Ep. #5, Software Autonomy: Owning Your Tech with Elliott Minns

John McBride: Welcome back everybody to the Open Source Ready podcast. I'm here again with Brian. How are you doing?

Brian Douglas: I'm doing fantastic. I have had a coffee, had a coffee drop and I'm ready to record a podcast.

John: Yeah, Brian's ready to rock and roll. It's cold here in Denver, but I'm very excited for our guest, Elliott, better known as Dreams of Code on the internet. How are you doing, Elliott?

Elliott Minns: I'm doing really well, thank you. It is also cold here as well.

Brian: Where's here?

Elliott: I'm actually based in northwest Arkansas of all places. I was saying to John before the show, the accent doesn't really match the locale.

John: Well, you run a very popular YouTube channel, called Dreams of Code, another one called Dreams of Autonomy.

Together, it's some 200,000 subscribers and I'm a big fan and love some of the educational content and kind of eyes on the open source ecosystem you have.

So, yeah. My first question really to you is what is your dream? What is your dream of autonomy?

Elliott: Yeah, yeah, so it's interesting. I do, as you mentioned there, I have two channels, Dreams of Code, mainly focused around teaching software development.

Then Dreams of Autonomy, which I need to give a lot more love to, but I've been real busy the last few months. But that's more focused around open source, self-hosting.

It's something I refer to as software autonomy, which is the ability to run your own software, host your own software, really to improve your understanding of how one's relationship with a computer can be liberating, rather than sort of what the common computer experience is these days where you're working with software that's proprietary, you have no real freedom or ability to make modifications to it to really understand it or to use it in the way that best suits you.

So, that's kind of the goal of or the dream I guess of that channel is to kind of teach more about software autonomy, which is something I personally am very excited about in my own life.

John: Yeah, I'm curious if you have a software background or is this something that you just started deep diving into yourself or where's the origin of some of this?

Elliott: Yeah, so I have been a software developer since, well, professionally since 2012. I wrote my first line of code in about 2007 I think was when I wrote my first line of code.

I went to university after that. I started on Windows first of all, which I think everybody does. Haven't used a Windows computer apart from gaming since 2008, I believe.

I moved onto like macOS before that or after that and always toyed with Linux at the same time, back when you could install it onto MacBook Pros.

I think I fully moved over to macOS in 2017, I believe it was when the Macs came out, the MacBook Pro with the touch bar came out in 2016. That for me was kind of a big wake up call of, I don't want this, I don't want this on my computer, I have no ability to remove this from my computer.

What are some alternative options? As well at the time, Intel's chips were pretty awful at that point. I think 2016, 2017 was when they started becoming worse and worse.

So, I spent thousands of pounds at the time, I was living in the UK, on this laptop that was not very high performance. It would overheat. I couldn't do anything that I really wanted to do with it.

And so I started looking at options and I actually bought a desktop computer and installed Linux on it for the first time in a while. I'd been using it on a laptop, but fully went into the Linux experience. I haven't really looked back since.

John: Yeah, absolutely. One of the reasons I was excited to have you on was just, yeah, you have a great sort of broad overview of how the industry, software, hardware sort of works, but you had this great video about Elastic and ElasticSearch going back into the open source.

And this was even before the Linux Foundation, announced the OpenSearch Foundation, which is Amazon's whole fork of that, now being supported by the Linux Foundation.

But your video was like three minutes long and was just a really excellent, concise explanation of the licensing and like what, the shenanigans in the market were with, how Elastic was reopened sourcing this thing apparently.

I'm curious if you could give our listeners, an overview of that and maybe a hot take or two in there about it as well.

Elliott: Yeah, okay. So, the video was interesting to me, because it's been a general theme of a lot of companies over the past four, five years where they've originally embraced open source.

Elastic was one of the big ones as well. HashiCorp is another huge one. Redis recently is another big one that actually stung quite hard.

So, these companies have embraced open source and open source licensing, which has allowed them to have many contributors come on and their success has been built off the back of that open source licensing in at least in my opinion, they also build in great products as well, which helps.

However, at some point, usually after these companies go public, there's a desire to suddenly increase profits and open source as a business model is incredibly challenging. I don't think anybody can deny that and I think that's part of some of the problems that these companies face.

You look at Docker as a really good example, hugely successful company, open source company, not making nearly as much money as I believe they should be, given the amount of value that they've actually brought into software developer community.

They're still remaining to be open source though, which is really, really, at least from in my perspective, it's, I'm blanking on the words,

John: It's serendipitous honestly for a lot of us who build container technology or containerized applications I guess.

I mean, what's your take on like the, OpenSearch foundation then? Because I think when you originally, we're talking about this, this was even before AWS had donated this, which I was pretty surprised about.

I think it surprised a lot of people in the industry that, Amazon was sort of divesting from this fork maybe, or in some ways like reestablishing

Brian: It seemed pretty clever. Well, clever might not be the best way that Amazon wants present themselves us, but like they kind of rug pulled the elastic product and then built this giant ecosystem with OpenSearch and then said, "Cool, we've done enough here, "let's go donate it and be like, have goodwill."

So, that was surprising. So yeah, curious to hear your take too as well, Elliot.

Elliott: Yeah, I've got a few opinions on it. I don't want to give too much of a hot take one, but I think it was from the point of view of keeping the product relevant, the best move that Amazon could have made.

Because I think having two forks, ElasticSearch going back to open source, which is still a little bit of a gray area, I think having both OpenSearch and Elasticsearch is like two competing products which have kind of diverged from one another.

So, giving it to the Linux Foundation to own, in my opinion, I think was a good move to make sure it stays relevant in the market, being compared to ElasticSearch as well.

I think that's probably my opinion on it. It's not a well-formed opinion, but that's kind of where I lean towards.

John: Yeah, that's fair. It's one of those things that seems to be ever evolving as well and yeah, I'll be curious to see how, I know the goodwill of developers, rebuilds around some of these things and if one or the other will become ubiquitous, who knows.

You mentioned in your little intro, how Intel chips at one point were a part of MacBooks and then kind of became cruddy and I did want to hear about your thoughts around Pat Gelsinger this week stepping down from Intel, which I think was pretty big business news, but I think we'll have big ramifications for the open source ecosystem.

Especially depending on what maybe they divest from or if they start trying to rebuild their GPU business. Curious what you would think about that.

Elliott: Yeah, that was pretty surprising news, but not surprising given Intel's, I guess last 10 years of history. I think there was a lot of hope when Pat came on that he was going to right the ship of Intel and correct it.

And in some ways I think he has been trying to do that, especially with like the CHIPS Act and bringing fabrication to the US, although that is a huge long-term investment and a four year tenure, I don't think was ever going to succeed at that.

I dunno how long it takes to build fabrication facilities in the US.

John: I guess 10 years or so, that seems about right. Funny enough, I was at VMware when Pat Gelsinger left VMware for Intel and it was definitely seen as this kind of like, oh thank God somebody who is a technical leader and who helped build Intel initially is returning, like going to say this whole company, this whole like mission.

I've found it interesting, even just the last few years though, kind of the resistance to engage in the GPU market where Nvidia just seems to still have this stranglehold and AMD seems to be going more and more into the CPU market, shipping enterprise CPUs to like what is it, like 190 plus cores on some of these things is crazy.

Where do you see that going from an open source perspective? Obviously ARM and some of the stuff with open chip sets has been really exciting and fascinating. Do you have much eyes into what's happening in sort of that sector of open source?

Elliott: I actually don't have that much of an opinion in that sector of open source to be honest. I mean in all honesty, I'm not too sure whether Pat leaving is going to impact open source and Intel's relationship too much.

I think Intel has always been very good with open source, no matter who's been at the helm. So, I'm hoping that's part of the culture at Intel, especially when you compare it to something like Nvidia who have been very turbulent when it comes to open source support.

AMD as well on the other hand are also really good when it comes to open source. For me personally, I think in order for Intel to remain competitive with AMD, well at least I hope they can't falter on the open source commitments that they currently do have.

John: Right.

Elliott: But otherwise, when it comes to the GPUs, I think their GPUs, I actually did look at one of the Arc GPUs recently. I wanted to install it for media transcoding.

They're really well supported from what I understand within Linux, which has been pretty great.

I haven't looked at any of the new generations there, just because they haven't been that appealing apart from media transcoding in my opinion.

John: Yeah, I think that was kind of the general consensus from, yeah, just some people in gaming or AI or whatever you're using GPUs for. It was kind of a, they're cool but you could get similar value for cost with whatever.

So yeah, definitely hope that Intel can turn it around and continue to innovate, 'cause very, very deep history of some really cool stuff that Intel had done.

Brian, do you have any thoughts on Pat Gelsinger or the Intel shenanigans?

Brian: No, I mean I don't have any deep thoughts. I'm intrigued because I think there's like this, I mean like Steve Jobs did it, like there's like having the originals come back and like pave out a path, you hope for the best, but I don't follow Intel that close.

I think the Nvidia stuff has been in more pop culture as of recent, so that's just been fascinating watching Jensen and team, continue to figure out what to do with GPUs, besides Bitcoin mining, which is maybe that's coming back already. I don't know.

I don't know if you follow the crypto markets but yeah, they're back.

Elliott: I actually spent a period of time working in the crypto industry, mainly because before COVID it was the only real industry that embraced remote work and you could go work for any company or any projects remotely.

So, I kind of fell into that, because I wanted to do remote work at the time. I was kind of migrating between the US and the UK quite a lot and I couldn't have an office job in the UK.

So, remote, yeah, I fell into kind of the crypto community through it, which is interesting. I don't work in crypto anymore. I haven't for a number of years, but it's fascinating.

It's kind of such a weird industry to be in, but you get to build some cool stuff, especially if you like doing GPU programming or anything like that.

Brian: You mentioned in briefly an intro, but like why Arkansas?

Elliott: Yes.

Brian: You got the accent, yeah, what's the deal there?

Elliott: Yeah, so Arkansas is actually home to Walmart, believe it or not. So, Walmart is the-

Brian: Yes, Bentonville.

Elliott: Yeah, Bentonville, Arkansas. So, at the time I was working for a startup that was doing indoor positioning systems.

So, basically it would read in data from mobile devices, gyro data, accelerometer data, Bluetooth data as well. And you could detect with some level of confidence where an individual was inside of the indoor space.

So, the biggest companies that were interested in that were airports and supermarkets. And so came out here to work with Walmart who were very interested in the technology in order to kind of roll it out into their stores.

They ended up not just because the sensors inside of mobile phones aren't the best, so you have to deal with drift and a few other things and the sales team that sold kind of the technology, set a high standard of requirements that just couldn't be met basically in the physical world.

So, it is a really hard problem to solve. I think Apple are like one of the biggest leaders in that at the moment. They still haven't really solved that problem. Yeah, that's what brought me out to Arkansas basically.

Brian: What about like Niantic and Pokemon Go? Because I'm thinking like I can compete against my gym, my local gym without, you're talking about drifting, but I can compete at the park, without getting outside my house now.

John: Yeah, yeah, yeah, yeah. That's funny, I haven't thought about Pokemon Go in like, three or four years now. That was the whole thing wasn't it?

Brian: Just like Bitcoin, it's back.

John: Yeah, I think I'd be remiss if we didn't talk about, 'cause this is another big thing I love about your content and your channel and some of the education stuff you're doing.

Talk about Nix and talk about sort of the Linux ecosystem of operating systems in general. How'd you get into Nix? What's your view on just like personal computing with Nix?

Elliott: Yeah, so I actually got into Nix probably at the beginning of this year.

John: And maybe for people who are unaware, let's describe what Nix is.

Elliott: Yeah, yeah of course, of course. So, Nix is a package manager but it's very novel in the way that it works. So, rather than being an imperative package manager, which you typically find with Unix-based systems, such as APT or Pacman or even Brew I guess on macOS, it's a declarative package manager.

So, you use it by defining what you want the state of your system to be. Either what packages should be installed and then I dunno what the actual Nix package manager command is 'cause I use NixOS, but I run the NixOS free build command and it will set up my system the way that I've defined it within my Nix configuration file.

NixOS uses the Nix package manager underneath but allows you to configure your entire operating system in this declarative manner.

So, rather than installing packages ad hoc, making changes to your system ad hoc, you define what you want your system to look like in a configuration file or multiple configuration files, run the NixOS rebuild commands and your system provided everything set up correctly, will have everything installed and have all the configurations defined that you've specified in your configuration file.

This brings huge number of benefits to the Linux experience, but as to like how I got into NixOS, I'd say it was like a bit of peer pressure from my Discord community.

For some reason they absolutely love NixOS, I mean I know why they love it but I seem to have attracted a lot of NixOS fans, before even using NixOS and so a lot of them have kind of encouraged me to take a look at it over the time and I have.

But before looking at NixOS, I looked at Guix back in I think 2021, which is the good new version of Nix or NixOS and Guix is really cool but I couldn't, I had issues with it when it came to like disk setup and things.

So, it only really worked in the most simplest of setups in my case and I was running a big workstation with multiple disks and I just couldn't get it to work.

NixOS has solutions to that problem which is why I was able to make it work as my kind of full-time distribution.

John: Yeah, interesting. I've only looked at it briefly and sort of more as a function of time abandoned it, but it does kind of get a weird and I guess a lot of things within Linux just generally, it gets kind of this reputation as being like, "Oh you got to have like a PhD level understanding of this configuration language or how the declarative nature of flakes and et cetera work and stuff."

How have you seen from your educational content perspective maybe or just in the open source, I mean how have people started approaching this?

Do you feel like, you've been paving the way to like enable people to have, just a different understanding of it or is it really that complicated?

Elliott: I'd say there is a huge learning curve initially, especially coming from a standard Linux distribution. I mean first of all, Nix is non-standard.

It uses a non-standard file hierarchy system which makes it difficult to understand where files should be placed and in some cases makes it so that existing software won't actually work, especially if the software assumes your commands are going to be in the /bin directory or anything like that.

So, there is absolutely a learning curve associated with it. I don't think there's a PhD level learning curve associated with it. It does require you to learn and understand how Nix differs in such a way. That being said, it is just Linux at the end of the day, so it's not like it's a complete departure into a brand new operating system.

It's Linux and so most of the principles still apply but I've spoken to other creators that have just tried to install it for the day and they've quickly gone back to Ubuntu or something else, because it's so different from, you can't just install Ruby on Rails or any other packages and just get going straight away.

You kind of have to understand how that works or how it works in the context of NixOS.

When it comes to like my content and paving the way, I have no idea if I'm anywhere near that yet. I actually did a video on installing NixOS from scratch and the total length of was about 40 minutes and I just thought I can't do this a single video, this is going to have to be a full course which is what I really want to do in the future.

I want to have like a full course dedicated to Nix, make it really, really affordable for people. But the amount of content associated is just huge.

I think there is definitely a time investment as you mentioned in order to understand how to do things in Nix and how to be productive and most of all how to solve issues or how to work around issues if you encounter them when you need to be productive.

I think that's one of the biggest hurdles that Nix has. Python for example, is almost impossible to use when it comes to Nix. I found there are probably ways you can do it using poetry or poetry to Nix or anything like that, but for me I just end up using Distrobox.

So, if I need to do something that I dunno how to do on Nix, I just spin up an arch Distrobox image and then just work through there.

John: Yeah, that makes a total of sense.

Brian: We spent some earlier time talking about like Elastic and OpenSearch and I was curious like what the correlation like with Flox and I know you mentioned Brew. Workbrew is another one of these, like these companies that are now building structure, around these open source package managers.

So, I was curious, your take are, if you could explain that correlation between Flox and Nix, I guess if you don't have any context, that's fine.

Elliott: Yeah, no, I don't actually.

Brian: Okay.

John: That was actually going to be, my exact same question Brian funny enough. My understanding is they've sort of tried to abstract a lot of the like gnarly parts of like Flox and trying to understand the configuration language, wrap that all in their product essentially and sell it.

Is that right Brian would you say?

Brian: Yeah, I mean I don't use Nix, I definitely use Homebrew instead because I'm on macOS, but yeah, there's a company that's I guess standardizing the relationship between NixOS and your machine.

That's as much as I know, like I've met with this team before but because I haven't used Nix, I don't really know what the features are. So, I was actually really curious of like what the deal is there.

Elliott: Yeah, I've thought about this quite often so it's really cool to hear that there is a company doing this, 'cause I think NixOS in itself, provides a lot of benefits to the Linux experience but as we talked about the difficulty is, actually understanding how to use it.

So, if you can kind of abstract the core Nix language and how everything works into creating kind of an operating system that has the transaction log abilities of Nix, I mean one of the biggest benefits I find is if you make a mistake in your system configuration or break something.

On your boot screen, there is a list of all of the generations of every version of your operating system that you've built from your config. So, you can just roll it back to a previous version and you can get back to being productive again, fix what the issue was that caused your machine to break later on once you're no longer on the clock or anything like that.

So, that's a huge benefit that's provided by the package manager or by NixOS in general. So, being able to kind of abstract, some of these core features, I think is basically a money maker, if someone can do that correctly and create an operating system that is as easy to use as Ubuntu or macOS, but has those other features that are provided by Nix.

I mean one of the other things that Nix solves is the ability to install multiple versions of packages onto your system, which is such an annoying problem to run into whenever you do, like I need to run Postgres version 15, but I've got Postgres 16 installed, it's just not going to work.

I have to now have a different binary name or a different service name, find it to a different port. Nix kind of, I mean you still have to do it but bind it to a different port.

But Nix allows you to have these multiple versions of packages installed on your system, because everything is kind of isolated together when you're installing something, you're not relying on your entire system dependencies to use that.

It will kind of create a dependency tree, just for that package. So, the drawback is you install a lot more stuff onto your system but everything is isolated and working well within each other, which is really nice.

John: Yeah, it's definitely an interesting paradigm. I'm curious to give it a try more and read some more of the documentation but speaking of reads, Elliott, are you ready to read?

Elliott: Yeah, yeah, yeah, yeah.

John: Alright so we got a few reads this week. I'll go first since I have two and it looks like Brian has one, but there was an interesting read I had from Substack and I think I pulled this off of Reddit at some point, but it was kind of a essay I guess about open source and legacy software and how maybe it's really just like people problems ultimately.

A lot of the things we have around code becoming legacy, being difficult to maintain, eventually getting super crufty and just unmanageable, is it maybe just a people problem ultimately.

And I've definitely experienced this in my own life managing some open source software in the past. It oftentimes just becomes like difficult people and engineering org problems.

I'm curious Brian or Elliott, if either of you have experienced the people problem, inside of open source before.

Brian: I mean you kind of alluded this a little bit, Elliott, when you're like, I use Nix because of Discord community.

Like there's like the cultural thing of like, this is the new hotness and I know like I was watching some videos, about like VPS servers and I know there's like top of mind for folks in the last couple months between like Vercel and all these other hosting providers that you could circumvent with a $5 box instead.

But yeah, I definitely like and I've also been a junior engineer before, so like I know where the craft of like, oh no I read that PHP was the best thing since sliced bread.

So, like we should be using that 'cause actually the phrase that I use is like PHP has used that Facebook, we should use PHP here at this company and that's it.

So yeah, I love thinking broader and not getting too like stuck in the mud, about like software decisions, 'cause you don't know what's coming out, the reasons why we took this path.

Like a lot of that context always gets lost with like poor documentation. So, I try to be empathetic as a developer.

John: Yeah, one of the sections in this read that's about that specifically, I think it's maybe a little clickbaity, but it's a don't innovate quote unquote, "Innovate" where it's like oh we're adopting all these cool, new, sexy, innovative technologies and like we're going to deploy this cool new service or something.

But really like that quote unquote, "Innovation," maybe is just a bunch of like cruft or things that will eventually like get really difficult to manage in the future.

Especially if those people who know those new technologies are things, like leave organizations. Yeah, what do you think about all that Elliot?

Elliott: Yeah, it's interesting the idea kind of legacy, I guess legacy code, legacy systems. I dunno what the definition of legacy is in the article.

John: It's pretty gray.

Elliott: I think, I mean that that's pretty much every company, I've ever worked at as well, right? Legacy is something that is generally seen as undesirable, but it's often still used in production.

I mean true legacy would be something that's no longer used. There's no point in having it around, right? But generally legacy is often used. It's still relevant in such a way and I'm kind of like skimming the article at the moment, the article on like don't innovate.

It's kind of interesting, 'cause it talks about using different languages to, or rebuilding or rewriting with this different language. I think go is the example that it mentions in the article. I'm not sure how that relates to innovation, but it is an interesting concept of like rewriting or anything.

But to get onto kind of the original point of the article, about the problem is people rather than code I think people problems are the harder problems when it comes to building software and I mean there's a huge spectrum of what that means is how do people use your software. That's more of a people problem than it is an actual code problem or a technology problem.

There are technology problems of course. I mean we deal with computers and we have to figure out how to scale things and how to make things work across different machines or different architectures.

But people problems I think for me are always one of the harder problems when it comes to code because code is kind of there, you can solve it, but people are so varied and so different and there's different levels of understanding, different levels of communication, different ways a software will be used.

So, the bounds are huge when it comes to dealing with your actual code or computers.

John: Yeah, this is exactly why it always kind of tickles me when people say like, AI is going to take our jobs and AI are going to like take over, engineer orgs and stuff and like, they're good tools but like ultimately they're not going to be very good at solving people problems and they're not going to be very good at like inferring relationships or like navigating orgs of people.

I'm sure there's things to be automated away, but I know that's my spicy take for the episode.

Brian, you had a good one from Paul Graham. This is titled Cities and Ambition. What's this one about?

Brian: Yeah, yeah, kind of related just on the other side. I was actually on the call two days ago with my manager and was chatting through, like legacy, tech debt in the product that we're touching today at the Linux Foundation, John.

And this article came up and I remember actually like folks when this came out of what? Quite a few years ago.

John: 2008?

Brian: Yeah, it was the concept of like and this was like right before the boom of like people moving to San Francisco, 'cause of like Web 2.0 startups.

Paul was like, he writes about like, he wanted to move to Berkeley because innovation, UC Berkeley's there, a lot of AI ML stuff's coming out of Berkeley even back then.

So, he thought if I can get out of MIT or sorry Cambridge, like you get better weather but you get just ambitious people and then you realize that's actually not the case.

Like there's a whole other like spectrum of people who are in Berkeley that are not just there for the college and interaction and ambition.

I was actually thinking about this, even after my conversation a couple couple days ago, 'cause like I moved from Tampa to the Bay area, because I want to work at startups. I wanted to like learn as an engineer faster in whatever.

And like funny enough you mentioned Elliott, like you've been remote in Bentonville in Arkansas, but you wanted to work in a remote culture. So like, I'm not sure if it's like Cambridge and outside of Walmart, but I was just thinking about that, 'cause like I've now got three kids.

I'm out here in the Bay Area. Do I still want to be here? Do I want to move somewhere else? Like, well I still have access to the knowledge that I came out here for.

So, I thought it was a great read and recommended for anybody else. And if you guys have thoughts. I have you guys read this paper or it was article?

John: I actually think about this one a lot, partly because I'm out in Denver, Colorado, which it feels like a place where people like to go to like have a lot of fun. Like if there was a mood like he talks about, in the city or like, what a city says.

Denver is a place where it's like I'm getting in the mountains, I'm going to go ski or climb and maybe when I say fun it's this kind of like suffer fest fun. Like I'm going to go hike at 14 or it's going to be tons of fun and it's going to take me 16 hours or something crazy.

I think even Brian, last year sometime when I went on a big ski trip you were like, is that going to be fun for you or are you just going to be tired? I was like, yeah. But I've always noticed like from the tech perspective, Denver doesn't feel very.

I noticed that shift I guess with like ambition or like what people want to do with trying to innovate with things.

There's like startups here, there's all different kinds of things, but it's a lot of defense. It's a lot of more, I don't know, I guess enterprise kind of contract businesses out here.

AWS has an office out here but most of those are like, working with government contractors and stuff. So, definitely a different mood.

Brian: Yeah, what's your take Elliott out there in Arkansas?

Elliott: Yeah, definitely northwest Arkansas is slightly different from say London, right? Which is where I spent most of my kind of 20s, during my career. Which was really nice because you have kind of tech meetups.

There were some big companies doing a lot of kind of, oh yeah, huge companies doing kind of technology in London, although not as big as I imagine the Bay Area would've been or kind of that startup scene.

I was always very envious of, the kind of Silicon Valley startup. I think were like, that would've been a wonderful experience to spend my 20s in, especially surrounded by such intelligent people.

London is still pretty great though for that, but it's much more diverse when it comes to kind of the different technologies that are out there.

So, Northwest Arkansas on the other hand's, very different, but we do actually have quite a few companies that have technically minded people, but there's no real kind of culture around it. There's no meetups or anything like that.

Walmart I think are bringing, all of their technical stuff over to their new campus that they've been building. We also have like JB Hunts, which have kind of a big technical team as well.

I actually flew to KubeCon in Denver I think in 2022 and that was actually a plane full of software developers, which is something I was surprised to see. But yeah, I think it's interesting. It is kind of what John talks about.

I think with remote first culture, there's that balance of, do you want to be kind of in the place where you are surrounded by other people that have a similar profession and you can go and learn from other people, which I think is really, really important.

The internet now provides that avenue and you can have a platform where you can help to do that as well. But I think it's not accessible as going out and actually meeting people at meetups or going to talks or anything like that.

Brian: Do you think that you would have Dreams of Code and your other channel if you were in San Francisco or in London? Do you feel like that you kind of built your ambition in your city separately?

Elliott: Yeah, I think that's actually a really great question. I think there was more of a desire to do so, given that I'm less surrounded by individuals.

I think living in London and stuff, there's like a lot of big social pressure to actually go out, see people hang out, drink after work, have a meal after work.

Then since working remotely and kind of living, I guess I don't live in a rural area, there's like 300,000 people here. It's still a reasonably populated area, but I think living in here, I think working remotely, you do kind of lean towards the internet more in order to share ideas, which one of those ways is through YouTube.

I think the goal of my YouTube channel was to kind of teach things, especially having worked, like I went back into the workforce at a remote company in like 2021 and then worked there for a couple of years and I decided I wanted to kind of teach what I had learned to other people, just because the technology landscape had changed so much from when I first got started.

The world of writing software now is so much easier than it used to be back then. But I think there's a lot of things that have been lost through that process of making everything easier. I wanted to kind of create my channel in order to help teach some of those things that I also learned and to make it a little bit easier, and yeah, to share the platform.

So, long way of answering your question. Yeah, Dreams of Code probably wouldn't have been there if I was still in London or if I was in San Francisco or somewhere where it was much more populous with developers I could just go out and meet.

John: Yeah, well you've had a huge impact already.

Elliott: Thank you, I appreciate that.

John: The last read I think Elliott you'll like quite a lot is actually something from the Pulumi Blog, about the cloud container iceberg and it's like the hundreds of ways that you can run a container on the cloud and how, honestly just mind boggling it is, the number of services that can do this.

Some of these range from sensible all the way down to like, just absolutely ridiculous. Like I think at the bottom of the iceberg is Jupyter Notebooks, which I would be shocked if anybody was actually, running Jupyter Notebooks with containers inside of them, but I better not say anything, 'cause I'm sure somebody's doing it for some reason.

But in the middle is the Kubernetes multiverse and a bunch of stuff like the Swiss Army knife of container hosting on just like Azure VMs, et cetera, et cetera.

But yeah, Elliott, you've had a lot of content about running like your own VPS and being mindful of like, saving some money alongside running your containers and maybe not always needing, one of these crazy container services.

What does the iceberg make you feel right now?

Elliott: It's incredibly interesting, isn't it? I'm looking at it and the Jupyter Notebooks is interesting. AWS device farm that's also-

John: There's some crazy ones in there. I've looked at GameLift before when I worked at AWS and that's more along the lines of like, oh, we're a game company and we're trying to stand up, obviously a bunch of like game servers for things and stuff, but you'd really never run a container on that platform, but raise your things have been done, I'm sure.

Elliott: Yeah, this is such a good read. I'm going to probably read this stuff to the podcast, 'cause I like the title of, now you see me now you don't for AWS Lambda obviously, 'cause Ephemeral Containers, which is, I mean that's actually something I've used a lot in the past and there's definitely benefits to doing it, but it's such a, it felt very antithetical to the whole point of containers to then running them onto a Lambda. But I mean, it makes a lot of sense.

But yeah, as you mentioned kind of the VPS, I guess series that I've kind of been doing at the moment, I think, I mean those videos are so popular and people want to learn more and more about how to host your own things on a VPS and I think that talks a little bit about the difficulties of building software 10, 15 years ago, compared to what it is now.

John: Yeah. How would you define a VPS.

Elliott: I guess it would be any sort of VM instance that you have access to and is running kind of a full operating system, so you can pretty much do anything within the limitations of that virtualized instance.

I would assume that to be a VPS, that definition I guess could be changed a little bit. I mean, does then like a container that you have access to on a Kubernetes cluster, feel like a VPS?

John: Yeah. It gets very muddy even with like firecracker VMs or like small, I don't even know what you call it these days, but like fly.io kind of is a weird middle ground where it's like, not really a VM, but sort of acts like a VM, but you can run containers on it and all different services and things.

Yeah, I think that series has been really interesting and I mean I've loved it as well. I think we'll end it there. And remember, audience: stay ready.