the customer was, They're like, yeah, well, I want to be able to customize it, which was me. I'm still not on an iPhone blast. And then there's people who are like, nope, just give me this thing that'll tell me how to do it. And I just, don't want any options. just want the technology. So yeah, it's, it's interesting because I think for, you know, somebody like you and me who wants something more customizable, it's like, well, now the options are going to be pretty endless. Yeah. I don't know. It seems. Seems like things are going quick. So I'm really curious to see this time next year what we're talking. No. It may not even be us. No, well, I mean, we're not cyborgs yet. But if it doesn't happen by next year, I'm to be pretty pissed. I'm still waiting for my car to drive me around. still hasn't happened. Still no hoverboards. I know. There's a lot of work to be done. No jetpacks. No jetpacks. AI would probably be like, jetpacks seem pretty unsafe. We're actually not going to make those. Artificial Breakdown (32:48.556) We've determined that you're not smart enough to handle this. This is outside of our community guidelines. You're not allowed to do this. So when are we thinking we'd start up again? Well, when we're in September. We deserve a little break. Yeah. By break, audience, I mean we're going to learn more to discuss. But yeah, mean, maybe the new year. Yeah. think that sounds about right. Anyone you were hoping to or any kind of topics you're thinking about for next year? There's so many people I want to have back on. I'd love to have Scott back on. Scott Irwin. Irwin. Yeah. I'd like to get Nagy back. I think there's some new stuff he wants to talk about. Dave Nagy would be great. Obviously Mike and Sammy would be cool to see what they're up to because I think they've been doing a lot of really cool stuff. yeah. He met Harry Potter, well I shouldn't say he met Harry Potter, he met Daniel Radcliffe. He did a magic trick for Daniel Radcliffe. I saw it and liked it. I like, you're so cool. just want to hear that part. You're so cool. Yeah, it'll just be mostly a Harry Potter themed episode. But yeah, I would love to talk agents. I'd love to have some people want to talk agents. I'd love to talk more about education, maybe like younger childhood education. think would be really interesting. Yeah, I agree. And obviously I love having the artists on. I'd love to talk to more people who are incorporating this into not necessarily their art, but how they're maybe running their art business with it. Or people who are incorporating it into art. that would be very cool. What do you think? I agree. I'd like to talk to more business businesses to understand a little bit about how they're incorporating it or not. Government even though it sounds a little boring. I think it would be kind fun to understand What they're thinking? Yeah, god. Yeah. totally or quasi government that freaks me out a little bit I'm so freaked out. gonna be like what? But it was interesting I did talk to one guy and just you know as far as like there's gonna be some stuff that I think is No matter how quickly this stuff goes. There's there's things that are embedded especially in government Artificial Breakdown (35:15.277) But you think of like FinTech, right? I had a good chat with Scott about this where it's like, okay, if things go down for a day, it's millions of dollars on the line, right? So the idea that it's kind of just instantly replace a bunch of like very complex software that has had like tons of stakeholder engagement and testing and all kinds of like authentications and all types of things that are going on, like to replace that instantly with AI software, I think. is a fair ways off just because the risk involved and the actual hoops that you have to go through to get this stuff through is not just build it and launch it. Totally. And stakes are high. Right. Like, and I mean, that's why we've seen that too. Like you have a smaller company or a smaller agency. Um, it could adopt these things so much faster, but yeah, when you start getting into these huge corporations and, and global, it's like, how do you make sure that this is going to be used and make sure it's going to work? It's like, well, maybe it's better if we just don't. I'm starting to understand that a little bit more. Not that I think that's the right answer, but now when people are like, ooh, I don't know, my eyes are a little risky, I do have some grace for these bigger corporations where it's like, no, it's a pretty huge risk, because if it doesn't work, we're screwed. Or it's just a smaller place. It's like, if it doesn't work, we just pivot. Right. Yeah. In a lot of cases, still software development and software development comes with risk all the time. True. So it is interesting to see where it all goes. But yeah, I don't know. Okay, what's your, let's say we take this exact clip a year from now. And what are you, what's your... What's your prediction? What do you think is going to happen in a year? Yeah, that's a good question. It's 2026. we look back on and laugh. I do think e-commerce in AI will be next to my thought. They talked about, you can give your AI agent or bots or whatever money to play with. So you could essentially be like, OK, I'm hungry. Artificial Breakdown (37:32.703) And it goes out goes well. I know on Thursdays you tend to eat pizza, so I'll just go buy pizza for you and get it delivered to you and it can arrive at your door in 20 minutes. Like that sort of thing where it's got the autonomy to go buy stuff. And that's huge shift from a marketing point of view because the second that happens, which again, I don't think is far away, then our audiences as marketers is people and agents. Yeah, right. You you talk B to C, B to B, and now people are talking B to A, B to A to I. And then eventually then really you're talking AI to AI. Yeah. So it's a really interesting moment. We've been beating around that whole search ability because if it is AI looking for stuff, then search is going to be a AIO. Big deal. That's a lot of my research this week. Yes. So yeah, interesting. And then, yeah, I don't... I don't know. Like I think the worry of course is like, okay, we've just got ChachiBD5 come out at this point, which is PhD level. They said, although there were some hiccups with it. Yeah, I'd say there's some hiccups and to me it felt a bit like, you know, it's been shooting up like this and it felt like it went like this a little bit, but it is still obviously growing and it's yeah, PhD level thinking. So we're on the verge of next. which is, PhD level is as high as we get as humans. So what is the next going to outstrip human intelligence, which is crazy to me, right? So if it gets to that, and then the snowball effect of like AI making AI could potentially happen. Yes. But I guess my question with that is if we're training these tools on collective human intelligence, how can it possibly get smarter than humans? even the most smartest of humans with the most information and all the data. Artificial Breakdown (39:38.411) Yeah, I don't know if we know how to answer that because we're not smart enough. But I think it's probably doable. Yeah, I think it's just, it's such a runaway because as a human, you've only got your own life experiences and your age, and time that you get with your life to learn things. Right. So you can like, if I wanted to read every book there was, I couldn't, I could maybe read, I don't know what a book a week or whatever the pace would be. That's the limit of me, right? And I can't be everywhere on the earth. don't know what lots and lots of countries are like. I don't know what tons of history that's going on. There's so many things that I can't do in the lifetime that I've got. Whereas this could potentially get it all, right? Like it's got all the literature ever written, every website that's ever been out there, every video, like it can consume everything. So its ability to like... grow and become more intelligent is infinitely higher than any human could possibly be. And you thought we were too stupid to answer that. That was a great answer! Well, I don't know, but I just like it's funny because that's when I when we that whole talk about creativity is like is creativity just a regurgitation of your experiences in your life that you are okay I like these these ideas because I've seen them somewhere in my subconscious and then they come out again but again limited to my own life experiences. Right which is why it's art right? Right. Because that's coming from such a personal place and then if other people see that and have some sort of connection with it and then they create something you know that's like this whole thing of art and then yeah you throw AI into it and it's like You know and it's funny because, Sorry this chair is unreal Some of the best I think creativity that I've ever seen comes from limited resources. So when things are in a small box. So a good example of this would be Commodore 64. I don't know why I had to bring these things up. It's not an episode of Artificial Breakdown. If Peter doesn't say... This might age me but... Artificial Breakdown (42:07.085) So Commodore 64, think had, well, it had the limit of the... I'm sorry, I do have to... What is that? It's an old computer, right? It was one of the very first ones. It was a VIC-20 and then there was the Commodore 64 that came out. It was a gaming, kind of presented as a gaming console. Interesting. But it created, it had floppy disks that you could put in and the floppy disk could hold 1.4 megabytes. What? Right, so... Is this Oregon Trail? Which? Oregon Trail? I don't know what that is. What? Okay, we had these old computers at school. They were all Macs. yeah. you put a floppy disk in and play. yeah, same floppy disk. Same kind of floppy disk, right? So, but, so every game developer had to create a game that could fit inside that 1.4 megs. And now you look at games and they're like terabytes. Terabytes, yeah. But this is 1.4 megs and some of the games there are still to this day I wish I could play again because they were so much fun and they're so limited by the you know the amount of pixels they could have on a screen the color palette the actual processing speed the actual file size so many limitations on them and they created these really great stories and storylines and stuff like that so sometimes I'm like I'm wondering if sky's the limit you know Sometimes you need to be thrown into a box to come up with kind of neat stuff. I don't know where I'm going with that. But you know, that's part of the whole taste thing where it's like, yeah, like you said, if you can do anything with these tools, what do you want to do? Right. And that's where I get stopped sometimes even in my personal life where I'm like, I'm a writer and I know I do all the, was like, do I want it to write stuff? Do I want it to help me brainstorm that? Like, where do I? Where do I stop, personally? Yeah. Yeah. Yeah. I think that's the trick. So it'll be an interesting year to see how we all fit into things. know. Yeah. Yeah. I think we're going to start seeing it in just popping up more and more in every space. Not that it already hasn't. But even all those, I think there's still these people who are like, well, it has no place in welding, you know? It's like, well, it does have a place probably in the business process and in that. Artificial Breakdown (44:24.663) planning and in your administration and that kind of stuff so yeah. It'll be really interesting. Running the business is always a spot for it I think at the moment. Alright, well, good work. was a lot of work. This is also our first episode of us being in the same room together. Right, yeah. So we've come a long way. I've come from down the hall to here. Yeah. It's very exciting. Well I can't wait till season two. Me neither. Yeah, excited. until then, if you have any ideas, our two listeners, feel free to... Yeah, this is for Heather and Rolly. If you guys want to hear anything specific, you let us know. Sounds good. I'll talk to you soon. Bye.