#456 Colin Fitzpatrick on AI Agents, Vibe Coding, and the Future of Web3, Robotics, and Work

[00:00:00] 
 

Mehmet: Hello and welcome back to a episode of the CTO Show With Mehmet today. I'm very pleased joining me, and I'm excited when I say from the same place. Well, I am in Dubai, Colin Fitzpatrick. So Colin, thank you very much for being here with me today on the show. The [00:01:00] way I love to do it is I keep to my guests to introduce themselves to do their own intros.

Mehmet: So tell us a little bit more about you, you know, your journey and what you're currently up to. And then we're gonna take the discussion from there. 

Colin: Amazing. So great to connect with you again for starters. I know we spoke many, many months ago and I remember the very interesting conversation, so delighted to have it back.

Colin: Uh, sure. So just a little bit about me, I'm Colin Fitzpatrick. Uh, I'm Irish, I'm from Dublin. But like yourself, I live in Dubai and here about six years. Uh, I suppose I've kind of always lived my life on the edge of technology. I've worked for a lot of the big software companies like Oracle, Salesforce, HubSpot, et.

Colin: In 2020, I, um, I've always been into crypto as well since about 20 16, 20 17. Got myself into the business like more, more properly. When I founded my own blockchain business in the Metaverse and NFT space in around 2021, uh, built a business there that we were doing concerts in the Metaverse and [00:02:00] NFTs for celebrities like Snoop Dogg and Alicia Keys, et cetera.

Colin: But really as soon as chat GPT came out, I thought it was, you know, the most game changing technology that we the world had ever seen. And I said, that's what I wanna do. And I've been working fairly solidly in AI since then. Um, done a couple of different things. Built a digital human, like a conversational 3D avatar for customer service and support, which I think is gonna be a, a massive industry and there's loads of people doing it now.

Colin: Um, for the last while I've been building AI agents. For Web3, uh, with Griffin ai. So that's what I'm doing right now, which is, um, providing a platform for developers to build, host, and monetize autonomous AI agents that can actually transact on chain. And I'm excited about it because I think it, uh, AI and AI agents specifically are come along and they're gonna solve a lot of the problems around complexity and learning curve in technology, and especially in Web3.

Colin: Because a lot of it is very, very difficult [00:03:00] to learn and to onboard too. But agents will come along and just take all those problems away. You know what you want done, the agents will do it and you'll get on with your job. So, um, I write an awful lot on LinkedIn. If you wanna follow me on LinkedIn, please do.

Colin: I sort of cover a awful lot of, uh, emerging technology, futurism, et cetera. And yeah, that's just what I'm very passionate about. 

Mehmet: Great. And thank you again, Colin, for being here with me today. So. It's very exciting times because yeah, just for the audience to know. So usually sometime what I do, I speak to my guests like time ahead and you know, this time it was quite some while ago and I spoke with Colin and since then Colin, a lot of things have changed, you know, both on the AI side, uh, but you mentioned something, you know, which stopped me.

Mehmet: You said when you saw Chad GPT and everyone talks about the Chad GPT moment. Now we are seeing this moment. I mean, almost repeated very, very fast. So [00:04:00] I can, you know, mention some and you might help me also. So we had, you know, after chat g pt, many things happened, but just 20, 25. So we had this, you know, uh, deep seek moment.

Mehmet: We had the man moment and I don't know, like the, the vibe coding moment. So among all of these, right, so which one stopped you or like. You said, okay. Like this is something again that I think you know, it gonna take us by storm. 

Colin: Well, okay. I would say yes. People talked a lot about the Chachi PT moment for deep seek.

Colin: Okay. And I think it was a very, very important moment, but it wasn't as big as Chachi. Manus is just one of the many, many, uh. Vibe coding, let's call them applications whereby you just have to. Tell the AI what you want and it writes the code for you. Now, I am not a software [00:05:00] developer, but I do like to keep close to what's happening, and there's many of them.

Colin: There's Cursor, which is probably the best one. There's Waterfall, there's a bunch more. Um, I. I would argue that we actually haven't seen the proper chat GPT moment yet, and I think that's coming this year when it really fully knocks our, our socks off and takes our breath away with its incredible abilities because the people I know that use them, um, cursor and Chat G pt, they're very, very good.

Colin: They automate a lot of the basic tasks, but. You still need to be a developer to figure out, to build a proper, you know, robust, scalable enterprise class application. That's just not there yet, but I absolutely believe that is coming. Um, you right now, I can go into Claude or chatt PT and I can get it to ride a little game like Snake like was on my first mobile phone, the Nokia fifty one ten, like 20 years ago, whenever it was.

Colin: And that's pretty [00:06:00] remarkable. But. The, the way I always like to describe it is if you look at text to video, so we've seen the, the, the Chachi pt moment for text to video was Sora. Uh, we had these models that came out like, will Smith eating spaghetti. And you know, you could see it was a video, but it was awful and blah, blah, blah.

Colin: And then I come Sora and it's just like, well, this is indistinguishable from real life. Realistically, that was a Chachi PT moment. Many more of them are still coming. So when it comes to this, uh, application development, vibe, coding, whatever we wanna call it, we've made massive strides. But I really believe that, uh, we are only scratching the surface.

Colin: I think for the whole of ai, we're at various different places on the exponential curve. And because of the way AI works, coding is something that it is going to be able to be extremely good at. Um, I've watched a number of very significant people this year. Who ha uh, sorry, this week who have said that by the end of this year, uh, AI [00:07:00] will be better than humans coding and that, and that's the end of it.

Colin: And then, you know, we go on, it's a little bit like how AI became better at chess 15 years ago with deep blue or 20 years ago, whatever it was, um, that will actually happen and I think. What that will mean is that some people like or software developer is gonna be outta jobs. Absolutely not. If anything, I think it will be the opposite because when these technologies essentially get democratized, um, the.

Colin: The demand for them goes up very significantly because the barrier to entry to building an app goes down very, very significantly. I have lots of active of ideas in my head, but I don't really wanna spend 50 grand on them to figure out if they're gonna work or not. But I might spend two grand or one, or you know, something like that.

Colin: So I think what's gonna happen very, very soon is that everyone's gonna be able to essentially just. Build any app they want, it'll stream it to you in real time. They will be very dynamic. Um, and also it's gonna [00:08:00] affect the existing market quite significantly. I come from enterprise software where every single large company is essentially a giant mess of horrible applications that are pieced together with cell tape and lollipop sticks and I think.

Colin: That will change because now we're gonna be able to vibe code our way into replacing old applications and into building better integrations and into improving software, which is something that's been very, very difficult. 

Mehmet: It's exciting time indeed. And yeah, so it was the end prop CO himself, DDI who, who, who mentioned like he thinks like.

Mehmet: Um, in 2025, 90% of code would be written by ai. Yeah. Um, honestly, Colin, I tried, I'm not a coder by any means, right? Yeah. So I I, I'm like any other one who graduated, uh, 20 years ago from college in computer engineering. So this is how much my knowledge is in coding. Like, I, I, I don't call self, call myself a coder.[00:09:00] 

Mehmet: Yeah. But you are, you are pretty much right because. What I have seen so far, and even so, people who follow, for example, I know like the vibe coders, they hang out on X, right? Yeah. Or on Twitter previously. So majority of them, they have been in coding for quite some time. Yes. So this is why I think they are doing fantastic jobs.

Mehmet: Um, I tried the tools. Including many, right. Without mentioning any names. Some of them Really? Yeah. They bring you a working app. I'm not saying no, but if you look at the code, you know, and I'm, again, I'm, I'm not a coder by any mean. Mm-hmm. I can see. It's, it's like a heavy code, right? Like it's writing thousands of, of lines.

Mehmet: And I'm not sure like, really did, we wanted s of lines, but yeah, to your points, the, the bar is getting lower and this is allowing people to, you know, so anyone can now at least, I would say they can develop their own MVPs, let's say. 

Colin: Like, let's, let's say exactly and, and, and [00:10:00] Isn't that amazing? You know, isn't that amazing?

Colin: I mean. You can now, instead of going out and hiring an agency and in a team of six or 12 people and taking months, now you can knock something together with, you know, my buddy Matt, Matt who has a little bit of experience from a while ago, and we can just spend the We weekend kind of knocking something up.

Colin: Um, I also follow quite closely this vibe coding community on x. I don't know if you know Manis McKay. I've been following him for a really long time. Manis, I think it was three years ago. He didn't know how to code and he started teaching himself kind of right along time as some of these tools started coming out.

Colin: And now he's knocking out some amazing, uh, things. And I think that, you know, I have a son, okay, he's five years old and in a few years. He's gonna be building his own apps, you know what I mean? He is a little bit young now, but it's just gonna become so accessible. Uh, and I think things will change a little bit because, I mean, we [00:11:00] have apps for everything already, but that's just gonna go into a different level.

Colin: But what I'm more excited about is these dynamic apps where if you don't like the way you it works, you can ask the app to change and do something else. That's a dream for me. I. Bang my head off a wall all the time about how a certain piece of software works and I don't like it, and why did they design it like this?

Colin: They're gonna be very, very flexible and dynamic and I, and I think that's amazing, but it will enable, like you said, it will enable people to knock up MVPs very simply and then potentially launch businesses. A soft launch way without spending a fortune, without spending six months, because the most important in the startup wall for a wall, the most important thing is market validation and product market fit.

Colin: And too many people, and I've, I've done it myself as well, you think you've got the best idea in the world. You put in a ton of work over many, many months. You build something and then you can't sell [00:12:00] it. What people are doing now is they're getting an idea, knocking up a prototype, getting it out there onto X, seeing the feedback, okay, they don't like this.

Colin: We're gonna change it that way. And you're able to dynamically build this product according to what you know, the market reaction is and things like that. And just while it's in my head, yesterday, Elon Musk tweeted, what don't you like about x? What do you want changed or improved? He said, uh, give as much detail and data as possible in the replies.

Colin: And obviously, you know, he got tens of thousands of replies. It's doesn't take a genius to work out what he's gonna do. He is gonna scrape all that data. He is gonna use AI to analyze it. He'll probably do it a couple more times. Then they will use that to see the real world feedback of what the important people, uh, on X who are the vibe coders or [00:13:00] are the I or whatever it is will do.

Colin: And that is what they will do. And to me, I think that is incredible. 'cause I think that's kind of what the way all software is gonna go. I haven't been involved in a huge amount of software projects, but I have worked for software companies and I kind of know how it works and sometimes it can be.

Colin: Terribly done and really awfully put together where the people building the software don't actually really truly understand the mentality of the people using it. And I think all of that is gonna change. 

Mehmet: Absolutely. And you know, again, people might say Grok is another moment where talking about Elon Musk and X and all the things that's happened because Yeah, it's really.

Mehmet: It surprises me, like when I ask someone, which tools are you using? And someone tells me, uh, only charge GPT maybe. And say, you're missing a lot now. Yeah, yeah. Now another, another thing which I. I'm gonna go from [00:14:00] it to the Web3 world with you, uh, Colin agents, AI agents, and you know, like this is actually what you are focusing on today for a use case, which is, you know, related to the blockchain and Web3.

Mehmet: But AI agents now, we saw attempts before and I remember, and I was still new to this podcast as well, I was doing solo episodes. Um, I. I saw, you know, the first attempts of auto GPT and, uh, baby a GI and all these, yeah. Other open source and community-based tools. And I said, okay, it's very early, but these guys are up to something and all of a sudden everyone starts to talk about AI agents.

Mehmet: What's your take? You know, what, what are we trying to achieve with these agents? 

Colin: So I've been actively writing about agents for quite a long time. Um, I. A year? No, actually it must be at least 18 months. 'cause I'm in my current role a year and. [00:15:00] I've noticed a massive turnaround in the last three months where all of a sudden now people are really taking agents seriously.

Colin: So let me first define what I believe as an AI agent is an autonomous piece of software that can operate independently to complete meaningful work. Um, by, by that I mean goal oriented tasks. There are lots of bots out there who can do one. An AI agent can do multiple tasks, but you give it a goal, it, you know, like Chachi, people go, here's all the different things that need to be done.

Colin: Often do it. It can use logic and you can use reason. It can problem solve and actually learn from its mistakes, but then it can use tools and that's the most exciting part of that. So you will in, by the end of this year, I believe there will be agents where you'll connect all your apps, your Gmail, your Slack, your.

Colin: You know, WhatsApp, your all your different messaging, and then probably lots of different business apps as well. And you will tell it to [00:16:00] go and do something and it will go and do it, and it'll come back when it's complete. And sometimes that task will take a minute and sometimes it could take an hour and eventually they'll be doing tasks that'll be taking days, I believe now.

Colin: So to me. Uh, I believe that AI agents will bring AI to life, and what I mean by that is that AI is fantastic. Now, we've had AI for 10 years built into applications and things like that. Then we have Gen ai, which brings in chat GPT, and is remarkable and transformational as it is. It is still an empty box that you type a question and you get an answer in text.

Colin: And as incredible as it is, 'cause it's got the whole worlds of, of human knowledge in there, that's kind of all it is. It doesn't actually do anything now when you, uh, ask it to make an image or a video that's kind of doing something a little bit more, but when it's going to be able to. Um, reserve me a hotel, book a flight, and then tell me the best restaurants to go to within [00:17:00] my, uh, budget and, and, and taste.

Colin: That's a bit bit better, and that's coming very, very, very soon. But also it'll be able to automate a whole load of things in your actual life. Okay. And I'm writing a book about this at the moment, uh, because I believe the people who will stay ahead of the game here using AI agents. I believe it'll take away a lot of the drudgery, the repetition, and the, the stuff that we don't like doing.

Colin: Like say for example, yesterday, I spent a really long time updating all my notes in CRM, and I hate doing that, and most sales guys do, but I think that'll all be automated. I. Um, you will have a digital personal assistant who will listen to everything that's going on, on your phone, on your computer. It'll know everything that's happening in your business.

Colin: It'll know every call that is happening, and it'll prompt you to, you know, to, to follow up with that guy or to go through your task list. Or it will, you know, um, write that brief that you talked about on the, uh, call and send it to the customer automatically. So. The [00:18:00] world's, the oyster here for these sort of things.

Colin: It'll start slow and it'll start to really, you know, get better with its abilities. But AI agents are very, very close now. They're really coming and it's gonna be transformational. The other one I would notice that, uh, I would mention, which I'm sure you've used, what I think is amazing, is the deep research.

Colin: So D three deep research is built into, um, grok, uh, uh, Chachi PT and uh, deep Seek and a couple of others. I've been playing around with this a lot. I. And I find it not perfect, but pretty damn amazing. And yesterday I just wanted to buy something and instead of, you know, I was just looking for the best price on a very simple item.

Colin: And instead of me going to, to 10 or 20 websites and I just got, um, I just got gr to do it. And it went, and you know, it, it's fantastic. It's showing you what it's doing and what it's thinking, and then wanna find family the best place to purchase this item at the best price with free delivery where it can be here tomorrow.

Colin: That I think is really, really cool. And that's kind of [00:19:00] the, the, the first iteration of these, you know, highly usable agents that are super simple. It just goes off and do it in the background, the whole computer use thing where you're, it will actually take over your browser and start using things that will get remarkably better in the next couple of months.

Colin: And by the end of the year, we'll we'll be using it. That's my prediction. 

Mehmet: Yeah, I'm happy you mentioned deep research because, and all of them, like I I, I tried the three exact ones you, you just mentioned, Colin, because I think this is what is available at least, um, you know, for, for us. So, um, when I was a little bit skeptical at the beginning about what I was reading for people who get the early access first, and then once I started to use, you know, grok.

Mehmet: And Gemini surprised me how good it is, and I think this is the edge of Google having all [00:20:00] the web content under their fingertips, right? So, um, 

Colin: this is exactly it. I mean, you know, when when Chat BT came out, everyone went, oh, Google's dead. Google's dead, blah, blah, blah, rubbish. Because guess what? Google has all of our data.

Colin: It's got 20 years of my search history, my photos, my um, uh, maps, uh, but like, it has everything. It knows everything about me, whereas Chachi PT only knows what I've actually told it. Now they're two very different pieces of data. I, I would say. But dear God, I think, I think everyone knows that Google probably knows you better than your bloody wife does, about your preferences and your likes and dislikes and where you like to spend your time.

Colin: I mean, YouTube, if you wanna get to know me, just look at my YouTube history. I mean, you know, it's not very surprising. It's all ai, uh, 'cause that's all I do. But that's the advantage that they have. I use Gemini because [00:21:00] it's built into my phone. I have a. Like a one plus, and I just hold this button and it does this.

Colin: And then I go, oh, tell me about the latest, you know, things going on in AI today, whatever it is. And it does that. It's amazing. I use Grok a lot because Grok is built into Twitter, and if it's something up to date I need, it is brilliant at doing that. And I would find things if I wanted to. Ask about one of your previous guests and the statistic that they brought up.

Colin: You know, it can, it can search previous transcriptions on YouTube. It's, it's amazing. I, I, I don't think people realize, uh, the stuff that it can do, but then again, maybe they, they wouldn't find as much value in it as I do. 'cause I do these things all of the time. Haven't used debrief. Sorry, I haven't used, uh, deep seek.

Colin: A lot. Just a little bit. Yeah. Um, the problem is, is with chat GPT, I'm. So locked in because I've spent two years pumping it full of my data. I record all my call. I think I told you [00:22:00] this last time. I record all my calls. I transcribe them and I put them into chat, GPT, and I create a database. Of my business so it knows everything and then I can ask it to spit stuff out.

Colin: I'm working on kind of multiple startup ideas right now, so I just brainstorm for 20 minutes and stick it in. And then I got on a call with my buddy and we do it, and I put it in again, and then I get it. Output new ideas and suggest new ways of thing or new, you know, it's remarkable. It is out. It is outsourcing your brain to it, a bigger digital brain and.

Colin: I don't think most people have fully grasped that yet, but I think it will become much more obvious. And the last thing I'll say is I'm wa, I've been waiting for, I dunno how long on the delivery of something called the limitless pendant. So it's a little thing that goes around here and it's just, it's essentially a microphone, but it's on twenty four seven.

Colin: It listens to everything. It. What that means is that the AI will have full context of your life. So instead [00:23:00] of you typing into an ai, okay, I'm working on the business idea and we're gonna be doing this thing, and you know, what do you think? Or, uh, Hey, I'm, you know, looking for this, whatever it is, it already knows and it's heard all the conversations and it's like, well, I heard you mention that thing there.

Colin: So actually I went and I found this, and I put this together for you. That is what actually makes me most excited about where this is going. And I've said this for a long time. There will be a race to essentially get as much possible data of your own personal opinions and data into an ai because the more you do, the better the outcomes will be.

Colin: Right? Of course, we're 

Mehmet: gonna hear a lot of, uh, people having some. I would say constraints with the privacy issues. But honestly, you know what I tell people, guess what, like already you gave your I'm not, I'm not advocating for That's 

Colin: already listening to you. 

Mehmet: Yeah, yeah, yeah. But I, I tell people, okay, but guess what?

Mehmet: Like [00:24:00] exactly the, I give the phone example. I give the Netflix example, I give the Amazon example. I said, guys, like really already, you know. Majority of the companies, they know what you do. Like, you don't even have to talk sometimes, you know, like your activities on, on the internet would, would, would put that.

Mehmet: Now put saying, you know, all these things about, uh, you know, the ai, the deep research and all this, how all that intersect, Colin with, you know, Web3 and the blockchain. 

Colin: So I believe that, uh, AI is coming to solve an awful lot of problems in Web3, but I also believe that Web3 is coming to solve an awful lot of problems in ai.

Colin: They're, they're essentially a very happy marriage of, uh, two technologies that will come together to create something significantly better. So what we and [00:25:00] Griffin do is we, um, enable any. Web3 app, like a Wallet or a DAP or a Dex or anything like that to come to our marketplace to find a whole host of different AI agents and to integrate it into their own app or technology to make things significantly easier.

Colin: So it can help with customer onboarding, customer journey to. Uh, customer service, but it can also do significantly more, uh, complicated things like make transactions or automatically make transactions. So you might, you know, create a trading bot. Uh, you can, we have a product called Chat, um, called, uh, the, uh, transaction execution agent.

Colin: And it's literally like chat PT for your wallet. You can act your wallet and you can talk to it. You can just. Ask it to, um, make transactions or do some research on where the best staking opportunities are, or help to, you know, decide what coins to invest in. So there's a lot of really cool stuff there.

Colin: And then there's an awful lot of projects that are out there in the world of Web3, just bolting on AI and [00:26:00] calling themselves an AI project. And I think everyone's pretty clear on, on, on that. But there's also a lot of really fantastic AI projects that are out there trying to solve the problems of. Just the sheer high learning curve to getting into defi, into becoming proficient with it and just the bad usability.

Colin: Web3 is known for bad usability. 'cause the entire ecosystem is way too focused on the underlying tech. Oh yeah. We're using the blockchain and we're on this layer one and our TPS is this, but no one gives a shit. People only care about, you know, why is this piece of software better for me to use? Does it solve my problem?

Colin: Does it provide me a better user experience? And that is where AI is gonna come in and fix an awful lot of things for Web3. 

Mehmet: That's interesting again, like, and I was checking the website. So you have these kinds of, uh, personas or like avatars, uh, um, you know, within the agent. So you can give the agent kind of, uh, a personality to act on a specific topic, which is super cool.

Mehmet: And, and [00:27:00] again, like it's bringing a character or let's say. Uh, something that we can visualize, right? So it's like we, we can visualize this agent because when, if today I'm speak about, let's say menace, right? Uh, other than the videos that we've seen of people using it and seeing, you know, people just, yeah, typing a prompt and then we see something working by itself.

Mehmet: So it's like. Having it as if it was a, you know, robot maybe. Uh, or, or, but at least in, in a virtual sense, I'm speaking like, like this. Uh, so, and speaking of this, Colin, like, um, so. You mentioned there is a lot of problems in blockchain that AI can solve and vice versa. Now, one of the things, usually people, especially when it comes to the Web3 and, and that domain is, you know, the amount of, you know, and especially you talked about NFTs and [00:28:00] we saw also like recently in, in the blockchain domain, or let's say like the crypto domain, these, uh, uh, what they call the, the meme coins and all this.

Mehmet: Yeah. Is there anything like AI can help. You know, also people, uh, when they decide, for example, to, to go for a specific new project or maybe if they're gonna go and, uh, I don't know, support a specific meme, coin or something like that. Is that something also that, uh, would, would AI would help in. 

Colin: Yeah, I mean, at the end of the day, what does AI do very well.

Colin: It ingests large volumes of data. It summarizes and understands it, and then, you know, spits something out the other end. So that's very, very, um, applicable for the world of meme coins because I'm not a meme coin guy. I'm not a DGN defi head or anything like that. But I do know lots of people who are, and they will spend their time on.

Colin: Six or seven main websites to pull together the [00:29:00] various different data. Uh, and then they'll have lots of other different places. And then they'll be on X and they'll be on Discord, and they'll be on Telegram and things like that. AI can essentially pull together all of that data. And provide some outcomes to you in a specified format that is helpful for you.

Colin: So we're building something called the Meme Coin Hunter. So it goes and, you know, it can look at, um, it can look at trends. Meme coins tend to be flash in the pan. You know, most of them are spike up and they're down within 24 to 48 hours. There's a few like Dogecoin and Dog with Hat and Pepe and a few others that have kind of lasted, but.

Colin: The ones that are launching on Solana every second these days, they tend to be, you know, up and down so it can look at certain trends that are going on TikTok, and then a, you know, a a, a meme coin that's launching that, you know, has certainly got support from some reasonably well followed influencer, and you might be able to ride the wave there.

Colin: But you know, it's degenerate gambling and that's kind of it. [00:30:00] Uh, I, I sort of do hope the meme coin thing will pass. It's not gonna die. It's not going to go away. It's part of the nomenclature now. I mean, that's it. But I sort of, kind of hope, and I hoped this before the meme coin thing came off, that projects with real solid utility and use case will be the ones that'll do well.

Colin: I was proven wrong on that. Um, my hope is that eventually things will turn around and we'll see the good projects, you know, sail 

Mehmet: right now on another important, uh, breakthrough I would call it, which happened. And for the sake of, uh, transparency with the audience we, we recording on the 18th of March. So, um, last week, uh, we've seen Google bringing Gemini for robotics.

Mehmet: And you know, people, again, maybe they didn't focus on it, but I see it's something really super cool there. [00:31:00] And I know, like Colin, you, you wanted to share also something regarding the robotics in general and what AI is bringing on the table. 

Colin: Yeah, I mean, go look at my post from yesterday on LinkedIn. I, I kind of, I like to update everyone on what's going on in robotics because every week, like Chachi, bt, there's another extremely impressive moment of, uh, an autonomous robot doing something pretty amazing.

Colin: Now, the latest one a few weeks ago, you might have seen was figure from Brett. Brett Adcock. He had two robots. He came up to them, dumped a bag of groceries in front of them and said, sort this. And the two robots worked together. One of them grabbed an apple and then he handed it to the other one, and then he opened the fridge and he put it in.

Colin: I. And then another guy grabbed a box of cookies or whatever it was and put it, you know, in the drawer over here. And it was, and there was one amazing moment of the, the, they were handing the two robots actually looked at each other [00:32:00] as if to acknowledge that they're passing this thing very, very human.

Colin: It was quite crazy. Um, yes, Google are getting into it. Uh, Nvidia are building essentially an entire operating system. For robots to understand the world because like self-driving, you know, it's easy to get a car to self-drive on the highway. It's very difficult to get it to operate on the back streets of Bangladesh.

Colin: There's just, you know that this, this like curve you, you can't program that so you can't. Program a robot to do every single task. 'cause there's an almost infinite amount of things that it might do. But what figure have done is they built this, uh, operating system based on a neural net. And I've been saying this for ages.

Colin: The future of computing is neural net in, sorry, data into a neural net and then output. That's it. Okay. It'll, it'll do the whole thing. And they're releasing this like robotic, uh, understanding of the world neural net so that it. Can do anything and you can literally show it [00:33:00] to, to how to do something yourself, like you would show a friend and it will learn.

Colin: And let's not forget. With a robot. Once one robot learns, they all know how to do it. So we all have this vision of AI being this like supercomputer in the sky that, you know, will know everything and then with agents can do things. No, because actually where a GI or anything like that is really gonna crum to life is through.

Colin: You know, uh, humanoid robots, it's the embodiment of this. A GI. And I was driving along in the car thinking about this yesterday. I was like, you know, we are gonna have, I think Elon Musk says he plans to be outputting a million robots a year by 2027. And if he's doing that, then there'll be several, you know, other companies doing.

Colin: Not as much, but, but similar. So they're gonna be everywhere very, very fast. Okay? They're only 3% the weight and complexity of a car so that they [00:34:00] can be produced an awful lot. And we're gonna go from, oh my God, there's a robot. Or I saw another robot the other day. To them being everywhere in a a few years, that's what's gonna happen.

Colin: But if you think about these robots, they're gonna be. You know, uh, dextrous and able to do anything and help you and blah, blah, blah. But they're also gonna be the most intelligent person on the planet because they'll be able to answer anything for you and they'll be able to do anything. And they've been trained on the entirety of YouTube and things like that.

Colin: So, you know, if I go on YouTube and someone goes, oh, here's a new way to redecorate your room. And we can, you know, get some wood and we can build this and we can build a banner and we can paint that. And I'm not very good at that. I can't do it, but hey, my robot can. And that's just one stupid example, but we're going to very, very quickly enter into a world where we're gonna have these robots all over the place and they're gonna be able to do kind of almost anything we can do, you know, within a subset of that, maybe eventually [00:35:00] everything.

Colin: And it's a. Terrifying thought because we're about to enter a new paradigm of a world that we can barely imagine. But I really do believe, I try not to be a DOR when it comes to ai. I think it is AI will take jobs. It will, okay, it's undoubtable. It will also create more in a certain way. But I believe that eventually in 20 years time, we're heading to a place where.

Colin: You know, uh, the industrial Revolution took away the ability, the, the, the need for us to have strength and muscles and stuff like that. And then we have, you know, machines to do things and then we have computers and, and things like that to do things faster. But it's all very task, uh, uh, uh, uh, relent now, you know, knowledge work is going to become.

Colin: Democratized over the next kind of 10 years and embodied into humanoid robots. And it could, I hope, bring a area of complete abundance. I mean, if you think about [00:36:00] 500 years ago or a thousand years ago, you know, someone had to go out and catch some birds or rabbits to cook on the fire for dinner. Now we don't have to worry about that.

Colin: You know what I mean? Uh, it could be the same for many other things. 

Mehmet: Absolutely. And of course, like this is debatable, you know, the way from two sides, like, I mean, um, how fast it gonna happen and you know what will happen actually. Yeah. Uh, I'm like you more on the optimistic side of, of the, uh, of the house.

Mehmet: I would say. Um, I, I see I'm not seeing it very far, by the way. I still believe that in my lifetime. I gotta start to see some of what you mentioned, Colin. Yeah. Um, and 

Colin: it's not like, I mean, look in, undoubtedly in everyone's lifetimes, I mean, look, you know, the way we talked about following the vibe coders on Twitter, I follow, follow a lot of the, what's called e [00:37:00] slash aac and it stands for effective acceleration.

Colin: And what they mean is, is that there's no point in putting on the brakes of trying to control AI or trying to regulate or trying to moderate it. We may as well just go full pelt and see where it gets us. Okay. And most of these guys are basically going. Uh, well, a GI 2030 is kind of what I think, 20 30, 20 32.

Colin: But a lot of them are like, no, 20, 27, you know, and they're also, I mean, you know, every week something happens where we're like, oh my God, I, I did not expect that so soon. We didn't expect soa, we didn't expect deep seek. We didn't expect Manus. Every month or so, we're coming out with a new, I call it a cliff.

Colin: You know, it's not, it's not just a gradual one like that. It just goes B, and then all of a sudden we're there and then well, and we're up again. Right? That's gonna keep happening. It comes down to, let's say, the scaling loss, and if anyone doesn't [00:38:00] know, it's basically that if you get a data center and it's got 10,000 GPUs in it, you'll get X, Y, z.

Colin: If you multiply that by 10, put a hundred thousand, it goes up by an order of magnitude and. There doesn't seem to be any clear limit on, well, if we keep getting bigger, then it'll slow down. It will. But not only that, the models are getting more efficient. I've been saying this for ages and then out comes deep seek and proves, I mean, we don't know how, how much, but proves that they could definitely, they did this change of thought, steady using tokens that kind of use entire sentences, I mean.

Colin: They talk about, you know, 700 billion tokens and things like that. If you take a PhD who's 60 years old, he's probably ingested about a billion tokens in his entire life, and I think that ais will get the same. The models will get smaller, they'll get better, they get smarter, and everything is just going to accelerate.

Colin: We're on the exponential curve now, so. It's not within our lifetime, it's within this [00:39:00] decade. We're gonna see things radically change and the next decade, it's really hard to tell because we're just not very good at predicting the future when it's gonna change this much and this fast. 

Mehmet: Absolutely. And, uh, for me also, I'm not relying only of course, like, uh, maybe I start to act like, uh, like an algorithm by myself, you know?

Mehmet: Mm-hmm. Analyzing previous data and trying to see like if I can expect things. So I'm not relying only on the breakthroughs that are coming from the technology companies alone. So I'm following. You know, different kind of people who are in different kind of sciences. Like I, I'm following people who are in robotics.

Mehmet: I'm following people who are in, uh, genetics. I'm following people who are in, you know, like, uh, working in labs and laboratories. And so, and I'm interested much in people who are [00:40:00] researchers also as well, right? So the PhDs and. Yeah, so, so, uh, because you know, the logic says that if we keep doing what we are doing, the cliffs, as, as you described them, Colin, um, some of them they believe that a GI depends how you define a GI, of course, but some people they say it's coming this year, right?

Mehmet: So some of, and, and they are like. You know? 

Colin: Yeah. I mean, some people think a GI is here already because we have a, you know, box that has all of world knowledge, but that look at the end of the day, it really, really depends on what your definition of a GI is. Because there's no set definition, uh, OpenAI define it as.

Colin: A system that can do, I think it is the, the majority of meaningful, valuable work. Human work. Okay. That's a pretty high bar. Uh, Sachin Nadela came out the other [00:41:00] day and he goes, I'll tell you when a GI is here when we're getting 10% GDP growth every single year. Okay. Uh, that's a very, um. Uh, that's a very set of numbers and, and, and growth phase based one for me, it, it, it's, it's, it's when I'm gonna be able to have a computer used style thing that'll sit on my desktop, watch it, and start doing my job.

Colin: Okay. And, and, and, and start doing lots and lots of other people's jobs in the fact that. You know, but what happens then? What happens when someone releases a piece of software and hey, OpenAI are already talking about $20,000 a month agents that'll do the job of a PhD researcher. Okay. But that's the thing.

Colin: But you and I, the job we do, or anyone like us, it's a very different job than a PhD researcher where they're just like working through a series of, you know, problems and putting a piece of thing. I, I don't know. You know what, what do we do? Thousands of different things every day. It's, it's, it's not really easy automatable, [00:42:00] but it, someone will just come out with a thing and it's like, right, here's your marketing team in a box.

Colin: Five grand a month, 10 grand a month, a hundred grand a month, whatever it is. And it will take your startup and it will just do everything because marketing is quite a established process. And you know, there's all of the different things, there's all of the different tools and so that's kind of do a ball, you know what I mean?

Colin: Um, is it gonna take the job podcaster? No, I don't wanna, I don't wanna watch an AI podcaster, you know? Yeah, 

Mehmet: of course, of course. You know, 

Colin: behind me here, I, I, I used to be a dj. I still play around a little bit. Okay. AI has been able to DJ for 10 years. Are you gonna pay money to go and see a computer play songs?

Colin: No. We still need the human element in this. Uh, that's kind of the important thing. But there's a lot of people in government jobs and low end administrator jobs who are just [00:43:00] pushing pencils around and, and, and, and doing administration and bureaucracy and things like that, that AI will be able to do a ton of their jobs.

Colin: And so what happens then when it's not like everyone's gonna lose their job at the month? If, if, if we got 15% unemployment over a period of six months. The world will turn upside down. 

Mehmet: Yeah, right. This is why I tell people about, you know, back to, to the point of, you know, getting skilled and getting skilled using these AI tools.

Mehmet: Um, there's another thing which I say is I think we have abundance in AI currently that sometime I feel people they are not even aware where they can use it. So, um, probably, you know. At some stage, I gotta, I gotta, you know, leave this as a, uh, thoughtful food. So, at some stage, we remember back in the late [00:44:00] nineties, beginning two thousands where people, they used to put on their resumes, like as proficiency in, uh, typing Microsoft Word and Excel and using Windows.

Mehmet: I don't know, it was XP at that time maybe. So it's gonna be like. You know, they should show a proficiency in, in, in using AI tools. Yeah. You, you wanted to say something, Colin? 

Colin: Yeah, well it's, you are absolutely right. I mean, people might put a prompt engineering on their CV today because it is a skill that not everyone has that.

Colin: If you're good at it, you can get the maximum amount of ai and it's exactly the same as, oh, you people used to be able to use word or, or typing. Going back a bit further. I believe that AI has to change and evolve to become more human. Mm-hmm. And what I mean by that is if I called you up and I said, uh, uh, you're an expert in, um, pitch decks, uh, [00:45:00] uh, for for startups, what are the elements of a, a good pitch deck?

Colin: What's the first thing you'd say to me? Well, you would ask me some qualifying questions. You'd go, well, hang on. I sure. I don't know. What's the business? What's a series A seed? You know, it has to qualify. AI really doesn't do that very much anymore. It just gives you an answer. However, I will say I have noticed that open ai, deep seek, sorry, open AI deep research.

Colin: Does come back now and ask you some clarifying questions, right? Gives you a thing like, oh, it asks you, okay, are you, are you, do you mean about that or do you mean about that? And I, I've got caught a couple of times 'cause I hit enter and then I go off and do something else, come back 10 minutes later and it's just asking me a question.

Colin: So it's starting to do that because, you know, I, I, I've always been a big proponent of voice in ai. I think. Certain circumstances, certainly not all, and not even, you know, just a small point. [00:46:00] But I think voice will be amazing because it is the. Preferable mode of communication for humans and has been for thousands of years.

Colin: You know what I mean? The, the conversation in my LinkedIn, I call myself a storyteller because, you know, I, I think it's a really important trait. Um, the conversation has gone back thousands of years before the written word, uh, and I think AI will become just like your digital friend and your digital buddy, and you'll ask it a question and it'll, it'll.

Colin: Ask you questions again and it'll confirm before. Instead of giving the answer right now, it just spits out an answer the best it can with the little information it has. So that's a change. The other thing that I think. We're gonna have AI personalities, because if I had a question about pitch decks, I'd call you.

Colin: If I have a question about financial analysis models, I'd call, you know my other friend over here. Now, obviously I know that AI is AI and it has all of the data, [00:47:00] but sometimes you want a different answer. Depending on the type of person you, you wanna ask, if that makes sense. So, uh, it's like the way I go to grok for one question today and I go to OpenAI for another.

Colin: And even the different models, I think we're gonna have like conversational avatars that one guy will be, Hey, if you, uh, just want some. Uh, superf information on the stock market. And another one will be if you wanna have a deep philosophical chat about the meaning of life. And then if you want to talk about celebrity gossip, you go talk to Paris Hilton.

Colin: And if you want to talk about sports, you'll talk to Tom Brady. I used those examples 'cause Facebook about a year ago came out with exactly that, but they never fulfilled on it. I do think that is common. And little by little bit by bit, we're just going to ingrain AI into everything we do. And the next step, maybe not this year, but certainly next year, is you're gonna have a pair of glasses, almost [00:48:00] identical to the ones you are wearing right now.

Colin: And you'll just go tap, tap, hey, um, remind me to do that thing, but also send that email to me met and, uh, go do some research on this new thing and, and it'll do it. It'll be listening all of the time. It can see as well and we will become like cyborgs. 

Mehmet: Right. It's exciting time. No, no, no doubt that we are living, uh, you know.

Mehmet: Like, every time I witnessed, I would say another breakthrough. I was so excited, said, yeah, this is the moment. But you know, this time I can feel it really, really different because it's, it's simplifying, you know, life. I'm, I'm going a little bit philosophical here, and I, I think what AI is doing, I mean, all, all the use cases that, by the way, even the vibe coders, you know, they're proving, they're proving a.

Mehmet: You know, like, oh my God, like how much have we complicated [00:49:00] things on ourselves? Like because, you know, ai, one of the things which is, you know, because of the way it's programmed, it tries to optimize, it tries to get you the shortcuts of course, and tries to automate and then you figure out, oh my God, like you just mentioned, for example, people who are like still working as clerks and you know, filling things with pen and paper.

Mehmet (2): Hmm. 

Mehmet: Like, look, look how much we could have had fun instead of, you know, hammering ourself in these repetitive nonsense tasks sometimes. And now, you know, this is why I'm excited about the age of, of, of AI and, you know, everything, uh, AI is bringing to us. Um, Colin like. Really, very, very, you know, rich conversation.

Mehmet: I would say. A, as we are coming to an end, any final thoughts? Any, you know, you mentioned about LinkedIn. I'm gonna put the link there, but if you want also to guide, uh, the audience also as well, uh, somewhere. 

Colin: Yeah. The, the only final thoughts would be, [00:50:00] the question that always comes up is something to do with, do I have to be afraid of ai?

Colin: Is it gonna replace me? Is it gonna take my job? And things like that. Okay. Uh, the answer is maybe if you bury your head in the sand and do absolutely nothing. Okay? But I can promise you that if you do try to stay ahead, uh, or, or close to what's going on, be aware of some of the tools that are coming out that you can use.

Colin: And, you know, there's basically very, very few right now. Apart from, and things like that, I mean, they, you know, agents or whatever, but they're coming on mass very fast, and if you just try to get an understanding of what's available. Play around with them and start using them. Start, you know, exploring and experimenting for, to save you time, to automate things that you do.

Colin: If you get into that mindset, you'll be so far ahead of 99% of people out there pretty much that you really have nothing to worry about. This is, uh, myself and a friend of mine are [00:51:00] writing a book about exactly this right now. Uh, because I think this time is coming, the, the, the gap between people who use AI and you and, and don't is gonna widen really fast.

Colin: Uh, so be on this side of it. Don't be on that side. That's kind of what I would say. I. And Okay. It's easy for me 'cause I'm a big nerd and I just sort of love this stuff and not a lot of other people are much more concerned about watching football than, you know, AI videos. But it will be fairly ubiquitous in, you know, coming out into software applications.

Colin: AI is software 2.0, I call it. It's just, it's gonna take over absolutely everything. So from your businesses' perspective, I mean the CEO of IDM came out and said the future of, you know, successful businesses is those with the best adaptation of AI strategy and tools. So if you're a small business or anything like that, whether you're employee or whatever.

Colin: Uh, my advice is to just do your best to try and keep on, uh, keep up to date with what's going on, using these [00:52:00] tools and experimenting and integrating them into your own process, workflows, and life as you can. And, uh, you, you definitely won't regret it. It will be a really fun, uh, thing to do. It will save you time and give you more time back to do the things you want to do, your hobbies, spend time with your family, et cetera, not, you know, working 60, 70 hour weeks.

Colin: And that's what I hope it comes to in the end, that we can all kind of work less and get more done and just, you know, get these, uh, AI bots and computers and robots to, to do most of the work for us, where we focus on actually 

Mehmet: living. I. Absolutely. So you mentioned the LinkedIn profile at the beginning, Colin, I'm gonna put it in the show so people can, uh, follow you and, uh, thank you.

Mehmet: You know, reach out to you also as well. And really it was, you know, very, uh, engaging and interesting discussion with you today. So thank you very much for your time. Always a 

Colin: great conversation with you. Mohammed. Let's catch up for coffee in Dubai at some stage as well. [00:53:00] 

Mehmet: Absolutely. And this is how usually I end my podcast.

Mehmet: This is for the, uh. Audience, guys, if you just discovered us, thank you by for passing by. I hope you enjoyed if you did, so please give us a thumb up and subscribe. Share it with your friends and colleagues, and if you are one of the people who keep coming again and again, thank you very much for all your support.

Mehmet: I'm repeating it since the beginning of this year because of your support, because your. Following what happened this year is something magical for the first time. I'm been doing this for the past two year and a half, almost now. The podcast is trending in the top 200 chart in multiple countries at the same time.

Mehmet: Amazing. So this is, thank you. And also we. We've been selected as one of the best to listen business podcasts here in Dubai. We are in the top 45 and ranking around 15 now. So thank you very much. So this is also couldn't happen without all your support, and as I say, always stay tuned for a new episode very soon.

Mehmet: Thank you. Bye-bye. [00:54:00]