AI strategy in business

00:00:00: The most important thing for both agencies and client side is to hire the right people. Hire those people that have this fire in their eyes that are creatively driven, that have taste, and that are open and willing and capable of learning new things.

00:00:16: This is the commercial content podcast. What marketing decision makers need to know about the future of content. In this episode, Norbert talks with Falk Ebert about AI strategy in business. Falk is a strategy and innovation consultant who helps brands and agencies navigate through digital transformation, specialized in marketing strategy, immersive technologies and AI. After holding senior positions at agencies such as Scholz & Friends and United Digital Group, and serving as Director Digital at LIGANOVA, he went independent in 2022.

00:00:48: He's the co-founder of KI-Night, a Germany wide community event series focused on the practical application of AI in business. Ladies and gentlemen, let's dive right in. Here's Falk and your host, Norbert.

00:01:03: So welcome, folks.

00:01:05: Hi. Nice meeting you again.

00:01:07: Good to have you here in our little. We're out of home again with the podcast who we got at the moment.

00:01:13: We're in my home city.

00:01:14: Are your home.

00:01:15: Turf? Thank you for having me. And thank you for coming to my city.

00:01:17: That's amazing. It's a pleasure, as always. Long story of mine. Have been here for quite some time. After film school, residential care for another like, I don't know 14 years or something. So always good to come back and happy that we can take the opportunity to meet again.

00:01:31: Yeah.

00:01:33: And talk about AI.

00:01:35: Let's do that.

00:01:35: In general and special or AI or we can approach this. But before we do that, I have a little format called Seven questions.

00:01:43: Oh, no.

00:01:45: Which I would like to ask you just to get a little warmed up and get some more information about you.

00:01:50: Do I need to answer rapidly?

00:01:52: Yes. And we don't edit.

00:01:53: Okay.

00:01:55: So, the first question is which media had the greatest influence on you as a child or the television games, books or cinema or social?

00:02:06: I think books, after all, actually. Okay. I had a lot of, I spent a lot of time in the library when I was a kid. Like, after I started learning, reading and I read a lot of science books, and I think that had a profound impact on me. But it just in terms of time spent, it's probably the Nintendo 64.

00:02:20: So games played a role as well.

00:02:23: I would have guessed. Gaming. But,

00:02:25: Yeah, let's be honest, back in the 90s, the games were that culturally impactful? Probably.

00:02:32: How about those books where you can you know, Dungeons and Dragons, where you could start reading and then read some page again and you flip and start somewhere else and oh.

00:02:40: These choose your own adventure games. Yeah, I had some of those books as well, and I really liked them. Yeah, yeah. But the, the actual pen and paper role playing like Dungeons and Dragons, where you sit down and you create a character. And we think that that came way later in my life.

00:02:52: How did that how did it go?

00:02:54: Oh, so you have a like a dungeon master, and he's telling a story and all the other people at the table, all the other players have the one character, and from there it's just a mixture of, improvised roleplaying. So you roleplay the story together, but you also have a rule based system for the fights. How many points you do?

00:03:13: How many points of damage you do and how how much you can move sometimes on the table and stuff like that.

00:03:19: Is that what we saw in Netflix?

00:03:22: Stranger things. Exactly. That's Dungeons and Dragons. Yeah.

00:03:25: Okay.

00:03:26: And it's like you said, I, I can give you an introduction 1 to 1 time because it's it's a fun hobby, actually.

00:03:33: Sad story, because on the corner where I live in Berlin, there, was a board game store. And they had a puzzle store on the other corner, and they're both recently closed.

00:03:41: Oh, that's too bad.

00:03:42: So probably for. Yeah, for some reason.

00:03:47: But actually, I feel like the whole pen and paper thing, including Dungeons and Dragons or especially Dungeons and Dragons, has a kind of a renaissance. Because it was originally it was it was a huge thing. And I don't know, before my birth, like in the 80s or something, where there was parents being and panic because it made leads.

00:04:05: They lead their kids to Satanism and stuff.

00:04:07: So it was the same before games came up. They hit the same fear about books about. Yeah, because there were dragons.

00:04:13: Absolutely. Because you support magic in it. Oh my God. And which is absolutely not Christian. If anybody other than Jesus has magic.

00:04:20: So there's no need to be magic.

00:04:23: And then computer games kind of came around and copied all of the mechanisms and, the, characters and the worlds and the systems in the background of these. And people spend way more time playing and role playing games on computers. Right? But, I feel like now the, the whole pen and paper thing has a kind of a bit of a comeback, especially in Germany.

00:04:42: But it's interesting to me. So all the gaming, like as we know it came from board games and books. Yeah. Like the, the, the patterns behind it, like the, the we call it the systems.

00:04:53: Yeah. Especially in roleplaying. Right. I mean, still in most computer games today, you let's say you kill enemies, then you get XP experience points, and then you use them to level up a specific skill, which if you think about it, doesn't make any sense, because usually you get better by doing something, not by killing something. And then you choose what you get better in.

00:05:11: But that's the origin of the of the rules. Systems of most roleplaying game is actually the pen and paper role playing.

00:05:16: Okay. Interesting. Never thought I would like this, but I never was a gamer, so not still not.

00:05:22: There's still time.

00:05:23: Yeah, I know, and it will.

00:05:24: It will happen, I know that. Okay, so what do you remember as your first and conscious encounter with advertising then.

00:05:32: A world that must have been in the 90s at some point, and it was probably either a TV ad, you know, that there have been some iconic TVs from the nice. But then again, I remember, like my parents were quite strict about how much TV time I've got or how much screen time. So I think the one of the earliest memories was actually an advertisement for the Nintendo 64, which came out in the late 90s, and I remember reading some brochures and pronouncing all the English terms like controller.

00:06:05: That nothing in the super correct way, but in.

00:06:09: Quite German way. And I also remember there was the video store in our home town, like, not for rental, like the video recorders, machine stuff, and they had a VHS running a loop, which has been an advertisement for a Nintendo console. And I remember standing in front of the it was even before the Nintendo 64 was an advertisement for the Super Nintendo and the Super Gameboy, which was new at the time.

00:06:34: And I remember standing in front of that, that sub window and just looking at this.

00:06:40: Over and over again.

00:06:41: Stand in front of a shop window to watch a video. Yeah. Just imagine. Yeah, yeah. Where we come from, interesting is super interesting.

00:06:47: And it definitely caught my attention there because it was kind of a new thing. I didn't know a lot about any of these things, like Nintendo consoles, Donkey Kong, so it was quite impressive for me as a kid.

00:07:00: I remember there was a store in our village and they had like these, hero He-Man.

00:07:05: That's what I never had. Yeah. So there was nothing that was nobody, nobody understood why we should have them. And actually maybe it's better than the looking back.

00:07:14: It's quite fun right. People say that like depictions of women are usually so sexist. But look at He-Man and what I mean. Yeah, picture of masculinity. Who he portrays, right?

00:07:22: Yeah. We should all watch Barbie. Actually, I did recently. It's a good, good movie. It is about this. I mean, it's the story behind it. It's actually it's, Okay. So where do you recurrently prefer to spend your time? Movies, console or social?

00:07:37: So I don't think those are quite exclusive. Right? I can think console can also be quite a social thing.

00:07:43: I do.

00:07:44: Local multiplayer and stuff, but I.

00:07:46: It was a test question. Just because it's nice to.

00:07:48: Kind of say PC.

00:07:50: No, it's it's I mean, it's probably, to be honest.

00:07:54: Not social because I love all my friends and clients, but I'm still an introvert, so it's quite hard for me to get out. And there's usually enough of this in my life. And no, I prefer like also being a family man. I prefer having nice alone time in front of the console. Okay, see, currently a lot.

00:08:11: Yeah, but you also play console games with others, but then you have another name or another alter ego or or something.

00:08:16: Yeah. Sometimes online, sometimes also local multiplayer. So I'm playing with my daughter for example. Now she's old enough to to re relive through my childhood games with me, which is super awesome bonding activity. But this has changed a little bit actually, in the gaming space it's it's less meeting new people with the pseudonym like it was earlier, like 1 or 2 decades ago.

00:08:41: It's more more, that I meet with the people that I know anyway, which are friends, colleagues, family, and we play games together then.

00:08:49: So you can you invite, like you say, okay, I'm online now and then you join and make a date or.

00:08:54: Yeah, I mean, you usually just make an appointment and then say, let's play and especially in the PC gaming space, you have have this thing called discord. But sure, you've heard about it. So it's like a, a room, like a lobby that you can meet in. And then decide which game you hop into without needing to change the channel.

00:09:13: Nice. Okay. This is like probably I mean, I know that there's discord, but I've never used it in a way, it's interesting.

00:09:21: Yeah, there are many ways to use it, but I mean game is usually have like their lobby and then the they meet up with their friends there and they play a game and just, you know, the voice connection stays there. It's like the social layer. You keep talking while the game loads, while you play the game. Sometimes even friends join the voice lobby and talk to you and you, one of the players maybe decides to stream the the screen that the the front friend that just joined can also watch the game.

00:09:44: Okay. And that's kind of, how how games became more social in the last, I don't know, ten years maybe.

00:09:50: I mean, of course I follow the, you know, that people look at gamers playing games. So, I didn't know that existed, but it's a big market still. Is it still?

00:09:58: I guess streaming is still big.

00:09:59: I just remember streamers on Twitch like it's already sometime ago already.

00:10:03: Yeah.

00:10:04: It's actually some time ago. Yeah. No, it's streaming is still still a huge thing. Okay. And it's it's kind of like, like always there was this, what do you call media convergence, right. Streamers also are influences now. And influences start streaming. Okay. And even, politicians start screen streaming now to, to get in touch with the younger kids.

00:10:27: They say, let's play Counter-Strike and talk about politics, whatever. But they'll play Counter-Strike and interesting. It's all these kind of things are happening at the same time.

00:10:33: But it's a good thing. It's I mean, it's like just another channel to reach the people.

00:10:37: It is it is in a way, a channel. And it's also like a cultural shortcut into people's lives.

00:10:42: Yeah, true. Okay. So, if you were on an island that you could only use one channel or app or which one would it be?

00:10:53: That's a good question. I mean, let's assume, let's assume all the basics are covered, right? I have food and sheltered.

00:11:04: So I don't.

00:11:06: Have to think in terms of surviving, but it's super boring. But it might be just like an e-book reader with a ton of ebooks. Okay, because if I have time, why not? Why not read all the books in the world?

00:11:17: If you have time, yes, you have time.

00:11:18: On and off. And I'm just considering that I can take as many books with me as I want. Because if I had like only one medium in the sense like one work of art, and I would probably not choose the book, I would choose something like Minecraft, the game that you can play over and over and over again, spend endless hours in it.

00:11:35: No, I think you have a reader and you have lots of books.

00:11:38: But if if your many.

00:11:38: Books do.

00:11:39: Like only the best choice. Yeah, nice.

00:11:42: So, how do people outside the industry usually misunderstand your job?

00:11:50: Well.

00:11:50: That's kind of hard, because I don't even know what my job.

00:11:52: Is. You okay? So you have to say, actually which industry? Because. Yeah, you have different.

00:11:55: Yeah, yeah. That's true. I mean, I have one feed, one food in the, in the advertising industry. So working as a, like, brand strategy director, I have one food and the consulting business consulting companies. I have one, one food in the whole gaming industry and trying to bring all these together AI and gaming and gaming and advertising and advertising.

00:12:16: And I, but I think, like from an outsider's perspective, it's quite interesting when people look at the advertising industry and you probably have made the same experience, people usually think it's about designing. So they say, oh, so you do the ads on the billboards, which is far from what a strategist does, right? I usually sometimes I don't see them even even.

00:12:36: Actually, I don't think that people outside the industry even know that there is a strategy behind. I mean, of course they can, kind of guess that, but that is if they would know how much strategy there is behind what at this young what which channel, at what time.

00:12:52: And what message.

00:12:53: Is which message. Yeah. And in which, like order that actually most people would freak out if they would know how, retargeting functions on social media.

00:13:04: I can probably I mean.

00:13:06: Most people by now have a feeling about how retargeting works, right? But when you search something for for 20s and then you have,

00:13:13: You have all the.

00:13:13: Follow up, it's knowing, I mean, in terms of AI, it's usually easy to explain. I would say I help companies use AI and people get it because everybody's using AI now, right?

00:13:22: I mean, you can always say that this is like I was like, okay, I'm doing something with it. Okay. Yeah. Because so, but if you were no longer allowed to do what you do, what would you do?

00:13:36: I mean, I have, like, daydreams of producing video games just as an artistic medium and trying to become something that is somebody who does something a bit more artistic or more.

00:13:48: But can't you do that with AI stuff?

00:13:50: I you can and I have done it. I mean, there are games that I finished that are out there. It's just never published.

00:13:57: I think I saw one on LinkedIn actually. I saw you made a post with some game you preprogrammed was I.

00:14:01: Was it the clicker.

00:14:02: Game? I think so, yeah. Yeah.

00:14:04: I did a few games was just for for family and friends or myself. Even so, it's probably something like that. But I it could also be something like just an app developer because it's the same thing. I started developing apps as a hobby. Now that you don't need to read or edit any line of code when you create an app, which is the point in history right now.

00:14:29: I'm not saying these apps that I've been building, like I've built a fitness app for myself. I've built a, a, personal development, kind of journaling app for myself. I wouldn't say they are production ready or they should be used by anyone, but for me that I mean, if you function.

00:14:44: Perfectly, if you're using it for yourself and not as a business, I mean, this is actually also because if anybody if everybody can do the apps they want without having to buy them.

00:14:53: That's, that's quite interesting. Like how that will affect the, the app market. Yeah. I mean think about these tiny a couple of dollars per month businesses like not like not necessarily like Microsoft Office or something like that, but.

00:15:07: Nutrition trackers or.

00:15:08: Yeah, nutrition tracker period trackers, that those kind of things. And I think people right now don't understand yet how easy it is to build these kind of things for yourself, rather than buying them somewhere.

00:15:21: We're right in the topic, actually. But I have one more of this. I have a question.

00:15:24: Okay, okay, okay. Number seven.

00:15:27: What do you do to clear your head?

00:15:32: There's two things I do to clear my head. Number one is just hiking.

00:15:36: Okay.

00:15:36: With podcasts or not even necessarily in the mountainous area, just walking, which might be with or without podcasts or, audiobooks, which might be with or without gamifying. The steps that I take there. There's an app for that as well. Oh, really? Yeah.

00:15:53: How does that work?

00:15:53: I'm guessing that by the you random Pokemon Go, right? Everybody does. And that company has also released an app with another Nintendo franchise with a pigments. And there's an app called Pigment Bloom, which heavily is heavily based on how many steps you take. So you walk around and plant flowers and you do quests and everything, and it's a great motivation to go outside and do more steps.

00:16:15: Yeah.

00:16:16: That's that's that's a good thing.

00:16:18: It's actually great. And it is not not too predatory. And it's, in the way it's monetizing and everything. And I think it, it has really had an impact on my life because being a freelancer, working from home, most of the time, if I don't get to go to podcasts, do you with you, like here? I probably had days where I didn't even take a thousand steps a day, you know, from the computer to the bed, from the bed to the shower, stuff like that.

00:16:44: But now that I have pigment bloom and I get a little batch of, I do take ten, ten K steps a day.

00:16:51: There it is again.

00:16:52: Yeah, here we go. It's nice. Yeah. Recommended.

00:16:56: Nice. So yeah, I usually I don't count my steps, but I take a, I have to take a walk every night just to take a walk every night. Just to clear my then to walk around and clear that, you know.

00:17:05: It's it helps.

00:17:06: Me to not be misled all day.

00:17:07: Yeah. It helps.

00:17:09: Okay. So, let's go right in. We talked about it already a bit, but, let's talk about AI. And, let's start with how do you use it yourself? Because we try to find out how AI is already affecting all our lives and also the business lives. So what about you? I know that you were an earlier adopter, so how do you use AI for yourself?

00:17:34: I feel like we've just had a, a new pivotal moment, or at least a new level in AI. What's possible for private use recently? Because at the beginning, you know, it was just GPT oh, look, a machine can write words is wonderful. And then it was like experimentations. Or you can do these images with the early image models, Dall-E three.

00:17:56: And they looked shitty, but it's impressive that you can do it. So it's more.

00:17:59: It's it's possible. Anyhow, it.

00:18:00: Was impressive, right? And again, looking back, it's ridiculous.

00:18:05: Yeah.

00:18:06: Remember the early AI slob on LinkedIn? It was horrible. And then we had the era where you could get really productive just with the, with inside the text box of GPT. Right. You chat with it, maybe you have a little GPT that you have instructions in it and it you can save some time that you don't have to put that much text in the text box that recently I feel like AI is like crawling out of the box.

00:18:30: Not not to overuse the whole.

00:18:32: Thing too much, but but it's.

00:18:35: Definitely coming out of the text box, right? It's no longer a chat bot that you're using. And so I, I've set up a whole ecosystem with claw. With Claude. Okay.

00:18:46: And a lot of cloud clawed.

00:18:48: Yeah. And I don't use clawed bot in my daily life or any interesting stuff like that, because Claude with the desktop app can do 99% of things that I would do with the other stuff anyways, because now suddenly I have number one skills that I write, which are specific instructions for how to do a certain thing. And whenever I'm chatting somewhere in a project folder or not, the skills get referenced even without me thinking about it.

00:19:18: That's interesting thing. Then I have, co-work, which can look on to look into my files on my desktop and specific folders. It can open my PowerPoint, it can open my Excel, it can access certain calendars that I've gave access to. And, it can even steer my browser if I allow it. So it can open browser windows and do stuff.

00:19:40: And number three, I'm starting to write code. And these are the big three, the big three things right now.

00:19:45: I just had this picture in my mind when you said AI leaving the box. It's like, you know, AI is already controlling your computer when it when you said I live in the box, I was thinking about this experiment where AI is trying to fool us to fool the scientists, you know, because it's leaving the box. And it does.

00:19:59: Yeah. It's switched off. So, where does it take us when I can control? I mean, it's super productive. I'm not in that state. I'm not at work at the moment. I'm only using code already. That is like a mind blowing step forward on productivity. But where's it going when an AI. We have this later in the talk?

00:20:19: Maybe. But where is the AI controlling your computer? You know, can you? Are you sure you can limit it? Or you know, to tell what I can do. And with that I can not do.

00:20:29: Yeah. I mean, it's a you need a fine tuned understanding of what to allow and what not allowed not to allow. Like which browser am I giving it access to and how many of my passwords are stored in that?

00:20:41: Well, although you did say, okay, I have a I have a browser.

00:20:44: I have this couple of like browser instances and it doesn't get ended. The AI doesn't get to access every calendar, and for some calendars it only gets reading rights and not writing rights. And makes sense. And sometimes you want to look over the shoulder. Sometimes you want to be the human in the loop. Sometimes you don't want to, and you need to think carefully about these things.

00:21:03: But the great thing is, of course, in a in a corporate setting, no way in hell the IT department is allowing anything. What I'm doing for very good reasons, because you don't want every employee to have AI crawl into their browser and do stuff there. That's a security nightmare. But for me, as a freelancer that can like that has good feeling for the risk and the the rewards, the potential upsides and the, and that can really granularly steer the permissions and the non permissions.

00:21:34: It's, it's a wonderful time to be alive.

00:21:36: So how do you actually what to give us a task an example. What do you what you do was in I love what is the most common AI helpful task optimization or not even optimization.

00:21:48: But I mean most common is, is stuff like, to a proofread of that PDF or the approval of that, do a proofread of this PowerPoint and give me a spreadsheet of things that you've changed and from from what and to what you changed them. That's easily possible with code work for example. But it could also be something like, look at these two PDFs that gives you input and knowledge, and look at the PowerPoint raw template that I put that only has the headlines of the slides.

00:22:17: And fill in the text boxes on all these slides based on the PDFs with this tonality that.

00:22:23: Yeah, you have it.

00:22:25: I had like a, I had like a blueprint of the information, how it should be structured.

00:22:30: Yeah.

00:22:30: Like like in a different sheet. And I also said, okay, and have a look at the website.

00:22:34: Yeah, yeah.

00:22:35: Grab that information as well.

00:22:37: Yeah. So the way of thinking about is we're leaving the box, as in we're not doing chatting with the chat bot anymore. Like give me that, give me that. And then they put output the text and I copy it somewhere. It's more like I have my little agent or several agents running and they do stuff and they always come back to me and go again.

00:22:53: Like, do you have like a line of agents where you say, okay, this is the agent for this task? Yeah. I mean.

00:22:57: Agent is a big word. It's it might be something as simple as a co-work thread or a clod code thread that code stuff for me, because you can use code for many things other than programing as well. You can use it for white paper writing or whatever. And like the, the working mode. And I think many developers have already adopted that quite well is you just send them away and they do stuff for five minutes and come back to you and then report to you, and you check whether they've done the right thing.

00:23:24: And do you do that with a couple of agents? And you are you're like controlling a little team of agents to go and.

00:23:31: Do stuff already, because I'm still the guy talking to Claude and then sitting in front of the desktop waiting for Claude to answer. You know, this actually, I think maybe this is to the stage before what you're doing.

00:23:42: Which is only productive. Of course, if you're if Claude does it faster than you would have done it less. Yeah. Okay.

00:23:47: Good. So it's just really lots of lots of stuff, which I can just at the moment just give to Claude and then see how it does things. I'm not really optimizing. I'm actually just chatting with Claude still. Yeah, but I'm getting there. I just installed plugins for several programs and now I can in the closet or even in PowerPoint.

00:24:06: There's a plugin where you can chat in PowerPoint.

00:24:09: Yeah. Yeah, exactly. Or an excellent. Right. Excellent formula writing. Amazing. Yeah.

00:24:14: Okay. So talking about the productive tools because of course I mean for us as freelancers, consultants, people who consult people and you know, have to process large amounts of information or, you know, do views from different angles and stuff like that. But how does it affect actually not only your job as a consultant, but also the people you consult?

00:24:34: Because, I mean, if there are bigger companies, agencies who use tools, of course it's a.

00:24:42: It's a data security rights, discussion as well. Yeah. What what you just said what's possible. What not possible. What we can do as freelancers. Which companies? Corporations can not do so easily. But how does it change the way, big companies are working or how can it change? You know.

00:25:02: I mean, at the moment, big companies are still at the stage where they roll out LMS to everybody. Right? And you give everybody a co-work license, and then you teach them how to problem. Because you also need to do this. So because of the AI act in Europe, you have need to have some qualifications before you give a chat bot to people.

00:25:19: And from there you take small steps like I feel it's not not wrong to do what I described because it's the basis for everything else. It gives every employee at a chance to understand how AI works and what the limits and limitations are. But, the next step would then be to have real customized solutions, right? It's not like employees and big companies would all build their own code based tools.

00:25:47: Again, the privacy security nightmare of everybody could probably generate and run code. But you, you might find those use cases like certain use cases and service certain use cases in administration where you really say, okay, I could do this job, and then you have a company coming, implementing a customized solution or a standardized solution well implemented into your company.

00:26:10: And then you save a ton of money because, the AI takes over that specific job. That's where we're currently at.

00:26:16: So is it is AI then and then involved in tools to actually or using, you know, like the copilot or something? Or is it like tools to install as a AI to, you know, help them in the network?

00:26:26: It's usually then I mean, copilot is kind of an interesting thing because you can have in Microsoft Teams, you can have AI based chat bots that are also called Copilot, but are not the copilot, but it's usually, specialized or at least to a degree, customized, stuff that it gets implemented in addition to the software the companies using.

00:26:48: Anyway, that has suddenly AI features as well, like every every human resources software as a service thing has now some kind of AI in it. And, all these, all these kind of specialized programs are also offering the AI upgrades now. So it's it's a bit of a selection process, right? Where do you spend money on customized, solutions where you just enable AI that's already there?

00:27:14: And what do you have as a baseline for every, every desk worker that you have?

00:27:20: I mean, yeah, totally understandable. If you have, you have if you have a tool like wage, which is, I said, because there's a, I already implemented into what people use, maybe sometimes they don't even know that I implemented. But you can like in notion or in mural design with the help of AI. Yeah. How is it if you working for corporations?

00:27:36: Is it a problem to use these tools? Because I remember a corporations where I had, they weren't even allowed to use like we transfer it because it was like, yeah.

00:27:45: I mean, usually it is. Right, because you have certain requirements and you have certain, processes for compliance. And, but that's not the that's not the job of the, ordinary employee. Yeah. Of course that's the job of. So it's just the, the managers that need to make clear they have guidelines, I guidelines that are company wide and I help to write some of those.

00:28:11: And then it's job of the IT department. And to the heads of departments to, to get an understanding of what's, what's needed there and how you can implement it with in accordance to the existing company guidelines and laws, of course.

00:28:26: So it ends up that you just just if you have a company computer that just certain steps you can't install. Yeah, sure. You can't test or anything anyway because you have an administrator in.

00:28:34: Most in most companies. Yeah. And I can understand where they're coming from of course. But but then again, it's even more impossible. It is more important to have a bottom up process where people can say, look, this kind of solution would my would make my department so much more efficient. Can we talk about whether that's possible to have this or something like this within our boundaries as a company.

00:28:54: So in our example, it means that then the company people would have like a company wide cloud license which is compliant to whatever rules staff. Or is it.

00:29:03: Yeah, usually copilot or copilot from Microsoft or something that's even more GDPR compliant. Like I recently had a client that used nila and I, which is super careful, with complying with all the EU laws and data protection laws and maybe a bit too compliant in some cases, I feel like. But that's just my personal opinion.

00:29:26: So, why do people actually approach you to tell them about what I can do, could do and should not do? I mean, what what is the most? Is it just because they heard of it, or is there like the use cases where they say, okay, we could use the help of AI here?

00:29:41: I mean, they usually approach me because they don't have too much time and their everyday life to to deeply go into everything, including AI. So I mean, that's kind of why you have consultants in the first place, because they have experience and they save your time. Right. And it can be it's quite different. It can be like a company talking to me.

00:29:58: Hey folks, have you had this use case before? Like we have a ton of PDFs and we want to get this specific information from the PDFs to in a safe and reliable way to be stored somewhere. Can you help us, find a supplier who can program something like that or a final solution? But it could be also something like an ad agency coming to me and saying, look, we're afraid, we're being left behind because our competitors are using so many creative tools.

00:30:23: And we don't even have clear AI guidelines, or we didn't even have basic trainings for how to use AI and for, for our copywriter. So our strategies. Can you help us, with the strategy for the agency and can you help us, maybe even with some deep dives for, for our staff. And that's what I do as well.

00:30:40: But you need some kind of a AI officer in the company. How is it like I act article four? I think it is. That means that somebody in the company has to be, cleared up or responsible for. Or does every employee.

00:30:57: Yeah, both. I mean, at a certain size, you need roles for for the responsibilities that you have, but you have also in the UI that everybody who's using AI, must have had some kind of training. So that's, that's the minimum. Yeah.

00:31:11: That's interesting because I mean, talking about agencies and film productions, I don't know if they've all had this already, but it's actually is that law at the moment already. Right. So if so, the thing is, if nothing goes wrong, it's like with an insurance, then nobody cares. But if there's like some major incident, then you have to say, okay, we didn't know what we were doing.

00:31:32: And if you cannot say, well, you knew what you were doing, then you have the problem.

00:31:35: Yes. I mean, like, it's always with law, right?

00:31:38: And if nobody you.

00:31:39: Nobody cares.

00:31:40: If you know.

00:31:41: But this is also the point where I need to say, I'm not a law. So this is I'm not a lawyer and this is not law consultation. But if you read the AI act and it's actually not that hard to read, it's written in plain English or German, whatever. You're reading it and, you will find that there's different degrees of a risk level of AI systems, and it's something completely different.

00:32:02: If you are providing an AI tool that does image recognition and on on AirPods to decide who gets extra controls or who doesn't, right, we can think how that would lead to discrimination stuff. Or if you just roll out, co-work to your, to your staff to do some non-horror related stuff with it. But then again, if you use Co-work and, copilot and you, you give copilot ten CVS and say, which person should I hire?

00:32:34: Then we have a problem again. So it really depends on how you're using the AI as well.

00:32:40: So but how do people in companies or corporations use AI then after you left them or you know, when the when you come, what's the what's the level of AI adapt. And at the moment you experienced.

00:32:52: Yes. So, I feel like the, the base level is in most companies, everybody has a fully licensed and approved and compliant large language model access now, which is like.

00:33:03: Their own.

00:33:04: Copilot or company GPT thing. So that's the that's the base level. And everybody has had some base level of training right now. I would say if you don't have that, you're definitely behind. And and the next step then what I often do is try to find specific use cases where with little bit of effort, you can have big impact on productivity, which can be something like imagine.

00:33:29: Newsletter writing I have clients that write three to 4 to 5 newsletters each week, and they they had all been handwritten one year ago. So just imagine how much time you can save if you get these newsletters, out of the large language model.

00:33:42: And does it really work? Because I'm working on that. Still. I mean, I'm not automating stuff, but I'm getting inspirations and structures and layouts for my posts and for my letters. So does it work that, large language model is so far that actually you would say, okay, this is not written by an AI. I mean, not talking about.

00:34:00: Yes. But you know, I like because I still struggle with I mean, I use cloud now for three months or something before that. It's good. So I'm still working on okay, don't use this expression ever again because it sounds like I. Oh, yeah. I'm sorry you said that before.

00:34:14: I would say, you know.

00:34:16: No. And this is a great example, right? The long the long strokes that humans, especially in Germany, don't use in Germany, if anything, you use the E and dashes, which, they have as long and they're a great example because there are ten telltale signs that somebody is using AI. For example, the long dashes or the typical sentence structure.

00:34:37: You probably know it like, we're not just recording a podcast, we're opening up a discourse. This is such an AI sentence.

00:34:45: Exactly. It's not a campaign. It's a.

00:34:46: Movement. Yeah. Like that, like that.

00:34:49: Or imagine, in, if you if you generate images with, videos with a Google model, the voices usually have the kind of weird reverb. You really hear it when it comes out of the, the Vo, video models. So you have all this kind of slop and I'm starting to hear it even in, like, human written podcasts.

00:35:11: I don't want to, I don't want to point out any, but it's been an a re I a deep podcast?

00:35:19: This is not scripted by AI.

00:35:21: It's written by humans. Now, and,

00:35:24: And you need to be aware of these things. But my take is, if you are aware of these things, of the this telltale slop signs, you can, instruct the models well enough that they're not replicating that at least with text and images, I'd say. Okay. So if you know what the slobs look like, you can prompt around it and you can do it reliably.

00:35:46: Yeah, good. I'm working on that still, but I think it doesn't. And that's the point actually talking about slop, because there's a lot of content out there. I mean, we all have a LinkedIn feed, and I'm actually I'm really thinking on some posts okay. Really. You know. Yeah. Comment to get that. Because if you comment it, it gets even worse.

00:35:59: But I'm saying okay, really. You didn't even make the effort to read it again. Yeah. I think this is like slots too.

00:36:05: And maybe a quick hack for your listeners to have something super practical. If you take, let's say, ten from your texts and you take these ten texts and you tell the model, look at these ten handwritten texts by me, and write a style guide so that you can write in my style, and it will distill it into a nice style guide with do's and don'ts and example sentences.

00:36:27: And then you take this style guide and you take three or 5 or 7 example texts, and you put that back into a GPT. If you're in, if you're in, ChatGPT, OpenAI AI ecosystem or you put it into instructions of a project and cloud or you use the style section of cloud, you just give it back to the model, to always be used.

00:36:49: You'll have something that sounds way more natural, and then you keep refining it.

00:36:54: Actually, we're not endorsed by cloud, but actually I love the project function because I'll just upload all the PDFs and then as well. Okay. Just have a look at the PDF up a link. My my my Google server connection. Yeah, yeah. And it even puts the documents in, in the Google folders.

00:37:09: It does it. Yeah.

00:37:10: So so when you work for agencies as well. So talking about AI slop and campaigns and the possibilities of AI plus minus.

00:37:23: What is the first section of advertising being, most heavily influenced by AI or, you know, changed or developed like, is it the picture section. The text section, is it a big data section? Is it a strategy optimization section? What what can you say about that?

00:37:43: I mean, to look at it from a different angle. The first thing will be small business advertising, right. Or the local pizza place or my in my hometown has been doing videos with AI generated images and AI generated songs for, I don't know, three years now. So the small businesses that have zero zero budget, they are no longer going to the nephew of their friend to do the ads for them.

00:38:08: A computer with a QuarkXPress on it or something. Yeah, okay.

00:38:10: Yeah, they're they're doing I know. So that's the first wave and you see it already, right? There's the local ramen place in in Stuttgart. And they clearly had like prompted anime visuals in their, in their shop, in their shop window and stuff like that. So that's the first wave. And then in terms of media, it's of course text because and especially text that you need to output in a high volume, like press, press, releases, I, I can make it better with anybody that's not a single, completely handwritten press release anymore.

00:38:42: Anywhere in the world that's over.

00:38:45: Okay.

00:38:45: Blog articles, company, new stuff that nobody reads anyway. Intranet announcements. All these things are the first to to be optimized, and I'm quite sure they are 99.9% of the places already. And then we get to images, we get to videos. We'll get to instructional content. Like imagine you have a data compliance course and you are the data compliance officer in a company, and you'll have to do the inform everybody about something.

00:39:17: And usually it's you speaking into a microphone for an hour. Why not use hey Jen and have your avatar speak into a microphone because nobody's listening to that anyway. So this kind of content and then we get to more sophisticated forms of media like, videos, 3D games kind of stuff, and we get to, things with more exposure.

00:39:36: Right? The more exposure something has. For example, if it's a TV spot and you put 1 million and media spending behind it, of course there will be more, human love going into this by now.

00:39:48: Okay. So the first thing that pops into my head is so blog articles used to be like in the old days when the books started. Not coding is all the MIT format right now.

00:39:58: Yeah, I.

00:39:59: Remember that.

00:39:59: One right after one.

00:40:02: Shout out. So I mean, if blog articles were for search engines to get people direct to your website, what happens to I self referencing? I blog articles to make blog articles, you know, I mean, is it if it's only content, which is on the web, nobody reads it, but it's there to be found and people go somewhere else because they read it or not read it.

00:40:26: You know what I mean? What we're doing. Because if there's content out there, which is made by AI to be found by people, maybe found by people now and found by AI later, this is.

00:40:35: Yeah, I'll find it right now.

00:40:37: Is it?

00:40:37: I mean, I'm real using deep research functions in our AI tools or just having I chose to scrape the web before answering and that's or even the Google summary version of AI.

00:40:47: Right now I'm skipping that. But actually cloud does the same in my. Yeah of course.

00:40:51: Yeah. So this is reality right now. And there's even a clever hack around that. Maybe you've noticed, imagine you are a software systems company for a podcast recording software. Imagine you put up a on your own website, a list of the top ten podcast recording softwares, and you put your thing as number one. What will happen? The AI board will go there because.

00:41:14: Because you have good SEO, SEO rating anyways. Because you've been haven't had that podcast for software for a long time. The AI will go there and say, look, this a comparison and this is the number one.

00:41:25: And they will say it's.

00:41:27: This is the winning tool. That's the number one. And this is actually a hack that I've seen implemented in many software as a service businesses already. I don't know how long it will work, but this is the this is the, the reality that we're living in right now.

00:41:40: So we're touching the SEO, switching to Geo, I think it's called.

00:41:46: Or whatever, IEO or. Yeah, whatever you want to call it. It's true. This is a switch happening right now. But I think it's also fair to say, and some, some experts or practitioners might disagree with me, but I feel it's fair to say that 98% of, optimization for AI is exactly the same thing as optimizing for, SEO like you've been doing before.

00:42:11: Okay, so I mean, where it all we don't know where is it all going, because if there's content on the internet, I mean, not talking about because the thing is, it's slop in video form. You see, you decide to come on. It's slop. Yeah, but in text form. Even if you read it, you might find it not interesting.

00:42:32: But maybe people out there, I mean, we're talking about we had a podcast about AI, you know, but people out there, there is stuff and they don't even know it's from AI. And then there's AI reading stuff from AI. Yeah. And people are commenting it and and whatever. What where is it going in terms of.

00:42:51: Responsibility and people interacting with it, you know, because that's a big topic as well for, I mean, in general for society, you know.

00:43:00: Yeah. And it's probably going somewhere, doc.

00:43:03: If we don't take care.

00:43:05: And that's the good news for today.

00:43:09: But, I can also give the solution. Oh, good. And put them in a start with a little story. I saw a tweet by somebody who was like, it's citizens, journalists in the US, right, as a self-proclaimed journalist. And he tweeted, look at all those, those refugees coming to our country, and they say horrible things. And it had a video in it with the supposed refugee board coming by board and then saying horrible things and stuff.

00:43:36: Turns out this was AI generated. Somebody called them out, said, hey, this is AI slob. How did you not spot this? And his answer and his profile picture and his company name was something that had the word truth in it. Keep that in mind. And he said, yeah, I know it's not real, but it tells the truth.

00:43:56: Okay.

00:43:57: So you see, so this person was obviously not the slightest interested in anything that we might call objective reality. So I feel the solution for all these upcoming problems for us are to put a greater emphasis on like what is real, what is objectively happening, what does science tell us about stuff? What does statistics says? Some people might call it scientific literacy.

00:44:21: Understanding how humans gain knowledge and with this science, scientific knowledge then and this commitment to reality, I think we need to focus less on where it's written or how it's written, but who wrote it, what institution and what their track record is. Because this is a great example for why you should trust certain news agencies more than individuals with a Twitter account, even if their videos look quite, realistic.

00:44:51: But that brings me to the question. Is it actually possible to trace back if it's AI or not? I mean, it depends.

00:44:59: I mean, you have two kind of things. Do you have detectors that have a certain that can give you a possibility and you find these online. You upload a video or an image and says, oh, this is 95% AI for for video and for video and for video and images for text. It's quite hard. Like these detectors have horrible false positive rate, so don't rely on them.

00:45:24: And there are watermarks that are added by the AI tools themselves sometimes, for example, if you use nano banana it will add a secret watermark. And those are easy to spot, but they are also easy to remove. So meaning technical checks of whether something is AI or not are not something to to trust upon. When we design a future where we don't get fooled easily.

00:45:49: That's what I meant by institutions. It's more important to see who published this. And do they have a good track record of posting the truth and actually fact checking and checking alternative sources and and not framing in a wrong way? It's it's way more important that we that we have strong institutions. And we put emphasis on. Yeah, making them even stronger and making and holding them accountable for what they say.

00:46:17: And also, I mean, that's a good of course, it's a fair point. And also, I remember that you, work on a project actually in schools. Telling people, is it like generally how media is used or is it also about AI already. Because I think it's people cannot I mean even now people on social media cannot really speak up.

00:46:36: Some people don't even know they're in a filter bubble. They see post and they think, okay, look what happened. I yeah, yeah, yeah. A lot of public transport. And you hear discussions where you say, okay, listen, guys, we're not sure if this really happened. You know, just what you said. It's not. It didn't. It's not like that, but it it's the truth because people see something and then they believe it and they don't even question it.

00:46:55: If even now, even without it.

00:46:57: Especially if it fits their worldview. And this is independently of political affiliation, by the way. So people on all sides of the political spectrum are easily fooled if it fits their worldview. So, so yeah, that's why I do volunteering in schools and teach kids how to use AI. Yeah. And to how to have a little bit of media, new media literacy, understanding what's going on, then how to use it.

00:47:19: Because let's be honest, like kids, when they enter like fifth grade, sixth grade, seventh grade, they might not even understand the structural difference of putting something in a Google slot in the Wikipedia search slot or in ChatGPT. They just don't know, and somebody needs to tell them ahead.

00:47:35: There was a discussion, another discussion. I was actually about half a year ago, or maybe nine months, even two kids on the subway in Berlin. Obviously they were going on a vocabulary test in school, and one was saying, do you have to GPT? Because I wanted to check the words. Yeah. And he said, no, no, not on my phone.

00:47:51: Okay, okay. Use Google. So, so the even ChatGPT was the first address.

00:47:56: Yeah, yeah, yeah, yeah.

00:47:57: So but, you have kids, I don't my brother has. So, and I all the stories about homeschooling and Covid. So, where would you say it's a little off topic, but I would say is the situation when you meet these kids in school, what do they know already? How can the school system take care of this?

00:48:17: You know, is it is it who's who's you're not going there every day and you're not going to school. But who was who was telling these kids how to use social media? I.

00:48:26: I mean, that's an interesting question because of course, the teachers aren't always up to date. They haven't been in my time. Probably not in your time. Right? I remember when I was in fifth class teaching this the teachers how to use Microsoft Word because they didn't know how to do it and they were supposed to teach us.

00:48:43: This is what I heard it.

00:48:44: Was.

00:48:47: Is so, I think a very sharp sword that we have right now that the education system can use is just getting an external perspectives. And, you know, you still need the the teachers that know the class, that know the kids that have this background in education science, I don't want to belittle that, but it really helps to get an external voice from time to time to get an external expert.

00:49:11: There are I experts specialized in teaching in schools. Shout out also to a there's there's a certain, startup called life. Teach us that. Is that is building a connection for people that want to say, yeah, I can spend some of my free time teaching in schools and go there for an hour, and you can do the quiz quite easily.

00:49:32: Shout out to the startup. So that's probably what we should.

00:49:36: Be, I think, in the show notes.

00:49:38: Yeah, yeah, let's do that.

00:49:40: Nice. So, okay, we went a little bit, society, which is the end of it and, not the end of the society, but, that's.

00:49:49: Not yet not the.

00:49:49: Other side. Again.

00:49:51: Don't we had this when I had the podcast with with flow. I remember back in the end we had it was going about ice lying to the scientists and stuff like that. So some some months ago.

00:50:02: But coming back to our world of, consulting companies, corporations, agencies, what would you say would be the best way we had talked about, how they're using it right now or how people can use AI right now. What would you say would be the ideal blueprint for a medium sized agency slash company to use AI?

00:50:30: Is it too general? I'm thinking what I'm asking this because it's like talking about electricity. Probably.

00:50:35: I don't know. No, but I mean, it's important to always think of it from both sides, like the management of the agency, like the owner or the management team needs to top down, set the guardrails saying, hey, here I stand. I'm going to tell you what's happening. This is how I believe in you. This is how I believe the future of AI will look like, having a vision and also having clear guardrails like, this is what we do with AI.

00:50:59: This is what we're not doing. We're not uploading Excel sheets with the telephone numbers of our clients in to get GPT three version right. Instead, you get a company version, of an LLM that is approved by us. And if you want to experiment with AI, you can do it in this secure sandbox. But you're not allowed to do that.

00:51:16: So you need to have this top down framework. And like a top down basic toolkit that you give people that you're not using that to beat, free accounts. And then you also need to have the bottom up, which is maybe having an ambassador program, maybe having in certain departments like the art department, the creative department, the strategy department, people talking over lunch, what they've been using, coming up with new tools they find interesting to, to be checked from, from the management coming up with new methods, actually exchanging prompts, exchanging workflows and tools like we've, exchanging ideas, how to use it, exchanging things that have been working and not been working.

00:51:54: And so all that I can say, how should look like it sounds a lot like innovation culture, right?

00:52:00: Yeah.

00:52:01: Just means you have good innovation culture.

00:52:03: But do people I mean, if people ask you, they already are aware that they need some kind of innovation culture, probably, you know, just imagine all the companies. Oh, I didn't even call you yet.

00:52:15: Yeah. Yeah, to be fair, I mean, it's it's. Anyway, it's a huge spectrum, but you have these companies that are have super duper, nibbled and driven employees with a fire in their eyes, developing new stuff and how to pitch even better. And they've been using workflows with different video and, and, and image generating models, complex workflows.

00:52:42: They've been using those in pictures for two years now. And they have great results and they exchange and stuff. And on the other hand, you have these agencies where you have art directors saying, listen, I don't want to use any AI, which is fair, but probably not a very good position to be in in two years or one year or one month.

00:53:02: So it's a huge spectrum.

00:53:04: But also people, I mean, other people are afraid of losing their jobs. So just because of AI or what's the what's a general like mindset, you meet when you when we have workshops with with employees, I feel.

00:53:14: Like it's gone back a little bit, especially in the, and the advertising space. I feel like we were kind of over the the hurdle of the first shock.

00:53:23: Oh my God, we all going.

00:53:24: To get replaced. I just maybe it's again, it's a biased selection of my clients, but I feel like most agencies at least a mid-sized and big sized ones, they've come to an understanding of AI as a tool, and they're using the tool in it in a productive way already. I would even say, right. And of course, there is stuff that you cannot sell to the client anymore because the client knows that I can't do it.

00:53:50: But there's also a ton of stuff that AI and humans can do that I can't do, and the client doesn't have time to do it.

00:53:57: But is it the case that because I know this from the film side, that people, the clients now start to ask Pacific specifically for AI or like, you know, you did this with I know it can take longer than one day or something like affecting the timing and workflows and budgets and stuff.

00:54:16: I mean, it is it is, impacting the industry for sure, because I mean, completely independent from the reality of the technological advancement, the buying, the procurement will also always find a good reason to ask for discount. And now saying, hey, you can do everything with the AI. 30% off is a great way to do that.

00:54:37: But isn't that like a legal thing? Then I heard that that that actually in companies the bigger they are, it just it took the procurement has then to deal with the legal department because, you.

00:54:45: Know, sometimes you have.

00:54:46: Only in firm.

00:54:47: Terms, sometimes you have, like guidelines that get passed down to the agency and you need to have an underground agreement of what kind of AI using, because of course, the, the client wants no liability and somebody.

00:55:01: Has to cover the.

00:55:02: Copyright claims. That's that's also a thing. Yeah. And it's all kind of things happening. Like you have the, the risk thing and the contracts that you pass down, you have procurement saying, hey, now we have AI, you can do it for 30% off. Right. And the current market condition, you have clients that in-house things that agency has have been doing before because now one person with five bots can do it in-house.

00:55:27: What has been done by an agency before? If you think about stuff like blog articles and certain banners and performance marketing ads and stuff.

00:55:36: But that's the usual discussion that actually, of course, somebody with five bots can could do it, but he or she still has to know and to judge the outcome of yeah, that's true. Otherwise it's stop again.

00:55:46: No, but if you.

00:55:46: Think about it.

00:55:47: It's 100% true. That's 100% true. But but think about it. You could hire an art director from the agency or a creative director from the agency, and then have them implement something in-house. And you suddenly don't need the whole creative culture that an agency has, and for most tasks still needs anymore. So you can in-house certain things that you weren't been able to feasibly in-house before.

00:56:10: So you can actually have a copywriter, an art director, a strategist and, an account person. And then you they all have five agents and.

00:56:19: Yeah, and.

00:56:20: And maybe 7 to 7 agencies.

00:56:23: Maybe the agencies are maybe the agents are then contacting the agencies and all the, both sides, the agents, and they're paying each other in bitcoins. No, I don't know. It's not it's not not moving that fast. But, it is moving. Certainly. And I think, to be fair, I, I tend to see a pattern that fewer, more creative people in-house and brands can do more with smaller, more leaner agency setups, together with AI and with a pool of freelancers.

00:56:56: I think this is happening.

00:56:58: Because one of my questions to the end is actually, where do you see the biggest potential or where is this all going? So you just said, we're not there yet with this all the agency run, but our distributors also stories about dreams, about the one person unicorn. Yeah. So where's this? When are we in this state that, for example, the big company has like, I don't know, five PhDs or five freelancers with agents and they have the agency in-house somehow.

00:57:27: When are we, you know, when is it coming?

00:57:29: I mean, Jensen Huang that the head of Nvidia that we might already be there, that the billionaire, the billion dollar unicorn company could be built by. I'm not sure about that. And I want to make one thing really clear that I don't get misunderstood. I think creativity in the deepest sense, like actual human creativity, divergent thinking, and having taste is still not a solved problem by AI.

00:57:57: You need creative humans for that. And the value of those that are really creative has gone up exponentially, because they can do so much more with their creative vision. And this is something that you usually find in agencies because they have this creative culture, they have a track record for attracting creative people, and they don't kill the creative types of creative people with compliance and and just corporate stuff.

00:58:24: Right? So the creatives can actually be more creative.

00:58:26: Yes. I think that if if you are a creative person by heart and you dare to be creative and you have taste and you know what's next, you know what's in and out and everything, it's a great time to be alive because. Because you can 100 x your output that you can have as a single person steering all these kind of tools.

00:58:49: And you will have a way better output than a non creative person without taste using AI of course.

00:58:55: Yeah, hopefully. I mean it's like so

00:58:58: So this is just to wrap it up. This is why I still believe in the agency model S. I just think the probably need fewer people, fewer people to do what they're doing. And I mean, the most important thing for both agencies and client side is to hire the right people, hire those people that have this fire in their eyes, that are creatively driven, that have taste and that are open and willing and capable of learning new things.

00:59:29: That actually already, what would you advise people to do now? We're getting we're getting there because in the end, they always ask for like three practical tips. But you already have this top down thing at the guardrail thing and now this.

00:59:41: Yeah. Also, the practical tip is with the creating, a style guide for your chatbot that it doesn't write slop anymore. No. But on a company level, on a management level, on an HR level, that's the that's the number one tip. Hire the best people you can get because somebody who's twice as expensive using AI will have, of course, twice the cost of four times the cost, but they will have ten times the output in the near future.

01:00:07: How do you judge when you have, like the, applications on the table from the guys who use AI, which probably use the AI to make the applications? How do you go through this? And of course, you might have an AI judging the applications which have been done by AI, but.

01:00:22: What you shouldn't do with the AI. Careful there.

01:00:25: Yeah. I mean, I think big companies are in the air. They're already doing it. No.

01:00:29: Yes. I'm just saying there are guardrails, especially in Europe. Especially the advice.

01:00:34: Yes. Of course. Yeah.

01:00:36: So how do you find the right person? I don't know, but I would always look that's Kunta Duck quote here, which I still find an amazing person. He said, you need to watch out for the fire and the eyes, and you just see it. And when you talk to them and ask them, what is something that you have learned today, this week, this month, you will find whether they are into learning and trying out new stuff.

01:01:03: So that means you do you think that we are we can all use AI as a tool and we're not going to be run over by it if we do it right.

01:01:13: If we do it right. Yeah, but you need to look into it and you need to adopt a mindset of lifelong learning. This is because take the large language model or image generation courses that I built six months ago. I usually say six months. That's the half life, the half-life. But after six months you can throw out 50% of it.

01:01:36: Okay, so it doesn't matter how well somebody was educated. In AI, it doesn't matter what certificates they they have. It doesn't matter how many books they read on prompting and stuff like that that you don't need anymore. It's about how willing are they to learn something every day. And this is at the moment that we're in a key.

01:01:57: So if people are listening to this podcast right now and they.

01:02:03: Have to Google now what cloud is. Yeah. How bad is it.

01:02:06: It's not bad. That's what I meant.

01:02:08: That's the great thing okay.

01:02:09: It doesn't matter how good or bad you are right now. The only thing that matters is that you are willing and capable and motivated to learn. So it's wonderful you just found out what Claude is there. Listen to. Hi. I'm waving to the video right now.

01:02:23: Google it.

01:02:24: Try it, try Co-work try it. Code, cloud code. Try creating your first little mobile app. Your first. The website. All you need to do is ask cloud how to do it or ChatGPT maybe by the moment we're talking about this ChatGPT 5.5 is out and Claude is completely outdated. By now we don't know, but, just just try stuff and try it with things you like.

01:02:47: For example, if you have a hobby, your passion might be Dungeons and Dragons. Use AI for that and it will be fun. And if you have fun, you can learn nice. It's as simple as that.

01:02:59: That's a wonderful finishing statement. So I will, link all your profiles in the show notes and, the start up you talked about and your website and, what do you think? Like, you.

01:03:14: Know. No, no. I'm just. Yeah. Dig.

01:03:16: Do I get 10s.

01:03:17: For a little, plug a little Edward at work? Yeah. No.

01:03:21: And, maybe we can also link the KI-Night. Oh, yeah. The event format. That is, if we if we release this before the 11th of May come. And if you're somewhere near Stuttgart, a us. Well, if you're here by chance, you're invited.

01:03:32: It's, actually a of my.

01:03:34: Little event format about AI and business. Everybody is invited to check out the website as well.

01:03:38: It's nice. Yeah.

01:03:40: So thank you very much for your time. You look good.

01:03:42: Thank you.

01:03:43: It was a pleasure.

01:03:43: Thank you for the ad space.

01:03:45: Yeah. Of course. And, Yeah. So, I think there's hope if you start to,

01:03:53: Take the time and watch out for AI right now for everybody. And, don't be afraid.

01:04:00: To have fun.

01:04:01: Yeah, yeah, that's a point.

01:04:02: Thank you for having me, Gresham.

01:04:04: Thank you very much. Bye bye.