Filip Dřímalka AI for Independent Professionals: Becoming Superpowered

Nigel Rawlns: 0:37

My guest today is Filip Dřímalka He wrote The Future of No Work And his first appearance here became the show's most downloaded episode. Since then he sold both his companies and now he runs his business through AI, finishing his third book, Superpowered, On a staycation in Costa Rica. The gap keeps widening between people who work with the best AI and everyone else. Filip calls it a parallel world. Most of us use AI as a muscle. Do this, shorten that, write this email. Filip uses it as a brain. Here is my goal. Here is my data. What do I do next? That shift is the difference between using AI and and becoming a superpowered professional. He's my one hundredth guest. This is the Wisepreneurs podcast. I'm Nigel Rawlins.

Nigel Rawlins: 1:24

Filip welcome back to the Wisepreneurs podcast. The last time I spoke to you was an episode sixty-seven, fourteenth of February last year, and you were in Sri Lanka and traveling. This is the 100th episode, so I've brought you back because you're my most popular download. And four of the five top downloads, believe it or not, are from Brno where you live. So welcome back. Filip, just to remind people of who you are, can you tell us something about yourself and where you are today?

Filip: 1:55

Yes. Okay, so thank you, Nigel. Thanks for having me and congrats because I think 100 episodes is really amazing. so good good job. So my name is Filip Drimalka and currently I am in Costa Rica surfing and finishing my third book. So this is what I'm trying to do here. and basically what I do is that I help people become superpowered. That means I helped them to use technology and I helped businesses to discover how technology, especially AI, nowadays can help them to do better work, to redesign what they do and how they do it.

Nigel Rawlins: 2:37

You know, reading your last book, in there you said, I wrote the book for people who enjoy work and who are always seeking new ways to do it better. And I think that's really, really important. I think The Future of No Work gave people the idea that you don't have to do any work. But really I think what you intended was seeking new ways to do work better. And I I think that's really, really important.

Filip: 3:01

You you you know what some some people say that the name of the book should have been the future of too much work. Because what we found out is that when AI gives you new capabilities, people who love what they are doing, they they do more. You know, so so this the I I can now develop new feature and new service and maybe you know create this and that. So basically the the whole No-Work concept is not about not working, but it's about doing something that we like, we enjoy, we are good at good at. And when you do something like that, the work doesn't look like work at all. So I think most of the principles, I would say almost all the principles in the book, you know are still valid, but I think what what we did not imagine is that how AI gives us really So many new options and new possibilities that we should be aware of what not to do. So this Jevons paradox, you know,

Nigel Rawlins: 4:06

Mm.

Filip: 4:06

that really really tend to be true. And I think even the new book I'm writing is for people who who just love doing what they what they do and they want to do it better.

Nigel Rawlins: 4:20

Yes. We're going to get into that a bit. I'm I'm tossing up whether I should read you this quote from 2010 actually from a fellow called Mason Cash about extended cognition. And I look, I think I will at the moment. It's a bit of a word thing, so have a little think about this. He's talking about an autonomous, intelligent agent. Now for us it's a human, is one who Critically engages the cognitive tools around them. See, think about that, the cognitive tools around them. One who selects or endorses and uses effective cognitive tools. This is our thinking tools. This was 2010, this guy wrote it. And then who replaces, refines, or augments less effective cognitive tools, and who selectively incorporates these social, relational, technological, environmentally, and bodily resources into their sense of who they are, what they know, what they want, and what they can do. And that's exactly who you are. I mean, as an agent, you have taken, and I'm sure you've gone through a number of different AIs. I I know I have and I'm, God, what am I on to? About my fifth one, and I'm using multiple ones now. You know, it's changed who we are as humans.

Filip: 5:46

Yes. Yes, absolutely. I I think AI expands human capabilities. And also I think AI is the very first technology that can, you know, even teach us how to use it and it can improve and it can be proactive and it can I I think it is much more than we thought. and that some people, especially those who are using just the basic AI tools and maybe even the free versions, they cannot imagine what AI is capable of. And one of the quotes that I really like and this is I think really connected to what you said is one of the founders of Anthropic the company that stands behind Claude Jack Clark he said that in summer 2026 People who work with the top AI models will feel like they live in a parallel world. And I think we are there. So I think that when you work with the top tools and models and you redesign the way you work, you can imagine and you can compare your experience to the quote that you just mentioned. But for some people, it's just still some, you know, some chat, some chat in in. Google Chrome that can just you know make email shorter, but this is very short sighted. And I think we already live in this these parallel worlds, and I think this will be this will have huge impact on on businesses, on organizations, on on people because we already see that when when when people use the best tools, they become impatient, they don't understand how you know how come that everything is so slow

Nigel Rawlins: 7:35

they become impatient, they don't understand how you know, how come that everything is

Filip: 7:40

with with the the the others. So yes, you're right. And I think now AI really takes most of the comments you you mentioned. So it you know most of the Chat Bots we are already there.

Nigel Rawlins: 7:55

Yeah, look I I've I've moved from Chat GPT and basically all I was doing was prompting there. You know, you put in a prompt and you'd you'd find a prompt and work the prompt. Then I went to Claude and then I noticed you made a comment. Now you speak in the Czech language LinkedIn translated, which is really

Filip: 8:14

Yeah.

Nigel Rawlins: 8:15

nice, and it said, Claude Code and Co-work, and I thought, what's that? And then I got stuck into that and built a whole lot of skills. And then somewhere I saw somebody talking about Hermes. And I use that within Obsidian, and we won't go into all that. So I asked my Hermes this morning, how am I using you? And it it basically summarized it it was saying that AI was something I picked up and put down. Now it's staff, a process, and an environment I work in. From wrench to workshop. That's what it said. So it told me a whole new things. But anyway, we should get into some of the things that have been happening with you. when we last spoke And you were travelling through Sri Lanka with your daughter. Now you've sold both your your Digi skills and AI ability. You wrote the book on the future work, then you lived it, then you exited. What did your actual experience teach you that the book didn't capture when you did that exit? You know, was it easier? Was it harder? Was it different when you exited?

Filip: 10:02

Okay, so so I exited Digiskills I still have I still have stake in AIbility So I'm still part of the team, although it has two co CEOs, and basically I'm not part of the internal team, but by I'm still talking about the strategy and product with the co CEOs. But what I found out is that I really love building things. And I consider myself as a creator. So what I don't like is deals and investments. And I I under I completely under I totally understand that some people like it and then they enjoy it to negotiate and prepare a strategy for you know selling the stakes in the company and buying companies and so on. But I found out that. It's it's not the game I I like and I enjoy. But of course it gave me some freedom. and I I really now like working individually or in very small teams, and I I believe that this is the future of work, really small teams with AI, because then everything is fast and the road from idea to execution to reality. It's just, you know, it it's just a couple of days or hours or even minutes. So nothing is slowing you down, no corporate processes, no red tape. And this is really something that I I I just like doing now, building new things.

Nigel Rawlins: 11:40

You know, it's interesting. There was a thing I read about normal traditional firms, the e economics of it in the past, it was much, much cheaper to have it all in-house because it was cheap. But

Filip: 11:51

Uh-huh. Yeah.

Nigel Rawlins: 11:52

now that we've got internet and we've got extended stuff, we've more got what we call nodes where you can sort of link into somebody who's an expert there and somebody there. So the cost of actually being a one person business or having, you know, several people in your business is so much cheaper today than running a big organisation that requires buildings and HR departments and all these

Filip: 12:16

Mm-hmm.

Nigel Rawlins: 12:19

different things. And that's what you've just said, that you can do that fast and it can scale if it needs to be. And that that's huge.

Filip: 12:26

Uh-huh. Yeah.

Nigel Rawlins: 12:28

And th that that research is nineteen thirty seven saying, you know, that's the benefit.

Filip: 12:31

Mm-hmm. Yeah. Yeah. And and and I think you know, you know, we we can see examples in nature. you know, when any animal is big, it requires more and more energy to to you know sustain the living and so on. And I think it the very same principle applies for corporations. So the bigger the company is, the more people just need to work on to Put everything together. And also, there is there is a law called Metcalf's law that means that the more people you have, the more connections you have. But it's not like five people, five connections, it's exponential. So there are so many different ways how I need to run any project when I work in in a in a large company, and especially now with AI. I think what's really now easy is to build one-person system where you basically build a system where you connect your AI to different apps and so on, and you build your workflows and automations and so on. When you have two or two and more people, it's you know it's much more difficult because there is no one single way how to do it. There is no one best practice how to do it. Everybody is just now discovering. How to build a second brain for a team and so on. But for large corporations, this will be super difficult. So now I really appreciate that I don't work in corporation anymore. And another problem I have now with large companies is that we help these people, we help these organizations to adopt AI. And the trouble is that they offer their people just the basic tools, sometimes even free versions like Microsoft Copilot.

Nigel Rawlins: 14:23

Mm.

Filip: 14:23

And they tell them you need to adopt AI. And here is Microsoft copilot-free version. But all these people, what they hear, what they see online is this Hermes and ChatGPT Codex and Claude code. So they think that there is only one AI. So how come that all these people are talking about agents and my copilot cannot do anything? But the trouble is not that you know they are frustrated, the trouble is. That when these people will go to another interview, they will be completely lost. We already see that the differences between people are enormous. So that's why we really encourage people, no matter what your employer gives you, no matter what they allow you to use, you really need to learn how to work with AI. In private life, you need to pay for AI from your own pocket. Because the next time you go, you you come to an interview, you have two options. You can say, I'm sorry, but my employer just you know gave me simple tools or I couldn't do anything. Or you can say, at work I couldn't do anything, but at but at home I made agents for my finances. Now, you know, with my wife, I built this website and so on. So, you know, this is now really troubling, and I really encourage people to think about their career. Differently than before because it's their career, it's not their employer's career. So, you know, I I, you know, got to a different direction, but this is the difference

Nigel Rawlins: 15:56

Yeah.

Filip: 15:56

between smaller companies that can use any tool and they are using the top AI tools and models, and large corporations that you know need to work on governance and so on. And you know, that's why I like working in in a very small team now.

Nigel Rawlins: 16:12

that that's what I love too. I can't think of ever working in an organization again. I I can see how I can mm fit in, but I I prefer to work with one person businesses anyway. But one of the dangers I am seeing and and I had a client do the other day, they used Claude to tell me how to redesign the their websites. And unfortunately that turned me into what we call a vendor, which anyone can do. And so I was a little bit annoyed about it, but it doesn't matter. I need to come back to them and explain that, you know, if you've just put this stuff into Claude, it doesn't have the context that we've been working on for the last ten years. And and because I'm into positioning and how an independent professional positions themselves, you know, you do have to be very careful. You just can't chop and change and then stick it on your website and hope that people will actually understand it. Because you've got to

Filip: 17:07

Mm-hmm.

Nigel Rawlins: 17:07

work with clients. And I spent several years building up the understanding. And then suddenly Claude says, Do this and they want it done. So I'll come back and rework that with that particular client. But that's the danger, is you can run with it and it's going to mislead

Filip: 17:22

Yeah.

Nigel Rawlins: 17:23

you.

Filip: 17:24

Definitely. And I think you know the the most important word is the context, right? Because I was now working on my own not only positioning but also ICP like I I ideal customer profile. And I built a system called PACT projects agents context tools, and it's a Basically folders on my Dropbox, like something like Google Disc, completely controlled and organized by ChatGPT Codex because I was I was using Claude 99% of the time. I spent you know just this year $12,000 on Claude and Cloud API. And last month for last month I have been working with the ChatGPT Codex because it is now for me the best tool ever. But it doesn't matter what tool are using. What's more important is connection to your data and then building your system of the second brain. And when I'm talking about context, I talked to our new marketer about the ICP. And what I did is I asked AI to go through all the transcripts of my workshops, presentations, to analyze all the feedback from our clients. in my programs, what we do is that I collect feedback what people are working on each week. So I have thousands of messages from people, and we ask AI, but it got the response from all these data because it's able to code the analysis. And you know, it came up with amazing description of what I do, how How I do it, what customers appreciate and so on. So I think context context is something what differentiates very good result from something that looks good, but it is not connected, is not personalized to what you are really doing.

Nigel Rawlins: 19:29

Gee, that sort of context is pretty powerful because if you ask just anybody who's a professional to sort of explain what they do, it's often hard because they can't examine that. But you're getting that sort of feedback from masses of people and that's really important. that's positioning and and really clarifying your position. That's very powerful.

Filip: 19:49

And one more tip. you know who knows you best nowadays? Your AI, but not in terms of just talking to AI about yourself. If you use Claude, there is a there is a feature called memory. And one of the really eye opening moments I had was when I opened the memory, and it was basically. The way I work and think condensed into six paragraphs. So we people think that we are super special, we are unique, and that AI will never be able to basically work the same way. But you can describe yourselves in a couple of paragraphs. But now what you can do, especially if you are using cowork and chat tools. You can analyze all your sessions with AI. And I think, you know, this is actually what you do now. This is how you think, this is how you create, how you build. And sometimes it's really interesting to even export your ChatGPT or Claude conversations and ask AI to analyze it to tell you not only who you are and how you think and how you work, but how to become better. You know, so so this is this I I loved it. And before, I would say a year before, the data analytics was my weakness. I was not patient to to learn about data analytics and all these tools were super complicated. But now I can ask AI to do the research about how to analyze anything, and then it codes, it programs the analytics. So basically now what I do most of the time, I just analyze data and I play with it. And this is this is I I I just love it. It's you know I have so much fun with it because I just control it with my natural language. So this is what I recommend. Talk to AI about what you do, put together everything, transcripts, sessions, discussions with your colleagues, and then talk about your positioning and ICP and everything in between.

Nigel Rawlins: 22:10

Mm, that's a very powerful approach. Okay. You've said your whole operation runs on IA and AI and hundred and thirty plus agents and recent projects built without writing a single line of code by hand. That sounds like the future of no work in action.

Filip: 22:30

Ha ha ha.

Nigel Rawlins: 22:32

But you know I'm sometimes worried that AI agents do break. You know, they break.

Filip: 22:40

Yeah. Yes, yes, you are right. actually now I have less agents because since the introduction of skills, some of the agents became skills. And now I say I would say that I have like 10 or 15 agents, and this is also something that changes that we should be comfortable with the fact that. We will need to change the way we work every couple of months. So I had 130 agents, and then with the latest models and skills that are now part of each you know good AI tool. We just analyze all those agents, and AI said, okay, but you need 10 or 15 agents, and then most of the stuff can be just skills. So, but but back to your question. I one of the Chapters of my new book, it's called 90-10. And I think that people who seek to have AI teams work in 24-7 without any human intervention, it's not working. We are not there yet. But we are still we have a very important role, and this is the last 10%, this is the last mile. So I think that. You know, AI can do the 90%, maybe 95%, maybe 99%, but I'm still the one who approves or who says that this job can be done by AI without human intervention. Like for example, today I saw a LinkedIn post from one of the startup media in the Czech Republic, and they said. We introduce events. We have a new page about events. So if you have any AI related event, just send us an email. And and what I did, I just took a screenshot and just pasted it to ChatGPT and said, okay, send an email about our event. And then you know I I know that the top model will analyze the email. What is it about? And it sent perfect email and I don't have to check it out. But of course, if I prepare important strategy for a client, I Work with agents all the time, but I'm still the one who is in charge. So, yes, agents break, but if you and these are the top three things that you should have if you want to have agents that don't break, or they don't break most of the time, is the top tool and top model. So for me is now the GPT 5.6 and Claude Opus its connection to data. And it is the system and workflows. And when I talk about system, you can design the agent so it knows where when to act, and it knows where to just give you, you know, a question. Filip shall I continue? So I think it is all about building a system that prevents any agents to break. And if agents don't know. It should tell you, I don't know, but this is really about the system design.

Nigel Rawlins: 26:02

Now we should explain to people what we mean by an agent and skills, 'cause I think they're two different things, aren't they? Yeah.

Filip: 26:07

Yeah. And this is interesting because, you know, agent can be a prompt, it can be a text file with instruction. Agent can be a tool that works in agentic workflow. So there are different descriptions of what agent is. But the I think the most important description or definition is it's an AI that plans, then acts, and in the reflects. So, from my point of view, it's the combination of the tool with this agentic workflow like Codex or Claude and so on. And then it's just the text description, like job description of what the agent should do. And skills. Skill is a repeatable workflow. So let's say that I would like to create short videos out of a podcast. So I would like to Have a subtitles in our design. I would like to add a logo. and I do it every single time after a podcast. So what I need to do is that I do it for the first time, and then I just ask AI to create a skill. So the next time AI does it autonomously with the very same result because the skill is the definition of how the result should look like, and so on. So it's just repeatable workflow. It's like human skill, right? So you do something and then you might improve and so on. So is the combination of the way the AI tool works with the definition of what it should work on and how. These are the agents and skills.

Nigel Rawlins: 27:50

Yeah, so the agent's like a little robot with a little job and

Filip: 27:54

Yes.

Nigel Rawlins: 27:54

the skills are the master master document behind it. So where do you stick your skills? 'Cause I've got about well, I've I think I've got about a hundred and twenty skills tucked away. I use mine in Obsidian so it can

Filip: 28:07

Mm-hmm.

Nigel Rawlins: 28:07

come and go. And obsidian is is saved in the cloud, so I'm not gonna lose them if I have them on my computer. So where are your skills kept?

Filip: So I have everything except of the web projects that are in on GitHub i I have everything on my Dropbox. So Dropbox is a file system, it's cloud file system like SharePoint or Google Disc. And there I have everything: 28:16

text files, spreadsheets, and everything in between images. But what's important is that I don't go there most of the time myself.

Filip: 28:45

Because it's the system I talked about, it's called PECT, and I have PECT assistant in my ChatGPT codex. So imagine that someone, some some of my colleagues sends me a folder, like Filip here is the new campaign, and there are some images, there is there are descriptions of the campaign and so on. So, what I do is that I ask my PECT assistant to analyze the folder and to put it You know, to to marketing. It knows where to put it in it, you know, creates a description and so on. And it it incorporates the folder within my system. So the next time I ask my agent about the campaign, it knows where to find it. So basically, I don't upload, organize and do anything with the folders and files anymore. My agent does that on my behalf. And this is this is the biggest difference between tools now.

Filip: Because there are three categories of tools: 29:43

chat, cowork, and code. And co-work, I think that it will not exist in a couple of months. But chat is about talking and it can create a document and so on, but it's for some basic, simple work. But the coding tool it can work with your local files or with the text files on Obsidian. So basically, I can have AI to organize.

Filip: 30:11

everything in my folders and that that's why I like having everything on my Dropbox because it's local, it's on my computer, and I just know how to work with that.

Nigel Rawlins: 30:23

You know, it's what we call distributed cognition that we're using a range of tools. But you're still in charge of all that, but you've you've given agents the job to extend your mind in many ways. So and I'm hoping people are hearing this, as you know, if you're an intelligent professional, you've got a whole lot of information. Well, sorry, information and knowledge

Filip: 30:44

Yes.

Nigel Rawlins: 30:45

in your head, but this can take you so much further. Okay,

Filip: 30:49

Yes.

Nigel Rawlins: 30:50

you've got a superpowered professional assessment. That gives people an archetype, a score, a certificate. Now tools do change. now your AI impact data, shows only about two percent of professionals are at agent level maturity. And, you know, tools are changing fast. So what skills matter most and how are human roles changing? And what

Filip: 31:16

Mm-hmm.

Nigel Rawlins: 31:16

what do you think separates the casual AI user from the superpowered professional.

Filip: 31:23

Yeah. Okay, so we now measure we now measure how people work with AI, and we have four levels of how people work. So the level number zero is what we call AI Explorer. So explorers are people who use AI, but they don't have any system, they don't have any repeatable workflows, they don't know What are the differences between models? So sometimes they use fast model for even larger operations and they are frustrated with the results. So basically, they use AI as a as a smarter Google. And the the second level, level one, is operator. And operators are people who use AI, they know most of the features, they know how to prompt, they know when it's better to ask AI to write a prompt on their behalf. The level two, which is basically three, is builder. And these people it's about building repeatable systems. Though so they have GPTs like assistants in ChatGPT, they have projects in Claude they work with the text files as you know you you describe, obsidian and so on. And the last level is transformer, and these people completely transform the way they do things. So I give you an example. many people in corporations and even individual professionals they create presentations. So an operator, what they do is that ask AI here, I have a document, create a couple of bullet points, and then they copy-paste it to PowerPoint or Keynote. A builder.

Filip: What they can do is that okay, so maybe I can have AI to program to code a presentation. And the transformer, what they do, they they completely differ, for example, they build an automation that after each conversation with a client, AI automatically creates presentation with the with the outputs and so on. So these levels: 33:11

explorer, operator, builder, and transformer.

Filip: 33:39

They are valid no matter what tool you are using, and even those tools progress, the what changes is the benchmark. So we worked with a company, we analyzed around 300 employees, and after one year, we did the assessment again, so we re-assessed them. And even though they work on AI adoption, we they organize some hackathons and workshops and so on. They had worse results. And when they asked us why, it was because the market went just faster. So the the the way they work is that they train people, but they didn't give them the tools. And after one year, all those other tools on the market they just progressed, you know, so much. So it's really about benchmark, but no matter what tools there will be, there still will be people who will be builders and transformers. So this is what we measure now. Still, there are just you know, you know, 2 or 3% of transformers in organizations. And there are, I would say, around 5 to 12, 13% of builders. So it's a huge opportunity for anyone who learns AI as a skill and who becomes builder because you know the the demand for these people. Just skyrocketed. So, you know, this is why I recommend professionals to to learn AI because when they combine their experience and knowledge and skills with AI, you know, this is this is this is where the gold lies. But when we talk about skills and roles, I really think it's the combination. It's I think the most important skill now is curiosity. Because you you you need to be able to play and to want to play with AI. If you have experienced senior programmer who is not interested in AI, and you have you know freshman coder who is in AI most of the time, in a couple of months the junior will be better than the senior. We already know that. So the curiosity and willingness to learn. This is, I think, the most important. And then I would say that it's really about finding way how to talk to AI and building your own system. How you prompt, how you document your knowledge so AI can work for you. So it's about system thinking. And the best part about these new tools is that you can improve them just by telling them what to do. So imagine when I talked about the the podcast publishing, we have that you know 100% in AI now. And what I can do is that I can ask AI, okay, so maybe every Friday go through the comments on my YouTube and create a summary for me. And then after a couple of weeks, I can then okay, that's good. So maybe every month create a summary of all those comments and create a recommendation. For my next podcast. So the more I use AI, the more ideas I have, and the more I improve the system. And this is, you know, the the improvement, this is amazing. So this is what now superpower professional can do. They can build the system, they can improve the system, and it's really about documenting my knowledge and building systems where AI can work not only on my behalf, but it can help me and serve me to do better work.

Nigel Rawlins: 37:30

And the most interesting thing about this is they're actually doing their job as they're using the tool.

Filip: 37:36

Yes. Absolutely. I I think you know it now your work is the classroom. the only way how to learn how to work with AI is learning by doing. There is no other way. And you know, before we said that when you want to become an expert in anything, you need to have 10,000 hours, you need to spend 10,000 hours. Now I believe with AI it's ten thousand iterations. You know, because all these

Nigel Rawlins: 38:05

That's a good point.

Filip: 38:06

experiments they give you better idea how to use it next time and next time, and it's really it's really the snowball effect, because then you have more and more idea. So the the more people work with AI, the better they get and then they improve the way and they expand the capabilities because they know how to use it for and how to push it further.

Nigel Rawlins: 38:27

I just couldn't believe it the other day and I said, Can you make an I video? And it did. I've still gotta work out the sound and the the things like that. But yeah, you can just ask it and it will either will tell you or not. But this is in Hermes, so I have a

Filip: 38:41

Yeah, yeah.

Nigel Rawlins: 38:43

a I I guess a a particular model will do that for me.

Filip: 38:47

maybe one more example. in in in codex or Claude code, you can just ask for automation and it just builds itself. And I would say limitation of AI before was that it has a pretty short memory. the context window, you know, is pretty wide. So you need to build a system around that. And what I built is that I built something that I called Work-log and work -radar So, what it does is that it goes through all my sessions each minute and it puts together what I'm working on. And it creates a work log that means what changed in every single project. So the next time I ask about, okay, so I just want to work on the book, it knows that the last time I I was working on a chapter number 25, and you know, and this is something that I I'm just asking. So now maybe you know, create a summary of. Every single session and create a Work-log And then every Friday create you know, analyze all those Work-logs and tell me what should be improved, and then maybe add some recommendations for automations because and then you know it you just ask and ask and ask, and it improves in time. So, of course, now ChatGPT announced a new feature called computer history. Because as we already discussed, it's all about the context. But you know, I just built it a couple of weeks ago or a couple months ago. And of course, maybe in in in in the future it will be part of every single tool we work with. But I believe that this system building gives me a huge advantage because I can fine-tune it for myself. You know, and I don't skip it because the more I use it, the more ideas I have. I think for every professional, it's really about imagination in their field. And this imagination is now, I believe, the maybe the if the most important. It's maybe even more important than curiosity. When I was writing, and sorry, maybe I'm too long, but when I was writing the Future of No Work, I love quotes. And there were two quotes I wanted to use. The first one is from Thomas. Edison, and he said that the genius is 1% inspiration and 99% perspiration. So it's about doing the work. But on the other hand, Albert Einstein he said that the only limitation is our imagination. And I was always fighting, like, okay, so which one is valid? Is this the inspiration or imagination, or it is the hard work? And I think with AI. It's really about that, you know, imagination because then AI can do the work. And of course you still need to sit and wait and give it direction and you know, give it a feedback. So it takes time. But I think AI does the heavy lifting now.

Nigel Rawlins: 41:55

Well, I I think the main thing is to have some expertise in in the first place so you know and I'd I'd put it a couple of different ways. One that you know that the output is actually good, that

Filip: 42:08

Uh-huh.

Nigel Rawlins: 42:08

you can actually evaluate the output. But the other one is to be a little bit careful about not letting your skills rust because the danger of of letting AI do everything is there's less Well, there's less of a particular area. But that's where you've got to judge whether you can do that or not. And I was thinking what you've just said is that you individualize your own apps in many ways or agents. Whereas in the past we'd go online and we'd download this and we'd download that and we'd have to

Filip: 42:40

Yes.

Nigel Rawlins: 42:40

learn it. This means, okay, what do I really need that's gonna help me and design my own little app just for myself?

Filip: 42:48

Yeah. Yeah. And the the evaluation is critical, so you need to know what good looks like. And yes, skills are rusting. And I found out when I started to write my third book. And it was super difficult to write. It was super difficult. I just you know, most of my stuff and emails and Slack messages messages and so on are written by AI. I just you know Talk, I I talk a lot. I dictate, so I dictate idea and then it writes in my tone of voice and so on. So I really lost my ability to write. And I needed to study to write again, and it's really difficult. So, yes, this will be something that we need to be aware of. but about these little tools, this is interesting because you know I'm now writing the third book and I built like five or six tools about that. So one one of the tools is that I read the book, the the first version and second and third version, and then I dictate my comments. And I have so many comments in in Google documents, and I dictate and so on. So what I did is that I put everything together to AI and I ask AI to build a an app. Where all these comments will be, and I will just you know dictate if to you know make it to to the next chapter and so on. So I built a system how to work with comments. Also, when I when I prepared my way to to here to Costa Rica, my flight was the first one was two hours and the second was 11 hours. And I knew that the internet is really bad. So what I did is that I asked AI. To go through everything, all the notes, all my sessions and everything, and build a database of stories and use cases and examples that I want to show in my book. And it built a catalog of comments that was searchable and everything. So I was able to work and it really helped me because I just I just asked my you know my system this catalog, and today what I did is that I wanted to run. some of the chapters and I wanted to discuss with my editor. But before what I did is that I asked AI to go through my Google disk and Google Docs and analyze all conversations I had with my editor in in comments. Because the last book what it you know we did it traditional way. So basically my editor went through the Google Doc And he put a comment and then I replied and he said, no, okay, so now it's good. And we had hundreds or maybe even thousands of these comments. So what I did, I asked AI to analyze it and to create a comprehensive document about how my editor thinks. So basically, what I did is I I I created a virtual editor built on how my editor thinks and works. And it was amazing because when I g gave him the first chapters, it really gave me very similar feedback how to prepare the you know what to change, how to make each other. So before I give it to the real one, I can work with the digital one. And I have you know different tools for different things. And it is amazing that it really the writing is still, you know, my job. But it can help me with everything, you know, during the process of of writing and editing.

Nigel Rawlins: 46:50

And see people are beginning to hear where you can go with this because one of the things we have to do as professionals is write and and put our knowledge and insights out there and this is a continuous thing. And if somebody's just starting out as independent, it is a couple of years of work. So this is continuous writing

Filip: 47:08

Yeah.

Nigel Rawlins: 47:09

ev even posting. Posting on LinkedIn, for example.

Filip: 47:11

No.

Nigel Rawlins: 47:12

Now AI does help me but gee, I have to spend a lot of time editing it. So I'm still not happy with with the output. Okay. I wanna move on to b you know come. Yep.

Filip: 47:22

Just maybe last comment, last note, I stopped using AI for writing LinkedIn posts. just because I want to have some friction. And when I was getting back to writing, I really needed to write. So for me, that was one of the decisions I made, and my LinkedIn posts are now 100%.

Nigel Rawlins: 47:45

Okay. I think that's a good approach. I mean that's a that's a good way back into taking back control of your writing. So what I was thinking about was how independent professionals structure their practice. Now, Charles Handy in nineteen eighty nine described companies. This is like what, forty years ago? No, it nearly forty years ago. And he said companies then This is pre-internet, they'd have a small professional core, have a contractual fringe of outside specialists and a flexible labor force. Now with AI, it changes things. And I I call this the Neo Shamrock. And and how I think a independent professional should structure their own practice. So I give it four leaves. What only you can do, this is your sovereign core. What Filip can do, the talent you bring in, the AI that extends you, or the automations, and the community that keeps you sharp. So there's four parts to it, I see, and that's very different from what Charles Handy said. So

Filip: 48:52

Mm-hmm.

Nigel Rawlins: 48:52

your dominant leaf looks like all your agents, your vibe-coded projects. And you'll fail. FAA so I should say this, we've got to describe this. It's not somebody failing. F A I L group. So

Filip: 49:09

Mm-hmm.

Nigel Rawlins: 49:10

what I'm curious is about how do you see those?

Filip: 49:14

Uh-huh.

Nigel Rawlins: 49:14

You know, is it just a program cohort, say your fail group, or

Filip: 49:18

Uh-huh.

Nigel Rawlins: 49:18

something you rely on yourself?

Filip: 49:20

okay, so so I I need to explain. Fail is the abbreviation for future AI Leader. And it started as a fun because I like it. Because what I say is the real fail is saying that I need to I just need to find time for learning AI after three years or four years when everybody is saying that this will be the biggest change that we experience. The real fail is when you slow down your team because you don't know how to work with AI. The real fail is when you ask your colleagues to do things that you can do yourself with AI. So we, you know, it it's it's not a fail. The future AI is an amazing program, it's an amazing community. And this year I completely redesigned because I vibe coded a learning platform. So just for the listeners, the Future AI Leader is a 10 weeks-long cohort program when we learn, you know, from symbiosis with AI, from learning how to work with the data, how to build any digital tool you can, how to create content with AI. So it's for people who really want to learn how to work with AI, or for managers and leaders who want to know. What to want from their teams and what are the possibilities. And this year, what I did is that I started, you know, I was always thinking about okay, so where to put videos and you know where to put notes and where to store presentations. So I decided to build a platform, and it was amazing because I was showing people how I did that. And anytime someone came up with an idea, I just asked AI and implemented that. So now the platform is. it's a portal where people see the program, they see the content, but at the beginning I ask people to fill in a questionnaire. And after they fill in this questionnaire or survey, AI creates a summary. I call it a system prompt for each participant, what they work on, what are they what are their goals, what are their levels of skills and so on. And then after each module, AI generates hyperpersonalized and individualized reports. So imagine that you go to a training when someone talks about the let's say, for example, prompting or how to use a tool. And then one hour after this session ends, you received a personalized. So, Nigel, you work on this and podcast and this. Based on what you just learned, these are the ways how you can use AI. And people can find individual prompts. So they don't even have to think about how to change what I said during the training because they have everything prepared. And someone said at the beginning, it would be really interesting if we could somehow share the use cases and examples. And I said, okay, so I built Something what I call collective intelligence. So every week I ask the participants two questions. One is, what have you been working on? You know, and with what results, and what are you struggling with or what do you plan to do? And then AI collects all these responses and puts it in one place. So there there are thousands of these examples and scenarios. So imagine you work in sales, and what you can do, you can go to this collective intelligence and say, okay, so how people are in sales are using this? And then the system gives it to you. So now I think I have the smartest portal for training, and I built it myself just with AI. And for me, this FAIL, this future AI leader is lab is a lab. What one person can do with AI is the community and People say that it really changes lives, not only because of the content that is now hyper-personalized, but because they are part of that community of people who are really open-minded and they want to change things. And especially when someone works in corporation or larger organization, not everybody is, you know, built this way. So it's really different when you work in an environment where Everybody or everything is slowing you down, and then you just become part of something that even a crazy idea is a great idea because you can do it with AI. So, you know, it's it's cohort-based learning, but it's I think it's more about the community and it's about showing people how I work, how I use AI, because most people they say that this is the biggest value that they get, that they can you now look. How I work, how I build things, and that inspires them. So now Future A Leader is the most important stuff I do. We'll see how the new book will turn out. But it's something that I really enjoy because I can do anything and I can do experiments. And even if the experiment won't work, people know that it's an experiment. And I know that some companies and some people they They want to have everything hundred percent working in their training programs, but I have the privilege to experiment because it's about showing people that the experiments is the part of the process. So I can play and I can do anything there and people just love it.

Nigel Rawlins: 55:26

Now that's a very different way of looking at training because I think when you think about learning and development in an organization, it's learn and pass this, whereas this one is continuously interacting, but it also gives you a huge amount of information about what are people thinking

Filip: 55:41

Yeah. Yeah.

Nigel Rawlins: 55:42

about, what are people doing? Are are you finding things in there that you wouldn't have thought of yourself?

Filip: 55:50

I gave the same data to all participants. So anyone can download the collective intelligence in markdown files or in CSV. And some people are using that. So they download all the anonymized data and they have the same data as I have, and it gives them a huge value. But what I found interesting is that There was one one one person that sent me a feedback. And he said, So after the second session, I just saw all these things and I felt frustrated because it just felt that it's so far away from me. And then I got mad at you because you were fast and you were explaining these things, and you were sending me all these personal things, but you know, it just didn't work. And then I Get mad at myself and I spent a weekend, and it was amazing what he was able to build in that weekend. And it's not a one case that people really need to understand that if they are lagging behind, they need to invest more time. So they need to not study but work with AI during evenings or during weekends. And I tell them, if you are lost in what I'm saying, markdown file and the model and harness and so on, this is not IT language anymore. This is new business language. And if you don't understand what people are talking about and what MCP is, then you need to make yourself a tea or maybe open a bottle of wine, sit down, upload the transcript of the session and talk to AI about that. And this is the biggest difference between people who are now power users. They invest more time into these experiments. So these small you know examples of people who just decided to invest more time and then they just you know just launch a completely different direction of their of their career. This is something that I always enjoy reading. And there is always someone who sends us an email saying, Hey guys, I really this is Too advanced for me. And there's just, you know, I I the last cohort was 550 people. So you know, it's pretty big. But the last one is just it was just one person. And we wanted to give her the money back. And I said, you know what? So maybe I will just, you know, help you a little bit. Just to maybe understand the basics and that you can learn with AI and so on. And she said, yes. So let's do that. And we always have we call that AI bootcamp, a holiday session where we meet in person. the first very first hundred people, we opened it for 200, so two sessions. And when I saw her, it was lady in her 50s. When I saw her, after like six weeks of this training, where at the beginning she was completely lost, and then she was sitting with a group of people, and she was so enthusiastic about. The way she, you know, built her assistant and so on. I was so happy and so proud that she did not give up. She did not give up.

Nigel Rawlins: 59:17

She did not

Filip: 59:19

And she said at at the end, she said, that completely changed my life because I really felt that it's not for me. But this is this was just a belief. This was not a fact. And she overturned it. And you know, I just love these examples and these stories.

Nigel Rawlins: 59:35

Yes, I think in your Future of No Work is the mindset. And we are talking about mindset here. But what I wanna now ask is you've predicted that AI is gonna shrink whole departments by twenty to forty percent within a few years. And your AI impact data is showing that sixty six percent of professionals are paying for AI tools out of their own pockets because companies won't. Now if that continues on What work's gonna survive? Not in the next five years, but say in twenty years, and you've now got a a teenage daughter, what's she gonna be doing when she's your age?

Filip: 1:00:14

Nigel, 20 years. I cannot predict five years. I cannot, you know, think about five to ten years. It's you know, I I I cannot even predict the next year. But Sam Altman recently said that he was too optimistic about the speed and the pace of AI adoption. And the the society and organizations have you know so much inertia and it will be slower, but we will not go the same speed. So there will be smaller teams that will be much, much, much faster. And I think that when I was talking about parallel worlds, there will be parallel worlds not only in terms of people and humans, but also organizations and teams. So we will see. You know, not not a two speed business, but maybe multiple, ten speed, and there will be AI first organizations that will be super fast. So when I think about the future of my kids, I think that entrepreneurship will be super important and agency. And agency is that I don't wait. Agency is that I find a problem, I deal with it, I resolve it, and it's done. Agency is something that my kids will have, I'm not afraid of their future because they will always find a way how to do anything. And if someone comes to our training and says, an HR and says, Okay, so we have

Nigel Rawlins: 1:01:54

Mm.

Filip: 1:01:55

a people here and they don't know the basics. And they need to learn the basis basics of ChatGPT. I always say that it's not a problem of their digital skills, it's the problem of their agency. Because after four years or three years, to not know anything about the most important tool, how come? So the agency means that I don't wait for HR to come up with the training and I just find out myself. So I really think it's about. Agency and entrepreneurship and also about talent. The whole book, The Future of No Work, it's about doing stuff you are good at and using your strengths and talents. And I'm not pushing my kids to use AI all the time, but I really want them to use their talent. And if it will be anything, I think that this is important to. really allow people to find out what their strengths and talent is or and what their strengths are and then you know let them to use that and I think this is the best future that they can imagine.

Nigel Rawlins: 1:03:08

Well, one of the biggest dangers I've just seen recently is some research which showed that kids who are using AI to do their homework when it comes to an exam can't do the exams. You know,

Filip: 1:03:18

Yeah, yeah, yeah.

Nigel Rawlins: 1:03:19

that's the biggest danger. They don't know much. And that's where

Filip: 1:03:21

It is, it is. Yeah, yeah.

Nigel Rawlins: 1:03:23

agency is you've got to know something. Okay, one of the things I've I've noticed the other day, and especially since I've been using Hermes, is I I clocked up a huge bill. Now, when we're talking about pricing, we're talking American dollars. You know, there's research out there now says, you know, the price of delegating is a black box, is how much you're going to spend on on your AI. and a cheaper reasoning model, now I use a number of models, it can actually cost a lot more than you actually know because it burns through a lot more reasoning than than say a stronger model. So this is gonna be a problem. you know, people who are starting out, as long as they stay with a cheap model to start with and don't keep putting money in, which is the biggest danger, is they have to learn about what that data or what that

Filip: 1:04:15

Yeah.

Nigel Rawlins: 1:04:15

model's actually charging them. So the price, you know, on the thing doesn't actually match what the work costs. Now some people are probably going, What the hell are you talking about, Nigel?

Filip: 1:04:26

Yeah, exactly.

Nigel Rawlins: 1:04:28

But what do you check before you delegate to

Filip: 1:04:30

Yeah, yes.

Nigel Rawlins: 1:04:31

AI. Now, with

Filip: 1:04:31

Yeah.

Nigel Rawlins: 1:04:32

my Hermes, I keep an eye on the actual cost. Because I might put in 150 US dollars and suddenly I'm down to twenty US dollars and I say, what happened there?

Filip: 1:04:41

Mm, yeah.

Nigel Rawlins: 1:04:42

And you know, some days it might only cost me a dollar a day because it's doing automatic stuff. So there you go. What do you check before you delegate something to AI?

Filip: 1:04:53

I don't check anything because I like to work with the best models, and only when I do some high volume tasks, like for example, when I told you about this personalization of my program. So when I build a workflow and I say, okay, so now analyze the transcript and create create this personalized reports. So the smarter model. Then gives the assignment to less smart models that create, for example, the first 10 reports. Then the smarter model reflects, create a feedback, and basically it's it's kind of a loop or parallel work when the model is like an organization, right? So you have a manager that is sometimes smarter, but then it has some, you know, the manager has some, colleagues, some people of the in the team, and some of the people are. You know, not so smart, so they do some you know less demanding work and so on. So it's really similar to organizations and to humans. And I really want to work with the best people ever. So I want to work with the best models. And I always say that if you want to earn three to four thousand dollars each month and more. You should be able to spend 100 to $200 for great AI. So now I use Chat GPT Codex when I have where I have $200 Plan Pro. I also use a Claude with the $90. And before, before Chat GPT Codex and Claude Code, I was using Cursor, and Cursor was just burning because you could pick different models. But I just I just loved working with Opus with the expensive one, but it created value. So I didn't mind spending you know thousands of dollars each month. But now I think with the ChatGPT Pro, you can pay $100, you can do so much work, and then you can use another tool for some less demanding, like Hermes or even Cursor with Grok. That can do you know so much work. So it's really like working with a team, and you need to find out how these models work, what they you know, how to make a good fit for the task that you are working on. But now I think for the majority of knowledge workers, spending hundred to two hundred dollars is is a must.

Nigel Rawlins: 1:07:36

definitely. And two hundred well, this is two hundred US dollars. So in Australia that's three hundred Australian dollars. So it does get a bit pricey, but that that's really we have to look at that as an investment to superpower you. And that's that's the next part of the topic, is superpowered. You know, you that's your next book, isn't it? That, you

Filip: 1:07:56

Yes, yes, it is.

Nigel Rawlins: 1:07:58

know, you're you're going to have to some s spend some money and delegate. But that was the interesting thing you were talking about there. It can use multiple models to do things and and that's what I do. You know, my base model does an assessment and then I ask it, okay, I want you to ask Grok it's opinion. And then when I want you to write it, I want you to use Claude Opus to rough me through and tell me how it goes. Then compare them. And and one of my tools is to compare, yeah, every two weeks, check the models, what's been happening with them? Has it been updated? Should we use that particular one? And that's all automated But let's get back to superpowered. And and we can finish on this because we've

Filip: 1:08:36

Well.

Nigel Rawlins: 1:08:37

been going for a little while. Superpowered. So this is your new book. So who are super powered professionals? And more importantly, how are we going to become one?

Filip: 1:08:47

That's a good question. So, superpowered professionals are actually builders. People who are building things with AI. And building is basically creating system where AI expands their capabilities and it just works for them. So it can superpowered can be used in life and at work. And it's really about building a system that knows you, creates context. So it's really not an AI, but it's your AI. The book is different than the last one, it will have a shorter essays, and it contains the most important principles from my experience how to become superpowered professional. So, of course, it's about experimenting and discovering and so on, but I think there are different parts of being superpower professional. The first one is a mindset. So again, again, the mindset is the most important part, but it changed a little bit because of the latest models and so on, because they are more capable. for example, one of the principles and one of the names of the one of the chapters is called muscle or brain, because many people they use AI as a muscle. So do this and then they ask, you know, five things. So, you know, now translate this and now make it shorter and so on. But I want to use, I want to use AI as a brain in terms of this is my goal, this is my intent, this is these are my data. So, how can we, you know, what should be the next step or analyze my intent and tell me what to do, and then I just approve this workflow. And this is a huge shift in how we work with AI. Because before, of course, it it was a muscle. Translate, make it shorter, write an email, but now it can analyze the data and even the AI labs, they say that with the smartest models, you should give it an intent and a goal, and then AI finds a way. So it has these principles like muscle. Or brain, or this, you know, 90/10, how to do the last mile, or how to build the second brain, and it's a series, series of principles that are just sort of essays because there is something that I called idea file. And idea file is an idea that is not super specific, and you can just give it to your AI, and it helps you to execute it. It's really about these principles that if you put it to your AI, it can it just just leads you and guide you through the process to and and helps you to implement it. So it has now 31 chapters. It will be not that big like The Future of No Work, but it's about the most important principles. When you are working with AI. And the the subtitle is How to Think, Work and Lead with AI. So it's really about different way of thinking, different way of working. And then the last part is about creating value and leading people, leading AI, no matter what you are leading, but even become a leader in your field. Because if you s when you said that AI is investment, I think you should be ambitious.

Filip: And $200 is nothing. You your ambition should be: 1:12:29

if you become superpower professional, you should be able to earn twice as much money as you were earning before. So this is the goal to not only to earn more, but to pick better projects, more interesting work. So it's not about automation, but the whole concept is about doing better work.

Filip: 1:12:56

And doing the work that just wow your clients.

Nigel Rawlins: 1:13:00

And it comes back to that quote I I I read earlier. You know, you wrote the second book, for people who enjoy work and who are always seeking new ways to do it better. And

Filip: 1:13:09

Mm-hmm.

Nigel Rawlins: 1:13:10

that's that's just fantastic. And that's just wonderful. at this point, how would you like people to connect with you? when's the book coming out and Does it come out in the Czech language first or will it be published in English as well?

Filip: 1:13:25

it will be published in English and then in Czech. it's another experiment. To be honest, I'm not sure about the whole book market in the future. Tim Ferris now published a pretty dark article about how the sales of his books just goes down like 50% each year, because People don't want to read about getting in shape. He wrote The Four Hour Body. It's a book about how to eat healthier and how to get in shape and so on. So people don't want to read 500 pages. They want to ask AI. And the benefit of AI is not only it gives you the advice from the book, It's more personalized and people just go to AI and ask things. So I'm not sure about the success of the book, but I I just feel the urge to to to to write it. But I wanted to write it in English first because I remember all those discussions with my editors, and we spend weeks discussing if this word should be you know translated into Czech and so on. And now I decided that the world is my playground and it is even easier for me. with AI to write it in in English. So it will be English first and I hope that it will be published in October.

Nigel Rawlins: 1:14:51

Okay. Well I'm looking forward to that. I will be purchasing one.

Filip: 1:14:54

Yeah.

Nigel Rawlins: 1:14:56

So how would you like people to find you?

Filip: 1:14:59

Yeah, so so I'm pretty active on LinkedIn, but I have my newsletter that is going to be launched in English in September. So if you go to drimalka.com or if you go to get getsuperpowered.com, then you can sign up for a newsletter and you can receive one of the first copies of the book. And also I'm on Twitter. And my handle is @drimalka so my last name. I was thinking about changing the name for the book, but then I decided that the the my name is so you know so unique that you can if you just type Drimalka it will always find myself.

Nigel Rawlins: 1:15:42

It does. You show up. That's great.

Filip: 1:15:45

Yeah.

Nigel Rawlins: 1:15:46

Well, thank you for being my one hundredth guest.

Filip: 1:15:50

my god. Let's celebrate! Thank you very much. it it's been an honor, and I really appreciate what you are doing, Nigel.

Nigel Rawlins: 1:15:59

Thank you very much.