#107: How to Have a Successful Cloud Career in 2024 | ft. Andrew Brown
52:33 Share

This episode is supported by Memento, a serverless cache that you can trust and only charges you for what you use. To learn more, visit goldmemento.co slash theburningmonk.

Hi, welcome back to another episode of Real World Serverless. Today, we're joined by Andrew Brown. Many of you may know Andrew already from his numerous courses on AWS, Google Cloud, as well as Azure. And yeah, I've been a big fan and follower of your work, Andrew. And so welcome to the show. I mean, also a fan of your content because you created very detailed, very detailed, I'm sure your audience already knows that, but very detailed content specifically for

serverless architecture and it's always been very fascinating to read that stuff. I think I used to work for a company, were they building like a social media platform? It was like completely serverless, purely serverless functions. And I can't remember if that's what it was, but it was just the, to see that level of architecture of serverless built was just fascinating to me. Sorry.

Thank you. It's an honor to hear you say that. One of my first big breakthroughs, I guess, in the serverless world was I was working for this social network where I sort of inherited this social network. It was all running on EC2. We had lots of problems and at that time, Lambda has support for API Gateway and integrated well with DynamoDB and other things. So I looked at what we had and I thought this is a good fit for Lambda. The fact that it

the traffic is quite bursty, but the base workload is quite low. And so, you know, I thought that we can make everything serverless and they will help us alleviate a lot of the scalability as well as operational challenges. Things like, you know, just scalability

EC2 servers keep getting on fire, we have to keep putting fire out in production. So that's when I kind of did this whole thing with Lambda and learned a lot about what it takes to build a production-grade server application. So I've been kind of showing, I guess sharing my lessons ever since.

From the serverless architecture perspective, I do like probably the most sinful serverless pattern, which is taking monoliths and deploying them to Lambda functions. So at our XAMPPro platform, our learning platform, it's a Rails application, and it's broken up into multiple engines. It can run as a monolith, but you basically break it up as smaller monoliths, and that's deployed to individual Lambdas

That was an architecture I had at one point. It didn't work out as well as I thought it would, but it was interesting to try it. So it's just interesting to see the different perspectives. So I do have a little bit of foot in the serverless world, at least in like when I used to live in Toronto, like we had a very strong community there. Like I believe you've presented it, the Toronto serverless group. So.

Yeah, I did, but remotely. I visited Toronto in May this year, but unfortunately, I think that they only do online meetings back then. So I didn't get to meet anyone when I was in Toronto.

Well, next time you want to go there, I'll come down. Again, I used to live in Toronto, but now I'm really far away. I'm in the same province. Everyone thinks when they come up to Canada, they're like, oh, I'll come visit you. I'm like, you sure? I'm near Niagara Falls. I'm near Toronto. And then they find out that I'm now 12, 14 hours at least away. But it's just I moved back up remotely because the

the business I'm in, the content creation business I'm in enables me to do it. And so there was no reason to be in the large tech hub anymore because when you're an online content creator, you can do it from anywhere. Actually, yeah, I'd love to learn a bit more about your online content business and how you kind of make that work, especially how you kind of mix that with your free content online. But maybe let's start by just getting, go back to the beginning and how you got into cloud in the first place and how you started making these educational courses.

Well, you know, I probably tell the story a little bit different every time. But the thing was, I was set out to actually build a learning platform, an LMS, a learning management system, because I've always been interested in education, specifically technical education. I was a CTO for previous startups, specifically in ed tech. Or if I wasn't, I was adjacent to software that was very similar.

And so naturally, it made sense for me to build an LMS. And the first thing I ever wanted to build as my own really big thing was an LMS. So I came back to that original dream. But it was kind of a side effect of me creating content because I wanted to build my LMS using the most modern tools. And at the time, AWS was the best choice for it. And I had already been really good at scaling applications, like millions of users on

Single virtual machines, if you want it to vertically scale, I'm the best at it. But at some point, you have to embrace the newer stuff. And it's not like I didn't use cloud for many years. I had used it back in 2009 or something because one of our...

our investors was very interested in the offering. And so I begrudgingly was pulled through eight of us. And obviously, GCP had App Engine very early on. So I felt at home at cloud. But I just started using study materials with a very popular provider. I won't name them, not to throw any shade, but the content didn't get me all the way there. And personally, I learn differently. I'm dyslexic. And

I need different learning tools, different means to learn. And so I just created the materials that I wanted to learn. And so I had rolled out a flashcard system for my platform, which then rolled out to practice exams. And then I started making courses. And so as a side effect, I'm in the content creation business now and I can't leave.

But I like it, so it's okay. Sorry. It's funny that you mentioned the GCP because I got into, I guess, AWS around the same time as you. And at that point, the company I was working for, we were building social games for initially Facebook, but then later pivoted to mobile because everyone kind of moved from Facebook to mobile games later.

And a lot of our backend was actually running either on AWS or on App Engine. I think App Engine was really ahead of its time back then. It was this whole, you know, platform-to-service thing. It was a really nice abstraction. They had good local simulations and all that stuff going. But then they never quite made that next step in terms of making what we think of, you know,

Lambda and serverless functions today. There's still a lot of, "Oh, you've got to decide your instance size, how many instances you want." But then after that, you can just deploy your Flask application or what have you, and it just handles all of that for you in terms of request routing, everything else. So App Engine was a really good product back then. It's a shame they never quite took it a step further and make it past what we think of as serverless today.

And so I guess in that case, you got into educational content creation as part of you building this LMS yourself. One of the things I noticed from your content is that you have a lot of content around the certifications. So, you know,

At first I thought, okay, you know, Antu must be a big fan of certifications, but then having listened to you, spoken a few times, it feels like you're actually, you know, of the mindset that's very similar to mine and many other people that the certifications are not a great way for you to gain a real competency. So,

What are your real thoughts on the certifications and the efficacy? Sure. So, I mean, not all certifications are built the same. And I think like in our industry, there have been certifications that have had rigorous testing practices and went through a process to which you could say this person is accredited or certified to have those skills.

that's not what certifications are, at least in the cloud spaces, that's not what they are. If we go back to, let's say, 2018, 2017, when I was starting to revitalize my cloud knowledge, I was finding that if you're just talking about having a certification, when all these companies were very interested but did not have a lot of domain knowledge, it was enough to get your foot in the door. And so as a means to get a job or get jobs,

interest as an applicant, I would say, yeah, sure, that made sense. But we're far from that now. And the reason why is that I think that companies are, they have that domain knowledge now. It's not a mystery anymore. You also think about, you know, the actual means to test people. So when you look at a certification, like AWS certification, what are they testing you on? They're giving you a multiple choice answer

But in the reality of your actual work, how would that ever measure your real skills, right? So, you know, if you need to implement anything like you're deploying serverless Fargate containers into a production environment, you know, you can answer questions on it, but can you do it? And that's a totally different story. And do you know the things around it? Because when you look at cloud certifications, they're focused on their services, not necessarily on the skills needed to actually

build, let's say if you're doing development, build those applications, um, uh, do project management, all the things around that. Um, and so, you know, um,

You know, we had a lot of people that have certifications and just simply can't do the work now. So we're over that period of mystery. And now we just want confident people. And these certifications have not evolved in a way to meet the demands of what people want to say, hey, this paper matches or makes this person an attractive applicant. If I'm not all over the place there. But, you know, I'll give you a solid example. When we did our free AWS Cloud Project Bootcamp, which we did last year, I think it was 2023.

which I did for free and I didn't get any sponsorship because at the time nobody wanted to give out money because they're all saving that money for the Gen AI, which now they're giving out money. But, you know, we surveyed and I did a, like I evaluated people's readiness to take our bootcamp. And we had people where they had every level of certification, like it'd be like, it'd be,

practitioner, all the associates, a professional, a specialty. They might have even gone across into other clouds. But when they came into the bootcamp, they couldn't make a GitHub repo. They couldn't write code. They couldn't use Discord. That's just what we're using to manage the project. If they made a pull request, it really messed them up. Their whole branch was a mess. They didn't know how to take screenshots. They would take

photos of their phone, it'd be blurry and they'd be like, help me out. I'd be like, I can't help you out. They couldn't use backticks to put code examples. They didn't know how to ask for help. And, you know, there was like, it's like, how did you get all through those certifications? It's because, you know, I think certifications is like a safe place to learn, but it never takes off the training wheels, at least again, the composition of them. And so I think that's the problem, right? And so that's actually influenced my content now. And so

I recently refreshed my AWS Solution Architect Associate course, and I produced 120 hours for it. And it wasn't that I was going in my head going, I need to make X amount of hours. I just kept going, what are the skills that are actually required to really learn? And then I made it and then I stopped. And then even then I had to cut it back to 75 hours because my co-founder's like, Andrew, you're just going to wear people down with the reality of learning cloud. And so, you know, that's,

That's what it is, you know, if that makes sense. Sorry.

Yeah, that makes a lot of sense. I have similar experiences with people that have got the certifications but can't really do the job because they've never been exposed to the messiness in the real world. There's no multiple choice. There's far more branching decision trees you have to navigate and there's often no obvious answer because those multiple choices always have one or two that sounds about right or you can maybe go through a process of elimination that the things that are

definitely wrong. And then you are down to maybe one or two choices that you can just flip a coin and maybe you get it right half the time. And definitely you don't have that the rigor of everyday developer, this kind of practices, like you said, in terms of working with Git, writing good commit messages and doing pull requests and being kind to each other with your pull requests. Don't delete like 10,000 files in one commit.

things like that. So yeah, I'm definitely much more in favor of project-based courses than the like bootcamps and things like that compared to certification-based courses where it's just all on paper. And yeah, definitely,

I'm not a fan of the way the certifications kind of work today, especially with how broad they are, but they're both too wide, too broad, and also not deep enough. It just covers too many things that you won't need in the real world, but there are not enough of the things that you are actually going to need to use every day. So I guess in that case, when you are creating, by the way, 128 hours is

because I've put all this together. You know, for me, it's roughly maybe 10 hours of work for every hour of content. So I can only imagine how much time and effort you put into your refresh. So, you know, kudos to that. Yeah.

So I guess in that case, what are some of the things you're doing in your latest courses to put more of a project-based focus on it as opposed to a certification-based focus, but still help people pass the certification exams? So, I mean, we can talk about AWS because that's what I've been mostly updating. And of course, I'm across the cloud, so each cloud has its own difference. But like, for instance, AWS has a solution architect associate and they've gone too broad.

They just have every service under the sun. They have information that conflicts with other information within their own exam guide. It looks like too many marketers got involved and some key people, there might've been a bit of a shuffle. And so they're needing some gardening, so to speak, to improve their exam guides. But when I see that, I go, I have to think of practice and go, okay, I need to touch everything in this thing just in case, even though only one third of the content appears on the exams.

But how am I going to get them to get the skills that they need? And so like, if we're going to go and this is what I did the other day, it's like we were going to do code pipeline. We need something to deploy. We're going to go build that app.

So we'll literally spend hours building out an application beforehand. And during that time, I'm going to talk about how I'm preparing that application specifically for cloud because cloud-first development is totally different than... Because I used to build it not the cloud-first way, like monoliths and throwing things up. And so I can, as I'm building out that app, talk about why you wouldn't want to architect your application code a particular way because then

then you're not going to be able to break those parts off to leverage managed or serverless services. Or just like if I'm using Ruby on Rails, we're going to drop the ORM and we're going to write raw SQL with the driver because if you were to use DynamoDB, Alex would probably agree with me on this, is that you don't want to utilize the ORM or some abstraction on top of it because

you really need to work with that database directly and to get it to do what you want to do. I'm not going to say that you can't, but I'm just saying that that's my preferred method. And so that stuff carries all the way through. And then, of course, we'll do things like commit messages and pull requests. But a key thing that I always do is documentation. So I always do heavy-duty documentation. And if you go to GitHub right now, go to github.com forward slash AWS examples, you can see I've been just keeping every piece of code in there

And it's just like, as we work through the project, I'm constantly pulling from the graveyard of other things that we're doing to be more productive in apps. And so we are seeing ourselves leverage other labs into other labs. And so we have a lot of complexity that I feel is closer to what real-world uses would be, as opposed to some other... No digs at other creators, but they will stage things and leave that stuff out because...

They want to just teach them the cloud part, which some people, that's all they wanted. They just want to pass the exam. They want to have the cloud part. But I try to show them all the parts now just because I believe that's what I need to serve for the community. Not necessarily. I'd probably make more money if I just didn't do that because there's enough people that don't want that. But, you know, you choose what you want to do, right, as a creator. Sorry. Mm-hmm.

Okay, so in that case, if someone was to try to... What advice would you have for someone who's trying to break into the cloud? Because obviously you need to learn about the cloud itself, but then also there's a lot of other skills around just general software engineering that you need to have as well. If someone who's...

I don't know, maybe who's not, maybe not come from a computer science background, trying to break into the cloud, trying to find one of these high-paying jobs. What advice would you have for them in terms of, you know, the learning paths and what things they should be doing?

Yesterday, somebody came to me that's a student and they already have a career in STEM as an educator for high school and they're very interested in that. My suggestion to them is that leverage your existing domain knowledge and use that as a positive where you might have negatives in your technical depth as you're learning cloud. The

The other thing is that whatever you do, because they were trying to say, like, do I go DevOps or do I go to data scientists? And I said, well, you're the one that has to do that job for the next three, four years. So choose something that you want to do. Because, like, yes, there are cycles of things being very popular. And so data scientist was very popular, I don't know, four or five years ago. And right now there's a lot of speculation that, again, Gen AI thinks it's taking everybody's job. Not necessarily it is, but that...

A lot of that stuff is going to be automated or assisted. And so to me, these jobs aren't going anywhere. The demand is still the same across the board. It's just that there's more noise. And you might as well pick something you're going to be really passionate about and you're going to like. Because I feel that most routes in tech, they all have high seats to sit on if you're looking for a good salary, if you want to keep working up that chain.

But I think that picking what you like is going to drive that passion to get you where you want to go. And obviously, on that road there, you can change because a lot of the stuff that you learn in tech, a lot of it allows you to pivot in any direction you want. So I think worrying less about choosing what it is that you want to do and just picking what speaks to you is important. That's just one point, right?

Another thing is setting realistic expectations. This really pulls at my heartstrings because I get emails from people often and their expectations are so misaligned because unfortunately we have folks at some of these organizations that are suggesting, I call it false hope marketing, that you can make it in

three to six months or nine months. Like you can become a solutions architect in nine months. And I'll get this message like, yeah, I put all my chips in. I'm not, I quit all my jobs. I'm living off my savings. I got my cloud partitioner. And in nine months, I'm going to become a solution architect associate. And that is hard because I believe, or I think that a more realistic timeline is three years.

Cause like, even when I, I started out, like I had a background in technology because we had a computer repair business and I had already had a lot of experience, um, uh, stuff, but it took me like three years to get the knowledge, three years, uh, working in an intermediate thing to get to where I actually want to go. Like I was building websites and things I didn't want to build. I wanted to build web apps. Uh, and then I got there. So like, even for me, the journey was very long and this is years and years and years ago. Um,

And, you know, I would say that you can expedite your learning. You don't have to take as long as I have, but you have a different problem, which is competing with more people, right? So I think that people need to set realistic timelines and say, I'm going to set for three years. I'm going to make sure I can financially support myself through this. And I'm going to choose something that I'm going to actually like to do and stick with because you're going to, again, it might take you three years before you get somewhere. What was, like, how was your experience

Hey, have you heard of Memento? The real-time data platform that's founded by the same engineering team that built DynamoDB. Memento delivers robust cloud primitives like cache, storage, and Pub/Sub.

that empowers developers to safely and reliably serve a massive global audience. It accelerates growth at critical moments through intelligent resource management that eliminates risk and downtime. Visit goldmemento.co/theburningmonk for more information. Now back to the episode.

Yeah, it was really slow and steady. Nothing happened instantly. You start with junior positions and then gradually learn your trade, make a lot of mistakes along the way and hopefully don't make them twice. I think like lots of people, I've had cases where I accidentally shut down a production server, things like that, which happened to the best of us, I hope.

Along the way, you learn from your mistakes and you try doing different things. I think one of the biggest things that has been beneficial for me is trying my hand at different paths. Even as someone who has been working in technology for a long time, I

didn't really know for sure what's life like as a data scientist working on data engineering versus building web apps versus doing more DevOps-focused kind of jobs. And one of the nice things about working in a smaller staff is that you kind of have to do a bit of everything. So you kind of get to expose yourself to different kind of challenges and

problems that you have to solve. And then maybe at that point, you kind of have a better idea in terms of what are the things that you enjoy. And so, you know, I spent most of my career patching machines and updating IEMs

AMIs and figuring out what's going on with this machine and kicking it and rebooting it and whatnot, fixing memory leak issues and problems like that. And I kind of learned over the course of many years that I don't really like that at all. I kind of like to just, okay,

take a customer requirement and then write the minimum amount of code I need to get it to work. And so that's why I'm really into serverless because even though it's not perfect, it gets me much closer to that world where I'm doing things I enjoy doing less of the things that I don't like doing. But I also know people that are really into, you know,

DevOps and figuring out all the different nuances in terms of setting up the cluster of EC2 instances and tweaking the network settings and things like that. Everyone's got different preferences, but I guess to your point about just choosing something that you enjoy, it's probably harder for you as a beginner to know what it is you're going to enjoy. So just pick anything, I think.

get your foot in the door first. And then once you're inside the industry, you can find different opportunities to work on different things. And again, smaller companies tend to offer much better mobility in terms of trying different kind of problems, different areas that you want to work in. And then over the course of many years, you may find out what you actually like doing. And then

Hopefully you can pivot and focus on doing stuff that interests you more. And I like what you said because it's something that I rarely hear from other folks, but it's something that I've said and it was my path, which is like cutting your teeth by working for small startups. And they still exist and they still want you, but the way you approach it is differently. So like the way I used to, when I wanted work, like I didn't use LinkedIn or resume. I avoided that for the longest time. When

When I wanted a job, I'd go to Crunchbase or whatever listed all the startups and I would just sort them based on when they got their seed funding or their Series A. Then I'd look at what the company is and I would get an idea like, okay, they're trying to build this thing and they don't have this and this and that. I would just go and I would cold email them very targeted. I'd be like, hey, I love what you're building. I wasn't saying I wanted a job, but I just talked about

The thing that they're built, like shown that I really understood what their thing is. And then I would go build a thing and ship it to them.

I would just go, here's some code, just have it. There you go. This was six, seven years ago when I was dealing with more bootcamp graduates in Toronto. They go, I'm applying to all these places and they all have this take-home homework. My thing is just get better at being able to rapidly build these things. I got to a point where I could MVP anything. Before I even knew about CloudFormation, I had my own YAML. I have it somewhere. I had my own YAML code where

It would just assemble various applications very quickly so that I could then get to the more important parts of what that company wants. And so I think you could just keep iterating, iterating, iterating, getting faster. And it became nothing for me to knock it out. And then I'd add it to my graveyard of catalog that I could pull from later. But that strategy was extremely effective and still works today. Like I'll give you an example that anyone could do right now. Right now, Memento. I got their cup right here.

Not to try to plug a serverless product, but they just raised money, right? They're not even a Gen AI company. It's crazy to hear a company that raised money that hasn't Gen AI, but they have a really cool...

infrastructure service, what we call it. It's like... I can't remember the name. But the idea is it's the development. It's not a SaaS. It's the infrastructure that you utilize, right? Memory, whatever, whatever. But you could go and you could go build a opinionated service framework around it and then ship it over to them and say, hey, I built this really cool thing. What do you think of it? You don't think they wouldn't want that and they wouldn't want to bring you in the doors? The first...

Company that I became a CTO for was was Teambox. And the way I got that job was they had an open source Rails app on GitHub.

and one of my clients wanted me to deploy it because they knew I knew Rails. And so I couldn't get it to work. It was in bad state. And so I just went Twitter and complained about it. And the co-founder, the founder at the time was like, what would you make better? I was like, I'll just show you. And then I spent like a week just like submitting open source pull requests in Forex. And I changed a large portion of the code base and

And fixed it. And they said, do you want to come to Barcelona and be our CTO? And I said, okay. And I didn't even know who they were. I just got on a plane and went there. I was younger, right?

And I saved the company because I guess they were, they had burned through their money and they, they, they were smart. They just didn't know. They didn't know what to do with that. I knew how to do. And that brought them more money in. And so I became this kind of person that would go in and save startups. I'd find startups where I felt like they might be struggling. And I would say just the right things they needed to hear, uh, and bring them the right things and show the productivity that it became, uh,

impossible for them not to want to hire me. And it was like three out of four places I would message and I would have my choice. It hasn't, again, hasn't changed, but you have to get to that level of communication and understanding what other people want if you go that route, right? If you want to go that crazy. Sorry.

Yeah, that's great. And well, completely incidental that this episode is also sponsored by Memento. So yeah, like I said, they just raised money and it's a really good team. I really love working with Quaja and Ellery and they're some really smart people working for them.

So yeah, it's good to hear that you are a big fan of them as well. And I love your approach to approaching companies. You're much more, especially I think now with all the AI generating CVs and AI reading and filtering out CVs, it's really hard for candidates to kind of be themselves and stand out and show their best self to potential employers. And so I think this approach of actually making a connection, first and foremost, I think is something that I think

I really like, I wish I had been more like you. For me, my career has been lucky. I took a few lucky turns because of just happened to have the recruiter give me the right, present me with the right opportunity to go into the cloud.

nothing was planned. I just happened to stumble on an opportunity that exposed me to serverless and my career kind of got shaped by a few lucky events as opposed to something that is much more so planned and methodical and I love the approach in

in the way you kind of approach them, but also the way that you kind of look out specifically for companies in the spaces that you want to get into. So yeah, you know, anyone who's listening, who's looking to get into the cloud, this is a really good way to kind of, you know, show yourself and show your best self to potential employers. I mean, and I should say it's one of the ways that you can do it right. And it worked for me because I came from an entrepreneurial background because the family business, right. And so,

I know people that are principal engineers for different services at AWS. And the way they got there was that they were just at the right school and AWS was like picking people out from there. Do you know what I mean? So like in Canada, it's not that you can't go to traditional schooling, you have to go to the right school. So in Canada, Waterloo is the big school that all the CTOs I know, there's no questions asked. They get their co-students, co-villagers,

co-op students co-op students and they pay for them and then they hope that they come back after they're done their co-op and they're done done their schooling because they set such a high bar uh for that and again coming back to certifications there's no reason we can't do that it's just that somebody would have to go and and um uh raise the bar so to speak but uh i would love to do that myself but um certifications is difficult because i think that

To have that level of accreditation, you have to be that organization. So when AWS is the creditor, it's very hard to say, I'm going to accredit you on AWS. But anyway, something I got to figure out. But yeah, certifications are fine. But it's a driving thing. And so I capture people in the certification thing and starting to steer them towards projects and other things. Sorry.

Yeah, I think that's a good way because a lot of people, especially for beginners, they don't know what they don't know. They don't know what they should know in terms of the actual hands-on stuff, the ability to actually do the work is what employers are looking for. Then they're using certifications in some ways as approximation for that, even though it's a pretty poor approximation as we've talked about already. There's so many people who have done certifications and they just

They got really good at passing certification exams, but it doesn't mean that they've got the necessary skills and aptitude to do the job. And what you've done in terms of being able to develop those muscles to rapidly MVP stuff and to be able to talk to companies and show them what you've got, I think that's...

That kind of attitude and ability is exactly what companies are looking for, especially for startups. It's all about getting stuff done. It's the attitude of, we can do it. So I think that for people that are looking to break into

Well, anything in the tech world, I think that is a really good way to look at things and kind of, you know, taking control of your own destiny as opposed to rely on what school you decide to go to or what certification you happen to stumble upon. Yeah.

And if some people feel that they need to do these things, like they might have like family obligate, like, oh, I'm the first one to go to college or university. So I have to do it. Like I went to, I didn't go to university, but I went to community college and it was more just like, I just didn't want to live on my own yet. And I wanted to live with my parents and said, if you're going to school, you can live here for free. And so I said, okay, I'm going to take this program called multimedia. I don't need it. I already know how to do Photoshop and video editing, all this stuff.

But I'm going to leverage that time. I'm going to get bad grades. And I'm going to go and leverage their technology. Because at the time, if you wanted to, like the mid-2000s or early 2000s, you had to have a server, right? So I needed a powerful Linux server. I didn't even have a Linux machine. So they had Mac servers. And so I said, OK, I want to be the person that in the summer I'll set up all the... Because I already knew how to do...

computer repair. So I said, I'll set up all your new computers. And I literally set it up in a day, like their network. And then I spent the entire summer trying to figure out how to deploy a Ruby on Rails app to Apache, going into IRC, like at the time it was IRC, so like in the chat and in mailing lists. And there was all these people that knew more than me. And they would be like, you're garbage, you're garbage. And I'd be like, yes, I'm garbage. Please tell me

your crumbs of knowledge, right? Give me the crumbs of knowledge and I'll make it work. I don't care. Tell me how dumb I am. I'll be dumb. Just give me the information. And that's how I had to do it, right? Or like I had to go through, because again, I was focused on Rails, but like there was big open source projects called Mephisto, which was a, I think it was a blog. And then there was the Beast Form. Don't look that up because that's something else. But the Beast Form was just a form system

And so I would go through the code and they were messy at the time because people were just learning how to use rails. But I had to go through and sift through the code and try to make sense of it and break it down until I could get something that worked. You know, and, and you still like, yes, we have all these assisting tools, but the stuff that people want you to know how to do is the hard stuff. And you have to go digging for that stuff still. And so, you know, if you can try to go digging, you know, into, into code bases or,

or push yourself into technical uncertainty, if that makes sense. Sorry. Yeah. And I also always find that when I, well, I mean,

Learning, I guess, how should I say this? Learning is not created equally. I mean, you can be on a job for five years, but maybe most of your learning will be coming from, say, six months when you're working on a specific problem where you're facing with a lot of new challenges, there are things that you don't know. I think that's where you have to actually exercise new muscles, develop new skills and learn new things. And what's that? 10,000 hours to become a... Bill Gates said, yeah.

Yeah, so that whole thing is around 10,000 hours of deliberate practice. What I find is that most of the time we are not exerting ourselves. So we are just doing what we know to do already. I know how to write JavaScript. So when I'm working on a project, chances are I'm just using the muscles that I already know how to use. I'm not developing new muscles. I'm not stressing myself.

But every now and then you come across a new project that actually you have a lot more you can learn within a very short space of time. And that's one of the things I actually quite like about being a consultant, at least part of it, sometimes that I often get exposed to new challenges, new domains, that the things that I don't, I wouldn't have to have to, you know, have to deal with if I was just building something for myself or working in my previous jobs. So in terms of learning,

One of the things I struggle with as a content creator is that you're always on this schedule trying to push out new content every week or create a new course or do something. So it's hard to find the time to do some of the deeper learning that I like.

And for you, you are creating content on not just AWS, but also Google Cloud and Azure as well. How are you finding, do you have a system for getting yourself, I guess, refreshed on the latest developments so that you can teach these things to your students? You know, I think the approach has just been all over the place because I thought that when you learn AWS, it's going to be the same way with Azure, same with GCP. Nope, nope, they're totally different.

But there is knowledge that does stack and there's conceptual knowledge that's the same. Like if you know virtual machines, you know virtual machines across the board. But the approach to how they might architect a service, like their networking layer, like GCPs is completely different from AWS and then Azure has extra components. And so I can say retroactively, the way that I think people should approach it, and now this is how I'm approaching it right now when I'm learning new things, is I will take three variants of something.

because I will find that I will learn whatever the base knowledge of it is, and then I will find the differences of those three systems, and then I'll just stick to one of them. For instance, I literally announced it today that I'm like, yeah, I think I'm going to do a Gen AI boot camp. I've entered the hype cycle. You know what I mean? Let's get the rave gear on. I know Bedrock to some degree, but when I'm doing stuff, I'm opening up Bedrock, Vertex AI,

and Azure AI Studio all at the same time, trying to do the same thing at the same time. I shot a video in, I think it was the developer advocate, like I'm just shipping this literally today. My team is shipping it for me, by the way. I don't do everything myself, by the way. I do all the video, but not all the stuff around it, so I can actually focus on. It's one of our questions, like how do you keep up with all the content? Because I got help. But in the developer associate, I had to show CodePipeline, and

Because I had, we were at the, we were both at the AWS Hero Global Summit. I was in the Claude talk and then the Lama talk. And at the time I was like, I don't really care about these. But then when the models dropped, it went, oh, I know about these now. And I didn't realize that they might be important. So I said, okay, I'm going to go jump in and utilize them. And so.

While I was showing CodePipeline, like how to build CodePipeline, I was using it to build out our IC, but I literally pulled up all three of them at the same time. And that's literally an example of how I learn things is I'll pull up all three and I'll run them through them. And then I will eventually go to a single one and then I'll circle back. But I think learning in threes is my effective way of learning. Sorry.

Okay. So in that case, I mean, one of the things I always sort of tell people is that, uh, it's better to learn the fund, the fundamentals because they don't really change. There's a lot of things that we build on top of that, that are always going to be different. Like you said, with the, the networking setup is so different in every single cloud, but the networking itself, the fundamentals of networking is the same everywhere. Uh,

And the same with learning a programming language, learning how a computer works versus some JavaScript front-end framework, which there's going to be a new one every month.

So in that case, so you say you're someone trying to learn about Gen AI. What would you say are the fundamentals that they should probably focus on versus the sort of more variations depending on different platform, the flavor of the month? Sure. So, I mean, like with Gen AI, you have fundamental things like what is a prompt document and you can feed that into any of them. And it's basically the same thing. There are knobs and knobs like temperature and top care, whatever the other value is, which changes the randomness.

They might have three knobs. They might have two knobs. The point is that they more or less do the same thing. They all have model cards. So if you want to learn more about them, you read the model cards of all three and you start seeing similarities. I started noticing that I didn't know that they tested these models against a bunch of stuff, like a benchmark, and they're all using the same similar benchmarks. And so you start seeing patterns across the board. All providers have a model as a service. I call them AI delivery platforms. They call them models of service.

And you go in them and you can see that they're very similar. Or it's like use tools, use tools, use tools. Tool use, tool use. They all kind of do that. Or regs, they all kind of do that. So, you know, to me, that's more fundamental stuff that can be applied to all of them. And literally in the next few days, I'm producing a Gen AI Essentials course, which is covering this stuff because I'm trying to, because I'm doing the Gen AI Bootcamp, right?

Not trying to promote anything. I'm just talking about what I'm doing. But I'm trying to lay that foundation and making it as a prerequisite before people come on to my Gen AI boot camp so that I don't have to cover that stuff. But then they have a basis where they feel comfortable. Like if they want to use Lama or they want to use Minstrel AI, they can do that. I don't care. As long as they follow the fundamental concepts that we want to do to deliver that end product. Sorry?

Okay, so by the time this episode airs, I think your bootcamp should be live already. So I'm just going to put a link in the description below so that if anyone want to go check out Andrew's new AI bootcamp, you can go and check it out.

So going back to, I guess, the way you talked about earlier in terms of building your own business, starting with building your own learning LMS, what are your thoughts on self-publishing? So if someone who is looking to become a content creator, wanted to publish their courses online and make some money from doing side hustles, what are your thoughts on doing self-publishing with things like Teachable, Gumroad versus using a platform like Udemy?

So there's the platform and then there's the publisher, right? So I built my own platform and that's because I already had it and I like building LMSs. So

I did actually use Udemy when I first started out because there was great pressure for my co-founder and my wife being like, don't give courses away for free and put it somewhere where people are going to see it and want to pay for it. When I published on Udemy, I found out that it's not a matter of just putting a course on there. They have an ecosystem. There's a way of getting your course to be popular. And that was going to affect how I produce my content. And I didn't want to do that. Udemy, it's like they're optimized for the first 15% of the course material. So you have to make that

part like pristine, perfect because it's going to pop up and they're going to have that thing that says, tell me how good this course was. And then you have to engineer for that, that feedback loop to, to be popular. There's other things with Udemy there, but I went, that's why you see Udemy courses and they're really good at the start. And then they kind of go like this because they're optimizing for that. It's like, if you produce for YouTube on your own, um, like I, and by the way, I'm not self-published. I published with free code camp. Right. So, um, uh,

which is like you could go with Pearl site, like they're a publisher technically, they're more than just a publisher, right? But you could go with them or you could be on Coursera where they have the platform and kind of the marketing with it. But I chose to go with FreeCodeCamp because it aligned with my company vision, which was I wanted to produce free content.

And, you know, you have to do what makes sense to you. Like any model works as long as you're willing to work with it. The model I chose is not the best for optimizing revenue. It's the best for optimizing what I want to do, which is community work and stuff like that. I just happen to get paid as a side effect of what I do, if that makes sense.

But I mean, I could talk about these in great detail all day, but yeah, sorry. Yeah, so with the free CodeCamp stuff, obviously you've, you know, you just...

updated your course probably spent hundreds of hours maybe you know if not thousands of hours on your on your free code camp boot camp which is which is amazing but at the same time you know you're not really promoting your courses your your your paid courses how do you strike a balance you know how do you keep your wife from from from saying hey come on angel you have to put put more effort onto paid courses and put food on a table big

because I don't promote it, everybody promotes it for me. Really? Because people know that they go, Andrew,

You don't promote it. I'll promote it for you. And I go, okay, great. So, you know, like if you serve the community, I mean, like I do put a little bit of marketing. So like if you were to go to free code camp and watch one of my certification courses, I mentioned my company once in the intro, I'll go, yeah, if you want to support this content, go over here. And then I just have a link at the top where, and I don't even tell them there's a pay thing. I tell them that there's a lot. I say that we have a free practice exam and we do give away a full free practice exam every single certification. So that leads people to our platform.

But for the most part, when people want to come back and they're looking for certification, like extra materials, they come across us anyway. Or again, it's like other people are promoting us. Because when you do that kind of free community work, everybody wants to talk about you anyway. And so that's my marketing, if that makes sense. Robert Leonard

Right. And it's like, also, like you said, you start with the certification stuff and then the DAS use that as a funnel to get them into the more project-based courses as well. So I guess in that case, your marketing strategy is you just give as much as you can for free. And then, you know, in terms of people reciprocate, well, reciprocate,

your goodwill and helping you with your paid courses as well. What are some of the biggest challenges you found with just creating educational content in general? Because obviously you have to do all of this work for a free step. Also, you have your paid courses and also other things you're doing around the community. I mean, I think there's external challenges and internal challenges. Internal challenges as a creator is like,

production like how much production do i want to put into it when do i call it quits and call this good enough um uh the other thing would be uh like like who am i making this for so there's a lot of things around the content creation process and so because like the way i produce content is it has to work for me so i'm dyslexic if i write something out and i read it it i sound like a robot

I sound awful. Like I was reading a book to my four-year-old or five-year-old yesterday. My wife's like, why are you reading it so weird? I'm just like, this is how I read. And I'm missing words. And it sounds really bizarre, but like I really have a hard time reading text. So I knew that if I was to make content, I couldn't follow a script. So I had to figure out a way to produce content where I wouldn't do that. The other thing is that I just didn't have any time to do editing. So I had to also...

make the content good enough that I could ship it without having to do any post editing. And the thing is, I make mistakes all the time when speaking, especially if I'm not making content, like technical content, like I'll just

Sometimes I'll be talking and then I'll become extremely incoherent. People are like, what happened to you? Did you have a minor seizure? I'm like, no, this is just the way I communicate. And so I had to work in those mistakes, right? Which apparently people actually like because it brings more personality into the content. So again, my approach is I will record something. And if it really goes off the rails, then I'm going to throw the content away or restart over.

Or what I'll do is I'll try to get it back on track and show the troubleshooting. So as a byproduct, I run into problems. And so I just solve them and I keep it part of the content. And apparently people actually like that because a lot of content creators do, like they try to make it as clean as possible so that people aren't complaining about inconsistencies or getting stuck.

But because I show those parts, people learn how to troubleshoot just the way that I would troubleshoot through a problem. But it also saves me on editing, right? So I think it's like that, right? And so when you say, how are you making 120 hours because of the pipeline that I've set up, sorry?

Yeah, that's actually something that I've heard from some of my students as well, that they actually enjoy some of the mistakes that are left in there so that they can see how I troubleshoot. But also that they don't feel as stupid when they make mistakes that they don't see people make on the course content that they otherwise consume.

So yeah, Andrew, thank you so much for taking the time to talk to us today. Before we go, is there anything else that you'd like to share with us? I guess we know that you've got your Gen 8.0 bookend coming up. Anything else that you are working on?

Oh, I'm always working on things, but I mean, hey, why don't we just plug the sponsors here? I bet you have like a preamble before, but yeah, if anyone doesn't know, GoMemento.com. I think it's .com, right? .co, actually. Oh, .co. They're like me, exampro.co. But yeah, GoMemento.co. They produce, right now I believe it's a serverless cache.

topics. And they also, I believe they have storage coming along. And I think people should check them out. And if you are trying to think of something to build as a really cool project to get into tech, why don't you build it around Memento, like your own opinionated framework, something like Firebase with their tools. Again, I think that'd be really cool. I just want to see it. So that's my plug for the sponsor, not for me.

Thank you for doing my job for me. But I should just check, it's goldmemento.com for the main website. For my personal link, it's goldmemento.co slash theburningmonk. So I'm going to leave that down in the description below as well so you can go and check them out. Okay.

But yeah, they're a really good team. I love their products. And yeah, especially for the topic stuff that they've been working on. It's quite refreshing how well that works and how easy to work with as well. Yeah. Thank you so much, Andrew. I'd love to catch up with you again, I guess, in a couple months' time when the re-event comes along. Are you going to be there this year? I think so. Yeah, I just got to go fill out my forms and make my way there.

Okay, sounds good. See you in a couple months. Take it easy, man. Ciao. Bye-bye. Thank you to Memento for supporting this episode. To learn more about their real-time data platform and how they can help you accelerate product development, go to gomemento.co slash theburningmonk for more information.

So that's it for another episode of Real World Serverless. To access the show notes, please go to realworldserverless.com. If you want to learn how to build production-ready serverless applications, please check out my upcoming courses at productionreadyserverless.com. And I'll see you guys next time.