Miniseries Tech for Good, Episode 4: AI and Agility

Welcome to empires of innovation. We're diving into the worlds of technology, education, impact and more with the dynamic Ozan Sonmes. Ozan has an impressive background ranging from finance to environmental studies, but is most well known in the startup world as a mentor, accelerator builder, corporate innovator, and entrepreneurial educator. Throughout this podcast, Ozan will be pulling on his decades of experience living and working as an angel investor, program director, and more throughout the Middle East, United States and his home country of Turkey. For this series, we're bringing it all together to talk about Tech for a Better Tomorrow, where we question the impact that technology is making today, how to fund and grow more of this impact in the future, and how founders in this field of technology can make it all happen. So let's jump into this episode. Welcome back. Happy Sunday.

Happy Sundays.

How you doing?

Good. Busy after Tunis, but a lot of travels. So of course, life I chose.

Yeah. Well, it's great to have you back for another episode. This is the miniseries that we've been working on about Tech for a Better Tomorrow. And today we're going to be kind of shifting gears slightly to talk about more kind of relevant topics to just the world in general. It would be wrong to really talk about developing solutions for existing world problems and not really talk about the resources that are there to support it.

Right.

So I think we have to address the elephant in the room, which is artificial intelligence, AI. It's a lofty word or lofty phrase, but it can really mean anything to anyone. Someone might think that it's computers taking over. Someone might think it's robots manufacturing something. But we're going to talk about it in the context of this miniseries, Tech for a Better Tomorrow. So, so I think we should just dive right in. I think first and foremost the most important question is what's your experience with AI up until this point?

So almost these days, last year, me and my team, we decided to use Friday afternoon for one hour. We booked in our own calendars to search what is happening in AI. We paid for ChatGPT and other tools. We said, okay, we're going to get the free trials and see what's happening. Every one of us was trying it in our own ways. Some of us are have tried in notion. We tried on Karma, we tried it on Mid Journey, we tried on ChatGPT and probably in about a month. And I kept reading about it. I mean this is, everything's around technology anyways, science fiction technology. So I, I, I'd already been super involved with trying to understand like Metaverse and The everything else around it, including crypto. And then after a month, my reaction became, okay, this is curious and useful to something more. Like, we are totally. Like we're. We're dead. We have to. That panic sets in because it became smarter and smarter with every conversation we had. We. We tried to feed it with characters like, this is me, this is how I act. This is. These are the videos that I have spoken. Then I think we started trying 11 labs, 11 AI with voice and then with images, with everything. And now we will just say like, this is not going to work out if we don't incorporate everything we do and if we don't become an AI first service provider there will be completely irrelevant in under two years. Of course we're exaggerating in a way, and I'm happy that we figured it out as well, because what we were looking for, how we were training ourselves, how we were taking this seriously, was maybe one in a million in every. People. Like, we thought everybody was freaking about this. We thought everybody was going to do this. We thought everybody's going to follow this. And people we've spoken with, our clients included our partners and everything else, they were okay. Like they were trying to read their. But they were not investing anything, so it's okay. Maybe we shouldn't panic that much because not everybody is trying to figure out with a belief that we have, like, this is going to make us irrelevant. And then we came up to another conclusion, like in steps. We realized nobody can beat us to it because we can actually detect when AI is bullshitting only because we have the experience. And we realized people will make fools out of themselves if they only trust AI and AI output, especially ChatGPT, Bard and everything that came after it. So we realized very quickly that things were getting as output in with requests of our own. We have the filter at a really strong level that we can filter out the bullshit. We can filter out the stuff that they're not making sense. One example is we had a client that we were creating reports and we tried to do this with AI, we tried to do it with ourselves, with partners in the US And AI started feeding us research that did not exist with results that are pretty convincing to our thoughts as well. So I said, okay, oh, somebody actually thought about the same thing. And they actually have a academic journal about it if we took it. And we were like, I haven't heard about this. I should have heard about this. Because we're in that field. Like, we're actually looking for everything possible on multiple sources and books and everything else. It's like, how come we didn't hear about this? And that skepticism is one we earned over the last 10 years. And then we looked into academic databases and everything else. Although it says there are references, it doesn't exist. So okay, now we said, okay, people will start making fools out of themselves. Unfortunately, that includes companies as well. And fortunately for us, we know what we're looking for. Like we will understand the impact of results and outputs and suggestions and we can filter it. And that made our life very easy and that kind of settled our panic a little bit more, thankfully. But now we are competing AI first service provider. Everything we do does go through currently not with Bard mostly, but with ChatGPT we're creating tools with ChatGPT we're writing things, having any AI tool, we use multiple of them in different circumstances and then having it filtered out again with AI. And we actually have a pretty cool presentation of how anybody can start using it and probably will share it with the show notes and everything else. We started presenting it in global events we recently did. My partner Vidiz and Alperan have created this really nice presentation showing people how you can double check AI with AI. So that's really actually very smart thing to do with it for everybody. So we are, we are developing ourselves and our tools Voice tool is going to make us really, really like impactful because what I can write, then my AI can voice out and it sounds like me very well. What it will add as well to our abilities is this language thing. It doesn't work quite well yet because when we tried our texts in English, our words or videos in English to Arabic and we sent it to our Arabic friends like native speakers, they said that's a very funny Arabic. We tried it with Italian, they said that's not how you talk in Italian. So the translation is not working yet, but it will. Like at one point we're going to be seeing a lot of really well translated designed videos and I'll be speaking Italian in a very funny accent or in Arabic in a very probably Lebanese or Tunisian accent with a little French involved. So. But yeah, I mean, to us last year had been a transformation from how can we create content, how can we do videos, et cetera. Like, how can we use AI to create content, how can we use AI to create videos, how can we AI use AI to build better reports, build better insights and kind of cross check ourselves with whatever is happening? And it's, it's good because people who are in the forefront of the technologies, somebody Will need to teach them how to use it really well and somebody will need to teach them how to use it really for the good.

Yeah.

Long answer to a, like, very, very, very simple question. But, like, I think everybody who goes through AI, like, oh, let me figure this out. We'll go through the same cycle.

Yeah. I mean, you would hope at least.

Yeah, I hope. Seriously?

Yeah. No, I mean, I think it's important, right, to get this. The context about, you know, I mean, also, you know, throughout this podcast, we've been hearing about your expertise, your experience, your background. But with something so kind of fresh to the average person, you know, it's good to know how other people investigate it and learn to use it and stuff like that.

Everybody is starting from zero at the moment, so my only recommendation is like, everybody should start. Yeah, design, like wherever they choose to start, either with images or with text or with videos. Like, start somewhere.

Yeah, for sure. Agreed. So how do you really, like, with your experience so far using AI over the past year or so, how do you see AI shaping these next generations of startups who have a social impact mission or have a desire to kind of fix a problem in their community or in the global society?

So I think I can divide it into two main categories. One, when you look into the SDGs, almost nine of them are related with things and five, six of them are related with humans. The domains where it's related with things like with nature, with water, with food, with land, with air, with ecosystems, things that are impacted by humans, but they have a different thing. You don't need to negotiate with them. I mean, that's probably how I can look into it. It's going to be absolutely mission critical to start using AI. If you're a startup doing something with forests, with something with land, you need to incorporate in your thought how to accumulate data, how to bridge and secure the data that's happening around whatever you want to impact, and then use it to feed the tools that you're using. So everything will be on a. Everything will be better going forward because you'll have some data you'll put in the model, you'll change your things and you'll be better. So I think that's a progressive moment. If startups start using it, my worry is not about that part. That part will be fine. I mean, I've seen startups using hardware with algorithms inside AI algorithms and data collection tools that they can put on a tree and it will cover a certain area, it will detect from past data of what happens during a fire. And it will correlate that with AI and it will try to multiply the opportunities and say, okay, there is a possibility of fire in this domain because some of the metrics have been moving differently, funny things happening, and they've been very successful, I hope they'll become more successful. So those things are absolutely important. So instead of these, just these rule based things has to be real time data, real time monitoring, real time estimation and prediction models with AI, wonderful with humans. That's where things are to me debatable and scary at the same time. Because the data that will startups will use to feed their algorithms contains inherent biases. A lot of examples are there already. So there's this beautiful movie called Coded Bias. It talks about how financial sectors start to use AI to hypothetically remove racial biases out of loan decisions that banks are making. And then they quickly realize, well, it's not as quick as they should, that they're actually making it worse because the data has been accumulated through biased actions, therefore the results are already biased. And once you use that data to bridge towards the future and do predictions because it's so misrepresented, because colored people, women, woman of color and like, like any other group than white males are misrepresented in every data in the US especially, I mean the rest of the world has different biases. But like in US is pretty obvious, then everything that comes through that AI is biased. And unlike rule based models and logic models, where there is a human building the logic, AI is a black box. So you write the algorithms, you feed it data. Most of the time I know I'm super, super simplifying it, but most of the people that are working with AI models don't actually have a clue on how that decisions are made. Why is this working that well, one example in medical science, I mean you feed the data, you feed algorithms with a lot of images that are done in scans and then you attach that to the actual review of physicians and practitioners and then you start feeding things that haven't been reviewed and then cross checking what the review would be. So AI does a really good job in both identifying false positive and false negatives at the same time. So when you ask like how come? Yeah, I don't know, like actually there's no reasoning, it just does it in most of the algorithms, especially when fed with pretty fuzzy data. So this is the part where it gets scary because for startups in the social space who are trying to solve human related problems, whether that's related to equality or gender or poverty or representation. They need to be very careful of the data sets they're using to feed into their solutions. They need to be very careful in acquiring that data and using that data. And removing bias from anything involving humans is a pretty daunting task to start with. So startups in the social space need to be extra cautious, especially if they're using somebody else's algorithm, especially if they're using somebody else's data set, because they might, they have the risk of harming them than making better unless they're very careful about it.

And I think even just looking for, you know, an average person who's interested in AI, you know, you can log in, create an Account, get into ChatGPT and start using it without really understanding what lies behind all of that, without the need to understand it. Right. You can just go in. You don't have to understand how it works to use it. And I think that's, it's really hard because people don't have that curiosity to learn more about it and just they, they think it makes their life easier so they can use it and it's just right there for them. And I think that's, that's really tricky, especially when you're looking at, you know, more complicated solutions, because it's easy to assume that you know, AI has all the answers, but it doesn't.

And especially on like areas where accessibility had been an issue. For example, I've seen a lot of startups in the recent two, three months that they're using different algorithms like open sources, especially to build support tools for, for people in stress, like young people that are having trouble with their families or women in areas where it's not culturally or physically able to get consultancy or get coaching or get psychotherapy. And there it's, it's, it's a brilliant tool. Like you don't need to go from one place to another, you don't need that. You, you can overcome the social stigma of going to a therapist. You can, like, there are a lot of good things about making it accessible without the need of a human, without the need of somebody that will be judgmental. So in some place removing the human element is necessary. That's okay, perfect. But like if that is, that algorithm is not culturally tailored or age wise tailored, the advice it might give you might not have actual control over what the advice will give. And again, you're feeding your data. Even if you have all the right questions to, all the, all the right answers to the questions that young people might have that are proven to help young People or the woman or whoever's in need. If it's not culturally tailored or age wise tailored, gender wise tailored, the advice will be creating more harm than good. Same thing with anything that's using open source to reflect back on human relations. So I think that's where it gets a little bit tricky. And that's where social startups, especially social, starts. Like startups with a social cause should be more careful than other startups.

Agreed. You know, before this session I was doing some research and obviously one of the big giants in the tech world is Google, Alphabet, whatever. And they've created a really interesting group called AI for Social Good. And it's a group of researchers, engineers, volunteers, and other people across Google who have a shared focus on positive social impact. And in their approach for this group, I think it's really interesting about kind of the caveats that they bring up. So I'm just going to read it out to you because then it kind of segues into our next question. So they say, what if people could predict natural disasters before they happen, Track disease as it spreads to eliminate it sooner or dramatically improve the lives of people with disabilities? AI can help, but it's not a silver bullet. Tackling these questions requires rooting innovation in research, responsibility and a concerted collaborative effort across all sectors of society. And I think that's really accurate, especially what you've just said as well. And I appreciate the fact that, you know, such a big giant in the world is saying something about that or at least commenting it's not.

Yeah. So the whole recent debates about created by people who worked for these companies in those initiatives leaving and then commenting on why they left, those are the moments where you get like, oh, my goodness, that's difficult. We don't need to go so far back. I think it's like, literally, I moved to the US six days before the elections in 2016, and then we know what happened during that elections after the whole thing. Some people said it's all Facebook's fault that Trump got elected. Some people said, oh, Cambridge Analytica has been doing that. They are in no fault. Some researchers said their impact has been exaggerated or not. But anybody who watches the great hack or social dilemma will realize decisions are impact. Human decisions are very easily impacted through carefully curated narratives. And if those narratives are built to influence buying decisions, it's one thing, not the best at all for. For anything. But that's one side of the capitalist world that forces people to consume things. That's okay, fine, not fine, but again, we know what it is. But on the other hand, when they're selling policy, selling votes, influencing things, so it kind of becomes, much like selling a car or selling diapers, it becomes selling votes of certain groups. We feed those algorithms with so much data, and the algorithms, the AI in those algorithms doesn't care who gets elected. And that's also one very common theme of all those dystopian movies where the machine decides it's better to remove the humans from the system. And it kind of locks us in fully controlled zones of humans against humanoids. And it kind of delivers the purpose. Like humanity will survive if we lock them in cages. And if we stop them from killing each other, humanity will survive. So is that the only goal or is there something like more important at stake here? So as I said, I got a lot of my thoughts and my futuristic reflections through science fiction. I'm a big buff for that. And I'm matching that science fiction literature with utopian and dystopian literature to figure out what will happen to humans. What will happen, what is human, what will happen to humans. And like, I keep on writing on different things and AI has already been in most of them because, like, it's easy to, it's a very easy character to fill in. You can make a super villain one super good one, naive one, and then take it anywhere interesting to see what will actually happen.

Yeah. So with all of these like uncertainties and obviously issues like the gender bias, what have you, that exists with AI, you know, it not being perfect, you know, should anyone who's interested in building something or a startup that's already in existence, should they be really investing their time in building AI into their solution, or should it be a resource or a stepping stone to their solution? How do you kind of feel about that?

So it's a good question. We were having this debate with a few investors currently. Like, investors are an interesting bunch to talk to because their hope, their dream is to get something out in five, seven, ten years. So they can't only look at today and say, this is good. They have to look into the future. Will this be good in the future? Regardless of the situation here. So they need to follow trends, they need to follow where this is headed. So there are two different classes of thought. One is because it's so new now, like resembling it to the first days of the social media, you can actually build a strong business on other others platforms today, much like a lot of companies were built on Facebook. Like one large example, that was part of my life as well. So farmville My mother used to play it along. Maybe she still doesn't play. I don't know. Like they got destroyed when Facebook changed what they were looking for. Same with many brand accounts on Instagram. So there used to be something called organic. You could actually have a conversation with your followers. You can lead them, you can follow them, you can have meaningful interactions until the algorithms change. So you won't be able to get there unless you pay. So these things shifted over time from, oh, it's free, do whatever you like, build your audience, and then all of a sudden, oh, now you have an audience, now you need to even pay to get to that audience. Things will happen that way. The difference might be like OpenAI and others have already started to charge for those services, so it's not going to be a big surprise if they charge more. So it's already didn't start free sometimes. I mean most of it is free, but you really need to pay to get to the cool stuff. So startups can today use these open platforms and their data sets, most importantly, and they learn like language models, especially on tech stuff, and then they can have impact really quickly, they can accelerate how they get to market really quickly, they can accelerate how they can present themselves to the right audience and create more impact really quickly. However, that's a short term plan. Every startup impact or tech needs to have a data acquisition strategy if they're going to put AI into their core business model, whether that is curation with customers, like dialogues with customers. This is very common now in most tools that we have already in our phones, like these chatbots. They were very generic and they were not actually meaningful up until the point where you ask a few things and then it just goes to, oh, should I connect you with a customer service agent? Okay, yes, because you don't know shit about this. And then now the customer service agents will become priority or customer service agents will be customized AI agents and they'll mean a lot, they will solve a lot of the problems rather than this rule based chatbot. So this is the evolution. We're going to be having a lot of them and that is only possible in the long run when you have your own data set, when you train the algorithms at your own flows and all infrastructure, that will be meaningful for a lot of people. But in the short term, yes, maybe it's going to not pay off to invest in creating data sets and priority algorithms, but then going forward they will need to incorporate a different strategy for acquiring what's theirs unique to them, unique to their customer set. Whether it's B2B or whether it's not customers, but beneficiaries. Like if you're building a business model where you sell something or you get donations or you get grants, or you get impact from corporates and then you have beneficiaries that you impact on a different level, then you need to think about both of these clientele to have a meaningful conversation with them. And to scale it will be really. There will not be a startup that doesn't use AIs when they're at scale. You can't scale without AI. Going forward, it will be extremely expensive to do it with the existing tools without AI. So that's a big challenge for sure.

For sure. I think we're really starting to see also a lot of discussions on a more international or national level as well about kind of these threats of AI and owning your own information and stuff like that, privacy, etc. Do you have an opinion about, as you're someone who's benefiting from AI and also so many of the innovative solutions that already utilize it, what role should public policy, if any, have in regulating kind of AI and how it's used also in when we're looking at, you know, solving problems for the future, what kind of role does that play?

So I think it's a, it's not a coincidence that a lot of these, both in European Union and in US and other places, most nations have started to have stable, concrete policy on data protection and privacy. GDPR leading the structures most of the time is fundamentally related to regulating AI will be more so because if we didn't have any privacy laws or data protection laws, then we couldn't be able to even imagine how to control or regulate AI. The tricky piece, for places where there's very little protection, data protection, enforcement and very little mind like work on privacy, then those places are going to be having issues. They won't be able to do anything, just controlling AI. It's not something you can control unless you control data. So this is very, very, very important. I think one recent case where New York Times sued OpenAI for using its resources to feed it to the algorithm so it becomes smarter. So again, copyright protection, data rights, individual data privacy is at the core of regulations and it's already been regulated. So it's not like, oh, wait a second, how are we going to regulate AI? No, no, no. I mean it started already because it had to start because people were using all kinds of data accumulation without any circumstances, without any limitation. Everything was used. So now I think it's becoming more important to enforce these, much like New York Times is trying to do and many others are trying to do. Similar debates were happening through Metaverse tools like NFT Tools. You would be really surprised to go to OpenSea and see a lot of art turned into NFT that's based on Damien art or that's based on open source art sent out by independent creators. So somebody would create it out, send it out, and then somebody else with NFT knowledge and crypto knowledge and currency knowledge will turn it into an NFT asset and put it on OpenSea. Like who owns this? Not the guy who has the nft. And all of a sudden this becomes a very complicated issue. But it all starts with data price. So who owns the data, who owns what? Anything other than that is going to be extremely difficult to regulate meaningfully because regulators come 10 to 20 years behind what technology is coming to. And you like, if you want to have a good laugh or like a very cynical laugh, like a drama, you watch the Congress testimonials and questions that are asked to Mark Zuckerberg or Jeff Bezos or Larry Page in Congress hearings on their thoughts about stuff. So the regulator's job is not to understand technology. And they don't and they won't. So any regulation will probably be overdone and irrelevant and unenforceable when it's created in a short amount of time. Not even the technology enthusiasts agree on what AI will create. Nobody in the ethics side like really can put their finger on what should be the ethical use of AI. A lot of debate is going on and it will keep on going on. That's how it's our new technology is being adopted. So when you look into regulation, I think the only thing that could have happened, and I'm glad it happened, is like data protection. Everything around will be enforced now because now people will start saying, okay, now that copyright will extend to feeding into algorithms. And it's easy to do it because there's infrastructure and they've been working on it for years and years and years. GDPR and data price law didn't come into effect in a short amount of time. But anything other than that, I think it's completely impossible to take it out from our sphere of influence. Major players might be maybe regulated. But again, most of this is like complete black boxes. Not even the people who kind of initiate this and send it out to the world understand why they does it. They do it that way. One funny story was like they it's sorry, as I read it's impossible to detect the gender of an individual if you only have their eye. Like, if you only have the retina with nothing there, it's just their only eye. Nobody can understand if it's a man or a woman. AI can understand that. They tested it and it's like freakingly accurate. Like, anything above 50% accuracy has already been more successful. Like, you can go random 50, 50 is okay. AI is very accurate. They don't know why. They don't know what it uses to understand from the looks, from the picture of the eye. If it's the gender, there's nothing. And people are surprised. Like, they don't know. They don't understand most of the stuff. When you ask people doing AI research and AI applications, like, I don't know. Some are saving themselves by saying, I don't know. I don't think that's honest. Most of them are 100% honest. But most people doing research, most of people doing these models actually don't know. They know what they fed in. They can understand what, why a little, but they don't understand the results. So that's where it gets tricky. And another movie for people interested is definitely AlphaGo. And there is a part that I keep remembering in the movie where the player, number one player, go player in the world said the move the machine made was out of this world was godlike. The ability to make that move transcends human perception, human analysis, human human cognition. And it was. And his words are like, amazing. He sees Korean. It's like, it's art. What. What the move was. And you feel that in his eyes and go like, holy crap. That that's going. And nobody who's doing The Alpha, the AlphaGo team, does not understand, like, it's a move that is not. Is not in the decision trees or anything. And it just makes it. I go like, okay. And then the whole conversation becomes, is AI sentient? Is artificial just a term? Is it like intelligent without the artificial part in it? Which is again, where we go to utopia, dystopia and science fiction. I can talk about that forever. So let me stop there.

It's very telling. I mean, I'm glad I read so much science fiction, especially when I was younger, to have grown up with all of this stuff happening. Reading Octavia Butler, Aldous Huxley, any of this stuff. And you're like, oh, it's real.

It's happening 20 years later. Oh, yeah, yeah, I know. I know the feeling.

It's crazy.

I know the feeling.

And I'm sure you know, you're in a really interesting position just with the work that you do. You know, you see so much of this happening on a day to day basis and I, I mean anybody could too if they're interested in this stuff and hopefully they are if they're tuning in right now. I think, you know something, you kind of touched on it a little bit about investors and investing in this sort of technology or trends or what's happening. So you being an angel investor and kind of also advising startups on like how to seek funding, how to go about the situation for those that are, you know, interested in including AI, what would be some unique challenges that relate to funding AI initiatives, especially those that are focused on social impact or have that kind of bigger issue in mind.

There are so many memes in the Internet and Instagram and Twitter on this. Three years ago it was, oh, I put this. And then they changed the word to crypto. And then all of a sudden investors are like, ah, it's crypto I'll need. And then it became Metaverse. Oh they, they do exact the same. But we're doing this in Metaverse and all the investors assumedly were running some, some of it's true and now it's like, oh, I do this with AI. All the investors are running through might be happening. I'm not saying it's not happening because even in the investment world there are people who are just becoming investors. It's a new thing like science fiction melodies. So what is happening in the investor side of AI is the real good ones are looking under the hood. They are actually trying to figure out, reaching out to people that are doing the research, people that are doing technology practice. Most investors now have a short list of technology people that understands the depths of the technology that's been built in order to get a reference of what's under the hood of these technologies, what's under the hood of the startup. The good ones are trying to shy away from investing in startups that are using other people's technologies or algorithms or open sources or LLMs and then building a layer on top of it. So everything we're doing and then we're using OpenAI. Yeah, okay, fine. But like maybe it's not going to be sustainable. That's where the investors are at this stage. And for impact startups, it's absolutely critical to understand what side are they going to be focusing on. This thing is not cheap. It is cheap in a way if you use others APIs, but it's not cheap to do it yourself. Not cheap to acquire the data, collect the data, sort the data, feed the data, fill the algorithm and tweak it. It's really, it takes a really strong technology team to build it themselves. So in order to start, they really need to, I think it's also my kind of advice, they need to start testing what happens if they use other people's platforms. But to kick things off, that's a good way, really fast way. But they need to have a strategy. They need to present to investors that they're not going to be relying on this. They're going to be owning the data that's created, owning the algorithm that's created and will have control over the black box so it doesn't become a black box. Investors will freak out with anything that they don't understand anyways. But now with an additional layer of technology that they don't understand, and not even the people who are building the technology understand, especially in the thought line, it's becoming more difficult to kind of say, oh, I'm an AI tool. Solving forest fire. We use AI to detect forest fires. Okay, how previously, how was, oh, just software. Okay, you write code, you use this. And the tax text was easy for investors. They'll you, they were used to it, they could understand it. Even in some of the really good VCs they had technology people who have built things before. So building a technology was common knowledge. Like the software east of the world was common knowledge for many investors who are not only coming from the finance background, but who are actually like operators themselves. But now there's a technology that not many investors have used in their life, have built things with. And now they need to judge it, they need to see, oh, they're using AI, not just a typical software on cloud, on ERPs and databases or something else. So they need resources, they need expertise from outside world. And I think those who are smart enough to just take the time to open and look under the hood will find really strong technology teams with impact on mind as well as well as the technology side. Because nobody is perfect at this at the moment. Like there's very few people who are really good at building AI tools and everybody's trying to learn, so everybody's trying to figure this out. Not that AI is something that just kept happening in our life. It's been in progress for the, since 1970s, people have been doing research around it, but the practical side is becoming more eminent now. So therefore, if you are an impact startup, make sure you talk to people, talk to investors that at least are spending the time to understand what's this whole thing about. But if you don't, then you'll find people who's probably going to be easy to convince when you actually show them something that works. Not on a hypothetical level, but here's what we do with AI, even though it's somebody else's tool, here's how effective, how it's going to be easier. And I just mentioned, like, they cannot scale without AI. No startup can scale without AI at the moment. Efficiently, cost efficient, at least now they're going to say, okay, here's what we're going to be doing to scale, here's what we're going to be incorporating tools with. And here, here's how we're going to kind of scale our impact. So they have to be very precise about where they will use AI, how they will use AI and how they'll control it. And I think that's not. Well, we're not going to be talking about in two, three years, I think about AI because it's just going to be, oh, it's, yeah, of course I'm using AI, what else could I be using? Like, this is not even going to be a conversation. It's like, oh, of course, of course we have a mobile app. Like, what do you mean? Do you have a mobile app? Of course we're using cloud services. What do you mean, do you have like, is there anything else? It's like going and asking today's teenagers, asking them what the phone is for. They're not calling people is not on their list. I was like, what do you mean you're calling? Of course you can call people. Like, yeah, fine, but like, oh, the other things are more important. So I think there will come a time it's just going to be ubiquitous and say, yeah, we use AI for this, we use AI for that, we use this for this and that for that, and then we own this, we own the data. I think a lot of people will reach to maturity quickly in AI and will be rewarded for that.

Yeah, I suspect as much as. Well, I think we're also seeing so many developments just in the span of days, weeks, months, ridiculous ridicules. So I think to really, I mean, you've already given a very broad look at what's to happen and also this topic. Yeah, could go either way. Who knows? What would you say is like your, your piece of advice to people who, who are really starting to, to work with AI or want to involve it into their daily life, like, what would you leave them with as kind of a piece.

So here's what I'm trying to do. Our habit is open a tab and google things. Now that's the habit. That's what we do. It's a word. Oh, have you not googled this? So it's surprising. Like it's like Kleenex, like in. In. In Turkey, Selpak, like it's a handkerchief. It's. It's disposable things. But it's called sa. It's the name of the brand. It's not the same thing. It's like you don't say give me cafe, you usually say give me Nescafe like it's a brand. But everybody understands what that is. So to me I think like have you not AI'd? This is some things that's gonna be like, have you not used chatgpt for this? Or have you barred this? Or have you co pilot this? What has your co pilot told you about this? I think that's going to come. So to get prepared for that moment where it becomes a word. Well, there are three things majorly competing for it. Have you barred that? This is a good alternative to have you googled this? What does your copilot say? I think it's a brilliant word. Like copiloting. Copilot is a really nice word. Like what does your copilot told you about this? Because like a companion at the same time. AI companion. Have you open AI? This is not going to. Have you chatgpted it? There's like difficult on the tongue. But it will work. One of them or all of them will work. So my recommendation to all startups is go to these tools first in your daily life. Go ask. Yes, it does not give you a precise answer. If you're not looking for a precise thing but like try it. I will give you extra two minutes to go back to Google if you're not satisfied and just Google the crap out of it. But again, that's how you start. Then my recommendation to everyone, whether they are in finance or in impact or in technology, doesn't matter. Use it for art. Use it for something that cannot be wronged. There is no wrong in art creation. There's a wrong in creation, in text, in search. Yes, this is right, this is wrong. This is like maybe true, maybe not for but in art everything is true, everything is created, everything is there. So the more you start using like we use mid journey now you can use Dall E, doesn't matter. Use visual things, use verbal commands prompts to create art and then figure it out. Like play More, play more, play more. You will have a really good back and forth because the feedback loop is super fast. You get what you ask for in a way that you didn't think possible because you're not able to visualize it. You're verbalizing it and you're getting a visual response. And then through that visual response, you're reverting it, you're getting back. So this immediate, instantaneous, unpredictable feedback loop will warm you to the possibilities of AI, not through your maybe daily work. Maybe you're not doing any creative work at all in your daily life, but that is going to start surprising you, pleasing you in a way that will trigger your creativity. Art is the path towards it. Yes, you might start using it for your business purposes, for your personal purposes, that's perfectly fine. But fundamentally my recommendation, start using it in art setting, in creativity setting, to kind of rewire a few things in your brain and then you can say, okay, wait a second, if it can do this, what else can it not do? What else can it do? So I think that's going to be very, very valuable in getting climatized with AI And I think that's a nice path to kind of see through the creative lens and then with new connections connected to your daily life or business life.

I think that's great. That's what happened to me. Yeah, Now I use it all the time. I love using it. It's also a form of validation for those who have imposter syndrome or otherwise. You can totally use it that way too.

It's pretty fun.

No, I think that's great. Thank you so much for your time. I think this is really, you know, an interesting conversation to have, especially in this day and age with so much that's going on. And yeah, I hope everybody starts to play around with it more and also learn the limitations and the ethics behind it as well. So always something to keep in mind when you're using it.

I mean, fear is a very human reaction to things, mostly to things that we don't know. And we shouldn't let fear cloud our judgment on anything new. In order to avoid that fear, you need to familiarize yourself. Then you might fear us more. But now, you know, much like we panicked and then that kind of. Okay, wait, what? Then you start having critical thinking. You can remain panicked. If you're in a profession that you see mundane tasks covering 50% of your time, you could validly panic. But then if you do, panic usually leads to action. You can be frightened and just shut the whole thing off. And try to put yourself in a cyber bunker. But most probably you will do some action. If you really understand how or visualize how it can impact your life, then there's action. There's nothing wrong about action.

No, for sure. Great way to end it. Thank you so much for your time also.

Pleasure. See you next time.

Okay, Next time.