Open Stacks, Federation, and the Changing Landscape of Software

Robert Jacobi: And welcome to another episode, uh, where I, Robert Jacobi, and co-host Dave Lockie, uh, will cover pseudo-random topics around open source and what it means for us today, tomorrow, and in the future. Dave, welcome.

Dave Lockie: Hey, how you doing?

Robert Jacobi: Good, good. I know there is a story of the day that I just wanna cut through so we can actually get on to more important business, in my opinion. Matt Mullenweg on a hiatus, I guess. I think that’s the best way to put it, from Automattic. And does that really change anything for the rest of the world?

Dave Lockie: Yes, it was dramatic. Sadly, I think most of the world heard about it at the same time as a lot of Automatticians. So yeah, whoever still worked at Automattic and keep confidential information confidential, I hope that that’s something that you stop doing. Let’s leave it there.

Robert Jacobi: But even before the show, we were just discussing, you know, still great products, still great service.

Dave Lockie: Right. I mean, like, Automattic today is very much Automattic as of yesterday. We still have products to build, customers to serve, money to make. And as I said, I mean, this is pretty much as new for us as it is for you. So right now, everyone that I’m talking to in the business is very much focused on, we still got jobs to do. And of course there are going to be some things that change. And I would guess that is what a lot of senior people in the business are spending a lot of time working on right now, but, uh, our mission objectives, parameters, values are of course greater than any one person. And so we persist and we will make meaning around this together and we will emerge no doubt victorious on the other side.

Robert Jacobi: Together. I love the British attitude. Almost Churchillian there.

Dave Lockie: Cup of tea and a biscuit, and then let’s get on with business.

Robert Jacobi: And that’s, you know, that’s not changing anytime soon. Still Automattic, still open source. So let’s get into open source and what’s transpiring in the universe. A lot of talk, of course, open source and AI, open weights, open this, open that. And I had the chance to be at WordCamp US in Phoenix last month, and a lot of folks are playing with tools which I would say are not open source. And I’m curious, you know, what that will lead to.

Dave Lockie: Give me some examples. What are these traitors using? I want names.

Robert Jacobi: What name? You want names? You want Hugging Faces and Claude and all those things?

Dave Lockie: No, no, no. I want the names of these traitors. So let’s go.

Robert Jacobi: Oh, these traitors.

Dave Lockie: We’ll hang them at dawn.

Robert Jacobi: We’re here to love, not hate.

Dave Lockie: I mean, I use non-proprietary, like I use non-open source tools all the time. I mean, I use Slack and I use Apple stuff. Like, I mean, not, everything in this world is open source, even if it is oftentimes built on open source. So anyway.

Robert Jacobi: But if we’re using WordPress, there’s that GPL handcuff.

Dave Lockie: Oh, right. Yeah. I mean, I use open source tools and systems where I can.

Robert Jacobi: Yeah. But if people start building apps, web apps with proprietary systems, those wonderful visions of open source may not come to fruition.

Dave Lockie: People have been doing that for a long time. I mean, you know, Wix is not open source. Shopify is not open source. As far as I know, Salesforce is not open source. You know, there are a lot of things that are good products and are not open source. So if everything was open source, then open source wouldn’t be our battle cry and mission, because you have to fight against something, fight to change something. Otherwise, you know, it’s pointless. Like, I’m fighting for the sun to come up tomorrow. It’s a nonsensical mission to have.

Robert Jacobi: But my concern is that open source is getting diminished as a priority with new developers. Let’s put it that way. Yeah. You know, the LLMs have had their feast on everything that’s open out there, and now do we just use it like air and water and don’t actually put more into the environment, for lack of, to keep the metaphor going?

Dave Lockie: I think it’s changing stuff, but how it’s changing it is, to my mind, kind of unpredictable so far. I mean, sure, people are like vibe coding a whole bunch of stuff. These agentic development tools, like coding agents, do seem to like open source stuff. They use a lot of Python, they use a lot of code repositories and then npm and all that kind of stuff. So I guess there’s a couple levels to it. Yes, I think probably people are building a lot of proprietary software using open source software components, but I think that’s always been true.

Robert Jacobi: Sure. I mean, we can look at even Apple’s fund announcements yesterday. I mean, that’s the that’s built upon BSD. So, if you look at iOS and macOS and all those things, you know, they’ve built a quite successful enterprise built on open source. I mean, and they do push back to the community. I mean, you have Safari, what do they call it? WebKit. As an open source project. So, that can happen. I’m just worried more about newer generations of folks coming in and being like, what is this open source stuff? I don’t need to worry about it. It’s a headache, it’s something goofy. There’s a bunch of, you know, long-haired hippies who are preaching this. Since this is audio, neither Dave nor I are long-haired hippies. Short-haired hippies, maybe, but not long-haired hippies.

Dave Lockie: Maybe not in my current physical form. Maybe, but developers like to build for and on stuff. So if you’re building a new thing and it is like your coding agent is like, oh, we found a bug in this package. Should we fix it upstream so other people can benefit from that fix? People might be like, sure, that seems like a good idea. Likewise, if you are building a bunch of stuff and it is in an open-source ecosystem, whether it’s kind of fully open source or partially, then if you’re contributing back to that ecosystem, I think that’s strengthening open source. Like, I mean, I’ve built a couple of Obsidian plugins that I’ve like pushed up and out and made available. Certainly wouldn’t have been doing that without Claude Code building them for me. I don’t see like people caring or not about open source as like a kind of near-term, kind of predictable trend. I think what we are seeing near term is that if your code is open, then the volume of bug reports and exploits and zero-day, like all of that stuff is going through the roof, especially if you are in a bounty program, right? Because now every kid out there that wants to make a buck can— come and probably find an exploit in your code if they are determined and have enough tokens to burn and everything else like that. And there have definitely been, kind of well outside of the WordPress ecosystem, there’s been a whole bunch of that sort of tension in open source of maintainers getting burned out from just having to deal with the sheer volume of submitted bug reports, which sometimes are correct and valid, oftentimes are not. I’m going to go out and say it, but Claude isn’t always right.

Robert Jacobi: How dare you? Oh my goodness.

Dave Lockie: Anthropic lawsuit incoming, no doubt. So yeah, I mean, that’s a real challenge. Because it’s sort of asymmetric warfare. It’s a small team of developers usually who are often kind of unpaid, now having to respond to an army of reasonably talented work experience kids, which is not really what they signed up for. And it’s changing folks’ behavior in terms of kind of are they willing to participate in bug bounty programs? And it may get to a point where it’s just too strong an incentive to make your GitHub private and to stop being open source. I think that’s maybe a more kind of insidious issue. And that’s not to say that open source software is any less secure than proprietary software. In fact, I think it can often be more secure. But if you can see all of the code and all the dependencies and everything else, then you’ve got that much more attack surface than if it’s a black box and you are just having to do kind of outside-in exploit discovery. There’s also been some incredibly bad supply chain attacks, which are fundamentally an open-source challenge. We’ve seen these npm worms, like Shai-Hulud and all these things, where they are managing to poison packages through various different exotic means. And that means that when developers pull these open-source packages into their projects, they are pulling in malicious code as well. And so that kind of whole dependency supply chain side of things is, I think that’s a bigger problem for open source than for proprietary, if we kind of forget about the fact that most proprietary software also uses a whole bunch of these open source packages as well. So I think it’s really difficult to, actually really difficult to draw them apart in a meaningful way because like, show me a piece of well-used proprietary software that includes zero open-source components in their supply chain.

Robert Jacobi: That would be pretty shocking these days. Yeah. I mean, it’s not feasible to write software that way anymore, that’s for sure.

Dave Lockie: I mean, you can do it, and maybe—

Robert Jacobi: Yeah, but no one’s going to. It’s time-consuming. It’s, I mean, if the software, you know, there’s always that it’s good enough to ship, and a lot of times, relying on open-source packages obviously makes that a lot faster to do. Boy, you’ve just made me think of, like, 15 different things I wanna talk about, and we only have, you know, 3 more days to be on here.

Dave Lockie: Get on with it, then. Let’s go.

Robert Jacobi: I’ll start with SBOMs, which I’d like to clarify for everyone, those are software bills of material, which you’re just alluding to, like npm, all those kinds of managers. And I know Europe’s trying to put that in place. I don’t know where it is with— what is it, under the Digital Rights Act, the Digital Something Act, ACT-E, ACT-E Digital? I forget where it lies. Do you have any insight into how that’s progressing, at least on the European side?

Dave Lockie: No, but I’m sure the legislation, when it comes out, will be tedious.

Robert Jacobi: I was going to say cumbersome, bureaucratic, unusable.

Dave Lockie: I mean, it could be done really well.

Robert Jacobi: Yeah. But with a kernel of, this could be a good thing.

Dave Lockie: Yeah. I mean, I think we’ve talked about this before, but I think often the EU has the right intentions, but ends up actually outputting a whole bunch of digital legislation, which is actually super unhelpful. Like, we didn’t have cookie banners everywhere.

Robert Jacobi: Oh God.

Dave Lockie: Like, there were so many ways to solve that problem, and we chose the one way that is actually the worst for privacy because it just forces everyone off the open web because it becomes so awful, which means that all the data flows into like the massive data processors, you know, big platforms. And then like, that makes it super easy for all the data brokers to like buy and harvest and sell all of that data. I mean, it’s just, and there’s actually kind of an interesting one there as well at the moment, which is this whole Iran war. They think a lot of the targeting that the Iranians have been able to do, why they’ve been able to inflict such a high price on the US and Israel is because they’ve had really excellent targeting data for their drones and missiles. And that targeting data, they’ve— and this is not a geopolitics podcast, but it’s kind of enabled by allied friendly—

Robert Jacobi: Mobile phones. I’m just going to say mobile phones.

Dave Lockie: Yeah. I mean, yeah, yeah, yeah. By technology that’s provided by, like, in some cases, China and Russia through SIGINT and surveillance satellites, all that kind of stuff. But a lot of the time it’s also just provided by the surveillance capitalism data issue, which is like our phones emit all sorts of stuff about us all the time, and military personnel have personal devices. And we’ve seen this a bunch of times, like people logging their Strava runs around secret Air Force Base A or aircraft carrier B.

Robert Jacobi: That is exactly the example I was looking for because, yeah, and posted publicly and but, you know, on a larger— yeah, that’s the best one is if you can find out that Dave is a runner on an aircraft carrier, just keep tracking where Dave’s runs occur. And if it’s all of a sudden in the middle of a sea or the ocean, you can pretty much bet that Dave is on the ship at that time and running in circles.

Dave Lockie: It does feel like that sometimes, Robert, lost on a ship at sea, running in circles. Maybe that’s the long-haired hippie coming out.

Robert Jacobi: But, but there’s also that idea of— you mentioned data brokers, and yeah, if I can get semi-real-time or real-time data about certain types of people, maybe all of a sudden I see there’s more people in a specific area. So if, you know, let’s— again, to avoid the geopolitical issues, but great example, today is, oh, if there are only 1,000 Americans in UAE 6 months ago and now there are 10,000, you could probably make a good assumption that something’s happening that’s not related to business militarily-wise.

Dave Lockie: Yeah, for sure.

Robert Jacobi: I mean, you can just look at these kinds of weird data points and—

Dave Lockie: It’s the pizza place at the Pentagon, right? What do they call it? The Pentagon Pizza Tracker. And it’s like, there’s a— I think there’s even an X account. And basically, before there’s going to be some kind of military action, it means everyone’s working overtime at the Pentagon. And so this particular pizza place gets like a lot of orders. And so Google says, oh, it’s busier than usual. And you’re like, okay, who’s getting it this time?

Robert Jacobi: Exactly. I mean, you can extrapolate some very interesting data points which you could not have done— well, you could not have done it as quickly a year ago. Now with AI, you can supercharge that. And frankly, the stuff just didn’t exist 25 years ago.

Dave Lockie: Jake and Nick, I think I want to segue back to Automattic for a minute here because when I got interested in crypto, it was partly because I thought the internet’s own money made sense. You know, it’s like a money protocol for the internet. I still think that’s like valid and interesting. But one of the kind of other threads I found kind of compelling to pull on is this idea that open source software isn’t enough on its own. And maybe that’s kind of heresy, but we’ve kind of just talked about a couple of examples of it there, which is that Meta probably uses a bunch of open source software, but does that make Meta an open business? No. And is WordPress open source software? Yes. But does that mean that if you’re running a WordPress site, then you are like as free as you could be? Also, probably no. Um, because you’re probably paying for hosting and you’re probably paying for like some central or paying for or using some centralized services, like whether it’s something free like Google Analytics or whether it’s like a Stripe or a payment processor.

Robert Jacobir: Jetpack.

Dave Lockie: Jetpack as well. Yeah. And I think that the very existence of Jetpack is illuminating in this regard. It is my assessment that there are, like, to provide contemporary best-in-class experiences, open source software is not enough, that it needs to be combined with cloud services because so much of what a good experience relies on is data. You know, I think like databases are kind of the original sin of an open internet. Because you can host code in a repo and run it kind of super cheaply, locally, whatever. Just running a bunch of code is easy. It’s free, especially if it’s lightweight, even just kind of markup like HTML. Where you end up with a perversion of open source is where you have databases, because a database definitely needs a server to run because you can’t all just run the same software. Uh, if it has a database, it needs to be a centralized database, right? You need, you go from like a peer-to-peer internet to a server-client.

Robert Jacobi: Well, we can talk about flat file databases and things like that. Those, those, I mean, there are—

Dave Lockie: We can do that, but I don’t want to kind of over-index on that. Are there any technical solutions to this? I want to talk about like the trajectory of what we see happen, right? Which is that until you got a database, you didn’t really need a server in the same way that you do now. When you’ve got a server, then that implies some kind of like ownership and terms and conditions of the corporation, and therefore money, and therefore power. And because data has its own inherent value, then there’s a kind of positive feedback loop, and you end up with bigger databases mean more infrastructure, which requires more capital, but generates more value, which makes bigger corporations, and so on. And so we find ourselves here with the Mag 7. Most of them are essentially kind of companies with huge data centers. Of course, they do lots of other things, but the data is absolutely critical to that. And one of the, like, going back to the kind of crypto threat, I wonder, and I guess like the premise of our conversation today is how is AI affecting open source software? And I would argue that open source software is important, but it’s not the only thing that’s important if we’re going to have free, open, democratic societies where we can build on each other’s work and we can avoid oppression and surveillance and all of these sorts of things, you also need to think about open data and open identity and open infrastructure and lots of things that maybe there are kind of niche technical solutions for, but either through, again, wonderful EU legislation or wonderful American legislation too, or just the Lindy factor. There are niche solutions out there, but they haven’t really been adopted. And because now we have these incentives in that having lots of data makes you powerful and rich as a company. And so therefore you’re incentivized to maintain the status quo rather than say, hey, what incentive is there for Meta to turn around tomorrow and say, you know what, we’re going to stop collecting all this data because we found this wonderful new technology, which means that we don’t have to have a centralized database anymore. And they kind of said they were going to do something like that with Threads and federated identity and everything else. But they haven’t really done anything with it. And so, yeah, like, I think we’re now at a point where, yes, AI affects open source code, but code is really like just another form of data to AI. And it’s also like a kind of a static and uninteresting form of data compared to like the real-time, like the content that’s generated or the output from sensors or the output from like their own research. That’s the stuff that’s really interesting here. What is all the data that’s sitting in these corporate repositories that these AI companies are now sucking up through? Okay, maybe they’re under some data processing thing, but I reckon they’re picking up a whole bunch of strays from people that are putting work data through personal accounts. And it’s a huge data play. And it’s like another vicious feedback cycle, right? Because the more data you have, the better you can train your models, which means the more people use them, which means you get more data. And it goes round and round. What I do think is interesting though, we’re seeing that it’s very difficult for intelligence to remain proprietary in the same way that code or databases did, were, are. And we’ve seen that because we’ve got GLM and Kimi and Qwen and DeepSeek, they’re still just models training on data. It just happens to be that a lot of that data is the output from other LLMs. And so that’s a really interesting, and I think powerful lever that did not really exist before. And we’re certainly seeing, it seems to be the case that the models themselves. Okay, so there’s lots to AI machinery. Like, when we use Claude, we’re not just using the model, we’re using the harness, we’re using the interfaces, we’re using the tools, we’re using the ecosystem, right? And the data processing agreements and the actual infrastructure to run it. When you’re paying for Claude, it’s not just about access to the model, it’s like everything that sits around the model to make it deliver a lot of value through to you. But just in terms of the models themselves, it seems to me that the Frontier Labs, certainly what they release is like 6 months, anywhere between 3 and 9 months behind, ahead of the open source, open weight, like the Chinese models basically. Are then able to ship. And I find that a really interesting dynamic, and I wonder how it plays out because do those open source, open weight models need the continued existence of frontier labs in order to be able to drive their own model quality forward? Or are we now at, have we reached, will we reach a point where actually the models are good enough that they can kind of self-improve just based on the data and the models? Like, you know, like, do you get to a threshold and it’s just like escape velocity? And even if Anthropic and OpenAI fire all up, you know, it’s too late.

Robert Jacobi: Then it’s just how many chips and data centers can you build to do that. I’m with you on that, that there’s going to be an inflection point of you have enough data and now we’re self-learning, quote unquote.

Dave Lockie: Especially because I would say that in terms of deployed fleet of sensors, all right, so Amy Webb, but South by last year, she did this really amazing talk where she talked about the, it was AI and it was maybe sensors. Anyway, there’s this convergence of 3 different trends. And the way that you can think about it is phones are already jam-packed full of sensors. There’s microphones, there’s GPS, there’s gyroscope, there’s all of this stuff, right? And then there’s all of the data that we emit as well. Humans talk and we write things and we do stuff, but essentially all that’s being consumed by the models already. The sum total of human recorded knowledge has already been ingested to all intents and purposes. And so when the models are looking to improve and learn more and go further, then they’re going to start bypassing— because humans are essentially filters, right? Our brain pattern matches, we filter stuff out, and then something kind of the other end of the sausage factory something comes out. And so when we would watch a football game and then we would write an article about the football game, and that would, 10 years ago, that’s a bunch of useful training data because you can look at who’s mentioned in the article and you can use that to do stuff with. The here and now version of that is that the players, football pitch, crowd, all of the environment is just like a massive big sensor fleet. There are cameras, there are microphones, there are heart rate monitors, there are all of this stuff, right? And the machines are just saying, well, why do we wait for the humans to tell us what happened?

Robert Jacobi: Write a story about it. Exactly. Yeah. We already know what’s happened.

Dave Lockie: Like a minute, an hour, a week later when we can see it in real time. And so her prediction was that we’re going to see more kind of face computers, like the Apple i-whatever it was, and like the Meta sunglasses, is it?

Robert Jacobi: The Vision, wait, Vision? Is that what it’s called?

Dave Lockie: Vision Pro, yeah. And like the Meta Quest and all that kind of stuff. And I don’t think we’ve really had a breakout product in terms of face computers, but we are seeing a lot of wearable stuff, right? You’ve got rings and watches and heart rate monitors and the growth in number of sensors is massively increasing and the quality of those and the connectivity of those and the ability of these data science and infrastructure and everything to process that data, it’s all kind of huge. And so if you look at how much sensor data is collected by American companies versus Chinese companies, like if I looked at just the stuff that I have in my house, you know, it’s a bunch of Xiaomi, like a robot hoover and a smart plug and like smart light bulbs and all this stuff that like I should really just take out the back of the house and shoot with a shotgun. I would argue that like China has, China is like, you know, this imaginary monolithic single entity, has a larger deployed sensor fleet than any other country in the world. And so it’s not just like the thing that matters here isn’t just like the current model and those frontier labs. I also think it starts becoming about like world models, which is a different class of AI. And world models definitely rely on sensor data more than they rely on anything that you or I choose to brain fart on any particular day. Sorry, that was a bit of a— that was a bit of a rambling—

Robert Jacobi: It was completely rambling, but I understand exactly what you’re talking about. I mean, and it’s an interesting construct to think about, not only today, but even in the short term. How does this change in 6 months? You know, we always underestimate the short term and sort of overestimate the long term. And even that short-term-long-term cycle has been crunched from 10 years to 5 years.

Dave Lockie: Isn’t it the other way around? We overestimate the short term and underestimate the long term.

Robert Jacobi: Well, not today.

Dave Lockie: Today we flip that paradigm.

Robert Jacobi: We flip that paradigm. I mean, and that really may be the case with AI just because we’re seeing things that we can’t figure out how to see. And the point in all this sensor information is no one forecasted that yesterday.

Dave Lockie: And where in that story, if you buy this story, if you liked my story, which part of that has anything to do with open-source software?

Robert Jacobi: So, I wanna take it back to the F-word, and that F-word is federation. So, I’m curious, you know, I grew up back in the day when we had things called modems, which ran at 300 bits per second. Bits. And the way to get good information back and forth was to make choices about what kind of information how are you going to reach out and spend, you know, 16 hours downloading something, downloading email for a bunch of folks? And there’s no silver bullet to, you know, sort of the concerns that you’ve brought up. But I think having many people think about it in many different ways and making that an open and available process is probably much more helpful than just saying, whatever. Apple has this, Google has this, uh, Huawei has this, and there’s nothing we can do about it. And which brings me to even the most strange topic for the day. Um, I was at Dutch Cloud Days and NLNOG on September 3rd and 4th, and the, uh, so the one session I saw is, uh, NLNOG specifically, that kind of really never heard of these tools or what’s going on is, uh, Meshcore. And so that little like chunk of Netherlands and Germany are using products called Meshcore and Meshtastic, which I guess are competing platforms. And they are like very strange semi-federated external internets. It’s, uh, it’s like ham radio meets the internet. And I know folks on Meshcore and Meshtastic will probably— it’s not ham radio, but it’s got this weird independent, distributed, but very open community that’s tiny. And I’m just pointing it out as there is stuff going on that we should be looking more for. And I’d never heard of this, but yeah, Meshtastic and Meshcore. And you can buy like a $20 radio and all of a sudden be connected to weird communities of stuff, uh, which really takes you outside of the, uh, almost internet universe.

Dave Lockie: It’ll be regulated soon, don’t worry.

Robert Jacobi: Uh, somehow, somewhere. It’s not, it’s not big enough yet. So, but that’s what, you know, open source makes me think of these kinds of things where like, you know, we’re gonna take whatever the cheapest thing we can find and all of a sudden like start building communities on it.

Dave Lockie: Yeah, I have seen some interesting stuff similarly around different radio, like LoRa.

Robert Jacobi: Yeah, this is exactly a LoRa kind of platform.

Dave Lockie: And there’s a bunch of other ones as well, but I even saw that you can buy, this is on my Christmas present list, if anyone wants to send it, my DMs are open on Twitter. You can buy a little modem that bolts to the back of your phone, and you connect to it via Bluetooth. And then you can use these like local radio instead of like the cell service to do internet, uh, in a kind of peer-to-peer way.

Robert Jacobi: So I’m gonna second Dave’s, uh, holiday request. I too will gladly take one of these radios. Um, DMs are always open. But yeah, this— I, I was— I never come across this until last week. And I’m like, this is gonna be my new, like, mid-January weird winter hobby, trying to figure out how to set this up and make it at least slightly interesting, if not actually valuable.

Dave Lockie: There’s a nice kind of rhyme here with freedom and open stack technology. So that’s the way I kind of think about stuff. It’s like open source is part of an open stack, you know, open data, and just to like finish the crypto loop, you can store data on blockchains and you can create identities on blockchains, which aren’t owned and controlled by any centralized party, right? They are decentralized and they’re arguably more free because of that. So anyway, yes, OpenStack. So open infrastructure or ownable infrastructure, ownable or open source data and software, and also services and networking. All of those things I think are interesting and important, but they are also one of those things, like a kind of go bag, like an emergency go bag, that it’s not important until it’s critical. I think that’s the sort of thing that we would do well to remember with OpenStack is probably just seems like a bit of an extra faff unless you’ve got a real strong use case. Until you like really need it to be open. Like, let’s say you, and this isn’t competitor bashing, I’m just looking for a kind of example of, uh, when you might suddenly need there to be a solid OpenStack solution to your problems. Let’s say you’re on Google and for some reason they like ban your account. Same with like Shopify or Etsy, you know, maybe they just decide your type of business isn’t the type of business they want.

Robert Jacobi: Or Facebook.

Dave Lockie: On Facebook, yeah, any of these, like, and I’m not saying these are bad people. I’m just saying, like, you know, they have to make decisions and sometimes they’re like, aren’t in favor of individuals. And so in those cases, like, what do you do? Do you like just stop selling online or do you look for a stack which can work outside of their terms and conditions? Uh, you know, so maybe it’s WooCommerce and you host it on your own bare metal in a data center and you use a high-risk payments provider or you use crypto or you don’t need these freedom technologies often until you really need them. And then if they’re not there, then you’re totally stiffed. And that’s the way I think about this sort of radio modem thing is I’ll probably buy it, make sure it works, put it in my emergency bag, and then hope I never have to use it again.

Robert Jacobi: Yeah, but I mean, you said put it into a data center. Why do you need data centers? I mean, back in the day—

Dave Lockie: Oh, right. Oh, on your own NAS or like, you know, home server or whatever.

Robert Jacobi: You could be running it off your iPhone. I remember running an Apache server on my iPhone back in, what, 15+ years ago when it was jailbroken and I could, uh, install whatever I wanted. And yeah, I literally had a web server running on my iPhone. There’s nothing stopping us from all having our personalized—

Dave Lockie: Did you hear that?

Robert Jacobi: Okay, that’s a long time ago. And on that note, I think we need to wrap it up for the day.

Dave Lockie: Bob, edit this out. Edit this out before we lose all our accounts everywhere.

Robert Jacobi: Bob’s like, this is the best part. It’s all— that’s what it’s gonna be.

Dave Lockie: This is the only bit I’m shipping.

Robert Jacobi: That’s right. Dave, thank you so much. A wonderful open-source chat with you.

Dave Lockie: Always good to connect. And I hope other people enjoy these conversations as much as I do.

Robert Jacobi: It’s a blast. Excellent.

Dave Lockie: All right, Robert, see you next time talking about random open stuff on Open Channels.