We Them Media Live Stream

0:00The one who trained Mac, this is my product. And now he’s an AI developer. Yeah, I take full credit. Next person is James. James. Jay? Where are you? Karibu Sana. Yay. And thank you guys for sitting here with me. As I said, I have a reward. I know you had. And it’s a very serious reward for tomorrow. For those who have seated me, I have noted all your faces So I’ll hand over to Alfred One, two, okay GM, GM, thank you guys for being with us Thank you everyone for the intro By trade, I’m a software engineer But by heart I’m an artist And I’m also building Sasa Sasa

0:59Sasa Sasa, like Now Now Which is a creator, community and ecosystem Focusing right now on music Yeah, this is a very interesting panel. I’d like Liki, founder of Questpanda, to start off with the intros, and then we’ll pass it on over. Okay, thank you, Alfred. My name is Liki Njeru. I’m a computer science graduate. So my passion is in building consumer applications. So after Compass, I built Pesabit, which is a platform where you can use your Bitcoin as collateral to get a Kenya shilling loan. So the product which I have today is called Quest Panda. It’s empowering creators in Africa, like TikTok creators, Instagram creators, and ex-creators

1:51to be able to earn money by using their influence to help small brands across Africa to market their products So thank you, and this is going to be an engaging session

2:11Hello guys, my name is James, I’m a blockchain software developer Yeah, so I work at Yamata, which is a company that is trying to bring in the centralized exchange goodies with the security of centralized exchange. So we are building this exchange where we are able to own your crypto and able to trade with all the advantages of centralized exchange with all the liquidity. Happy to engage with you guys. Yeah. Okay. Hello, everyone. My name is Mark Matakili, I’m a founder of healthcare, to make it AI powered, which means that we are introducing a brain in the midst of an application that is telemedicine, that is, you are able, the fact that you are storing your health on the same system,

3:19It being AI-powered, you are able to get human guided responses and analytics on the same. So within the session, we are going to explore more on how we are using that. Thank you. Awesome. So we are a bunch of nerds, right? We know a lot about the tech, but we also want the crowd to help engage because AI is not It’s not a nerd’s technology, it’s everyone’s. So we’re gonna, I think you’re gonna kick us off, right? I think you have a question that can start us off and we’re gonna ask the crowd to kind of engage with us as much as possible. Not at the end, during. Okay, please.

4:13So, I understand that AI is not a new thing. Everyone in the crowd understands what AI is or rather has used an AI product before. I’d love to try and maybe share something you’ve done with AI, be a tool, so that we can better understand you guys and engage with you on this. We want to make it a one-time conversation with you guys So if you can get someone to pass a mic around and try maybe two or three people to give us an example of a tool they have used before So just to summarize, what AI tool are you currently using, right, that wasn’t there in your daily driver before? Does that make sense?

5:10and hopefully it’s not all devs in the crowd, so you can just, it doesn’t have to be something crazy. For example, I use Perplexity. Perplexity, anybody? Shoutouts to Project Mocha for that free Perplexity subscription. It is an amazing research tool. Perplexity is Google’s rival, and this is how I do my research now. So for you guys, what is one AI tool that you just can’t get rid of. Is there anyone else? Please, a few volunteers. All right. We’ll take two from the crowd, and then we’ll cook it. Yvonne?

6:00Hi, my name is Dennis. Thank you. The AI tool that I mostly use is one’s called Decagon. Another one is Lexica as well. Lexica and Midjani. Those three are the AI tools that I cannot get rid of to build my art, as well as engage with the community, not here, but globally. That’s why they are going to help me, you know, structure some messages and all that. Yes.

6:33Yes. Hi, guys. So my name is Moses. So you can just call me the on-chain chef.

6:43So one AI tool that I use most of the time is cloud code. Yeah. So it helps me to ship features very fast within hours. I can launch a product within 24 hours. So that’s the best tool, I guess, I can recommend for developers and also co-pilot. Thank you, Yvonne. My name is George. I’m a data scientist and a builder. Like you said, I also do use perplexity for code, but I also use the so-called manners to help me with my writing. Thank you.

7:38Hello. My name is Francis. I’m Francis. My favorite AI tools. One is Copilot. I like it. I started with Casa, but when I switched to co-pilot, it’s best. And then the other one is Gemini, and the last one is a GPT.

8:30So, hi. Okay.

8:34The main eye that I use most of the time is Casa and Cloudy, which I find is Cloud course. Yeah, it helps with my work for them to make it more seamless or something. Hi, I’m Anandeb. Hi, I’m Aziz. I go by Motomoto in the space. I use Lindy AI to automate most of mundane tasks, cloud and ChatGPT, perplexity.

9:43From what I hear, Perplexity, Cloud Code, Decagon, Lindy AI, Co-Pilot, ChatGPT, Gemini, I’m sure there’s another hundred in the crowd that all of you know that maybe came out last week, right? The space is shifting so fast, and I think that’s a good segue to start, right? For the people on the panel, you guys are building, but how has AI transformed your agency and your ability to reach the objectives? I think you already started this conversation, but on an individual level, what level are you using the same tools to scale?

10:35Okay, once again, AI tools are coming up, like, okay, from the crowd, the way you people have discussed with us, there are so many tools out there, hundreds of them, and the way we are speaking, there are also other companies that are coming up with tools. I’ll dwell Well, mostly from a founder’s perspective, I’ve heard most of you have talked about the AI tools that are more specific on the developer side, but I’ll talk about it now as a founder or someone who is running a business, an entrepreneur, I’ve gotten to like this aspect of the AI agents that automate tasks, and we are mostly using tools like entertain, we have Zapier, and we

11:47have another intelligent one called pipe dream that you can just see also just give it a prompt and it also gives you another, it auto-generates for you, it creates for you another AI tool that you can use to automate your tasks. Some of the tasks that I use for I have one that is able to web scrap for me some of the leads within the market from social media, it’s able to go through the Twitter and be able to tell me grant opportunities that are out there, the remote works and remote works that are available that have just been been advertised maybe in the span of like 24 hours ago. So such an AI agent, I’ve automated

12:39it in such a way that every day at 7am it sent me those emails with all those updates. So from that I see that as one of the use cases where founders and entrepreneurs can use such such kind of skills to build tools that can give them information because you know when you’re a builder, a founder, an entrepreneur, information is the key to everything. You need information on the news, on the upcoming trends, and generally even regulations. So those are some of the tools that you can use and you’ll be up to date with the latest News, and it can purely align with the motives of your company, your goal, and everything

13:28that you need. And now, talking on the blockchain side also, these tools, they’re not just that you can, they’re user-friendly, you can automate them in such a way that it can send you information to even your WhatsApp and SMS, you don’t need a website or something that you need to log in, just user-friendly. But If You Are Trying To Also Build Something On The Blockchain You Can Incorporate Them Onto Websites If You Want Them To Push Some Notifications On Your Website That’s How We Are Using Them At Helcon To Push Updates And Also Anyone Who Is Visiting Our Website We Have A Chatbot That They Are Able To Chat With It Give Them All That Information And They Will Be Up To Date

14:21So, I think that’s one use case that we are currently using.

14:34I guess I can also share. This is the part where I want us to transition to the conversation of the hype, Africa, and where this ties in. So, it’s obvious that it’s adding value, despite these tools, business models not being built for our pockets, right? All these subscriptions matter to us, but the value we get from them compared to the amount of money we make in this economy is not the same. So my question to you as a panel is, as AI has grown, what are the shifts in the African space that you’re seeing are transforming our ability to actually take agency? We’ve talked about consuming it.

15:21I use AI at Sasa Sasa to help us extract data from images for things like KYC. It’s cheaper than most OCR scans instead of setting up your own OCR pipeline. But I see a problem where the data is skewed. The darkness of the continent makes it harder, for example, for the training data to notice a face that might have a different complexion, right? So let’s dive into, given the leverage and the utility, what are the gaps in the state of Africa as a region that this hype train is missing? I don’t know if Liki, you want to take that? Sure, that’s a very nice question. So we can begin by looking how AI applications are made.

16:20So, first of all, you need very expensive hardware. These are GPUs. And in Africa, getting such amount of capital whereby you can have a thousand GPUs in your maybe garage to train an AI, it’s quite a challenge. So, that’s one of the things that I think we are missing We are using in Africa our own trained AI models and this hackathon I was checking lately whereby they are building AI platforms for Africans who have resource-constrained devices, like most of us here, if you have been Zoom call before, you found out that maybe your video was kind of not smooth. like other users in different parts of the world.

17:17So when you are building AI platforms for Africa, they should first of all not be resource intensive. You should cater for the bandwidth so that it can be not very consuming. And also, we need AI models which are light, which can run offline in our own devices. like how you use a simple tool like a calculator, think an AI model which can run in most people’s devices, which can serve people in remote areas. So that’s what I’m thinking about. And as an example, if you look in northeastern Kenya, we have a lack of education for most the most of the children, because there’s no schools, there’s a small number of teachers

18:11around, and I think if we can be able to use AI platforms where the kids from that region can be able to access education from a simple platform like ChatGPT, which is running offline, where the children can learn. So that’s what I’m thinking about. Very interesting. Interesting. So you’re mentioning things around not just network bandwidth, but I guess infrastructure struggles. So also energy, right? Like the need for Africa’s energy to come online so that even these low bandwidth solutions can be sustainable. I think one of the struggles that we’re struggling with, one of the things I’m noticing is the coordination of data so that these local models can run efficiently. I don’t know if any of you had comments on

19:05I think all of us can agree that most of these AI models are trained on data that is not African-based. For example, you would ask a question, you want A2M in the Kenyan or African market, but it gives you a solution that maybe don’t really match what we experience on the ground. So this is a major issue we’ve got because most of the data that has been scraped from all these platforms like Reddit, Google and all that, these are most of the comments, most of the posts that have been more focused on the maybe outside of African way of doing things. So I think that’s one of the major issues that is really impacting how AI becomes

20:00that comes a benefit to the African market and to the continent as a whole. So, yeah, I think that’s one of the major things. Something we could do around that is to ensure we collect data that is more African-based. We integrate our local languages into these AI models, because I have some good friends of mine who are trying to build a model around the Lua language to try to make the AI understand the law language and all that, right? So these are some steps you can make. You can put us as the African builders to try and ensure that we are not left behind in this AI race as these other nations are trying to aim for AGI and all this, right?

20:46Yeah, I think that would be my comment on that. Okay, just to add on, I listened to the previous panel and they talked more about the issue of regulatory within the African ecosystem mostly and that is the issue of governance. I’ll come to this scenario that I think lately within the Kenyan ecosystem where we experienced people being abducted because they were auto-generating some content using AI and it was a bit challenging. So the issue of regulations and The government on how it’s handling these AI tools. It was some kind of a bit harsh. To me, it showed to me that we don’t have clear guidelines from the government, maybe, that shows to the users, like we the users, how can we use these AI tools ethically with those in authority.

21:50Yeah, so that can also be one of the challenges that the African ecosystem One of the things that is currently difficult to do is in our space where data isn’t protected, right? The communities and the people that are providing the information are gaining from it And so I really think that regulation question, as much as Africans might give their data, regulation is really important Could you just color that up and for us

22:26So what I would say for a smartphone right now you are able to see the current weather conditions of the place you are currently So imagine in different parts of Kenya we have some weather stations which are providing weather data to an insurance company So in case a certain part of a certain region in the country faces like famine or drought, the insurance company is able to automatically disperse the insurance composition to these farmers. So that’s one of the most interesting use cases I’ve found recently. And actually the problem was about AI and Web3 use cases in Africa. So when I look into Web3, we have like DeFi lending platforms which are providing access to credit to people in Africa

23:34So like how we are doing is that we look for capital, so like there is a global liquidity pool of investors who provide their money in a central place Then the money is able to be used in the real world The money goes to the farmer in Kenya, in Tanzania They use the money to produce agricultural products When they make a profit, the money is sent back to investors Who provided the capital All this is done by the power of Web3 So those are some of the use cases we are seeing locally and they are working actually I think we have the same question to you guys

24:31Is there a path forward for us to capitalize on this regardless of where we are starting from So that whether the industry’s hype or momentum, what you call it, is also Africa’s windfall Like we use that same energy

24:58Any time someone asks me if AI is a bubble or just something that is here for a short while, let’s turn back to the dev in story. I think every dev in the house remember when, was it dev in, I think it was last year or last year, when we were told that there’s this AI board that can go and get an issue and resolve the issue and all that. Many of us devs thought that work would be done for us, right? Because you have something that can build that thing that you were supposed to be doing, right? But I think we were all jealous when we learned that maybe there was some backstory to that.

25:43Yeah, so just that, by the way, but AI I think is here to stay. It’s not a hype or something, but we have companies like OpenAI, rushing for AGI, right? We have Anthropic trying to build the best AI audience out there for coding. So I believe when you bring it to the African market, what can we do with this? What can we try to build on top of this? Whatever people are trying to say is a boom or just something passing, right? So, let’s dive down into it, right? We are way back when it comes to AI development in Africa. We are, I can say we are marginalized when it comes to AI because most of the solutions, most of the data trend on this thing doesn’t take care of us, as I had said earlier on. And I think this is our time. This is our period to, like, double down on this because we’ve got nations, right, betting on AI. We’ve got nations investing heavily on AI. We’ve got China.

26:47trying to integrate AI into our development of students, right? So this is something that you can see. Development nations are banking on this, and I think us as Africans, we should not be left back. We should embrace AI as it is and try to push forward into it. I understand that there’s a very big knowledge gap when it comes to African development because we’ve got developers being signed like football players with hundreds of millions to just switch companies, right? And we’ve got devs here in Africa who maybe are just trying to use products that are built by these devs to try and build. This is our time to create our own AI models that are the challenges that we go through

27:35as Africans and build on top of this. I don’t think AI is a hoax or a boom. I just believe this is something just waiting to boom, waiting to accelerate what we have crypto in 2020 and 2021, but this is here to stay.

27:57Okay, talking about the hype, I think this is something that relies much on builders and and innovators. It’s us who will make it just a hype and disappear or like it’s for us to choose. We either make it a hype and let it disappear or make it look, rebuild it to result with our own culture, with our own native understanding and build along with I like giving an example of M-Pesa. M-Pesa has captured the mobile money market because its origin is from Kenya and it’s really hard for any other mobile banking company that is trying to come up. It’s a bit hard for them to take over Safaricom because it has that aspect of originality from

28:59from us. So if we can turn around this AI thing that its origin is from Europe, USA and those other countries, if we can try as builders, as innovators and make it actually resonate with us in such a way that if you’re using this tool you feel that this is something that we own it. That is the point where we will use it for a long term and it won’t just be there as a hype. It’s something that will stay. We are in a bubble. That’s my honest opinion, that we are in a bubble and it will burst and like all bubbles bursting, like you said, the builders will continue building even when the bubble is burst.

29:57So I think the message is, even with our problems, the hype or the opportunity is in the hands of the people like us, or everyone here who’s mentioned one, gives you the ability to feel like you have a team, or you can 10x your outputs like Moses and output something in 24 hours. That is the promise of AI. and I think for me it’s not hype because every task that is done by AI and now we can act as a higher standard of the minimum quality. Does that make sense? Like we can now do a better bare minimum and for us where every, like in the token economy where you’re paying, where you’re paying, in the token economy where you’re paying,

30:57The African-percent for each request. Who knows that better than the person paying for prefilled tokens, right, than the African who’s used to being frugal. So there is an opportunity in how we understand these systems from, I guess, from a struggle perspective. We’ve done so much with so little that if we can leverage this technology to create more opportunity despite our infrastructural challenges, then certainly this can’t be hype for us. because our bare minimum for everyone with a phone and an internet connection, just like with Bitcoin, has gone higher. So that’s my take on the AI part. Now, I want to segue into something you mentioned.

31:38And I, again, maybe it’s controversial, I don’t think we’ll reach AGI. I think Balaji says it best. This version of AI is augmented intelligence. It allows and enhances your abilities as a builder or as a founder or as a creator. But it’s using that leverage to be better. Without that, it cannot function well and it’s guessing really well. So I wanted to hear from the panelists. Do you guys believe we’re on our path to AGI? Because I don’t think that’s where we’re headed to with this interpretation of AI. Liki? Yeah, so, Ari, on our path to AGI, so, like, looking from the African perspective, we have so many, like, challenges which can be solved by AI and Web3 intersection.

32:36We look like, we look at raising standards of living, how many people in Africa and Kenya are able to get employment. or make money by using the AI and Web3 applications. So we have also cross-border remittance. You know it’s not easy to send money from one African country to another because of the high forex fees. We have stablecoins in Web3 space which have cut down the fees to more than 80% because sending money using web-driven platforms is almost free. So I believe in my own perspective, we are not in a rush to AGI. That’s my own opinion because we have core problems in Africa which can be solved by spatial AI models

33:39which have been trained to solve that specific challenge. So that’s what I’m thinking about.

33:50I love that you brought up the local model. I think that could be an interesting way for Africans to, and like how Mpeso leap forwarded traditional banking. Maybe we skip the large language models and go to small language models and see how that works for us. But please continue with the question. Yeah, so I think I beg to differ. So I’m a bit, I’m a sci-fi guy, so I’d like to bend on the AGI to be a reality thing. So we are headed there. So I can see it might take quite some time, but like deep down my heart I would love us to have an AGI, have something that we could all rest, have something, think for you,

34:40Right? And quite at a level where a human could do, right? Imagine if we could involve emotion into an AI, right? It could make decisions in a way that a human could do, right? Right now, AI is more of a good pattern matcher, but not a thinker, right? If you could make this AI a thing, I imagine all this possibility that we could unlock, right? Maybe we have all drug, sorry, all cures to diseases right in front of our eyes, we can’t see it. Or we could start intergalactic voyages and we can’t right now because we don’t have the knowledge or how an AI could think.

35:32If He Had Our Emotions, Our Intellectual Ability So I think these are things that would come not just to make this short, I believe in it, and yeah, let’s wait and see Okay, I’ll be brief on the same When it comes to AGI, I believe that, okay, I always believe there’s nothing that is almost impossible when it comes to technology But again, there is nothing that is perfect So we might get close, a bit close, but I’m not really sure if we can fully achieve the meaning of AGI the way we are defining it but if you can think the way by the year 2000 no one ever thought about AI but here it is. So you never know.

36:22Okay. Thank you. So, consensus is there’s possibility, but maybe it’s, yeah, maybe we’re not properly on that path with this technology stack. I know I’m out of time, so Yvonne, I will let you be, but my question to the crowd is the same.

36:43What he mentioned around the builders being the decider is us and these tools are not things we’re consuming, they’re things we can innovate on right now. How many of us have been on Hugging Face? Hugging Face? Right? How many of us have heard of Hugging Face? It sounds familiar. The things that we use every day are right there for us to tap it to. And Helcon is using it right now and they don’t need to cost much. Our challenge as Africans is to figure out how do we make this accessible at our price range and in our business context so that the Africans can use it without losing out on something else.

37:30I think that’s the challenge here, right? Whether it’s the use of data capture methods that you mentioned which are lower in fees or better efficient because they’re focused in a specific use case, or it’s making sure that as an ecosystem, we build data capture regulation that allows us to collectively keep this data and grow it so it scales. I just want to ask each of the panelists to just summarize whether, just your summary on this very, very long topic, because I think there’s two other conversations coming in on AI.

38:08Okay, just to summarize about the topic of this panel, which was AI and Web3, I generally say if you are a builder, if you are a blockchain builder, incorporating AI is like giving your your blockchain a brain, so it’s a cool feature that you can try to leverage.

38:40I would love to tell you a little note about this, whether it’s a hype or not, let’s keep building. When crypto was coming up, everyone said it’s a hype. When the internet came up, everyone said it’s a hype. But look where we are at the moment. We are reaping the benefit of internet, reaping the benefit of crypto and blockchain, you can buy anything you want at the moment with crypto, you can interact or communicate with anyone using the internet, so let’s keep building, let’s not be left behind, let’s acknowledge and take this as part of a technology that will

39:37So my final thoughts about AI and Web3 use case in Africa is that I see this is a very promising future because first of all as a developer you want to build a product maybe which uses AI for making manual and for easing the life of people. First of all, what I would like to tell you is that we have open source AI models which you can integrate in your solution. We have a lot of models which deal with the visual analytics where it’s able to understand video and image data. So you can use the open source models because we don’t have the GPUs to train our own AI.

40:36and that’s some of the votes which I have and to add on to something we should build AI models taking into perspective that majority of our users have simple smartphones which are maybe $100 or less so those devices should be able to run the product which you are building So, those are my lots. Thank you. Thank you very much. Round of applause. Thank you, guys. And you guys are amazing. You’re sitting through. It’s a very 2024 would never.

41:39So, next we’re going to have a start-up showcase, but before that we’ll have someone small telling us about one of our partners speaking about something. Not now, just a second. Before that, Jerome has requested that I get people to stretch, because you guys have been sitting And I know you want my reward. I’ll give you tomorrow. So we’re first going to do a stretch. If you could rise up. I’m going to entertain you. So we’re going to try and spell it safari with ourselves, yeah? We’re going to spell it safari. And if you can guess how, but do it with me. So we’ll start with the letter E. Now letter E, we do it like that.

42:36To the farthest you can. Letter E, everyone, let’s go. Are you feeling that stretch, yeah? Okay, let’s go to T. T, you put your hands like this and bend your head. So, T. Let’s go, T. Until you feel like a crack on your back, yeah? Let’s do H. H, you stand like a monkey, fetching something. Are you feeling it?

43:19S, we make an S like this, like this, yeah? That’s an S, right? It’s kind of an S, yeah? Let’s do A, and then you stretch to the furthest you can, yeah? Get to as high as you can, you stretch those legs, yeah? Let’s do F. F, you stretch your hands. You guys are feeling it? Is Is it working? Is it working for me?

44:06That looks good. You guys look good from up here. Let’s go R. R is...

44:18Yay! And I, lastly, I stretch to the furthest you can.

44:31Yeah, and what have we spelled?

44:39I can’t hear you, oh my god. Is it me? One, two, three. Thank you. So I’ll allow the person who did all the designs with Safari this year is apparently our partner. So I’ll let him speak for just one second, one minute, one minute, two minutes about the event they’re hosting in Rwanda. And then we’ll go into the Startup Showcase.