Ep#10: What’s Next For Product Design in 2025?

Welcome to Method, where we decode what it really takes to build and scale startups. I'm Andy Walsh, founder and strategist. Each week we'll bring you actionable insights and real stories. Let's dive in. Hello and welcome back to Startups Decoded. Today we're venturing into the world of product. I have a special guest today with Nick Sober, who has a deep history at IBM but is also the co founder of Avalifi. Nick, welcome to the show.

Thank you, Andy. Thanks for having me.

This is a treat for me because I am an advisor on a Valley Phi and I've known you, Nick, for probably close to 18 months now, I would say. And we've shared a lot through that time as far as getting a product, from having a MVP to launching to going through validation and a lot of things. So we've been in the trenches together and I think it's a perfect way for us to be able to set up a conversation today around what product looks like in 2025 because there's been a lot of seismic shifts that if you've not seen what's happened in product development with the evolution of AI, then you've clearly not reading the headlines because everyone is talking about it. So today what we want to be able to do is a little bit of a look back as far as what product design development validation used to look like and where we've shifted to now and potentially even where we think it may go to in the future. So to kick things off though, Nick, I'd love to be able to give listeners a sense of yourself and, and your background. So where did you start as far as study into your first tech job? Give us a bit of history to yourself.

For sure. Thanks, Andy. So I, I'm, I'm born and raised in Italy, Italian by design. I always say, I like to say this in, in all the forms and, and what it takes to be an Italian. I studied computer science. That's my background. And maybe the most interesting thing I should say at this point is that while studying computer science I also studied drama and a bit of singing, which was a bit of Dr. Jekyll and. Mr. I'd during, during my studies because, you know, I ended up coding in the morning and Learning by heart 60 pages in the evening to have a show in the, in the, in the forthcoming weeks. So that, that's where I'm coming, that's my background. I got my first job in tech at IBM actually and I had almost a decade career there where I sort of Started as a technical administrator, then moved very quickly into technical leadership, team leadership, service management, crit, sit situation critical situation management. Sorry and. And then I moved up into the data division and IBM Watson division started to do. I made a shift of career from service management to data, data science, data analytics. That led me very quickly to, as I said, Watson services and all the way through from information architect to, to essentially being a consultant in data analytics and also a prolific inventor in. Into. Into this area. I left IBM with the title of IBM Senior Inventor which simply means I got pretty prolific with patents solving problems for the first time in different industries, for different customers which led to new products, innovation and success.

It's a great background and I would love to have the title of something inventor. It would. It just feel, I mean there's probably a lot of expectation that comes with been that role as far as, particularly in a big company like IBM in developing something that for them has commercial return ultimately. Right. I had a guest previously, Neil Gosh, who's at 9.8 Venture Studio and he talks through his experience at Amazon, ultimately building an internal venture studio and the difference with that versus what he's doing now and I would assume for yourself working at IBM versus your own startup, there's some really sort of stark polarizations in the sense of how you operate and the ability to drive change. But I love the fact that inventories in there and patents are in there as well because as far as thinking about the future, which we'll cover today with product and they play a key role in that. So to get things started like from a product management perspective, obviously as I set up earlier, there's been a huge shift that's happened in the tools, platforms, expertise, et cetera for people to create products. And there's a lot of buzz at the moment around speed to launch. But looking back, like thinking about the formalities of a product strategy, what are some of the key things that go into that process?

Yeah, interesting question. So I would start from saying first of all to set up the context here that I believe, strongly believe that product management as we know it is over. We're not going to be back to the let's say the playbook that words worked so far because the world is a totally different place and the tools in our hands are totally different. You know, I would say I wrote many times that the this, let's say democratization, this super easy access to AI tools today, it's the same as the invention of the wheel. It's something that is changing everything. No One is going to be back to move stuff with square, with square wheels. You know, this is a round wheel and this is enabling everyone to go faster. So everything that we ever thought and done in a certain manner is changing and will keep changing with AI. So what's not going to change is probably the traits of PMs, product managers. And what's not going to change is that we still need to build products that provide value to users. These are the things that are going to stay. The rest is, I would say, all in transition, all in the evolution, and they will keep being like this for a while. To answer your question, you know, in product strategy to date, probably or better, without AI, we had to spend weeks in making researches, talking to users, validating markets, validating insights. We had to take days to consolidate feedback requirements, write the product requirements documents, we had to draw wireframes. We have to do a lot of stuff, put it in the hands of customers, do a bunch of AB testing and start from there and then start putting metrics everywhere, measuring stuff. All this somehow sort of remains true, but the pace of doing this from months will be, is getting to weeks and even days. And in the realm of startups especially, we have seen this in venture capital since 2021 onward, I would say, and since the general availability of LLMs and all this generative AI bubble and boom, I would say everything we have seen in the past years, it's been a transition in understanding how this new world is going to, is going to work and understanding how value and profit is going to be generated. And that's why we have seen investors behaving differently, founders behaving differently, the entire startup ecosystem adapting and reflecting all of this into how product strategies are designed.

Yeah, it's such a huge shift. And I was talking to someone the other day and their background is in AI and LLM and they've been doing it for 25 years. The technology itself is old. It's the democratization that you talk to there that puts the technology in the hands of everyone to be able to do things. So that changes a lot of things as far as how you build a product, but also the number of businesses that are in market with products, competing against each other, which is going to be a fascinating thing to sort of keep an eye on over the coming years about who survives that path of getting revenue, maintaining revenue, retaining customers and then scaling. So let's pause on that for a second and I'd love to sort of click in on one of the pieces that you talk to there as far as some of the tools that people are using now. So there's a lot of platforms that have a copilot solution. You can do a written brief to turn it into a digital product. What, what are some of the workflows that you're seeing in that early concept stage when you're building a prototype?

Yeah, so essentially one of the biggest change I've seen is that you can essentially start coding today, not even thinking about the low code, no code solutions, but literally by telling a story. You can start describing what are the user stories. You can start describing what are the requirements for a given feature in a given product. You can explain everything in natural language in plain language. I read something this morning around natural language programming. So LLMs that are able to translate the story into software specifications that are going to be translated into any code, the front end, back end infrastructure behind and deployed. I think we've all seen already things where you put there a prompt and you see a landing page being generated, like generate for me a software as a service landing page. And that's enough because the LLM may be strained on building landing landing pages. If you start to think about that and to scale this concept and what pertains in the realm of product management, for instance, you can do exactly the same by providing user feedback and understanding not only better the feedback, but also getting as an output the corrective action to address the feedback and how that should be incorporated into into a product. So what I'm saying is you don't even have to think about it. That can be automated. This kind of input can pass through a data flow that you architect for your product. That makes sense. You might have different channels from which you receive feedback and this feedback can be consolidated, provided to specialized agents that are capable of dissecting what are the the correct insights from there and essentially feeding the product architecture itself with a corrective action to incorporate the feature and put it there for AB testing. The same happens for the day to day work we all provide, maybe Async updates through channels like Slack or emails. And all these things can be already incorporated and become something for the product. Product work today takes less time, less Toad and probably less PMs, less product managers too. That's what I'm trying to say. We need to become an AI powered team now, if we haven't done it yet. And of course we have to think about automating ourselves. Everything we do has to become, if not fully automated, at least streamlined by 75%, 70, 80%. This is the ratio and then we have to grow as somebody who can do a tiny little piece of something else. I don't want to say that product managers should become data scientists, but maybe can learn a tiny little skills into that area as well and become the ones who are automating other pieces of this work. And so when we were few months ago, maybe, maybe it's already a year by now, Sam Holtman saying we will have a one person unicorn. You know that that is probably, I would say too much, a bit too early, but the direction there was set to say you're going to be empowered with a team of AIs, AI agents, AI assistants that can do different pieces of your work for you, with you, and you can be the orchestrator of this workflow. And the same is going to happen with a team, with the product team. Engineering, design, product. All these areas have different pieces that can be already fully automated right now. And the assembly of all this is going to lead to a completely different way to design and build products.

The democratization is interesting and my background in creative storytelling, right now there's so many platforms that can do copywriting, they can do static design, they can do video animation. It can get quite prescribed on what you can do. And in the creative sense, the democratization of creativity isn't necessarily always a good thing because you've got people that aren't qualified doing creative work that they push out and they're relying on that to do a certain function. Are there similar similarities with product design? If you don't have product design experience and you're using these platforms, are there risks in that or are they evolved enough to be able to prompt you on key things that you need to consider? If you're building a prototype in the early stage, totally.

It just depends on how accurate you want to be with your, how sophisticated you want to be with this, with this design, there are already several tools that allow you to build fully functional applications with a decent design that can be even articulated into specifics of logos and branding and have already a brand identity into it. It's not that sophisticated. And that's probably what AI shouldn't be working on. I think these are all the areas related to creativity that should stick with humans. So to certain extent, AI is creating first pass, which is building essentially the scaffolding, but then the rest of the application of the ui, of the interface, but then the rest, the user experience and the design itself that should remain with humans. And also, let's not forget about this. This is a very important, I want to make especially if we think about solutions that are LLM powered. These are what LLMs are essentially are, as I love to call them, glorified stochastic pirates. What I'm saying is they are so damn good at putting everything that has been put into their training data set and recalibrate re extract from the language model things that statistically will make sense and resonate within the context of the prompt. So what I'm saying is they can only answer to what they've already seen. And there is a technology limitation in inventing, in creating, creating things from that are, that are brand new, that were never put together. And this is instead where human excels. We're just not that trained anymore on this and we should get back to this truck. So I would say probably my background as an inventor is, is always bringing me to that direction and that's what is also always resonating with me. I know that I will never expect something new from, from an LLM and I'm always the one challenging even the tiny little details. And I know I will be the one pulling off things that were never meant to be part of a product by, from. From a statistical standpoint. But the human element in this system, system of machine will detect what is exactly the point where to anchor the novel stuff. The things that will lead the product to the next generation. The thing that will lead to the thing that a machine will never be able to come up with.

Yeah. And exploring that a little further. I'd love to look at the accessibility and democratization means that more people have access to create these things, right? Depending on the platform. If you have 10 people who create companies, they all see the same problem in market and they all decide to create a product in a business to try and solve that problem. Product management 101 is you need to solve a problem. You need to validate that problem and make money solving that problem. Like really simple foundations of product management, right? If you've got lots of people doing that, is there ability to have something that is unique as far as a feature set, et cetera. If all those 10 people were using the same platform, would they ultimately build the same product? Or can you be very custom with how you present your version of solving that problem?

Well, what creates there the, let's say variability in the system is the way we are asking questions. There are people who are very good at asking questions. There are people who are very good at triggering outliers outlier results instead of the expected from a statistical standpoint results. And so this kind of diversity of us using even the same brain, it's like we are querying to the same brain. But if we ask the question differently or set up the context differently, we are likely to get either similar responses or totally different responses. This is statistically almost, almost the same. So the point what, the point I'm trying to make is we should use these tools to help us think faster, like a debater. We should ask questions or we should provide ideas and get feedback on the ideas and use it more as a tool that can help us reasoning better, think faster, retrieving data in the context and maybe collaborate with other people. With an AI in the loop, this is probably better usage of this and that's how the democratization of this should I say be incentivized to be used. I can easily see this if I think back to my time in IBM and the challenges we were solving. I would have loved to have an AI tool in the room to interrogate with some, some of the nasty technical questions instead of spending time going through documentation to understand if we could have done something, we could have asked for this kind of help or we could have asked like, based on the conversation you've, you have recorded so far, what is that we haven't mentioned or are there any weaknesses in our strategy? Is there any point of view we haven't considered? You know, this is the way, the correct way for me to, to utilize AI. It should always be a collaboration. It's never brain outsourcing. It's not where what is what, what has to happen. We don't have to stop thinking. We have to keep driving the conversation. We have to keep the control of, of the things and just avoiding this. Yeah, this, this full outsourcing of our thoughts. That's not what AI is meant to be.

Yeah. And I can picture that scenario now in an office, if you've got someone who's maybe a fresh graduate out of university and they're in their first job in their first year and they use AI 100% because they don't have any experience and then they put it forward as their own work and someone's like, well this doesn't make any sense. And it's like, well, AI told me you can't trust it. To your point, there's from where it's sourcing the data to the output and it's, I heard a great example the other day where someone talked about it like a golden retriever dog where it's just got this like this native desire to want to please You. So it's always trying to be like positive and find the right answers and get something for you. But sometimes there's not an answer and it'll get to a point and I use it daily for me to be able to collaborate like you talked about. But there's times it will present something and I'm like that is way off and you have to course correct it. So if you don't have the experience to know what's right and what's wrong, you can also go down a bad path.

Totally talking about golden stuff. You know, it's not a golden hammer. And we should not be the goldfish of this, of the story. We have to get, we have to get back in control of the critical thinking and to make experience on our own. Especially for juniors and fresh graduates. It's as you said, the most important thing is exactly to. You have to go through a period where you just do stuff and hit the wall because that will build you. That will build the resilience. That will, that will build the experience. Simply. There's nothing like oh, can you. Can you can you had in my, in my head like a USB key with all the, all the, all the lessons you have given get in. In. In your lifetime. No, it's not matrix yet as far as I, as I know. So we can't transfer this to other people or, or better books should meant to be, should be the, the, the thing we used to do this because they should be a 1, 1, 1 transfer media. Like there's no, there's no teacher teaching you something. You're reading the book and there's no intermediary intermediary there. That should be the tool to help us navigate a tough period on a certain topic on a given topic or learn a new skill, whatever the book is, for instance. But AI is not the substitute of that. That instead if you remember, I remember a very famous speech by Steve Jobs where he's saying like AI will allow us to ask Plato for certain questions. We can read his book, we can read whatever he left. And the power of AI is that we can maybe keep asking questions, you know, and that's exactly what this should remain. This should remain a tool that amplifies what we have already learned in different ways. But it should not be an outsourced activity of the critical thinking because that's what we are doing willfully now.

Yeah. And it's buyer beware. Right. It's, it's definitely has an effectiveness layer to it. There's a sense of robust confidence that it gives you, because it is looking at such a large volume of data in such a small, small period of time that would take you days, weeks, months to do the same thing yourself manually. Right. So there is that sort of level of confidence that comes with it, but there's also that level of naivety too. And I'm not sure if you saw the experiment. It's something that I loved. Reid Hoffman, the co founder of LinkedIn, did an experiment where he created a digital twin of himself. And then I think he. I feel like it was the. He did it the year that he founded LinkedIn and basically trained it only on information up till that date. And then he had a conversation with his digital twin, with himself today to sort of look at his point of view, look at his insights, look at the things that he knew. And it was just a fascinating social experiment into how at a certain age he had X amount of experience. And today it's very different. And obviously the success of LinkedIn speaks for itself. But that for me is a great way to think about AI in the sense of like you can train it on a certain amount of information, but then there'll be a point where it stops and it doesn't have that full rounded approach. And if you're a researcher, you always want to have that level of significance of 0.05. Right. I did a lot of research in my university days and you want to have that fact. So people are doing shortcuts. So there's a level of taking it for granted that it's true. But then you also have to validate it to be able to make sure that what it's saying is right. And then how you get customers to buy your product off the back of that information.

Totally, and I didn't know about this, but it looks a very, very good usage, very smart usage of these technologies. This is what it can help us do. Retrospectives, maybe in that case on himself. But it should never be the tool for, let's say, envisioning the future. It can help us think about the future, but it shouldn't be the one designing it. And what I want to have as a last, I would say pitfall and thing to consider into this as a human element is also that especially if you're young or you focus too much on using those tools, what happens is you're simply becoming lazier. And that doesn't have to happen because if you getting back to the product management and thinking about users using AI and users expectations when they're using products and the way the users and customers are using those tools. You know, this abundance of choices and this abundance of options and I would say easy way of going through life and going through solutions in an easy way to achieve their tasks is going to create an effect in which the abundance of choice, the abundance of those tools and the way the tools are being used will create almost even a decision paralysis at the point that when they will be asked even for feedback or for what's the feature that is missing, they wouldn't know about that. Probably their first, their first thought would be, let me ask AI, what should be the next feature? What do I want? What do I need? What are the options I can provide? How can I sound smarter? You know, there are a lot of nuanced things that are weird thing, profound thing. There are things that are happening right now because if we haven't noticed, we are, we live in a world where for the first time we have things four to five different generations using this kind of technologies and they are all using it in, in their own unique ways for doing maybe the same thing or similar use cases, but they have different expectations and they have different experience and they have different way of seeing the world. And so what I'm saying is all this contributes to the mass of the, what the product manager should be doing today in this landscape. And in thinking about products, thinking about product design, commercialization, marketing, everything is impacted by how people are currently using AI and how AI can be embedded into everyday tasks and life.

Yeah, one of the things that you said there jumped out for me and I think that there's a real risk inherent in the fact of you can create an echo chamber in the sense of if you have a certain point of view and you're asking and prompting AI in a way to get feedback on something like you would in research, if you have a loaded question and it's steering someone in a certain direction rather than a neutral open ended question, the validation of that as well, you're basically asking it to prove something right rather than prove something wrong. So you can skew that yourself. And then depending on the data that it's trained on, you can also get a biased point of view as well. Like you're assuming that it's well rounded. So I think that having that sort of education and not getting lazy and always staying on top of what's happening in the world across competitor products, et cetera, is very critical. Right?

Totally, totally. Because this kind of, there's no AI without data and if the data is skewed by this kind of biases, this kind of issues, or I would say as you said, echo chambers that are being created around this. You know, we're going to get everything wrong and we're going to automate the wrong stuff. So what should remain? It's always this kind of human interactions. Of course I said before, technically this is feasible and at scale maybe you don't get these problems. But especially for startups, especially for the early adopters, especially in context where you don't, in scenarios where you don't have a lot of data, it's super important to understand what to measure, what you're looking at and making sure AI is infused in the right steps. It's a double edged sword. You really need to have cognition about how users are using things, how you're using it and removing from your analysis all this context behind what you're seeing.

Yeah. And you touch again within that. I would love to be able to look at scale and I've talked a lot about this previously. If you look back in history, history can repeat itself. And I go back to the dot com era and the dot com boom and in the late 90s there was a lot of money being dropped by investors into dot com companies. Some of them were six months old and getting insane amounts of investment at that point. And obviously there was a huge cliff that happened because the world wasn't ready for the volume of businesses there. And I fast forward to today and there has to be a level of risk associated there too. Right. If you're an investor, you assume say 10% of your portfolio makes it through. So you know there's a risk. But if you're a consumer and you're using one of those platforms, if you bank your whole product strategy on a business that is brand new, is reliant on investors or funding or retained revenue to be able to grow, there seems to be an inherent risk that that business could disappear in 12 months time if they run out of money because there's a competitor that offers a better product or has better brand awareness and people buy on that. So do you think that people are thinking about that risk in the sense of like who they choose as a platform where they build?

It depends from the generations and the experience. I would say first of all it's important to say that talking about scaling, probably we are. There's never been a better moment in history to build stuff or better. We are equipped with everything we need with just a laptop and desk and an Internet connection. We have already everything it takes to make first steps to start A digital business, for instance, and we have access potentially to all the tools to build whatever. But then scaling, scaling that, it's a different, is a totally different challenge. And it's also, I would say, hyper difficult to do. So one is because you really need to find the right things to measure to understand if this is something that can really scale as a founder, as an investor, as even, even as a consumer, when you're betting your own money to start using a new tool. That's what I'm, that's what I'm doing. At least I'm trying to understand like, okay, if I start using this brand new tool that I've seen just on right now, on Instagram, will this be around in three months from now? In three months from now? That's what I'm asking myself. And most of the times I have to say I don't think that will be around in three months from now. The reason is because I've already seen signals of that into other products I've already seen generally available somewhere else. What I'm saying is people are probably more risk takers right now because they can quickly build stuff at very low cost if you have the skills and a few friends that are, that are also good at what you're doing are complementary to your, your expertise. But scaling will be a totally different thing because the market in which you're competing is 10 times what it was 10 years ago, five years ago, three years ago. I don't have the numbers. What I'm saying is there are far more startups right now than a year ago, two years ago, three years ago. The reason is exactly because of this and even for VCs is a hell of a challenge to identify the winners into this mess. So it's literally a totally brand new word. It's a brand new word. We all have to adapt, we all have to remove signals, noise from signals, understand what's, what's going to work and, and making sure we are not outsourcing the critical thinking. Like I said before, we really need to have strong business insights that are saying, I'm betting on this because there is this signal and this trend and this other thing that will align very soon and will create a totally new market for the future around community building. That's also another story and that's also another skill, I would say, attracting the right people that are willing to stay. I'm also for instance, reading a lot of newsletters, reading a lot of, entering a lot of slack channels, discord channels, but I'm not Staying in all of them. So how do you attract the right people? How do you retain the right people? Over curiosity over longer period of times, this is a challenge.

Yeah. And on product management, like historically, there's always been tension with the team's expertise and what customer feedback says. Right. If you listen 100% to customer feedback, you might end up creating a Frankenstein of a product based upon what everyone's telling you that they want. When it may not be the best path for you as a product strategy leader to think about, okay, what problem are we solving? What feature does that? How do we build a product that's sustainable and continue to grow? So in the world of AI, that dynamic feedback loop that you get through really quick validation from customer feedback, if you're a product designer, product manager in those different roles, you need that expertise. Right? So that tension of dynamic feedback today versus historically is very different too.

Totally. And you know, among the things I think that will not change in the foreseeable future around product managers is that great product managers are always opinionated and probably principled as well, but for the right reasons and on the right stuff, they do have a drive towards a vision. They do, they do have a vision. They see the product of the future, they see the things ahead of the game, they see what's, what's coming and all they want in their, either they're conscious or, or not is essentially driving everyone towards that. So traits of a product managers are, they are opinionated, stubborn, probably principled on the, on the good stuff. They have to be leader and let's say, be able to drive not only the team, but also the users towards a certain desired outcome, let's say. And they have to be great communicator, therefore, so that they explain things clearly in a few words. And probably the other trait that will never change or should never change, is that they, they need to be doers, not only thinkers. So these are the things that will remain in the hands of product managers. They are doers who can communicate well, who can convince people to go in a certain direction, and they have to be stubborn enough to make it. All these things are things that AI won't have because they will trust the data, trust the feedback a hundred percent of the time without understanding the context of why that feedback was provided, because we don't always have that information. But from a human standpoint, we do grasp a lot of stuff that are not written or the signals that are not captured by text only or transcriptions only, which is the typical source of feedback that AI is using. So what I'm saying is human elements remains even in feedback, interpretation and for, for, for a very good reason. We have to grasp the human element in everything, in every aspect.

It makes me think of a book called Pattern breakers by Mike Maples. Within that book he talks a lot about the founders who are real visionaries being able to see into the future and think about creating a product that serves a future need. And there may be a problem starting to bubble up now that they're trying to solve, but it's a future state. I look at something as simple as Canva as a great example. Like when they started, they were competing with Adobe, which was like the pinnacle for anyone who's designing anything of any nature. So to be able to eat market share from those guys. Fast forward to today where the democratization again of creativity is almost a necessity, an assumption for people working. So that kind of thinking and planning ahead is really important. And AI can only give you data up to today, depending on what it's been trained on, et cetera. It can't look into the future. It can use data to try and glean some kind of pattern that may happen in the future. But you have to have a certain mind state to be able to think outside what everyone else is thinking and, and create something for the future.

Totally. Totally. And sometimes they might not be the ones that are believed to be right immediately. Those, those faults, those leaders, those patterns, breakers, typically they, they are right a lot later on. And there, there's a, there's a, there are different ways to get there, I think, and they are all super contextual to the product being done. It's about making sure the others can see the very same future, the very same trajectory of growth of a product or the value of the world around them, or how it, how it will be the world different with that version of the product, with that feature, and how that changes completely the landscape surrounding them. I think we have discussed something similar in the past when we, we made an example around like Uber would have never been possible without having GPS feature into our smartphones. So the real breaker, the real, let's say pattern breaker was to say, okay, let's put localization into these devices so that we know where we are on a map. And in the moment that happens, you can start asking what are the restaurants nearby me? Which is something that today we give for granted. But until the GPS wasn't in our phones or devices, you know, that wasn't possible. So there are things that are enabling other stuff and those things Are most of the times seen as, nah, you're crazy the first time you hear it. They don't make sense. Like, why should we make use of a military technology into our, into my phone, I just make calls, you know, but then you see this backwards. Some of the points can only be connected looking backwards, you know, and everything makes sense. But on the spot, maybe something doesn't make any sense or is not that widely accepted or acceptable.

Yeah. And I've heard and seen a lot of stories from VCs who have like sort of hitting themselves in a way that they missed out on an investment opportunity for a business that was presented to them, that has gone on to make billions on exits for them as well. Because they're not seeing the vision the same way the founder is. And sometimes it takes a certain way that your brain is wired to have that future vision. And for a vc, if they don't see that same vision, then you don't get that investment. And I think investment is a really curious thing to be able to look at too because if you're developing a product in a much shorter time and getting customers earlier, you're driving revenue. All of a sudden the requirement on capital changes significantly.

Yes. And you know, to connect to this, we are visual animals. We need to see things. That's why we say the vision, because we have to see things, we have to envision things. For we do have this powerful thing called language, but it doesn't always work. And think about my. I'm thinking about myself for a moment. I'm not an English native speaker. Every time I'm trying to explain something, I had to make an effort into translating. So I might not be using the best term possible, but just the nearest that is in my mind that helps me just to convey the message. So one thing that I would say, AI is probably going to enable us to do better or my wish is that I'm gonna, I will be able, I'm already able probably with AI to convey the perfect shades of an idea just by telling a story in my own language and then convey that fully and completely to somebody that is a native speaker in Korean maybe without knowing a single word of Korean. And not only this. One of the ways, I must confess I use a lot AI is to generate analogies. Because I, I, I believe analogies are very good, are very good tool to explain things to different audiences. So most of the times I, I struggle to, to concentrate a, a difficult technical concept into a short sentence. So I just switch into the analogies realm and Most of the times I find myself asking AIs, asking different LLMs for instance, to come up with potential analogies to explain that concept better. These are things that can help us see, envision things as humans.

Yeah, it's good advice for founders. If you're thinking about telling your story to a VC and your concept is very future state and feels potentially innovation for innovation's sake rather than a legitimate idea. Being able to have an analogy is a great way to be able to take someone into your own world pretty quickly with something that they are familiar with. Right. So I think that is really strong advice if you're thinking about that pitch state. So just to close, we had a conversation before we jumped on to record today about investment stages and what's happening and we touched on it a little bit there. As far as getting a product to market, getting validation, getting early customers who allow you then to be able to get momentum, what do you see as some of the trends in the coming years as far as investment requirements, exit timelines, et cetera?

Yeah. So to recap on what shared before building things today is easy, scale them is not. And so what I, I see and what I expect is that when there is a scaling potential that will manifest almost immediately in the very, very first weeks to months from launching something to the market. So first of all what I'm, what I'm envisioning is founders coming up with an insight, with an idea, coding, building that right away with all the tools that are available, with all the communities that are available. It's pretty easy to find even co founders if you have none. And in the moment there is an idea, there is a very small team working on something, they would launch something very quickly to the market, iterate super fast and in the moment the idea is validated, is already working. And so that's already very good moment to the next generation of venture capitals to enter the game, enter the picture. So what I'm envisioning is far less exits at IPOs or merging acquisition in later stages as we have seen in startups, and a lot more of super early exit. That's, that's what, that's the way I see it. So you build something super fast, you attract the interest of investors very fast and there are faster exits. Not all the rounds that we used to see precede seed series A, series B till IPO or merging acquisition. You can already Google this precede is dead. Most of the people are already saying that there is, there is no more pre seed funding, they will directly be seed and then probably you don't even get to serious a because the target for everyone would be to exit the business as soon as possible.

Yeah. So it's interesting the flow on effect of a change in technology with a change of how you make revenue, how you build a business and there's lots of positive flow on effects for that as well is from an investment standpoint it feels like there's a lower risk on a business that is already cash positive and starting to get momentum. It has early validation and you can get to that information so much faster than previously. Right. So that creates one element. But if you get early exits like you're forecasting there, it also means that you've got exited founders in market with capital that can then reinvest and help nurture the next generation of leaders and entrepreneurs as well. So it's a very positive flywheel effect that I see in that and I get excited when I start to see growth in innovation and changing because the last few years have been challenging across the board for investors and for founders and I think we've turned a corner in 2025 and I'm excited for the year ahead. And on that point Nick, we've actually at time I say this every episode but I could talk for hours and hours on every topic that I have. But I've really enjoyed talking to you today Nick and appreciate your insight. It's been very fascinating.

Likewise only thank you so much. Just want to conclude saying that this in this potential second third order effect of having founders with capital in their hands and using entrepreneurship as a vehicle to solve world's challenges. Well the future might be far brighter than we expected or we have seen in the past years. So looking forward and looking ahead to.

This brighter future, a great way to find finish up as far as thinking about a positive world and solving problems and making our lives better which I think we've lost our way over the last two decades a little bit and getting back to positivity, community, connectedness in a genuine way, I think there's a lot of opportunity there. So Nick, thank you for joining. You're listening to Startups Decoded. We drop episodes every week. Please follow Share with your friends. Give us a poll Positive rating. I love these conversations. Thank you for joining us today.