Ep. 12 | AI Literacy is the New Excel: What Every Professional Needs to Know Now with Sherry Jiang

AI is no longer the domain of specialists, it’s fast becoming a baseline skill, much like Excel was a decade ago. But what does “AI literacy” actually look like in practice? And how prepared are today’s teams for this shift? 

In this episode of f, Richard sits down with AI leader Sherry Jiang to unpack how the role of the AI engineer is evolving, and why that matters for everyone else in the organisation. 

Together, they explore what skills are quickly becoming essential across the workforce, how businesses can build AI-ready teams, and what leaders need to rethink when it comes to hiring, training, and managing talent in an AI-driven world. 

If you’re navigating digital transformation, scaling globally, or simply trying to future-proof your workforce, this conversation will give you a clear, practical lens on where things are heading, and how to stay ahead of the curve. 

Transcript

SPEAKER_00 0:01

Hey everyone, welcome back to Work Writers with Rich, the show where we get into the honest and filter side of what it really takes to make work, well, work. I'm Richard Farmer, your host and managing director at We have all been hearing a lot about AI, but not always from the people that are actually building with it day to day. So I'm really excited for this one. My next guest is an entrepreneur, founder, and pioneering AI educator. She's passionate about opening people's eyes to what is possible when AI is used with intent and understanding. So welcome to the show, Sherry Jang.

SPEAKER_01 0:41

Thank you so much. Um glad to be here, Richard.

SPEAKER_00 0:44

Yeah, really glad to uh have you on board. I'm really excited for this conversation because I think you're going to give us some tools and tricks and some ideas of what's really possible in this new world. But let's just start us off. Let's get to know you. Um, so can you tell a little bit about yourself, Sherry, about your background and how you first got into AI?

SPEAKER_01 1:03

Yes, absolutely. Uh so I am an entrepreneur today, and I actually run a couple of different businesses. I have one AI product business, which is where I do a lot of my engineering work. It's a it's a personal finance app that uses AI to help the younger generation save money with a little bit more fun. But a lot of my also my other work um in my businesses is actually teaching and doing the community work of bringing people up to speed on how they could actually build applications and do all sorts of this stuff with AI using some of the lessons I've learned building my own company. Before all of this, I used to actually work in corporate. I was at Google for a large part of my life. I worked on consumer products there, especially in the payment space. So I've always been really interested in fintech, uh understanding the behavioral science of how people make decisions with money. So that's obviously translated into the work I do today as well. But I'm really a huge advocate around um the joy of learning how you can use AI on your day-to-day. Uh, I would love to share anything that could be useful to the audience around how they can get started on making their lives easier using some of these tools.

SPEAKER_00 2:16

Ah, fantastic. We'll delve into many of those. I think you you just kind of piqued my interest, though, with two comments there. Is one, you've got obviously you've done corporate, you've got your own business, but then you mentioned two things. Two of the biggest things that are clashing at the moment, which is teaching and learning and a new technology, right? Because I think the world of education is going to change. Um, but you've adapted both, which is absolutely fantastic from a learning thing. But how did you first, you know, uh because you didn't start as a technologist, is that correct? You weren't a technical person. No. So how did you uh teach yourself or how did you which one came first? The teaching yourself or the technology? Which one kind of piqued your interest first?

SPEAKER_01 2:59

Yeah. So around 2024 was when uh I started getting introduced to these tools. Um, so I think the first wave of AI was a lot of this, like, you know, input-output. You you chat with Chat GPT, it does a little bit of web search, gives you an answer. But I started to notice something where people were starting to use it for much, much more, which is writing lines of code without actually having to know the syntax, right? It's almost like you write down what you want to say and you translate something to Japanese without having to know the structure of Japanese, but it kind of does it on its own. So I had this light bulb moment where I was like, you know what? I've understood how to build products for over a decade in my career. I know how the architecture of products work, but what had always stopped me was the syntax and the tools, right? Because that like wasn't something I learned to do in a traditional way, which is the software engineering practice. But that barrier kind of went away with AI, where it abstracted that part away. So I kind of fell into a rabbit hole where I started to build anything I can think of under the sun for fun, honestly. So I built like my my first app was actually a rent versus buy calculator, because in Singapore, uh, housing decision is a very polarizing and heated topic. And I used to tell people, look, sometimes renting is better than buying as country. So I think it is, yeah. Right. So I build uh an app where it actually pulled public data around uh value appreciation, um, the alternative, uh, your opportunity cost of buying a house, which is how much your returns you can get in the market, and actually gave you whether or not renting or buying for specific properties made sense. And that tool actually went viral for a bit. It got like a few thousand usages in a day. And it just made me realize, hey, all I had to have was an idea. And I just spun up a product and like that just that kind of clicked for me that in that moment. And ever since then, I started to build my own small apps until I got more and more confident to even working with like my pro now my uh company's entire code base, right? But that's that's how it started. It was like a how might I actually use this to do something, even if it sounds ridiculous, and then getting the first taste of oh my gosh, it's possible. And my mini apps actually have users, to now actually being confident enough to actually bring it into what I'm doing with my work.

SPEAKER_00 5:28

Yeah. Well, I mean, I you started off with a grenade there, really, basically, um, challenging the status quo, straight away to something that's historic, right? Between the the buying and and and the renting conundrum and then the like shares, where do I save my money? I think it's probably one of the most important things in Asian culture. Yeah. Um, one of my other guests, Georginia, she she started off with a very similar thing, and she does it with her son. She was just right, what problem can it solve us for today? And the more you're reading about AI is is don't start with a simple problem, simply one with it's challenging or a problem that really needs solving in your life. I think that the hardest thing, you know, for us, and we'll probably dig into Bradis, there's so many platforms to look at, and I'm sure all our listeners will be like, which one do I use? I know I think I'm using about four at the moment, I even forget which one I'm actually using half the time. So it's very it's it's it's a it's a very complex thing. So you touched there on the uh on the on what you built straight away, but what kind of moved you towards the teaching part? Yeah, you're like, you know, to actually have the confidence then to go, I'm gonna instill knowledge to other people. Because I'm a big believer in, you know, when you when you're in business, it's not just been a bit of teachers and being a coach, right? I think it's being able to share that information. So what took you on that journey?

SPEAKER_01 6:43

For sure. Well, I think I always had a part of me where I wanted to see how I could make the impact that I have on the world beyond just my own company. So sharing content on social media, whether it's LinkedIn or Twitter, was something that I had been doing for years. Um, one of the things I started to build into that, besides, hey, here are the lessons I learned as a startup founder that raised money, was here are the things that you can build with AI. So I actually recorded demos of like myself actually coding with AI because honestly, when I was telling people I can do this, it's a little bit hard to believe. Like you've really got to show, not tell. So I would actually record. Um, one of the first ones that actually went quite viral on LinkedIn, I think I got like a million views or something, was I clone Google Finance. Yeah, I can show you the post maybe afterwards, but um, I I use a tool called V0.dev, which is created by a company called For Sell. And I took a couple screenshots of Google Finance, which a lot of people use for portfolio tracking. And I just showed people how it actually cloned the app just by mere screenshots. And I think people thought it was black magic at that point in time because it was, I think now people are more used to it, but at the time it was just like mind-blowing to see that. So that that really started to take off, right? So I started to post uh probably a video or two every single week. Just again, I'm doing this already. I might as well use Screen Studio to record what I do. I share it and then see what people learn from it. And then a ton of people started to reach out organically and they were like, hey, Sherry, like, can I grab you coffee? Could you show me how you do this? Hey, Sherry, do you have a PDF? Hey, Sherry. So at some point I'm like, look, I cannot do one-on-ones with 200 people, right? Like I realized that people wanted more than just reading my content. Like they wanted something more in-depth. So I was like, why don't I just put out a class? I mean, I never really had a formal course business or whatever. I'm not a I, you know, I didn't work for General Assembly, but I threw out a wait list on Google Form and was like, okay, anyone who wants to sign up for this class, just like put your name down. And if I see enough names, I'll run it. So 200 people signed up um-ish within like a 24-hour period. So then um my my partner on the business and I were like, okay, let's let's uh try to book a space and we'll pilot a course for 14 people to start and we'll do it as a pay as you wish kind of model. Like, hey, you guys are all guinea pigs, so we're gonna figure out like if this works or not. But the the feedback for the first session was overwhelmingly positive. I mean, people were like just uh really mind blown of what was possible. Uh, this is 2024, by the way. So August 2024. Um, people were installing Claude for the first time. Like, that's like the that's where we were at. People didn't know Claude. So yeah.

SPEAKER_00 9:25

Well, some people still are sherry.

SPEAKER_01 9:28

Okay.

unknown 9:28

Okay.

SPEAKER_00 9:29

There's some people still put in Claude in, but like, yeah, 2024. I mean, you're the kind of tip of the iceberg there in terms of like actually getting across people, but there's always uh individuals that are the first adopters, right? They come out and go, I wanted to adopt this technology, I'm inquisitive, I want to learn. I think that's fantastic because you bring in your product knowledge, also the way that you see things from a different lens. Because uh when I look at code, I've always said code's the one you've universal language, right? Even even sign language isn't global, but right? They could have made it global, but they didn't. Um, so you've got all these different languages, then you've actually got one language that people can learn, but then with the advent of AI, do you now see it that it is uh the universal language, like in terms of the because you can get it to a point that most individuals can understand and code, yeah, right? Which has always been important, does always have uh stop people from getting into technology.

SPEAKER_01 10:25

Yeah, yeah. I mean, at the end of the day, code is just symbolic meaning, and language is symbolic meaning, right? I am talking to you, I is an object, you are an object, and talking is an action, and two is a relation, right? And language is a symbolic representation of meaning. Code is a symbolic representation of meaning as well. And so actually, what most of the time you need to understand how to build any kind of products is actually just a logical sequence of how information flows, how things are constructed. And what I found is in teaching our class, which now we've had over a thousand people go through, is it actually doesn't matter what your background is, if you're a 13-year-old still in school, or you're like 60 some, you're retired, but you're just coming back into the space because all this is exciting. We all have that generalized general intelligence to be able to understand these symbolic structures, right? And code is just a way to actually allow us to communicate with a machine, to take what's in our head to create, okay, in a rent versus buy calculator, uh, your relationship between input and output should be defined UIUX-wise by a couple of boxes. And then the relationship between the numbers are what you calculate to arrive at the numbers, right? That's all in our heads. We're all capable of doing it. I think for a long time we didn't reach the code part of things because it felt really unfamiliar, right? Because it felt like there's like a lot of layers of complexity and abstraction to get to being able to create something. But that that's taken away now. And a lot of us actually can express ourselves that way and be able to build products.

SPEAKER_00 12:05

Yeah, I totally agree with what you're saying. I think when I first like had my own business, um, I used to sit next to some coders, and it was kind of like the coders were the cool guys and kind of, but they had their own community, where I think what it's done is it's brought down barriers and the ability to do things and try things, which I think is fantastic for the whole world. Um, and being able to just, you know, have an idea and try and run with it, which is great. So um, if we take a step back though, so like you know, um 2024, you've you've just started in your looking at the AI space. What do you think uh you know has changed the most about the AI space since you first started? Other than the fact there's that many platforms, but like um what what has changed the most?

SPEAKER_01 12:48

Yeah, I think that the accessibility of the space and the diversity of backgrounds is is truly astounding, right? I mean, I I can't go and say like ever it's completely mainstream and everybody who um you know everyone feels welcome, but it's it's it's a step above any kind of like I guess point in time in history as it relates to software where you have way more people from all different types of domains that are actually building now. And I think that's that's really amazing. So I I will share like us some stats, right? Because I I look at it from um because I Well, it was a stats. I yeah, I'm a stats and examples person, right? Uh uh so I I run a lot of these hackathons and builder events now. Um basically they're their uh hackathons would be like hundreds of people coming together. They have 24 hours to like build something, pitch it, uh, demo it, and then hopefully win something. And it's it's meant to use that constrained time space to like really innovate, come up with wacky ideas, and maybe a company comes out, or maybe you just push the limits of the technology. Um one thing I I found that was like so interesting was um one of the biggest hackathons that um I ran last year in Singapore uh was one called Cursor Hackathon. I'm actually wearing the logo of the T-Shirt.

SPEAKER_00 14:01

Oh, yeah, the t-shirt. Okay.

SPEAKER_01 14:03

Fantastic tool. They're my day one. They're the thing that taught me how to code with AI. So I owe a lot to them. So of course I'll run their hackathon. Um, it was it was pretty amazing. We asked people in the intake form when they were applying for the event is this your first hackathon? Have you done it before? And 50% of the participants out of 500 have never been to a hackathon before. And really, what was even more amazing was there was actually quite a few finalists who like made it to the top 15 out of uh 200 or so submissions that were among those who never been to a hackathon. And one of the guys is a mental health professional who literally downloaded the tools on Thursday before the hackathon and then attended the hackathon on Saturday and then placed within the top 15. But he built something again that was a very empathetic product that really was trying to help people coach through anxiety. So he kind of approached it with uh, I have some domain knowledge because I have 10 years of experience, right, doing like uh mental health counseling. I had never been able to build a product around it until recently. And I think that like depth of knowledge and understanding what the pain points of the users, what the users have, allowed him to create a product that actually stood out and um, I mean, beat out submissions by people who are like experienced engineers actually. So I think a lot of this has shifted right around who's kind of building in this space now. I think two years ago, a lot of people thought people who are non-technical, might like myself, were just building like gimmicky, like weekend projects. Like it's not gonna go anywhere. It's not serious stuff. But I think 2026 now, people are realizing that actually people without that traditional background are actually building some serious stuff in this space that's getting funding, that's winning hackathons, that's like, you know, getting acquired by larger companies.

SPEAKER_00 16:00

Wow. Yeah. I mean, do you feel, and this is uh kind of going at a piece, but great to get your input. Do you feel that the platforms then have the rigor to then become a proper product? Is there certain platforms that so my reservation around it? So I've looked at it and go, right, want to build code, but then you're like, you've always kind of been brought up with like, okay, Google, Microsoft, SAP, Oracle, and like they have to have this certain rigor for compliance for this. Yeah. But do you see that the platforms now that are accessible to people in the public or new um engineers or developers actually have the rigor to actually run a company or run a platform on that they can take out?

SPEAKER_01 16:41

Yeah, uh I think that at this point, everybody has the capability with enough patience and willingness to ask questions. Like if you're like a solo founder who's like just learning how to do all this and you're like, I don't want to ask for help. I want to figure everything out in like one month. And if in one month I don't figure it out, I stop my business. I could probably tell you like that is probably unrealistic expectations. If you're somebody who's like, look, I recognize what I can do, what the gaps are, where I might need advisors or consult someone or ask my smart friend who's a CTO somewhere else, like how to do something, I think you'll go very far. And so I've already been seeing this. Um, there's actually like an unbelievably high number of 17, 18-year-olds on Twitter that build indie businesses that make sometimes five, six figures a month. And they don't come from any professional background doing this. I mean, they're just like picked up a computer and they're like, you know, math class is boring in school, and I find it much more interesting to build and make money. Um, they're able to do it, right? Because they don't have the same kind of reservations we as working adults have, which is like, oh, like we're not meant to do this. Like we, I don't know if we can do it. When they're younger, they don't have a sense of what society tells them they can or cannot do. So they just do it. And yeah, I mean, maybe there's some things that you have to learn along the way. Um, but I don't think there's there's anything that you can't actually do with enough patience and enough persistence.

SPEAKER_00 18:08

Yeah, I totally agree with that. And I think it's just having that open mind. And there's another point that you made about the the individual that came and got a 15th in the hackathon as from a mental health component. And I suppose where I'm seeing products come out into the market is you are seeing people that are coming at it from a different lens. So they're coming from a different industry, teaching, coaching. Um, they may have just come from a manufacturing but always knew it could be done in a different way. But they could never get something to market because they couldn't develop or they couldn't do it. Well, now you're seeing them come out, and it it's a different lens to a normal engineer that would look at because people learn differently. So, I've have you seen any other examples where you know, have you been blown away by a product that's come out that you just wow, that's so cool. Like, you know, it's and it's come out by someone that's completely non-technical.

SPEAKER_01 18:56

Yeah, so many, right? Um, I'll I'll list a couple. One of the hackathon projects that I've really enjoyed seeing that final the Google Gemini hackathon was one built by someone who's an architect. And if you're someone who's an architect, you'll know that one of the most painful parts of your job is taking your diagrams and trying to understand which parts could be at most risk for any kind of building code violations, right? So in in Singapore, there's all these like documents, regulations, like you need to have like maybe a fire hydrant here, you need to make sure like you know, you have enough uh spacing in hallways for people to rent out in case there's a fire, like all these things, right? It's it's super, super complex and like tons of human error all the time. And when you're an architect, like maybe the more fun part of it is like being creative, but then these are things that you have to get right, or else, you know, things like what happened in Hong Kong happen, right, with the bird, the building burning down, or there's a huge amount of risk, right, when there's human error. So this guy who's an architect came in and actually built a product which takes into account the diagrams. So he uses a kind of a what they call OCR optical character recognition, right? So they could actually read these diagrams, pinpoint areas of like each of these uh room diagrams, cross-check it with the really messy, ugly PDFs that Singapore has around building codes and everything, and actually list down here are all the potential violations in the orders of risk and what are some things you can do to address it. And I thought that was incredible because I don't think I could have built that, right? It's because I mean, sure, maybe like an engineer in the past would have to sit down and like, I don't know, try to empathize with the architect, sit down and look at their work, like basically study what they're doing and try to like mimic what their day-to-day is and build the product. But now that's not necessary because you have this person who knows it so much in their head, right? The IP is in their head, and they're able to like go and um build this out. So I thought that was pretty cool, right? And you've you've got probably millions and millions of ideas that never saw the light of day because that person who had the problem and the inspiration just didn't have the means to build, and that is what is possible to see today. So I'm pretty excited about that.

SPEAKER_00 21:13

Yeah, I'm supposed it's like go back to those books where you wrote me your ideas down and go, right there, this could come to life now because I could do it. Like, yeah, before I'd need like$10 million to be able to put this to market, and now it's like, yeah, now I don't. Um, so you feel like you're thinking about like, you know, you've you've mentioned two people that have come from a different works of like and these amazing hackathons where people are getting to, you know, get it get a first go at doing AI. So what what would you say is the uh key skill at the moment? Because you mentioned um AI skills can become as common as they excel and so forth, but what is the basic AI skills to professionals we start with?

SPEAKER_01 21:50

Yeah, there's two skills that I think are most important in AI: uh agency and communication, which I think all of us should be able to have. So let's start with agency, right? Right. In the previous world that we've all lived in, right, we we have a job or skill we trained for. And we're told, look, if you're a it, you need to be a lawyer to know law. You need to be a doctor to do to know medical science. You need to be X to be X. Right. And so it was very much siloed and like very kind of identity driven. So that's all kind of unraveling. We're like, you're seeing like stories of people who were able to take blood work of their dogs and run it through Chat GPT, figure out a diagnosis, go to a doctor and confirm it's the problem, actually save their dog's life when doctors who are busy maybe haven't gone through all that process, right? So that is an example of someone who has high agency where they're not waiting around being like, look, I need to have an expert tell me I have the right or permission to do something. You just go ahead and do it. So I like uh I always say, like in our classes, right? Um, people always ask, like, can I blink? Can I add a database here? Can I do this? I always say, let's reverse that question. Instead of say, can I, say, how might I? Because everything is possible. You can do everything. It's about how might you do it? How might you unlock that, right? And that's the approach I take with I don't think most things are actually impossible. Maybe going back in time is impossible at this current moment, but I think most things with enough persistence and creativity, like you can always solve it, especially now with AI, right? So I think you just gotta like figure out how to do things. Like if you have an intent in your head, like lower the time from intent to action. Just go ahead and do it. And don't care if you think you're qualified or not for certain things, like just go ahead and do it. Now, I will say for some things you may want to get a professional opinion. Like, don't think you could just go be a lawyer and like, you know, write your own legal opinions. But I won't lie that sometimes I will write my initial thoughts around what a situation is, and I'll go to a lawyer and be like, does this sound right? Another lawyer told me this. And they'll be like, Yeah, your lawyer is uh pretty accurate. So I say it's good. So it's actually pretty funny. I'm like, okay, like I'm not too at least my lawyer friend. Um, so so that's the first part is just like high agency, right? Um, the second one is communication. The people who are the best at prompting and using AI are good at expressing themselves. Now, expressing yourself comes from two things, right? One is are you good at clear thinking? Are you able to break down a problem or a statement in your head, understand exactly what context you need to provide to a personage with AI so they can fully execute on a task. Just like, you know, when you bring on someone new to your team, you don't just like shove them into a role, right? You you give them context on why should you talk to this person, what's been done. So that comes from clear structured thinking. Um the other part of it is how do you co-translate that structured thinking into some kind of articulation or expression, right? And in this case, it's now English, right, instead of code. So I actually find that some of the best people in our classes are people who you wouldn't think are like technical. They're like a lot of teachers are actually very good at using AI to build because they're constantly communicating and providing context and like trying to empathize with like the other party and like explaining things. So we've had some teachers who just absolutely crush it in our class because they're great, great, great communicators. They've had to sit down, lesson plan, break down problems, clear thinking, and then share. So I hope that's helpful on like some of the things that I think are important. It's it has nothing to do with technical skills, by the way.

SPEAKER_00 25:30

It's it's actually just like it's a mindset.

SPEAKER_01 25:33

It's a mindset, it is, yeah.

SPEAKER_00 25:35

Yeah, I like the agency and communication and and teachers are and same as engineers, right? They're problem solvers, they talk about these different things, and then it's kind of breaking down that problem. But from the um perspective of actual platforms, people are using Chat GPT for certain things, Copilot for others, um, Gemini for others. Like, if they want to start building something, where do they start now? They're like kind of going, well, where do I start even typing things in? What can it do? Where do I store it? Um, how often should I be using it? So even going down to that level, any insights you could share around that?

SPEAKER_01 26:12

Yeah, I do think ultimately your choice of first tool actually matters far less. I think it's good to just pick a tool and then the intuition that you've build around using it will translate across all of them. So it's okay to be like, and you don't need to start with the perfect tool. Like just if your friend tells you, hey, this thing exists, lovable exists, proper exists, don't let that be like just like an analysis paralysis. Just go ahead and use it. And you just your skill level is probably too low for that to even matter, right? When you're starting starting out. Like when I just started out, like I would kid myself if I think the decision of what I use matter. I was just like so new to it, right? That I'm gonna suck at all of them. So I'm gonna just try, try them all out. No, it's just true, right? I think only with competence does choice matter, is what I tell people, right? Like when you're starting to drive, like I'm not being like, oh, should I drive a Ferrari or should I drive like, you know, the Lambo, whatever. Like, you just drive a simple sedan, right? And if as you get better, sure the gears matter and all that stuff, but does not matter. Anyone who tells you like you need to pick all that and it matters is just trying to sell you something, which I never do. I'm just like, look, I no, no one's paying me to say any of this, right? Now, that being said, I actually think that the tools that you can consider, um I I find it more helpful to find either people who I know who are smarter than me in an area, or ideally people who work for those companies, right? If I'm able to access somebody who I can converse with, I tend to like to use that tool. So I've gravitated towards using Cursor and I've gravitated towards using Vercell because early on I met the team on Twitter. We were like online friends. I give a lot of feedback, they give me tips, they give me credits, I use the tools. And so a lot of my decision actually came from a relationship thing. So I it's it's funny because I think in other traditional businesses, people say relationships matter, right? There's a lot of different things out there, but you pick one based on who you like and who you've actually built a lot of trust. I actually do the same thing with myself as well, which is which communities have I like had a long-standing relationship with, who has been really helpful, which founders of these tools have been very responsive online when I give feedback. And that's the tools I end up using versus you know looking at a broader spectrum. Exactly. Yeah.

SPEAKER_00 28:36

Well, I think you mentioned a really important word there, which is community. And I think what you're getting from one of the main things people are saying that might change with AI and so forth is the lack of human human emotions and community, right? But I think what you're seeing is a lot of people, there's different people coming into the communities around these platforms. I think say so, took it from a personal perspective, right? What what I would see as my fear factor would be I've got a great idea. I start using, and I'm not yeah, using a platform, but then you go down a rabbit, you you build it to a certain point, and then the platform has limitations. You're like, oh, you you could have put built on this platform, it would be a lot easier, quicker, and less expensive, or you know, something like that. So, have you seen those kind of questions come up? So, questions around could that happen, or is it quite easy to change the the kind of code and the platform around that? Or is it more the fact that it doesn't matter as much because we don't know the limitations of these things yet? Yeah, yeah, because they're gonna just keep going, right?

SPEAKER_01 29:35

Yeah, yes and no. Um, so uh yes, and that the great thing about these platforms is that they're different than no code tools before that really lock you into the ecosystem. So you may have heard of tools like Bubble, uh I guess Shopify, even to some extent, where like it's no code, but like you like if you build something and you're like, oh, I really want to improve it, you can't really just give it to a developer who knows React, right? Like it's like completely different. Code is vibe coding platforms are actually somewhat interoperable in the sense that they're all using the same language. So you can actually port one thing to another usually fairly easily. Now, there are some limitations, I would say, where um I guess without getting into like too much of the technical details, there's sometimes choices of um services that certain platforms will use. So I'll give an example. Uh, I have a tool I to use called Manaus. Manus is its own native database, meaning if I wanted to build anything where I save, let's say, login information from someone, et cetera, MANIS uses its own database. So it's different than, let's say, Firebase, which is Google's uh database solution, which works across a lot of different platforms. So if all of a sudden I'm like, I don't like Manaus, I don't want to build this anymore in Manus, and I have a database, uh, when I port the code over, I had to do something called refactor, which is like take out the database from MANIS and then like slip in another database. It's possible, but it's a little bit more work. But um, you know, that that's kind of where things are at. So look, if that happens and you you let's say you try building on one platform and you realize you want to go to another one, that's fine. I mean, nowadays, like you can spin up something, see it, don't like it, and start over within a day. It's it's actually okay. And sometimes you don't know the limitations until you try. So you kind of have to just accept that imperfection is just part of the process and that uh you you you want to be in a place where you don't want to try to overly control the outcomes and just be be willing to be more creative and accept the ride, you know, things might be different, but you gotta like try things out, experiment, and see how it goes.

SPEAKER_00 31:44

Try, test, try it, as I say. You seem like you could be one of the most patient people, if you get things done really quickly, but then you're not getting flustered when it goes wrong. I mean, where where I can see, and probably this is it is it is the it's the juxtaposition and the conflict now between the corporate world and this new world of and what I'd say is technology entrepreneurs, right? Because the entrepreneur world is growing faster now than this beast here. So, you know, if you're a young person in a business or someone in a business that's got a great idea and could build code, how how would you give advice to them to maybe go like share that idea with a company? Because coming from our uh our business, we're looking at AI, we're using GitHub, you're using all these different things that we can uh try and come with ideas, but then how can it uh is it gonna be a compliant product? How would you get it to to run in that business? But also if you're coming, this is probably a double-ended question, but like you know, you're you're confident, but you might not be the confident person, how are you gonna pitch that to your CEO, right? Or your head CTO?

SPEAKER_01 32:46

Yeah, this is a really, really tough question. Um, I have had many students in our classes, uh, some of them work at like let's say Deutsche Bank or like a financial institution where it's heavily regulated. And they're like, how do I convince my boss like AI is important to implement, right? It's it's it cap is a lot. It's it's very challenging. We even like uh for the for conference that I have coming up, we even have something we call a leadership track where we're actually inviting uh leaders at corporates to come in to listen to folks from open AI, Gemini talk about how to do AI implementation at larger organizations with all these complexities involved. So it is a difficult question that, like, you know, I can probably fulfill one perspective of it, but it is an ongoing exercise industry-wide that is that is happening, right? With other people that might, you know, offer that experience as well. Um, I will say that there's a few things that can help, right? Um, let's say you're, you know, younger person, a company, you want to push this. Uh, I think um exposure to how other companies that are similar to you are implementing it can work because it starts to create two things. One is FOMO, fear of missing out, like, oh, our competitor or someone similar to us who supposedly has the same problems as us, is running ahead and doing all this. It creates this like it's that this fear of uncertainty, it's like a fear of falling behind, right? That kind of kicks in. Um, so uh I'll give I'll give another example, right? I was at a session last week um where they had a couple companies, and these are established companies, by the way, in Singapore, talk about how they're using agents now to automate their software development lifecycle. So this is not just like AI writing code, but AI uh writing like approvals to push code, AI deploying code, AI reviewing design. So it's encompassing like a much, much broader set of things to do than just like writing code, right, in a in an IDE. The company that was actually presenting on this, they're a fintech company. FinTech is one of the most highly regulated spaces, I would say, when it comes to like any sort of company, right? And the reason why these guys are, I would say, a little bit more forward-looking is because I know their engineering team is very like AI four. Like they're the ones that fly to the US, attend conferences, try to learn, you know, what's the latest out there, and they bring it back. So if I know somebody who, let's say, is working for a similar-sized fintech company that's frustrated with the pace of things moving, I would probably say, look, you should bring the example of what your competitor is essentially doing, how they're actually able to do it, and then your leaders will pay attention because at some point they're like, okay, well, we don't want to your fear of compliance is gonna like diminish in comparison to your fear of falling behind when someone else has kind of figured out a, you know, uh a way to actually do some of this. So I I that that's that's one that's one thought, right? Which is like just using the thing.

SPEAKER_00 35:47

No, I completely agree with it.

SPEAKER_01 35:49

Yeah, yeah.

SPEAKER_00 35:50

I mean, I think um you see it from a lot of things having gone through my career in terms of you've seen different changes. This has been probably the fastest one, but the FOMO one's a big one. But often when people have FOMOs, they get really frustrated and they have the fear of missing out, but they don't action it. So they don't do anything about it, they just go round and round in circles and talk about different things rather than implementing something, or you know, is doing something enough. Um, well, one thing I heard the other day, and uh, this is probably you've answered it slightly, but um if you were so to flip the script, so you're trying to get someone to pitch, but you've walked in today and I've gone, Sherry, right? You're the CEO of this company, this is your product. How would you start your AI strategy from the get-go? Other than stealing the developers from that concept, um, like the forward-thinking guys. Well, how would you like kind of go, right? This is my department, what would I do differently?

SPEAKER_01 36:43

Yeah, I I mean, I'm gonna take a probably a very contrarian approach to this, which I don't think everyone will agree with me. I will ask myself this serious question, right? Regardless of how much I have invested into building my company up until this point, if I were to start over tomorrow, would it be the same product? Would it be the same teams? Would it be the same processes? And if the answer is no, and the answer requires overhauling and changing huge pieces of it, I'm gonna do it. Right. And again, this is very contrarian because some people will do more like incremental changes, right? And I've seen some companies try to do this. I mean, I'm not gonna name specific names, but a lot of SaaS products of like that were very popular that really kind of grew in the late 2010s, early 2020s, um, are kind of struggling right now because they kind of slap on AI as a chat feature, or they they just slap AI onto the product without really rethinking how do you actually redesign the entire product to work LLM natively? And that's actually a much deeper thing than oh, chat GPT wrapper on top of your app. It's like, is your even your database uh readable by LLMs, right? Is it agent ready in every single part of the stack, every single micro decision that you've made in the last few years? Have you really looked at it and were like, I'm willing to let go of some of them to build for the future? And a lot of people are kind of afraid of doing that. They just go with incrementalism. But then what happens is like their old users don't quite like the product because the AI features aren't that great and they're charging a lot for it because it's hyped, but they're also not able to keep up to date with the series A startups that get to start from scratch that are outpacing them and able to kind of innovate much, much faster. So if I actually had a company today in that situation, I would actually if if the answer is it's not what we're building today, I would stop everything and maybe I'll okay, because there's probably some contractual obligations you have to fulfill your customers. I will I would probably build a what they what I call a skunk works team, like a separate, like, you know, like best of the best people, right? SWAT team, right? To build out an entirely new product starting from scratch, right? And maybe you only need five people to do it. And if that five-person team is building a product that outpaces one that's run by 500 people, then I think I have my answer around how I need to do things. Uh, but I know that like this approach obviously is is a is a much more blunt one than maybe.

SPEAKER_00 39:10

I don't get I like it. I like it, I like it. I I think I think you you you I think you it's kind of both. I really do. I think you've got to you you come in across it as the right way because you're challenging the state with describing the answer in terms of like look, couldn't companies have got to think like this. As you said, they've still got contracts, they've still got to support the some of those companies that don't want to change, but they can have these innovation things, you know, or Spongebob's teams that come in. What I have seen is that sometimes innovation more it's they don't really innovate, like innovation is actually talking about the the problems of today and improving them. It doesn't have to be something brand new, it could be, as you said, Sherry, you look at a product and go, we can get this better and to market at a different price that can supplement this one. But a lot of people, as you said, I've seen agents, I've seen technology, like it's kind of been joined on. I know exactly the kind of companies you're talking about. Um, but they're all trying to get through that because they it is it is the you know you being an entrepreneur world, there's probably other people sitting there at a boardroom table going, I want to make that decision, but can I? Because if I do, then I might be the one that gets pushed out, right? You know, the the Jerry Maguire old when he wrote wrote the the kind of con the document and then sent it out to everyone, the new way of working, but then you know it's it's the change. But I do agree with that. Um one of the conversations you had, they were talking about the future of work and how it should be. And they said they said, look, the way they see it is there's there's gonna be uh AI AI engineers, and they're gonna be probably about one percent of of the population because they're gonna be very technical. So, but then the rest are gonna be AI literacy. Is that is that something that you would agree with that you can see in the future it's gonna go down that route?

SPEAKER_01 40:56

Yeah, I think that is uh certainly a view that I think uh will will happen. There's actually a term uh that was in Straits Times uh called AI bilingualism. But I think it's literally getting at the same thing that you're talking about, which is like you have some, they have a skill, right? I am a school teacher and I understand how 10-year-olds learn maths, right? Now I know that. Plus, how can I use AI to create more adaptive, customized learning outcomes for my students, right? Um, I am somebody who understands how commerce platforms work. Now I'm gonna use AI to figure out how I can automate the seven, 10 processes of getting money from point A to point B and make it much more simple, right? So that literacy or that bilingualism is going to be very, very powerful in the future. And and what's great is those people no longer always necessarily need to have a software engineering representative that mimics and tries to understand what their worldview is and builds, that person can now build. And then at the top, as you mentioned, are the people who are going to be like, you know, your your applied re uh AI people, your research scientists, the people who are like trying to figure out how to create super intelligence, or you're the one, you're the people who are like, look, text to text with large language models is not the end-all be all. How do you actually like model the entire physical world in reality and create a generalized spatial intelligence skill to use? Like those people are gonna exist, and they're not necessarily gonna be the people who are, you know, the non-technical people, right? But they're there because they're on the frontier and that kind of work then will go down to I don't know, I'm a surgeon who now wants to apply AI to you know how to operate. But the tier work that came before it started with the AI engineers or the researchers who are much deeper, deeper in that kind of field. I think that that's a really that that is a picture of the world that I see see us moving towards.

SPEAKER_00 42:48

Fantastic. Um, but look, Sherry, absolutely fantastic to meet. You would love the conversation, but how one, how do people get on your course? Two, where where do they find you? Um and um when where tell us when this amazing conference is gonna be.

SPEAKER_01 43:05

Yeah, for sure. Um, so they can find the course at codewithai.xyz.

SPEAKER_00 43:11

Okay.

SPEAKER_01 43:13

You can find me on Twitter and LinkedIn. If you search up Sherry Jang, you should find the result pretty quickly. And then for the conference, it's happening May 15th through 17th in Singapore, and it's called AI Engineer Singapore. All the names of the stuff I do are very straightforward. So it's like, what's the class? Co with AI, what do you do? You co with AI. What's the conference? AI engineer, what do you do? You learn about AI engineering, and it's in Singapore.

SPEAKER_00 43:41

So fantastic. Well, I hope to be I hope to come to the conference. I'm in town at that, and um, I'd love to love to take the course as well. Think notes. I could uh be create creating. So I think I I might bring my son as well.

SPEAKER_01 43:54

Uh we've had a few, like uh, you know, mother, son, father, daughter. father-son duos come to class. It's uh yeah, it's it's it's it's a fun time.

SPEAKER_00 44:06

It's a fun time, yeah. And you're just getting shown up by your 11 year old. Uh but look, an absolute pleasure to have uh you have you on the show. I've really enjoyed the conversation.

SPEAKER_01 44:14

It's been yes, thank you so much for having me.