Matt and Sean talk to Paulus Schoutsen, the software engineer who started Home Assistant–the largest open-source smart-home platform in the world, running in over 2 million homes.
Chapters
- 00:00: Intro
- 01:20: Paulus Schoutsen
- 52:08: Feedback - Updates
- 56:20: Feedback - What Happens with the Chlorine?
- 59:24: Feedback - Climate Change
- 01:01:00: Feedback - Poutine
Transcript
Sean Ferrell: We're going to talk about homes that are smarter than Sean. Welcome everybody. To Still to Be Determined. This is the podcast that takes a look at technology. It is a follow up to Undecided with Matt Ferrell. And as always, I am not Matt Ferrell. My apologies to everybody who thought I was. I'm Sean Ferrell. I'm Matt's older brother. I'm a writer. With me, as always, is the aforementioned Matt. Matt, how are you doing today?
Matt Ferrell: I'm doing well, getting over a cold, so I may sound a little froggy, so I apologize.
Sean Ferrell: Well, the good news for you is that the talking is pretty much done because, yes, we have a long form interview for everybody today. We are going to share a conversation that we just had with Paulus Schoutsen, who is the software engineer who started Home Assistant, which has since grown into the largest open source smart home platform in the world. Matt is a big fan. He's going to fanboy over it a little bit during the conversation, as you will see in just a few moments. On now to our conversation with Paulus.
Matt Ferrell: Hi Paulus. Thank you so much for joining us today. I have to start off by just saying I'm a big kind of home assistant nerd and fan myself. I, a number of years ago transitioned my entire smart home over to home assistant just because it works with basically anything that you throw at it and it runs on any device you want to run it on. It's just been this rock solid backbone of my home and my home and my channel and everything I talk about tends to be around sustainability. And some people get confused as to like, why I'm so into smart homes. And it's like, you can make your energy efficient home so much more efficient with smart home gear. And so like, when somebody from your team reached out to me saying that I could talk to you about Home Assistant and its origins and where it's going, I immediately jumped at the chance to talk to you. So I just wanted to kind of like start with a little bit about who are you and what's your origin story with Home Assistant? Like, how did it, how did it start?
Paulus Schoutsen: Yeah, um, excellent. Well, first of all, I'm very glad to be here. Um, yeah, but origin story, it started like 13 and a half years ago where I wasn't even trying to build any of this. Right. Like, nobody wakes up and it's like, I'm gonna build like the biggest smart home platform in the world. No, I, I got the Philips U Lite because they were just released and I'm like, you know, a gadget nerd. And I like technology. And so I bought them. I realized they had a local API. I'm a Python programmer and like I was writing some Python and I was able to talk, talk to the local API and I got that working. So from my computer I was able to control my lights. And being able to control the physical world with like computers is just very, I know, satisfying, right? Because you know, drawing a box on the screen is like, sure, but now like something else is moving in the real world or light is being turned on. But I wanted to use that skill. And so I decided, okay, let me pull in the sun. So if I notice time of the sun and when the sun is setting and I turn on the lights, because that's when I normally would want to turn on the lights. So I built that next up I realized, oh, now my lights are turning on when I'm not at home, right? So presence detection was added. And then I realized, ah, you know, when I actually the sun set, especially in the winter, it's too late. You have to look at the horizon and like, you know what, let me just do an offset of 45 minutes. And basically that's how this whole ball started rolling, right? Because then that came and then that came. And at some point I was, I was applying for jobs and I was like, you know, I need to have something on my GitHub profile, right? Like, so let me put this out there as open source. And I did get the job, by the way, but nothing related with Python or anything.
But yeah, then I had an open source project and I was like, you know, on the home automation subreddit saying, hey, you know, people are asking, I have this problem. And I was like, well, on Home Assistant we kind of solved it this way. And Home Assistant became kind of my toy project. Every time I wanted to play with technology, I now kind of had an app where I could like put these things in, right? And I was doing front end at work and we're doing a lot of REACT stuff and I wasn't really that much of a fan of react. I wanted to try other technologies. So I was like, oh, let me add a front end to Home Assistant now. At the same time Google came out with Material Design. And so I decided, oh, you know what, I'm going to add Material Design to Home Assistant. And then all of a sudden we're on like the front page of Hacker News, which is this social media website for like Silicon Valley startup kind of people. And you know, that actually attracted more people that want to do something with smart home and saw like the system I was building and they're like, oh, this is cool. And I started to get some people contributing. And at some point, you know, like two or three years later, this person Pascal came around. He out of nowhere and he's like, I wanted. Because Home assistant was an application, you had to install it on your own computer. And that was too hard. He thought, he was like, no, people need to build and we need to have an operating system that runs your application. And that way it's going to be called has IO and that way we can just give people an image for their Raspberry PI and then they, we have basically as if you buy something from the store. Five years into of all of this, basically we were burning out, we were too successful. So we were.
What happened is that like, you know, we, I, we were doing this before work, after work, we're doing this in the weekend and just like, you know, people have bugs, people have issues, people have contributions. And so now we're doing release management and security issue triage and all these things. And so around like the five year birthday, we started a company. So me, Pascal, who create operating system and Ben, who was a contributor who helped like redesign the whole architecture to be more modern Python. We started a company called Nabukasa. And the goal of this company was to make development sustainable. So really the ideal is like, okay, we want to get paid to do this so that we can keep building it, right? Something was already happening in the smart home world, which is that Amazon had launched Alexa and Google had launched Google Home. And they had launched that in a way, as big tech always enters a field. They had created their own cloud platform that didn't care about any standard. They told everyone, you can integrate with us, you have to send your data to us. You can no longer get that data. You cannot get the data of other devices. If a user wants to do something with their smart home, they have to go through our software or our hardware. And all of a sudden I see like this whole future and dream of the smart home crumbling because anyone that goes all in with Google can only wait for Google to add the features they want. But Google has limited attention span. Google is based in Silicon Valley. They have different problems and different technology than for example, countries in Europe. And even every European country already has different technologies, right? Like how can you build something that is not compatible or works for everyone? And so when we started the company, we, we saw that trend happening in the world.
We actually decided like to double down. It's like privacy and local control were always the foundation of Home Assistant. So we started this company, we decided it's all going to be local. Now what we also did is that this company is not gating things away, paywalling off features that run locally, because that doesn't like that makes no sense. Right. And so we were like, no, you know what, we're going to build an extra service, a cloud service, which ironically, because we hate clouds, but we also realize that there are parts that can run in the cloud without taking away from Home Assistant. So one part is Home Assistant runs fully local. Well, if you're away from home, you cannot access your smart home unless you can open ports on your router or a VPN. And back then it was even more difficult than it is today. And so that's a feature that people are willing to pay for. But also if we have like a cloud connection with, you could actually integrate with Google Home and Amazon Alexa, which some people want that because of their voice assistants. That's great for us because that we will make that kind of things easier. But we always had this rule. Nabukasa was never allowed to have features. If they were able to run locally, they should run locally. Right. We were never going to invest the resources that we got into building something proprietary and wallet off. And this actually worked very much in our favor. Like for example, the Google Home integration people could set it up themselves or people could go to a Home Assistant cloud. It was using the same code. And so all of a sudden Nabukasa employees are working on this, which are really just paid community members, but also volunteer community members are working on this because they run it in their smart home, setting it up themselves. Right. So we're just all working on this technology together.
Everyone should really be able to get a private smart home now. Nabukasa kept growing and growing. Home Assistant was also very successful. We kept growing and growing. We've been in the top 10 most active open source projects in the world. In 2024, we were even the number one most active open source project because we had 21,000 volunteers that are contributing to Home Assistant. And these are all people that like as a programmer, and this is before AI, right? In the pre AI world, when a programmer came home and they wanted to build something or they want to use their skill, like they love programming, they come home and they want to do something with technology, right? Like what are you going to do? Well, you're going to work on your home, right? Like maybe you're going to inside your home. There's different avenues of technology, right? Like sustainability or like, and heat pumps and all that stuff. Or maybe it is smart home. Well, what platform are you going to use for smart home? Well, that's Home Assistant. Because you can run it yourself, you can tinker all the pieces, you can create plugins, you can work on it. And because Home Assistant is open source, why wouldn't you give your bug fixes back to the community, right? And so this flywheel has been like accelerating for years now. And so, yeah, that's how we get all these people work on home assistance. And then around two years ago, we launched the Open Home foundation. Because up till then Home Assistant was separate from the company, right? Like Home Assistant will was still my project and Nabukasa was next to it had like around 50 employees at that time working on the cloud services, working on a home assistant, and all the money is flowing through Nabukasa.
And the problem was, well, there wasn't really a problem, but there was still one kind of Achilles heel to this whole system is that I was owning Home Assistant, right? And so if ever anything would happen to me, it would like ruin everything. But what we have built together was so important and we had this, in a way, an alternative to what big tech is offering. So if you look at it like our stack is fully independent, right? We have our own operating system. Don't rely on the App Store, right? Like, if Google and Apple want to ban us from the App Store, Home Assistant will live. We, we will survive. We have a website for access. We can still have mobile access, access from anywhere, right? So we had this full independent stack and so we really wanted to make sure that this keeps on living forever. And so we started the Open Home foundation and I donated Home Assistant to the foundation. The employees of Nabukasa that were working on open source also moved over to the Open Home Foundation. The Open Home foundation became the owner of all the trademarks of all the ip. My company today, Nabukasa, is a commercial partner. They have a license on the name Home Assistant, so they're responsible for turnkey Home Assistant solutions, as we call it. So anything that can work, plug and play. Home Assistant Hub, these antennas like the ZBT ZigBee antenna or Z Wave Antenna, and Home Assistant Cloud. But it's also no longer the only commercial partner, right? So the Open Home foundation now has a second commercial partner, which is Apollo Automation, and they are focusing on our ESP Home brand.
But we now have two commercial partners and these licensed products, anything that is being sold, 60% of the profits go into the Open Home Foundation.
Sean Ferrell: Oh wow.
Paulus Schoutsen: Right. So nowadays the Open Home foundation is 77 people, right? And they are like working. But because the Open Home foundation now gets the money and it's no longer nabukasa. The Open Home foundation is also structured differently than Nabukasa. They have their own goals, they have their own organization, they have, you know, they have UX researchers, we have a whole product team, we have a marketing department. All these things that are now also only focused on the Open Home foundation by splitting it up, by having like a very dedicated commercial group, having a dedicated mission driven foundation. And the commercial group is also mission driven, of course, like their heart is also with open source. But the meetings are about different topics, right? Like in the foundation, we don't have investors, we don't have anyone owning us, right? Like the only thing we care about is our mission and that's to build a smart home around privacy, choice and sustainability. And that is the only thing. So every feature we add is in those that mission in mind. And when we set the goals for the year, we don't want to build technology where we can become the next big tech, right? We don't want to be the one that owns all that technology. No, we want there to be like a thousand companies making voice assistants. Because if there's a thousand companies making voice assistants, they choose technology, they try things. The best technology kind of wins. Everybody will win, right? And so we don't beat big tech by becoming big tech. You change it by changing the game and really just trying to commoditize all the technology that they put out there. Let's create a private alternative, right? Like a privacy focused alternative.
Sean Ferrell: To jump off of what you just said. Like you've made a, you've already made a very clear argument for you. Don't be big tech by becoming big tech. Which that should be on like bumper stickers all over the world. It's just like on a personal level there's the, there's the question of how do you, how do you avoid that moment of saying like I could, I could cash in, I could take this, take some VC money and walk away. You could, yeah, you make, you make your bank and then, and then you go do some other project and your next project is the one that's for the people, but this one becomes the cash cow and you've sidestepped that while letting it become a corporation that can become self sustaining through a corporate model. And you've got the nonprofit side and you're maintaining public ownership through this very open license, it sounds. As for me, I have a background as a writer. So for me, Creative Commons would be the equivalent if I was to put together work and just throw stuff out there and say like, you can do what you want with this. How do you avoid. On the personal level, was there ever a part of you, the little devil on your shoulder that was just like, this is it, this is it. Go open a bank account tomorrow. And then
Paulus Schoutsen: Look, if, if I think, you know, it's. It's ironic that. Cause I'm from the Netherlands originally and I was in America and I was on a Visa and so on a Visa, I was only ever allowed to make money for the company that was sponsoring my Visa. And so mind you, for the first five years, that was not the company that paid me to work on Home Assistant, but on something else, right? So I was never able to monetize the first five years. And you know, the things I saw, the trend I was talking about like the Google home and the cloud stuff that really kind of like, you know, soured me on the whole idea of like a startup. And then, you know, talking to people and hearing about like VC money and like VC money, people think it's like, oh, you get a lot of money. It's like, yeah, but now they are in control, right? You give up control, you're no longer allowed to do and build the best product. You need to serve the investors. And in the beginning, these things are often aligned. The investors, they want growth. So let's give everybody free ChatGPT, right? And then if we get everybody in our pocket and they're used to our products, right, like now let's start monetizing it, let's increase the cost, let's remove the subsidies and then we see the real cost. And all of a sudden like, oh crap, I just forgot how to program myself. And now I don't know how to. I need to keep paying these companies for things. But I saw this in happening there. I was like, no, this, I don't want to go in that direction. But I also realized that the other thing that you know, I really like is being in control and not having people telling me what to do, right? And being actually able to build the right thing in a way.
I exchanged like one time cash out and having to find a second dream in which I'm as good as I'm in now in like this project and the fit and doing the cool stuff which I don't know if that's possible. I don't think that's possible. But also now I'm like, no one is ever going to tell me what to do, right? Like we are. There's no two hats where I serve investors and I serve the users. No, we only serve the users. And I will have this job for the rest of my life. Right? Like there's no worry is somebody going to fire me? Is there is like, you know, we own. Of course we have to be kept accountable by our users and they don't if they stop paying us. Right. That is because we are not serving them. Right. Like we're. There's something else, but we are only serving the user. And if you build a platform where you only serve the user, you can make so much great technology. I don't know if you guys are familiar with the work of Cory Doctorow calledEnshitification where he's like, yeah, you build your modes, you get everybody in the door and then you squeeze the left side, you squeeze the right side and in the end everybody's kind of sitting around because the, the cost of changing is too big and everybody that works there knows that they could do better, but they're not allowed to. Like, that's terrible, right? Like, and now I don't ever have to deal with any of that stuff. I just, you know, I travel to conferences. We build great technologies. Like this year we've been doing a lot of great. You know, we've did like infrared proxies and radio frequency proxy. We're bringing a lot of old technology into home assistant. And like that really excites me, right? Because we are just unlocking all these new possibilities. We're letting people keep old technology. We don't have to throw everything away, buy everything new just because there's a new standard or something.
And so we're just doing really cool stuff.
Sean Ferrell: That leads perfectly to one of our notes, which is one of the angles in examining the kind of work you're doing is that your smart home shouldn't die when some company somewhere dies that you've got. You went out, you bought a product and why shouldn't you be able to continue use that product if it makes you happy with the results and the company that made it goes away, but you can keep it going. And I was wondering, do you have any examples in your, in either your lived experience or things that you know are going on within the community that you've created around this work where a product is still out there and People are still using it because the community itself is keeping it alive.
Paulus Schoutsen: Yeah, so I think a lot of like products that for example, they use open standards like zigbee or Z Wave, they will continue, they will forever continue to work because they all work locally and there are just companies like I have a water leak detector here from a company that doesn't exist anymore. Their website is not online yet. I can still add this to my smart home. It just works fully locally. For example, X10, like all these old standards, right? Like nobody's adding X10 support to Google Home, but in Home Assistant we have support for X10. And in fact we've been this year adding a lot of work into serial connections and bringing serial connections, making that wirelessly work with Home Assistant. Do better work on mopus for example, because we want to. It's not just companies that went out of business. There's also companies that don't care about the smart part, Right. They will add for example, mopus connector because that is required by law that there needs to be some local connection. But. But they don't make it easily consumer accessible. That's more for either pro smart home installers or people that are doing more elaborate constructions. Well, we realized like, wait a second, we can connect to any port, we can read out any port and if it has a remote control even that's enough to put it in Home Assistant. But yeah, there's companies today also where they make smart products. They're not even discontinuing them. They just don't want their users to access it outside of their apps. And for example, that is happening right now with Volkswagen, right? So Volkswagen has an API and this is cars.
You might be like, is cars part of the smart home? Well actually if you think about sustainability and when do I want to charge my car? When do I want to like, you know, do I have a lot of solar energy? There's a lot of reasons why you want to use smart home to control your car. And Volkswagen is actively blocking everyone right now, right? So they blocked Home Assistant, they blocked other open source projects because they want people to go to through their app. And their features that they offer are just, you know, they only barely scratch the surface of what is possible. And I get it, they cannot build, they're not a technology company that can build all those things. But we are, right? Let us access it. It's not like the people that use Home Assistant to access Volkswagen's API. They actually bought a car, right? They bought a Volkswagen car. They're maintaining their car, they're buying new spare parts. Like Volkswagen earns money from these users by them having that car. Nobody's just accessing that API for nothing. But somehow at Volkswagen they are like, nah, we can. That's not cool.
And so, yeah, we're getting blocked.
Matt Ferrell: I was gonna say, I have a personal experience with, I won't name the brand, but a garage door opener that requires you to pay a subscription so that you can remote control and tie things together. And it was like, no, I found a very open source, easy way to do it and just bypass them. And it's all in home assistant, controlled by me locally, and I don't have to pay a subscription fee. That stuff drives me nuts.
Paulus Schoutsen: I mean, that garage door opener, it's interesting because this is one of those objects that is so big and heavy that if your garage door opener is going to charge you a subscription, you're just going to accept the fact that you don't have remote control anymore. You're just not going to like, you're not going to go out and call an installer to disassemble your garage door opener because this thing is heavy and big and it needs to be aligned with the door and all that. It's too hard, right? So you need to now all of a sudden, it's like a thousand dollar thing and you're like, that's not worth it. And this company, of course, they know it, right, that this is why they make it. You pay the subscription. And so, yeah, luckily there are open source solutions, although there's this cat and mouse game now going on where one of the garage door companies made it impossible for us to locally connect to their garage door opener. And so now there is another company, and this is amazing, they created a device, a zigbee device that you put your remote in from that is like allowed that you bought from the company. And all it does is just presses the button. That's it.
Matt Ferrell: Yeah, so it's a little, basically a robot finger that just pushes the button for you. That's great.
Sean Ferrell: And of course that's good for them.
Paulus Schoutsen: Why not? Right? Like that is like, you cannot stop humans from like pressing the button. So of course a robot can press that button.
Matt Ferrell: I want to tie it back to what I said at the beginning, which is like, I'm huge in sustainability. I have solar panels, batteries, I'm driving an EV. And a number of years ago, you guys, integrated energy is like a, like a first party. Like it's, it's a priority item. Like when you install it it's like there's an energy dashboard and you just plug in a couple things and suddenly you got this beautiful dashboard that shows your energy use, water use, all this kind of stuff. It's fantastic. What was the motivation around making that a first class citizen inside the Home Assistant experience?
Paulus Schoutsen: So in a way it was a bit of an experiment in that like we, Home Assistant integrates with everything, right? This is a programmer first platform. Like it took I don't know, like six, seven years before we hired our first product at UX or something, right? Like it was like we were hiring programmers and we would hire from our community which are mainly in the volunteer world is more programmers. I mean we're trying to breed now also UX and product like people in our volunteer community, but it's a lot more difficult. But yeah, we were, it was an experiment like, okay, we have built this perfect toolbox but you open Home Assistant and it is this blank slate. It's like good luck, you have every tool in the sun. You have the ultimate power, but where do you start, right? And so we always had discovery. People would discover some devices. At least your dashboard start is filling up. But that was not enough. So we really wanted to see how far can we push it. And this kind of aligned with it was an intern, we actually only had one intern ever at Nabukasa and but when he started, he's like, I need a project. I was like, okay, let's build an energy dashboard. Because the idea was like, could we basically build a solution on top of Home Assistant ourselves where instead of you integrating a device and then coming up with the idea of a dashboard of which data points you want to see and tying it all together, could we just ask the user, what is your grid import? What is your grid export? What is your solar generation? What is your battery import? Your battery export? And then we have all the pieces and we can make graphs, we can make usage charts. We asked them for like their pricing as well. The intern was doing research and he designed some mockups.
And then together with the community, with Nabukasa back then, we built this and people loved it, right? Because all of a sudden, yeah, you just, you plug in a few data points. Like you know, in the Netherlands you would buy like a P1 reader which gives you a real time energy usage from the grid. And everybody was like, wow, I have a €30 device and now I have this whole dashboard out of nothing. And this has been growing over like the year since then. It has, I have to admit, not has gotten as much love anymore after like our first, like six months. In a way, this was driven because we wanted to build a solution. But also at that point, like I was, you know, I read that book, I think Bill Gates had put out a book about like, you know, the, you know, the climate crisis and how we can solve it and, you know, what is the carbon emission and how. I really like the book because it's very, in a way, engineering approach with what are the solutions and how can we solve this and what is causing it. And yeah, that was also like, okay, I really want this, want to help have Home Assistant help people make, you know, a dent, right? And I think that when it comes to being sustainable, the first step is actually measuring, right? Because maybe, you know, you don't know where your energy is going. Like maybe you have an old switch somewhere that is just using 40 watts in standby, while new ones would use 1 or 2 watts, right? But you don't know because you've never measured this. How can we help people with these kind of things? And so one of the things we actually did when we launched our energy dashboard is that there was this company in Denmark called Electricity Maps and they were tracking per region in Europe, I think maybe they had the US coverage too, but definitely in Europe that the energy would get from their socket.
Right? Now, how much of that is from sustainable sources and how much was being generated by coal, right? So all of a sudden we could give people a score saying, okay, you've used, you know, 5 kilowatt hours from the grid. 70% of that was during the, during the day. So that was all green, right? So the user could then see, oh, if I was only to cover that 30% with a battery, I could actually be fully on sustainable sources while still relying on the grid. Because my country or region has already done the work, right? To get like 2 carbon dioxide neutral energy generation.
Matt Ferrell: A lot of people who follow me and use a lot of the same stuff I use, they're all on Home Assistant because they can get it all tied in. See their dashboards, see where their energy's going. It's like that knowledge, with knowledge comes power. It's understanding where things are going and how you can tie it all then into your smart shades and the lights around your house and when you're charging your EV or running your dishwasher, it's like all that kind of stuff. Those decision points come from what you guys are surfacing inside of Home Assistant, which is why I'm kind of obsessed with it. And it's been really fun to see how the community that's around my channel also is got a little obsessed with it as well because it gives you so much information about your home.
Paulus Schoutsen: Yeah, yeah. I feel like there's some things I wish we could do better. Not about gathering the information. I think gathering the information we do well. Visualization. We also do well. We've now have this Sankey chart which is a really. That's this chart where like you see where the energy is coming from into your house and it splits it by floor, by area and by device. You can really see where everything is flowing. And so that kind of stuff we've kind of nailed down. But I think that's just the insight. I think for example, for automations there's so much room for improvements. I think we don't do a good job in like oh, we think that you're going to have 3kwh excess energy today from your solar. You should do something now. Right? Like that is a trigger. Like we only know after the fact right now. And we should have some predictive capabilities in Home assistant to really like make that better. But that's, you know, this is one of these things where like I wish somebody would do it, right. Like we're open source. Could somebody like stop by and like. Because like this is the problem we have within the Open Home foundation because we're working on zigbee and Z Wave and matter and energy and dashboards and AI and voice and like there's just all and then. And music and whatever. Right. We're doing all these different things and so where is the energy team? Right. We don't have a dedicated energy team. We have a core team. We have a front end team of Home assistant. And so they do all these different
Sean Ferrell: things to circle back to the intern who helped develop that dashboard. Did you hire him?
Paulus Schoutsen: We did not end up hiring him but he's still active in the community.
Sean Ferrell: That's very good. Very good.
Matt Ferrell: Good for him. That's awesome.
Sean Ferrell: Yeah.
Matt Ferrell: I do want to tie back a little bit to when you're talking about how giving companies the power to build on top of Home Assistant. What I find interesting is over the past few years I go to CES every year and I'll go talk to some smart home companies that have products that look interesting to me like oh, a water usage tracker and stuff like that. One of the first questions I always ask every single one of them is do you support Home Assistant? And the response I get is one of two things. One which is kind of a look of surprise of like, really, you want Home Assistant? Why would you want Home Assistant to. Oh, yeah, we're absolutely, we're working on it, or it's coming or it's already out. Like, it's one of those two things. And it always is very telling because the companies that have that first surprised reaction are like the big tech angle. They're trying to create a moat. They want you to use your app, we want people to use our app because we're going to surface special things, all that kind of stuff that they're trying to create. And then the companies that go the other direction, they seem to click and get it, that go to where the users are, not where you want them to be. And there's several companies over this past couple years that when I've asked them that question, like, there was a water heater company that's located here in Massachusetts that has a smart water heater. And when I asked them, do you guys support Home Assistant? Like, yeah, we actually just launched our integration and it's like. And like I was just like very excited. And there's a company I'm actually an advisor to that's making an energy storage device. And when I first met with the team and I asked, are you guys going to support Home Assistant? The lead developer was like, oh, I'm a huge Home Assistant fan Matt we're going to be doing all this stuff.
It's going to be Home Assistant from day one. And I was like, so it's really cool to see that it seems to be gaining steam among the companies because more and more companies that I talk to are saying, yes, like where? That was not the case just a few years ago. So it seems to be getting steam.
Paulus Schoutsen: Yeah, I think what's happening is that, and it's partially because of the Open Home foundation, is that people realize that, oh, we are not just a group of open source hippies. They are kind of throwing some code around. Right? Like, no, we are here permanently. We are also not a company that can be bought. This is the Open Home foundation. And in 50 years they are there still as well. Right. And this is also what's really actually surprised me when we started Open Home foundation is that the type of conversations that we were having with companies was changing because all of a sudden they got it right. Like, I remember having conversations with companies, they were like, oh, you want access to our data? Well, we can set up a commercial partnership and you can pay per use that wants to access the data. And we would be like, no, we don't do that. Like it is the user's data and they access their own data and we just provide them a tool. And like nobody should be paying anyone anywhere here, right? And so now these companies used to be like, okay, but who are, why would we give it to you for free? And be like well we're an open source project. And, and there was like, they were like, what is an open source project? Like, well we all our volunteers were doing this together and they were like, who pays your salary? I was like, well I get my salary from this company, Nabukasa. Oh, you're a commercial partner? Partner. No, no, no. And so with the Open Home foundation it's very clear now to everyone, okay, this is a foundation. They see kinds like the Signal foundation or Mozilla foundation making like, you know, Mozilla makes Firefox, signal makes Signal. They understand now, ah, this is how it works. I think the one thing I think we could as Open Home foundation, don't do it.
All right now is that we don't get any funding from the industry, right? Like we are completely funded by the user. So we, if a company doesn't want to work with us and they're trying to strong arm us, we just be like, nah, it's not interesting for us. It's like your users, right? And they will vote with their wallet and if you don't want to support Home Assistant, be my guest, right? Like your support people will know, your engineers will know.
Sean Ferrell: And yeah, Matt took the angle of like he's, he's in this world, he's built this community, he has used the product, he, he loves interfacing with it. And I'm going to come at it from the exact opposite angle. I'm the guy who doesn't know any of this. How easy is it for something, somebody like me to jump in and say like I think I want to try this but where do you start and what do you need to start?
Paulus Schoutsen: So Home Assistant is very easy to get started with. I think that we have from our historical perspective where we were an engineering focused organization, it was difficult, you needed Python, whatever. Today you can just buy a home assistant green. Which sadly because of memory and pricing and storage pricing now is around $200. You buy that box, you plug it into your house, you open the app, you have a working smart home. And, and we will make sure that we discover a lot of integration. So from the get go we will find your TVs, your receivers, your maybe your smart Lights that you already have from another brand, like from a brand like Philips Hue, for example, you'll find those, ask them. You want to set it up. You go through these wizards. We have a very good default dashboard now we're working on revamping automations. All of that stuff is very easy. Now the reason why we still have this reputation of oh, Home Assistant is a little difficult is because when it comes to Home Assistant, the answer if something is possible is never no, right? If you are an Apple user or a Google user and you're Google, you search online like, how do I integrate this with Google Home? And you'll find some form that is not possible. Go use Home Assistant, right? That is like the answer you get, probably. And that's how people get to Home Assistant. But some of these things to integrate, they'll be like, okay, you have to go to this website, open the console, type in this code, you'll get the token, then put that into Home Assistant. Or you, you have to create a developer account for this service. And it's all very complicated, but in the end it's possible. Now what happens if you are an inexperienced user? Then, you know, you see that it's possible.
You see YouTube videos of people that it's possible and you try to follow those steps and you fail because maybe they've changed the UI or maybe one step is slightly different. Now you feel stupid because you didn't do it, but everybody else is saying that it's possible. Well, if you were a Google user, you say it's not possible. You just stop and you have a different plan, like, oh, you know what, let me return this device, let's buy something else. The core of Home Assistant, the everyday use is very stable, it's fast, it is easy to use. We make sure we put a lot of love in our user interface and then, yeah, the harder things are possible, but might be difficult.
Matt Ferrell: That kind of leads into kind of a future question or where things are heading. But like, before I get to that, I gotta pay props to the whole community and the team for what you're doing for the UX. Cause I used to be a UI UX designer in my previous life and in 2018, when I first started using Home Assistant, it was very techy, Like I was into it, I could use it very easily, but I would have not recommended it to somebody like Sean. I would have not said, hey, Sean, you should do a home assistant. It would probably be a little too daunting. Today, I would not hesitate to recommend to Sean to use it. The UX is like an order of magnitude, like every six to 12 months it feels like you guys get even better than it was before. So you're making some huge headway in the UX side, but with LLMs and AI and where things are heading in that direction, I'm curious, where do you think this is heading? Because like what I've done, just a few months ago I installed a community MCP server for Home Assistant and now I can talk to Claude and say, this automation's not working the way I want it to. I want it to do this and it'll spin away for a few minutes and go, I can do that for you. I'm like, yes, please. And so suddenly I've got this very complex automation that it built for me. I can tweak all I want, but like it was, the effort was non existent for me. Where do you think this is heading?
Paulus Schoutsen: So Home Assistant is the perfect platform for AI because AI is built on top of your smart home. It needs to have the ultimate toolbox. Well, that's where we came from, right? Like we actually started with the ultimate toolbox, trying to make it easier for human users, but we actually always already had the perfect layer for the AI users. And on top of that everything is accessible, right? So you can just read and write YAML automations. You can. Our database format and tables are all documented. So an AI can just go open the SQLITE database, look at your historical data, it can query any data in a home assistant. And so being the perfect platform for that, combined with the agentic coding phase in which we just arrived, right, like it is going to be awesome. I think the almost in a way the interface that we're building today for humans might not be the interface that is going to be used most of the time because people are just going to talk to their AI. And I think I'm totally fine with that because I want people. Home Assistant should be in the background. You should be able to live your life and let Home Assistant automate automatically close the blinds when it's like the evening or when the sun is shining on that side of the house, or turn on the lights when the sun sets in your home and let you know if it's your hobby. Yes, you're going to dive in and you want to do all the cool stuff. But most people, smart home is not their hobby and it should just work. And I think the bottleneck used to be like, you need to have the idea which, you know, you go on YouTube or you read magazines or whatever, you find social media somewhere, you find ideas, you want to bring that to your home. And now you had to learn home assistant, right? You had to learn what entity IDs were, what are devices and areas and triggers, conditions.
And I think the next step is going to be that you'll be able to talk to your AI. You say, I saw this idea, I want this. And I will be like, okay, so you want the light to turn on in the hallway when you walk out at night, but it shouldn't be too bright. Okay, well you need, your light is not brightness capable. So that's the thing. You don't have a motion sensor. But then you buy these things and then the AI will just help you solve that problem and then you move on. And my, my kind of dream in that sense is that we even go one step further because you talking to the AI, what it should do is it, what if we kind of more have a rule book where let's say you have think of it as a Google Doc that you share with, you know, your partner at home. And in that Google Doc you've like, I live here. These are the people that live here. These are the areas that I have. These are the kind of smart home rules you just write like as if you go to an Airbnb and you have all these rules like put out the trash that day. You just, what if you just had a sentence there in English that says oh, I want the lights to turn on in the hallway at night but not too bright. And the AI would see those rules create automations for you. You don't even know how the automations work, right? This is in the way the mindset, how people are programming nowadays. There's a hundred X times amount of people writing code. You wouldn't call them programmers anymore, but they're building, right? And I think that the AI can make these automations. Are they the most performant, most optimized automations? No. Will they take your smart home down also? No. So then it's fine, right? Like, then it's like it's okay. And if there's an edge case, let the AI look at the logs and figure it out for you.
So I really think that we're going to a smart home world where maybe you're just writing a book about your house, right, with the rules. And then if you want to go to a different system, you could take your rules, right? You just be like, let's give them to the next smart home system. Let them figure out all these Things. Is that like, sad for our users? No. Is it sad for the people that are hobbyists also? No. Because they still have the ui, they can still call all these tools. It's just that we will be able to reach way more users, right? So if people are not technical today and they don't even want to get even a little bit technical, to kind of, you know, install a box and plug an ethernet or something, right? Well, soon the plug in Ethernet might be the last step. You just say, hey, I just bought Home Assistant. Go figure it out. And then it just works. And we have some community members that are not happy with this idea. And I think this is not per se for our community, but people have put in a lot of effort to learn skills and learn the hard way how to do this, right? Like, I spent years learning how to program and to be able to make home assistant, and now somebody could describe something like home assistant and get something that looks on the outside, like in a weekend, maybe similar or something, right? I mean, it would not be as robust and performant, et cetera, right? But like they can get somewhere started, but way faster than I ever got to that point. But I think that's only great, right? We as programmers, and I say we as an I as a programmer and other programmers, we have solved our own problems mostly right in open source. Like, we, you know, there's people that are solving problems that other people have, but most solutions built in the open source world.
Wherefore what the programmers that were able of doing the writing wanted to fix, right? Like, I had my smart home lights, I want to control locally. That's a script I wrote. Well, there are people that have all these different hobbies that are now able to build tools for those hobbies without needing a technical background. And I think we're just going to see so many cool things that technology could always have done, but nobody bothered to build it. And so I think it's going to be glorious.
Matt Ferrell: We talked about like not being big tech, but right now AI is kind of ruled by big tech. So what, what's, what's your thoughts on that of like, how do we have our cake and eat the two here without having to give into big tech?
Paulus Schoutsen: So I have, I run, I have at home here a Jetson, a $2,000 box, and I run like Quentin 330 billion. It's not good enough yet. But last year Quentin didn't exist. I was running Ollama or yeah, Madlama from Meta. It was even worse, right? So the models are getting better. The technology is getting better. The way I see it is that like, you know, people are raving about deep seq4 when you run it on like a $6,000 Mac with like 128 gigs of memory. And sure, that is a lot of money and that is like a lot of memory. The memory that you need, the computer that you need is getting cheaper and cheaper, right? And so of course now we're also the memory crunch, which is not helping with this. But I expect there to be chips being developed like Google's developing their cpu, right? Which is like chips optimized for running models. Those are going to come out also for the general public and we should be able to run models in five years or 10 years that are way more efficient than where we are today. I think today there's a lot of energy being used, there's a lot of power being given to big tech because we're all jumping to their solutions. But just like in the beginning there was no Linux, right? There was Unix and mainframes and IBM and Linux came in the long tail and all of a sudden it's like ruling the world. Home assistant, we're not that far yet as Linux, but our impact and our power is growing. And I think that we will see more open weights models, maybe like models that you could actually reproduce building yourself where the data sets are public. That will come. And I think we get to a point where the models are good enough that we can do locally a lot of the tasks. And we are, when we look at, you know, today we have Opus 4.8 entropic, right? Like it's like a frontier model, state of the art, et cetera, et cetera. But it's four point.
Opus 4.5 is already good enough, right? Like I was already at that point, I was like, whoa, I can do all these things I couldn't think of. And of course you get higher models, you get better. But if the open source world would get to opus 4.5 level like that valuation that these companies are chasing is gone, right? Because why would you be a trillion dollar company that is going to get the money from replacing all the jobs if people will just run it at home or in their office and just be more productive without spending thousands of dollars every month? Open source in the end will always win because we see that there's just many stakeholders, many people investing in it and it just gets commoditized, right? Like you can use the same line of source code again and again and again and one person stops by makes it faster and gets faster for everybody. They will spend their time on something else, making a feature there, making it faster. And we build like we stand on the shoulders of giants. And I think, you know, as Home Assistant is built on Python, built on Linux, right. Like there's just going to be also it just compounds.
Sean Ferrell: As we look toward the future and we've just talked about like the wide horizon, you know, the idea of open source AI, you know, is for me personally where I would want to see that, that go, but I want to pull back and look at the micro level and ask you where do you see yourself in the next 5, 10, 15 years? Will you be a 90 year old hacker who's still working on the open source code that you're working on now?
Paulus Schoutsen: I, you know, I hope so, right. This is what I'm aiming for. Like I can play with gadgets all day long and I think we are just barely scratching the surface of what is possible. So you know, we've only talked about Home Assistant. We actually have a project called Esphome where people can make their own smart home devices, right? Like you want to make your own temperature sensors, you want to make your own buttons, that kind of stuff is ridiculously cool. And as we've been bringing like radio frequency and infrared into Home Assistant, we actually leverage Esphome. And so I think in five years, short term plan for Home Assistant is that I want Home Assistant to be the foundation of everybody's smart home. I don't care if you are using Google, Apple or AI on top of that. But at the core we have Home Assistant. We collect all your devices, information about it, control about it and then you can do with what you want. You can stick, stay in the Home Assistant universe or you can use any other system. But I want us to be that ground level so that everybody has autonomy, right? And anyone can do with it what they want. If we achieve that, we will see a level of innovation that is like, well, like combined with AI where everybody can build tools on top of all this data that they can now access of every device in their house. It's going to be amazing. As long as we keep building that and users trust us, users support us and we just keep maintaining that, I think we'll be set for life and we will be there in 20, 30 years. I'll be hacking on my gadgets and whatever gadget is then, right? We'll make it happen.
Sean Ferrell: Well, Paulus, I think that you, you paint a, an optimistic picture at a time when I think Most of the news that we consume about this kind of thing feels like the bridges are all on fire and we don't know which way to turn. So I appreciate the optimism and the ultimate goal of doing all of this so you can hand control over to the user, the owner, the individual, as opposed to. And then we've got your data and now we're going to be able to sell our next product to you. So thank you for that being your vision and your goal. And thank you for joining us. We really appreciate it.
Paulus Schoutsen: Yeah, you're welcome.
Sean Ferrell: Our thanks once again to Paulus. And we have a running theme here on Still To Be Determined. I'm sure many people have seen it. In fact, I'm going to guess a lot of you dropped into the comments without me even bringing it up. Yes, this podcast, when we have guests, does appear to just be three bald white men talking about tech that is not by design. It's not a limitation we're putting on who our guest will be. And with Paulus, there's the additional unexpected layer. Matt, I don't know if you picked up on this, if we both took off our glasses and if you gave the lower half of my face to the upper half of yours. You have Paulus.
Matt Ferrell: We're Paulus.
Sean Ferrell: We are Paulus. Talk about open source. So thank you everybody for watching the conversation we just had with Paulus before we get into taking a look at the mailbag and what you've said about Matt's most recent episode over on Undecided. Quickly, just a reminder, jump into the comments, let us know what you thought about that conversation. Was there something you wish we'd asked, something that you would like to hear more about? We would love to be able to follow up, not just in our conversation with you, but also follow up with Paulus, who is. Who was open source Pun, who is open to follow up questions from all of you. So jump into the comments now if you have anything to add to that conversation. And now, just to visit Undecided for a few minutes, I pulled up a couple of comments, Matt, about your most recent this is the this 10 cent battery turns seawater into fresh water. Yes, we can all stop worrying about taking our iPhones to the beach because if you drop it in the surf, it's simply going to recharge. Isn't that right, Matt?
Matt Ferrell: Yeah, exactly.
Sean Ferrell: No, that's the, that's what Sean took away from the video. But there were some other conversations. There was quite a bit about updates in the comments, Matt, not related to this topic, but basically to all topics people have asked before and they have asked in this one in particular. Revisit some of the old topics that you've talked about and give us some updates about where the tech is. And you and I have talked about this before. You have said, yeah, I'm doing that with some of these things. So I'm opening the door for you now. What update videos can we expect from you that you might be already working on for the somewhat near future? Because believe it or not, everybody, the scope of planning from Matt and his team in his videos, it looks months ahead. This is not. Matt decided on Tuesday to put together a video about seawater battery. This is. Yeah, this video probably started back in. I'm just throwing a. I'm guessing February, maybe even late last year. So.
Matt Ferrell: Yeah.
Sean Ferrell: What sorts of updates can people look forward to coming up later this year?
Matt Ferrell: Well, some that we're working on, like Fusion is one that gets people all riled up. But I haven't talked about Fusion in a while. And there's a company I made a video on like four years ago, Commonwealth Fusion Systems. I'm actually trying to talk to them about getting to one of their facilities to see if I can get in there to see things firsthand, talk to their team about how things are going. I typically tend to wait kind of like chunks of time because some of these things don't move on a dime. It's not like you can talk about fusion of like every six months, what's going on. It's like it's going to look like it did six months ago. So you kind of have to wait a couple years, see how things are going. But that's an example of one of the things I'm doing right now. So it's like I do revisit. I will be revisiting. There's battery technologies that I'll probably be revisiting that I did probably two years ago that I'll be revisiting now to see where they are now. So this stuff is definitely in the works. And Sean has seen. He has seen my work in Notion, so he's seen my production board to see what's on there.
Sean Ferrell: Matt and I have an ongoing battle that isn't. It's not acrimonious in any way, shape or form, which is nice to say because when Matt and I were little, it would have been. But
Matt Ferrell: yes, it would.
Sean Ferrell: The Matt is trying to get me to use Notion and I want to use Notion, but when I click any link to Notion, it doesn't take me anywhere where I know where I am. And I literally feel like I just kind of pop into. I pop into a parallel world and I look around and am I somewhere where I don't know where I am? And then I tell Matt I, I couldn't find the notes. And then he sends me a new link, and it takes me right where I need it to be. So, yes, I've seen the work board, I've seen the plans, I've seen the calendar. It's all very impressive, and it makes me very, very sleepy. There was also a good number of comments there. There were a good number of comments about this topic, which is, okay, you're. You're looking at companies that are thinking about seawater, batteries and how you might get the sources for that and what you might do with byproducts of that. Effectively, people who are asking what Jazz says in this comment, so what happens with all the chlorine? Every reactions got to balance. So why I pulled this comment in particular as a model is there were some comments that were saying, like, oh, this could be technology. That could be a byproduct of lithium gathering. Where they have the large brine pools, they're gathering the lithium, they have the brine byproduct, maybe that could be used as the source water for this kind of battery. And then there are people like Jazzar who's saying, yes, but if you're doing this, you're going to get another byproduct. Yeah, what happens to that? So I'm kind of inviting you in your research. Did you see anything about the causal chains that are in place around all of this? And are we at a point where. And I'm guessing I know the answer. Are we at the point where the research is still too young for anybody to have said, okay, here's what we're gonna do, and here's what we're gonna do with the Bibro.
Matt Ferrell: You'd be spot on, Sean. The research is a little too early. So, yes, in chemistry, it's all about balance. You're gonna end up with things like chlorine. So what do you. What do you do? That's still being worked out. And we could have gone down a whole rabbit hole on this kind of stuff in the video, but we didn't because trying to keep it more succinct and just to the high level aspects of it. But to address the chlorine specifically, the thing to remember is that in a simple way, this is just like you put seawater in and then you get fresh water out kind of the middle. And then the sodium and chloride that you separate during the charging cycle recombines when you discharge the battery. And it leaves us kind of that concentrated brine. So you're not like, you're not getting just chlorine. So it's like. So most of the chloride just stays dissolved as dissolved ions, not a gas.
Sean Ferrell: It's similar to other battery technologies you've talked about. You've talked about other batteries which have moved from. They move from a solid to a liquid and vice versa. And they're doing all these things. Like the oxidation battery, I remember, was similar thing where it's like, yeah, it's oxidizing and then it moves the other direction and it goes back and forth with charges and discharges. So we'd be looking at more of a closed loop system bouncing back and forth, as opposed to something that now you have to say, like, now I've got all this chlorine, what do I do with it?
Matt Ferrell: But again, to oversimplify things, there's things that the researchers are looking into that weren't included in some of the diagrams that I showed. But, like, they're trying to figure out, like, there are things you could do with catalysts and things like that to reduce how much chlorine and stuff like that is. Comes out of the process. So it's like they're trying to find ways to mitigate some of this to answer those questions. Not all the questions have been answered quite yet, but they're very valid questions that like, jazz was bringing up like a hundred percent.
Sean Ferrell: Yes.
Matt Ferrell: It's like, what are the off gases or the leftover materials that you're left with? What do you do with them? Those are trying to be resolved right now.
Sean Ferrell: There are occasionally people who show up and say, matt, you're talking maybe about the wrong thing. Like this one from Miglout who says, I wish you could get past thinking that we need to remove CO2 from the atmosphere. You need to check out the science on that because you're behind the curve. Are you aware of arguments like this that are saying, this isn't something we need to be worrying about that are legitimately scientifically based? Or is this something where maybe it's a bit of mis or half information that maybe isn't coming through on that specific comment?
Matt Ferrell: It's hard to read because there's no specifics given. It's just read the science and it's like, well, give me a link of what you're talking about. There's two basic ways I read this, and I don't want to put words in this person's mouth, but how I would interpret is one is if he's talking about mechanical direct air capture pulling CO2 out of the atmosphere, I think that's a fair critique because those systems have questionable use. It's a lot of energy. It's very costly. What do you do with it? It's like, it's, it's just there's a whole bunch of like, I don't know about this. But the other way to read his comment is if he means that CO2 isn't a problem for the climate, the science points to the exact opposite on that. So I just don't don't agree with it. That kind of a thing typically comes from the misinformation side of things. So without knowing what his motivations were or the details he's referring to, it's hard to tell what he's trying to get at.
Sean Ferrell: Finally, our best worst comment. And there's nothing bad about these comments. I just like calling them best worst. And it's two comments that were right next to each other and both had to do with poutine. Todd Ferguson jumps in to say, cheese curds with fries and gravy. Not just cheese. As a Canadian, I cannot stress this enough. Yes, Todd Ferguson. Poutine, fries, gravy and cheese. And then Nazmus shows up to say, dang it, now I want some poutine too. And dang it, now I want some poutine too.
Matt Ferrell: Poutine equals sleepy time.
Sean Ferrell: Poutine does equal sleepy time, at least for me. But there's nothing wrong with that. So thank you everybody for your comments. Please jump into the comments here. Let us know what you had to say about the comments that you just heard and the conversation you just heard. We look forward to reading what you have to say and sharing your thoughts in an upcoming episode. As always, if you want to support us liking, subscribing, commenting, sharing with your friends, those are all very easy ways for you to support us, as well as direct support, which is by clicking the join button on YouTube or going to StillTBD.fm. Both those ways allow you to throw coins at our heads. We appreciate the welts. And then we get down to the heavy, heavy business of trying to stay awake after eating poutine. Thanks so much everybody for taking the time to watch or listen. We'll talk to you next time.
Member discussion: