C (2:24)

Correct.

D (2:25)

In my calendar, just to see if the history was there. And literally April 3rd at 10:30am Chris Benson on Skype. That's how far. That's how far.

C (2:38)

Year. What's the year?

D (2:40)

2018. Did I not say the year? My bad.

C (2:42)

No, you didn't.

D (2:42)

April 3, 2018. Chris Benson. 10:30am Skype. That's what you're doing.

B (2:47)

That was way back when we used Skype, you know.

C (2:50)

That's right. We had to.

D (2:51)

That was so wild.

C (2:52)

It was our only option.

D (2:53)

And that's what the. That was the original conversation that started the host. Co host. Practical AI. I think it was a data show back then even. I'm not even sure if it had a name.

C (3:02)

And it didn't have a name yet.

D (3:04)

The beginnings of practically I and this long history of relationship.

B (3:09)

It was funny because I know I had reached out to you guys and then like, so, you know, there was. You guys had go time and you know, you know, there was this kind of changelog, you know, family that was already there and I wasn't part of it yet, but Daniel and I were. Daniel Whitenack and I were both kind of the data AI people in the GO community at the time. And. And so, like, I was thinking, you know, listen to changelog and stuff and thinking, boy, you know, maybe it's. Maybe these guys need to start an AI, you know, focused podcast or something. And I was thinking, but like, I'd like to do that, but I was thinking, but I need. I need somebody to do it with. And I was thinking, I got to reach out to Daniel, you know, he's the other AI data. So I reach out to Daniel and he's like, oh, by the way, I just started talking to Jared and Adam about this and like, I was like, perfect. I just sent them a message. So the timing, yeah, it all just came together. The timing was perfect.

C (4:07)

You guys were so far ahead of the curve.

B (4:09)

Yeah, well, it was like, it was very clear if you were really plugged into the AI world at that point, it was very clear that this was going like, you know, like where it was going. Who know, you know, things change all the time. But, like, it was very clear by that time that the gas pedal was on and, you know, sky was the limit and there was some kind of journey ahead. And at that point, Daniel and I wanted. We wanted to be steering that. That journey for everybody. And that was how, you know, and you guys were awesome in terms of thing. This would be fantastic and we'd love to do it. And, you know, that was back in 2018, and here we are in 2025, late 2025.

C (4:48)

Yeah, things have changed, but have stayed the same as well. Here's a funny story that you might not know, Chris, I've given you credit for this before, but I don't think I've ever told you this, which is at some point the four of us were on a call and this is like post launching practical AI but pre chatgpt moment. And you were lamenting that we like missed Nvidia or something. Like you, you. We were talking about the run up. I think Nvidia had just had a huge run up with regards to first it was gaming, but then also, you know, machine learning was kind of starting to take off and you were like, man, I can't believe, like, look at Nvidia. It's crazy, the hockey stick growth on that stock. You're like, but, well, we're too late now. We're too late. And this is like 2019. This is like 2019.

B (5:31)

I was so wrong.

C (5:32)

Yeah. Here's the funny part, Chris, is I thought to myself, are we though? I said, I was like, are we? And I actually left that call and I went, I bought a little bit of Nvidia stock thinking, you know, if Chris thinks we're too late. This guy's always ahead of everything. So I think he's ahead. So I, I have to thank you for a stock tip that has paid off nicely.

B (5:51)

Well, you're welcome. You buy. Contrary to my advice. But that's maybe, that's, that's probably, you know, that's probably right.

C (5:57)

So I need to talk to you more often and kind of do the opposite thing.

B (6:00)

Yeah, there you go.

C (6:01)

So, yeah, thanks for that. That was cool. Unfortunately I didn't buy enough to like just quit everything else and retire. But you know, I'm happy that you thought we missed it.

B (6:12)

I'm glad I was wrong on that. They've done amazing things and like, you know, I think it's kind of funny, you know, just in AI in general. You know, AI has been around at some level. Even, even the modern form of AI has been around for decades. You know, it's not a recent thing because like I got introduced to it by my, my parents who were actually, who were technical, technical people, Georgia Tech and Lockheed and things like that. And they were doing stuff back in the late 80s and early 90s and stuff. And my dad introduced me to neural networks, which is still the basis of all this stuff in 1992. And I think like, it was funny, you know, the tie in here to Nvidia is like we went through another AI winter. There's been a series of kind of like where everyone gave up on AI for a little while and then circled back around. They're called AI winters. And so the last AI winter kind of happened at the end of the 90s going into the 2000s there for a few years before the modern era, if you will, picked up. But I think the difference is that the notion of modeling and the software basis of AI was, was there and there were a lot of great ideas. And a lot of the stuff we're doing today originated back then conceptually, but we didn't have the hardware. We couldn't actually do the thing. You know, we didn't have these GPUs and now other types of chips that enabled all this to happen. And so it was really like the hardware side of things had to catch up so that the software. And like, when people say, well, why did we have an AI Winter? And I think to a large degree it wasn't the lack of amazing brain power, you know, to solve these problems and create the models. It was the fact that you didn't have the hardware infrastructure to do the things that people were envisioning were possible. And it wasn't until Nvidia came along and became the AI, you know, really the AI hardware company. I mean, I know they do a lot of software stuff, but you know, that, that made the difference. And you know, Google came along eventually with TPUs and lots of other players jumped in, but both sides had to be there. So a little, little, little journey down memory lane there. Yeah, it's the, it's the benefit of being old.

C (8:23)

Seen it all, Chris. You have seen it all.

B (8:25)

I've been around. I'm old as dirt. So.

C (8:27)

So from your purview, what is dirt? This is not stock advice. But from your purview here at the end of 2025, and you have Nvidia, you have AMD, you have Google, you have Meta, you have these large players making huge investments. OpenAI, of course, I mean, the list goes on and on and on. Which single entity do you think is best positioned to like, succeed over the next 10 years if you had to pick one of the top contenders? Like, is it Google? They seem like they've really turned the corner, but I'm not sure if their capital investment on their own infrastructure is going to be the big win that some people are saying it is. I don't know. What do you think?

B (9:04)

I think they're so. I, I don't, I, I'm gonna cheat a little bit. I don't really have a one. You know, for a long time people would say open AI, and before that they were saying, Google, there, there is a, there's a, a top group and they are certainly doing well. And I think kind of the, at the risk of getting slightly in terms of social issues, you know, there's growing inequality between kind of those group of haves and kind of a lot of others that are have nots in that way. But I know one of the things I think is that, I really think that open models are becoming increasingly important because the difference, if you go back a few years and it wasn't coming out of OpenAI, there was a big performance difference in what you were able to do. And if you look at the closing of the gap between what's possible, I mean there are millions of open models out there and there are hundreds of them that are in kind of like they are nipping at the heels of the, of the leading ones and that gap between the latest, greatest thing from one of these big name companies and what's possible in the open world has narrowed dramatically. And what that's really doing is pushing, pushing model creation into something of a commodity, you know, area. And so like I'm, while I, and I think you've seen that in terms of what some of these big companies, you know, they built services and they're building separate businesses and they're going into verticals and things like that. But that's because the, just the model generation is not going to be the profitable thing, you know, for years and years going forward.

D (10:45)

Yeah.

B (10:45)

And so they're, they're, they're turning, they're turning from being AI providers explicitly into AI service providers now that are specific to different types of businesses. And, and I think they'll, I think they'll do quite well, I think kind of, I don't know, I'm afraid, especially after pointing out my horrendous last time. Yeah, I was going to say after my horrendous Nvidia prediction, you know, the last thing I'm going to go do is pick a winner here. So. But yeah, I mean they're making a lot of money by, by pivoting, you know, within the scope of what they do and they have that, the expertise. And I mean like meta as we're talking now, meta is just like just purely buying the AI talent, you know, like I don't care what Google is going to pay you, I'm going to pay you 10 times more. And there's no way you're going to go anyplace but us and trying to kind of catch up to that, that open AI, you know, which is still, as we speak, probably still the, you know, kind of the gold standard there, but with a few others such as, you know, such as Google as you mentioned and, and other, and several others that are kind of nipping at the heels there. So yeah, it's interesting times.

D (11:51)

So a long winded answer is OpenAI? Is that what you're saying?

C (11:56)

You have to go back and analyze what Chris said and tease out the truth of it.

B (12:00)

Oh, I tried to escape that. You know, Adam, that was not fair. You know, I worked really hard for five minutes to kind of squirm my way out of your question there. Yes. So very close when you say open AI, very close. You got the word open, right? How's that?

C (12:12)

Okay, so Chris's answer is the open models will catch, will commoditize the frontier models, so to speak. And these people that are just buying all the GPUs and just training, training, training and then of course inference as well.

B (12:26)

But I mean what you can do, it's requiring, we're seeing this progression where we're building out frontier models is costing less money. Like there's a ton of money in some of them. But the efficiencies that are now built into training from some of the latest research has made it to where you can build some amazing stuff with not quite as much as you might have expected a year or two ago in terms of relative, you know, performance against the hardware that you need to support that. So it might be, who knows, I mean, where the research is taking.

C (12:57)

Is there such thing as peak parameter? I mean, I think I read that xai's next model coming out whenever is going to have a trillion parameters or something. And it's like, how large is large, too large or is there no such thing?

B (13:12)

So, yeah, I mean, one of the things that we've, that we've been talking about for a while now is the fact that like it used to be in the early days of the, of the GPT series from OpenAI that, you know, you saw distinct capability differences as you went from 3 to 3.5 and to 4 and that kind of stuff. But there's also been, you know, we've seen kind of plateau. It's almost like you're seeing that a lot of the, it's not just a model thing, but also some of the infrastructure that's being built around it has given it a much, has made it much more accessible in terms of its productivity and its usefulness and there's less of a friction when we're trying to use models at this point. So I do think that there, that like there is no infinite rise in terms of the number of parameters you have to do I think, I think that that does level out and also like if you're going to have that many parameters, being able to use that productively from an inference standpoint, the world turning out to be a mini model world instead of a giant model world, you know. And so I, and I'm not sure that a lot of people in the general public, you know, that aren't people like us that follow this closely really realize that. I think when they think AI, they're thinking chat GPT because it's what they know and you know, one model to rule them all, one model to bind us. And like I'm not at all like that's not what I think is the world. I think the world is, is Many, many models contribute to solving a problem in various ways. And you know, we're here we are in 2025, deeply into the, the, the agent, the age of agents. And so it's no longer just models but now agents with models that are, they're acting on your behalf. And I think it's, the reality is it's a mini, it's a mini agent future that we're talking about here.

D (14:58)

Before we go there, I got to ask you, because we're talking about companies and predictions and potential here. Have you tapped into or heard of the next Jeff Bezos thing? Prometheus and the startup he's sharing, co founding, etc. Are you, are you tapped into that?

B (15:14)

I'm not, I'm not up to date on the details.

C (15:15)

That's like hot off the press, isn't it?

D (15:17)

They announced that yesterday's news basically today, Today's news.

C (15:21)

I think there's like a perpetual Bezos Musk pissing contest that goes on and this seems like the next one. He's like, you have Xai, I've got this thing.

D (15:29)

According to TechCrunch, Jeff Bezos reportedly returns to the trenches as co CEO of new AI startup Prometheus Project. Prometheus. So he hasn't done anything from a CEO aside from shareholder, you know, chairman, etc behind Amazon. He's been just, you know, getting swallow essentially getting slow and going to space on his yacht.

C (15:55)

Yeah, yeah.

D (15:56)

As you would if you were.

C (15:57)

He's been doing the space stuff. He's been, been doing Blue Origin.

D (15:59)

That's what I said. Like getting swollen. Going to space. That's what he's been doing.

C (16:02)

Yeah, yeah.

D (16:03)

So this is kind of cool that I suppose the next big thing could be from him. So maybe the next time we talk, Chris, you can give us your non prediction prediction.

B (16:15)

I can Slide out of that one too.

D (16:16)

Yeah.

C (16:17)

Like what we want. Do we go by Amazon right now? That's what we want to know, Chris.

B (16:20)

So, so I, you know, like, I'm probably the wrong person to talk to about this. Not only because of the prediction that we just talked about.

C (16:27)

Right.

B (16:27)

But also I want to point out, like I, I am honestly, like, this may sound really counterproductive as, you know, practical AI co host on this, but I'm. And I think Daniel's the same way. We're less interested in kind of the big, big names coming out with their latest big things because there's so much amazing work being done by like real people out there. You know, like, like there's the, there's the chat. Yeah. Plastic Jeff Bezos, like, hi, I'm Jeff Bezos, kind of, you know, like, and Musk and, and all these guys and I'm just like, they're always one upping each other and, and they do some big things. But like, like, I think like 99 of the press is going to this, these, these people. But I think 99 of the real productive work in AI is going to all these invisible masses of amazing people that are doing the stuff every day. And like, I, I like, you know, if I could, if, if I could get the mainstream press to kind of like refocus, I'd be like, they like, like look around. Like there's just, there's just astounding amazing things that are happening, but they're not happening by these like famous figures. And these guys. Yes, they have tons of money. They're super, super ultra wealthy beyond imagination. And, and they can throw their money around and stuff. But you kind of mentioned it's kind of the pissing contest, for instance, between some of them. And Yeah, I just like, there's so much cool stuff out there that's not, that's not the latest, you know, the latest Bezos, you know, Elon Musk. Yeah. Massive thing.

D (18:05)

I mean, $6.2 billion behind this thing is quite crazy. Quite a, an investment in there that he's raised for it. $6.2 billion.

C (18:14)

What are they doing? What's their deal?

D (18:16)

It's only speculative at this point. It's only got a name. Project Prometheus. Jeff Bezos, co founder, I believe is, is Vic. I would only mess up the last name. B A J. A J is the last name of Vic.

B (18:31)

Can you imagine being able to throw $6.2 billion at something that you don't really know what it is yet?

C (18:37)

Right.

D (18:37)

Well, I think he Knows, I'm just.

B (18:38)

Saying, I don't know if we know.

D (18:41)

I think you've already checked for 6.2 bill or you even raise those funds.

C (18:45)

You got the reason he announced it is to get better raises.

D (18:48)

Yeah, that's right. Some version of more money, get people interested.

C (18:51)

So Chris, you probably can't convince the mainstream media to ignore, you know, the 800 pound gorillas, but you can convince us. So here we are, we're ready. What's cool? What's underneath the covers or like what's the invisible stuff that people are doing that, that you and Dan and we should be interested in.

B (19:09)

So I think like, like it's funny, we just. I'm gonna say something that I said the other day and I'm starting to say it more and more, but like I think people easily look around wherever they are in the world and whatever their politics are and it feels like moment and it feels like, you know, that you, there's all these things you can point at and say we're going through a really tough time and it's tough and everyone's trying to figure out. But I'm, I want to offer a counter narrative to that. It. We're also at this moment where this stuff is, has, you know, this, the AI and there's a hardware revolution going on and there's a robotics revolution going on all together and they're all, you know, connected and they're powering each other. And I think we live in the coolest moment in human history right now. Like we are sitting in it as we speak today. And so I agree, what, what's happening right now is with all of these different relevant, you know, capabilities, you know, and the robot people and the AI people and the software people and the hardware people, it's all coming together and you can do amazing stuff today that, that even a year ago we couldn't do. I mean, it's like if you think about before now we would, we'd had kind of have several years of little software eras, you know, and we were, we were getting into certain ecosystems with a language or whatever and they kind of run for a few years. But right now it's changing so fast and the capability is coming so fast that like, aside from the big 800 pound gorilla types and stuff, like everybody can get into this stuff. And so I think we're at a moment right now where like it's really going to start being pervasive in everyone's life in a bigger way than it has been. Not just like, I'm going to open my phone up and talk to chat GPT kind of way because yeah, I mean that was unimaginable if you think about it just a few years ago. It wasn't. It hasn't been long since. That was an unimaginably amazing thing to do. But that's like we don't even think about that now. You know, we do it all the time. Don't even think about it now. But like physical AI and the fact that robotics have come so far in the last few years and that now there are. In addition to Nvidia there are many other chip makers that are coming on scene to support AI and some of them are doing more of the dedicated AI chips and others are doing more like combining different types of chips so that you have that. And some are great for data centers, cloud big cloud data centers and others are great for edge devices and tiny little constructs. And I think like you're going to see so much happening in the marketplace right now that are coming from startups. They're not coming from the 800 pound gorillas. They'll have their fair share at 6.2 billion. They better. Yeah but.

C (22:04)

Or do something with that.

B (22:05)

Yeah, you're going to see amazing capabilities coming out of fairly small companies and like and to speaking back again to Daniel Whitenack, my co host and part of our family in this. He started his own company which is kind of supporting that and that's what I like seeing. He has Prediction Guard which is kind of supporting open model approach and I think that in general that approach of anybody can go whether you're using a cloud environment or startup like Daniels or something like that, you can go productively pull down models from hugging face, you know which I liken to GitHub for AI, you know the way GitHub has always been for software. Combine a bunch of different fairly sophisticated open source software packages and do some amazing things. Without 6.2 million you can do it as a, as a college student in the dorm, figuratively speaking. And that's, that's like what, that's the thing that really excites me is that is the ability to everyone becomes a maker, if you will. Everyone out there can become. Once upon a time we were kind of like, you know, hey, we have the Internet. Everyone can be a software developer. You know, all the stuff you need to learn is online and there's all these resources. A lot of it can be done for free. It doesn't matter where in the world you are. Well now everybody can Become a maker. Everybody can take, can, can access these different things and go do something great. And I think that's that the fact that, like, we all have these, like, Roomba type things, you know, rope these, these vacuums in our houses and, and everybody is now completely used to that. But I think we're right on the cusp of having lots of little devices like that in our houses and our businesses that are doing all these things which eventually will get us into this. This notion of swarming that we're, that we're going to talk about.

C (23:55)

I'm ready for the little robots. I don't want the big scary robots, but I like the little robots that help you do things. The Neo thing is weird. We don't have to talk about that. But that was kind of strange.

D (24:04)

Was it Neo? It wasn't Neo. John Mnemonics, you think Johnny Mnemonics?

C (24:08)

Yeah. What's Johnny Mnemonics?

D (24:09)

Well, Johnny Mnemonics was like he had, Man, I can't remember this one, but it was the same actor, Keanu Reeves, and I believe he had like. Oh, he had something in him and he was carrying data. And it was vaguely.

B (24:21)

I recalled this. Yeah, it's been a while.

D (24:23)

The idea of a mule but not drugs. Yeah, it was.

B (24:26)

That was back when he was young.

D (24:28)

Yes, yes. I thought you were talking about Johnny the Monitor.

C (24:31)

You jumped right to the Matrix. Which makes sense, Adam, because most of my references are the Matrix. But I was actually talking about this new robot in your house that cost 20 grand and is controlled by a human currently.

B (24:42)

I saw that, but I still don't think that's going to be the thing.

C (24:45)

No, I don't think so. I'm going to say that's kind of weird. At this phase, it's a general purpose. Like, it does laundry, it does your dishes, and it's like a humanoid, full size, similar to what the optimists think they're building. And yet it's at this point because they need data to train these models better. It's not at all autonomous. It's controlled by a human with what I imagine is like a sophisticated joystick, you know, probably overseas.

B (25:08)

It's kind of creepy when you think about it, isn't it?

C (25:10)

It's super creepy.

B (25:10)

Your grandma's in there with a stranger in the form of a robot.

C (25:14)

The Wall Street Journal did a great video about it and it, like, you know, Joanna Stern told it to do the dishes or something. It took like three minutes to load a cup into the, you know, into the dishwasher. Which is a 15 second task anyways.

D (25:27)

Let's not do that.

C (25:28)

Yeah, it's not there yet. I feel like that's being too big in general purpose, I feel like more specific, small, like sing like the Roomba, you know, it's gonna vacuum.

B (25:36)

The Roomba is the future. Is that like that, that was an early, you know, thing. But like it's, it's purpose built for a very specific thing. And it's, and there's a whole bunch of them on the market, you know, a bunch of different makes and manufacturers and stuff on the market. And you can, we can go through and debate what's better and all that kind of stuff, but I think you're seeing that times many, many, many things across all sorts of tasks. And like, they're cheap and like, even Roomba type, you know, the vacuums are too expensive right now. I think as I think with the cost of robotics coming down and accessibility, then it's like if you think, you know, outside this and just walking into a retail store or getting online to Amazon or whatever and just buying something, you once upon a time might have been expensive and now it's 30 bucks, you know, and I think that like, you know, in this day and age, that 30 buck purchase, I think that, you know, getting a robot that'll do this and that and the other, and the fact that they have eventually, you know, you have families of robots that can do different things and you can put it in swarming mode and just say auto, my house in swarming mode as we'll get into. And they just like coordinate and do all the stuff. They're sensing you, they're moving around you, you're doing the thing and, and that's, that's real life. You know, you have, you know, you know, aside from just the vacuum, your, your lawn and garden care is getting taken care of, your security around your house, your roof and gutter inspections, you know, it's integrated into your smart home stuff. You're like, you know, you don't have to worry anymore about where your packages were left by the delivery driver because the, the, those robots or the swarms that are managing your house are just doing that. And it's not insanely expensive. People are like, yeah, yeah, where am I going to get the 6.2 billion from Bezos to buy my swarm for my house? And I'm like, no, no, it's not. You're going to have the Christmas deal. You know, we're coming up on the holiday time and you're going to get Online and you'll have all the different packages. About what level of swarming do you want? This one is an 18 accessory Swarm package that you can come, it's going to handle your outside, it's going to do this and you're like, you're trying to choose, you're like, well I don't know, you know, I'm gonna spend more, I'm gonna spend more for my kids, you know, on that. But you know, there's, there's, there's great aunt, you know, Louise and we only talk to her once every five years and I send her kind of a token thing. So I'll send her the four item swarm package, you know, that she can add into whatever she's already using because it's all open stuff and like that's not like that's going to be normal and we're not that far from the opportunity and it's not the 800 pound gorillas that are going to bring that. It's going to be the billions of startups out there that are each doing a little piece of it and they're and their swarm components and stuff are able to communicate. That's the future that we're going to build.

D (28:32)

Well, I'll tell you one thing, you've definitely put a lot more pressure on the idea of Home lab, that's for sure because that's all Home Lab. Those are a ton of DNS queries out there, probably a ton of telemetry being tracked. A lot of things you may or may not be concerned. Those are things I think about when I think about adding more and more devices to my home. Gosh, man.

B (28:55)

So separate. I have a slight side story, but it contributes to that. So about a year ago now, almost exactly a year ago, we bought and moved into the house that I'm in now now. And the guy that we bought it from, he and his wife, he was a fanatical home automation person. And so like we moved in not because of the automation that was incidental but like it's had, it's helped me move from just like thinking like more of a professional kind of thing. Like you know, we're talking AI and a professional kind of to thinking about stuff around the house with all the sensors and the cameras and stuff. And we have all the, you know, the, the various types of home automation stuff that you see out there combined. Costas here, every like we have many, many, many dozens of casas devices all over the place.

D (29:47)

And CASA is the brand that from Lutron, is that right? Am I picking that right?

B (29:50)

It's from TP Link, actually. TP Link, but that's just one. There's a whole bunch of them. And you know, Apple Home and Google Home.

D (29:57)

I was thinking Casita. Casita's from Lutron. Those are the light switches.

B (30:00)

But yeah, yeah, the Lutron does the light, does light switches. But like, they all, there's some common protocols that they all work on. And I'm starting to see, like, I, like, because I didn't have to go start it from scratch and because I inherited what this guy had already kind of put together and then had to figure it out and make it work. And suddenly I'm like, oh gosh, it would be really easy to add this. Like, and we're, we're talking about this robotic future even in our homes. Not just a commercial or industrial or whatever thing, but in our homes. Like, it's so easy for me to see that now because, like, I realize I already have a good bit of infrastructure here. And it's not expensive and it's not, it just takes a little bit of effort. And if they can make that easier for people to get into, it's a done deal. Like, you know, we already have, we, we, we already have WI fi and all the other things. And then you start adding things to plug in. It's, it's like Legos. It's like home automation. Legos in your home.

D (31:12)

Well, friends, you know that feeling. Your team has solid ideas, you got some good stuff in there. But there's this gap between these brainstorm sessions and actually shipping the stuff. Weeks of back and forth scattered feedback. You're in chat, GPT, friends and colleagues are in Claude and somebody else is in Claude code. And it's just everywhere, right? You're throwing AI at it. It's not solving the problem just gets messier. Well, Miro, they flipped the script. Miro's innovation workspace. It is powered by AI and it turns work that normally takes weeks into days. Here's the magic Miro. AI sidekicks aren't there to replace your thinking, they're there to extend it. And custom sidekicks that think like product leaders, agile coaches or product marketers, they review your materials, they highlight what to double down on, and they catch gaps for you. So you're not spending weeks, hours, days, months, whatever, going down the wrong road without details. It integrates directly into your workflow as an extension of your team's capabilities, not a replacement. You can generate insights faster without tool switching. Miro AI sorts through everything, sticky notes Research ideas, different formats and combines them into structured research, summaries, product briefs, sentiment analysis, you name it. Test 20 concepts rapidly with Miro prototypes. Generate instant prototypes right from your board so you can iterate on variations and feel confident about feasibility before you get into to the heavy builds. And this opens the door for teams to move at a brand new speed. The speed of their ideas, the speed of enhanced thinking. Spend time building the right things, not digging for information. Get great done. If you're ready to move faster, check out our friends over at Miro Rhymes with Hero m I r o.com Once again, miro.com.

C (33:09)

Speaking of Legos and home automation, IKEA just announced a whole new set of like 27 smart home things coming from Ikea. I saw that talk about bringing it to the masses. Like that's the kind of thing that IKEA brings to the masses now is they make it very simple and straightforward and Lego esque in order to. And it all runs on matter, which is I think the open standard for communication between these things.

B (33:31)

And so like matter's in an interesting place and that like I only buy things that, that are matter, that have matter integrated in and for, for listeners and viewers. The matter is a protocol that allows different makes and models of automation to work together over a common protocol. And it's local based instead of cloud based. And so like, but not everything does it yet. So it's still kind of working. It's been very slow, it took a long time to kind of come into, into play, but it seems to be having a second wind right now because of all this new capability that's coming about. And so like every new thing I buy, whether I'm using matter yet on that or not, I have to have matter so that as I go forward I can integrate into that. But like, yeah, you know, everything is, it's local, it's matter. And I'm finding with today's craziness out there that I'm moving more local and a little bit more out of the cloud. And so matter is becoming increasingly important from my standpoint.

C (34:26)

Well, from the startup perspective and the swarming perhaps at least the droning perspective, you'll be happy to hear, Chris, that we do have a startup coming on soon, Zipline, who are now moving delivery drones into production. They actually have a delivery drone system that is started off delivering medical needs in Africa and vaccines and stuff like that. And now they're moving into the States and they're doing food delivery, small item delivery, small package. So you think your Chipotle burrito, that kind of thing. Yeah, your, you know, your.

D (34:58)

8 pounds or less.

C (34:59)

Yeah, 8 pounds or less. It's super cool stuff. And they've got it to where they're actually rolling out into, into commercialization now. So startups are making moves in this direction. And now there's our. I assume in each city they have a fleet of these delivery drones, but obviously each drone is, is operating on its own, I assume eventually autonomously. It actually seems like a simpler problem than autonomous cars because the airspace is just pretty open. Right. Like you got problems like wind and snow and stuff like that. Birds. But it's gotta be easier than cars.

B (35:34)

Yeah, generally. And so it's a different problem. So it's easier. There's easier. It's a little bit of both. It kind of depends on how you're looking at it with cars. And like we were just talking to Waymo again a few weeks ago ago on practical AI about this. So this is very top of mind for me. With cars, yeah, there are a lot of challenges and you have the, the notion of the, you know, the child running out or the ball bouncing out and, you know, there's a lot of stuff that's right there. But also your, you know, what you're, how you're navigating is very well defined in terms of the streets and stuff like that. Air becomes more three dimensional and so the challenges are different. But so long as it's not highly congested, you're. I would agree with you that it is generally easier that you, you can kind of move from here to there. And so long as you have good collision avoidance and some other, you know, capabilities for navigation there, then you're probably doing okay. Though that changes with swarming because swarming brings in close collaboration. Yeah.

C (36:33)

So define swarming then, because I think of killer bees when I hear swarming, and I assume with drones you're talking about a bunch of drones nearby each other, then you are.

B (36:42)

And, and it's in whether. And it's not just a physical distance thing, because what. Distant. What is physical distance is a relative thing depending on what it is you're trying to do. But it also, it's really more about behavior. And so we can dive into that. But before you say that we can, we. I think that's a, A line of thought we should go down. Is that as. As you guys know, I'm really into animals. We were making jokes earlier about bazillion dogs and stuff like that. I'm a licensed wildlife rehabber and I study animals. And in the context of swarming, Mother Nature has perfected not just swarming, but there are many different types of swarming from different species. And so I have a set of species that I tend to look to for swarming purposes and say if I want to swarm with this type of technology or this type of platform, like how do we get started on that? You know, how do we get inspiration or look for some insights on the technology where you can look to certain species that are, that are similar to the technology platforms you're interested in in terms of how they move around and do stuff and say, well, how, how has nature solved it there? And I definitely do that a lot. I, I, it's not uncommon for me to go into tech meetings and start off with lots of pictures of animals and stuff and people like, what's, what's.

C (37:59)

Going on with this, who is this guy?

D (38:00)

Are you thinking like fungus? Bees and bees bat.

B (38:04)

Like I do a lot of bees, bats, birds, starlings, you know, those huge, what are called murmurations of starlings. Ants are awesome. Ants are awesome. When I'm thinking about robotics on the ground, meaning what we would call a ugv, which is an unmanned or uncrewed ground vehicle. Ants are amazing in what they can do and so they're an awesome thing to look at. But for the. I'll start with the definition that I use. Given the fact that I work in the, in the military intelligence space. My definition sounds kind of, it uses that jargon, but it really don't get caught up in that. It can be applied to residential, it can be applied to commercial, it can be applied to industrial. So don't get caught up in this specific wording. So I'm going to read it in front of me. It's one really long run on sentence that's very specific in what it's trying to imply.

C (39:02)

Okay.

B (39:03)

It is Swarming occurs when numerous independent, fully autonomous, multi agentic platforms exhibit highly coordinated locomotive and emergent behaviors with agency and self governance in any domain which could be air, ground, sea, undersea or space functioning as a single, independent, logical, distributed, decentralized decisioning entity for purposes of C3, which is command, control and communications with human operators on the loop to implement actions that achieve strategic, tactical or operational effects in the furtherance of a mission. So long, long, long sentence. But it hits a bunch of very precise concepts and integrates them in together.

C (39:52)

I can tell each word was selected there.

B (39:54)

Yeah, a mission might be, instead of thinking military, a mission might be getting a package to your house that might Be the mission. And that does have command, control and communications involved. So like you can put the, you can, it doesn't have to be the, the military esque jargon that we're talking about.

C (40:13)

Yeah, yeah.

B (40:13)

It applies to any of these, to any of these, you know, commercial, industrial, residential, military, whatever. So. So that's a lot. That's a lot. It's a lot. And if you want, I can kind of break down high level what some of those mean.

C (40:29)

Yeah, I think takeaway, we can talk about the individual words because I know they're very specifically chosen, like independent, logical, distributed, decentralized decisioning entity like stuff like that. I can tell each word was selected for a reason. But I think my grand takeaway of a swarm is kind of the e pluribus unum. Like it's like, okay, all these things are individual and autonomous, but they're all acting as one. They're acting with one purpose.

B (40:56)

That's a fantastic insight that you have. And that, that is the key to it is like swarm. Such a buzz, you know, we, we always have buzzwords in, in this AI and software space. There's always the buzzwords of the year. And swarm is certainly a huge buzzword right now. And, and almost without exception, I will turn around until I can go back to my definition, assuming that you want to accept that as the definition of swarming. And I can defend that fiercely.

C (41:24)

I cannot attack it. Can you attack it, Adam?

D (41:28)

Wouldn't try it.

C (41:29)

You're the expert here, Chris.

B (41:30)

But I would say like all you people who are talking swarming. No, you're not. It's not swarming. What you're describing is all sorts of things that lead to swarming. There's a whole bunch of incremental capabilities that, that would eventually, as you add all those capabilities together, they culminate in swarming.

C (41:48)

Right.

B (41:48)

But, but the chances of somebody saying that what they're doing out there is, is consistent with Chris Benson's definition of swarming.

C (41:58)

Right.

B (41:59)

Is pretty low. So you, what you said was right on. And that is that, that just as you see in nature with those ants, all, every little ant has its, you know, neural capability, shall we call it, you know, and what it's doing. But at the end of the day, they're functioning to get a mission done, a job done, something productive for the colony. And they are all lending themselves to that greater good. Even if some of them may not survive that kind of thing. They're, they are, they are functioning as a single entity and it is the entity that's trying to get the thing done. Not the individual ants. The individual ant may be like, we have a crack in the ground and we have to get from this side to that side. And they build an ant bridge. You know, we've seen pictures of that and like, you know, that one job, one little ant may have the job of I'm holding on to the ant on this side and the ant below me is holding on there. And then they have that going on as well. And they were all creating this ant bridge over a chasm that none of us individually could span. But by working together for that swarm approach, which is make that accessible, they are doing something well beyond what any of the individuals can do. It's. They are super anchored in that way.

D (43:20)

Yeah.

B (43:21)

And that's what I'm getting at is that that ability to, to give up your individual identity as a member of a swarm for the purpose of the overall swarm's intent. And that swarm itself has an intent. That is a swarm level thing. Kind of to your point, Jared, is, is like, you know, it's not, that's not the thing on any one brain. But when you put all those brains together or technology that represents that there is a, there's a thing that the overall thing is, is trying to do as a single entity.

D (43:53)

It's powers in number. It's like I saw my kids love ants, animals, you know, all the things essentially, venomous plants that kill things, you know, that stuff entertains them dramatically. Venus fly traps, things like that. And we watched this show. It's kind of a documentary, but it's also kind of dramatic. And John Cusack was the narrator. And it's a movie called the besieged Fortress from 2006. And there's an ant type that I want to mention to you. It was actually, I ruined the plot a little bit, but it was ants versus termites, essentially. And it was very, very well done. If you've seen it, Chris obviously say so.

B (44:35)

I have not. But I'm going to check it out now.

D (44:37)

It's, it's 100 worth is phenomenal. It's probably going to visualize for our entire, entire audience the things you're talking about because the particular ant, I guess you would call it the name of the ant, I suppose is how you describe it, were driver ants. And these driver ants are, are so swarm, like, you know, they don't think little. They, they, they create rafts for themselves. The entire colony can, can float. I mean, you can put them underwater. They won't die. They will like create this Bubble. They are just basically resilient to the nth degree. And if you're in their path, you're dead. Like no matter what you are, snake.

B (45:16)

A rat, a bug, they're gonna overwhelm you.

D (45:19)

Oh yeah, they, they drive in numbers. They're called driver ants and they are truly, truly incredible. And this whole entire dramatic documentary narrated by John Cusack is phenomenal. The Beseech Fortress. I would highly recommend it. 2006. Amazing. But these drive rants probably elicit a lot of the, a lot of the qualities and characteristics that you're mentioning because they act like if you're in their path, it's not as if they're one, it's they're many and they act together and it's wild.

B (45:49)

Yeah. I mean it brings a whole capability like whether you're talking the ant or whether we're humans with our technology doing this, you're basically inventing a whole new category of what's possible by, by, by, by introducing this. And like, you know, while, because I, because the conflict of interest and I stay away from my employer, Lockheed Martin, and, and generally in delicate, on defense and intelligence stuff anyway, when we're talking in public. But the notion of if you were to look on the military side for just a second at a high level, there's the notion of, of mass, you know, and if you go back and warfi, I mean people would say, okay, let's build up mass to win against an enemy. And then as things progress forward, we learned that maneuver could, could kind of out. You know, you could go around mass and you could hit it from different ways. And so maneuver as a capability started trumping what was possible with mass. But swarming becomes like a whole new thing, is that you're, you're kind of getting the best of mass at individual small scale, but you're getting mass and you're getting hyper maneuverability and so it's able to trump that, you know, so in that domain, in that kind of military world, it brings about a whole new capability that it never existed before and you're in similarly when you move into commercial and industrial and you know, we talked about this, this, this super automated house a few minutes ago, you're bringing about things that just were not possible before. You could have little pieces of it that were possible discreetly from a, from a source. But the notion of this integrated solution that would just kind of go attack a real world problem and overcome it, you know, kind of going back to your driver ants is, is a new capability that, that that the world will enjoy going forward across all different types of domains. And so I think that's, I mean that's the magic of swarming right there is. It's not, it's different from a fleet. I think a lot of the times where people throw up a whole bunch of things like drones, that's the thing everyone knows will throw up a whole bunch of drones in the air. But it's not really a swarm, it's a fleet of drones. That's what it is. And each one requires individual programming to go do this or do that. There may be some communication between them potentially depending on what they're doing. But they're not thinking almost like a brain, like an abstract brain themselves. They're not looking and dynamically handling what's happening in the real world in real time and saying this is changing right here, right now as a swarm, I'm going to go do that. They can't do that. They're fleets, they can respond. But it's going to take inputs, it's going to take some collaboration between them, but it's going to take a lot of guidance, you know, from afar to make that happen. And that that's the difference in mass numbers in a fleet versus what a true swarm would be is that, that, that capability and that intent and, and that emergent behavior is, is really key to identifying a swarm. And you do see that in mother nature.

C (49:01)

So let's take a recent phenomenon which is the drone light shows, you know where they go out and let's say they're making a Dragon.

B (49:08)

This is like not a swarm, but. Yes.

C (49:10)

Well, I was going to ask. Depends on the intense, right?

B (49:13)

See how I did that? Just say I did that. Not as warm.

C (49:16)

Well, I was going to ask. It depends on how it's implemented, isn't it? Couldn't you swarm to accomplish a Dragon?

B (49:20)

Absolutely could, but nobody has. So what like these days what they're doing is, you know, you may see these, these light shows that you know where they have thousands of drones in the air, but each one of those is following a pre programmed path. There might be some limited communication when they're very close in case of their winds and things like that, that in terms of anti collision. But, but the, but what I would say is like if you were to do the big dragon that you talked about as a swarm, the swarm would figure out how to do it in real time. It's actually using that, that, that decisioning entity that we talked about in the definition and saying my mission is to Produce a dragon over this area for people to watch and it would go do that. Like, it would go figure out where all the pieces need to be for that dragon to come about. That's true. Swarm, like, because if you think about animals that are getting out and doing something, they're not producing dragons, but they're going out and doing something in a swarm. They're not. There's no external thing saying, you know, swarm of bees, I'm telling you to go do this and you need to make an adjustment there and all. They figure it out in real time in the swarm and make whatever it is that those species are trying to achieve, it happens. It's emergent behavior that's real and in real time that supersedes the individuals. And that's what I'm saying. The light shows fleets of drones that are being, that are being provided instructions, often, you know, essentially a three dimensional vectoring trajectory on what me as an individual drone would do, regardless of what all these others are doing.

C (51:01)

Okay, so even inside of emergent behavior, let's say in an ant colony you have roles, you have leadership. There's some sort of, like, there's some sort of mission that comes from somewhere.

B (51:15)

There is.

C (51:16)

And I assume now we're getting, we're getting to the part where it's like, okay, how do you make these things? Because as a guy who's makes fancy websites his entire life for a living, like, this sounds really hard. I just like, if I had a new, hey, Jared, your new job is like build a swarming technology of autonomous whatever is. I'd be like, nope, not going to even try that. Because that just sounds very, very, very difficult. Where do you start? Like, how do you, how do you, how do you do it?

B (51:43)

That's a great question. And not only that, but you've identified the thing that you just said in your, in your vulnerable moment there in terms of like, I don't even know if. Nowhere to go. That's what almost everybody. That is why it is a problem yet to be solved. And there are many, there are many groups, companies, individuals out there working on it, including me. And, and, and you know that this is my passion. And all of us at some point start, Some of us might have had the benefit of coming from robotics, but just like many other skills that also carries some baggage with it that you have to unlearn to do it. And that's, that's one of the pros. So when I talk to people like I've been doing drones for 20 years, I know Everything. And I'm like, well, I'm like, that's good in some ways, but not a swarm. Yeah, not a swarm. And not only that, but sometimes it's the, it's.

D (52:35)

Let me get you a T shirt says not a swarm.

B (52:36)

That is not a swarm.

D (52:37)

Yeah, ask me anything. Not a swarm.

B (52:39)

That fresh. That fresh learner's mind though often does it. And so it's a complex problem and you have to break it down into its constituent parts and there's a whole bunch of layers because there's like, there's things that have to happen at the member, like if you talk about the individual ant, you know, at the member level, there's a whole bunch of things it's got to navigate. And that's kind of like where we are on like drones today. You know, in the sense of like if you go buy one, you know, we're going to go out, you go out to the toy store and you buy a drone today or order one online these days because toy stores are not, not so common anymore. So we ordered the drone online and like that has basic navigation and there's a whole bunch of tasks associated with that. And that's where most of the robotics world has been obviously over the years. But as you move into communication between them and what kind of tasks happen, you kind of move up to a level. There's a local, there's a local drone level in a larger swarm and then there is the how do all those locals operate together? So you're, you kind of steadily move up in abstraction until you have that, that notion of this emergent thing which is really, it's really quite a challenge of like, because there's not a master member, there's not the boss. Some, some you may have a queen.

C (54:01)

Wouldn't it be easier if there was a boss though?

B (54:03)

It would. So it depends on what you're trying to do. I would say that's like the step below if you're doing almost everything a swarm can do but you still have some centralized control. There's a couple of levels below that. And I, and while I swarm, I can't share it. Today I'm gonna try. I invented, I, I, I created a document that allow it doc that, that helps people at my company evaluate these technologies at different levels. It's called a maturity, a maturity model towards swarming. And, and they can look at anyone else. Somebody has put something out there and we can evaluate it based on that criteria about what exactly it does. And I need to see if they'll Let me release it publicly because I think it would be useful.

D (54:44)

Let me see if I could maybe break down an idea and I don't have your depth, but if I were thinking about this problem and obviously when we compare ants, so in the case of the driver ants, just because that's my example that I have some clarity on, at least they do have a queen and the job is to protect the queen. It's like if the queen disappears, they will elect or attempt to elect a new queen, but there's always somebody in charge essentially. But if that's not, if that's not a swarm, then the way I might try to create a boss would be through consensus. Because if you're a controlling entity that's connected and so you know, all your parties in this connected mesh network or whatever you want to call this, then you know, player B versus Z over here has new information the swarm needs to know to consensusly, that's even a word, to have consensus on the next decision. And so we may as a swarm elect a new, not so much boss, but a primary information source that changes the way the swarm acts as an entity. And so it's sort of self evolutionary.

B (55:55)

You're hired because you're on the right track. That's it.

C (55:58)

So aren't they just making their own boss then?

B (56:00)

Basically, so, so that's the thing like, so the queen, like in the case of the queen, the. Yes, there's a queen who is the, you know, the general, the one in charge, but at the same time she's actually not making all the decisions. You know, a lot of it is instinct, you know, that is being played out.

D (56:19)

It's preservation at that case.

C (56:20)

Right.

D (56:21)

The queen is not the boss in terms of leadership and knowledge because the drones have the knowledge.

C (56:26)

Right?

D (56:26)

The drone ants out there doing the work.

B (56:27)

That's right.

D (56:28)

She is the, the preservation system for the entity. It's a necessary component of many.

B (56:33)

So she's not a master, like a master direction giver, you know, that's not. Her role is, as you said, perpetuation of the colony versus she's not driving the specific actions of the, of the drones. Those are built in. You know, Mother Nature has imbued the members with that and they understand how to do that. But, but to your point, Adam, that notion of kind of consensus, you know, there's different approaches to it, we can use some different words because they're different algorithmic approaches of consensus election, things like that, in terms of saying, well, we have, we have a distributed compute grid that is Our swarm that, you know, that is imbued in our swarm. And, and how do we arrive at single overarching directives that perpetuate themselves downward through the swarm and which change as they go down? Because this is the overall, this is our, what we need to do. There's a mission, there is a high level sense of abstraction about. Well, to accomplish the mission you must do A, B and C. But A has 10 steps to it, you know, and some of the SOAR members are going to be, are going to take the assignment of doing those and others are going to say, well, I'm going to go off and do these other things that are, that are part of that, that might have been part of the B category. And so they have to self organize in the way to do that in real time. Because this is a physical technology. So it's, it's one of those. And there are sensors coming in. Things are changing, you know, constantly without, without. And so you're, you're acknowledged, you are with your sensors. Whether you're a biological being or whether you're technology, you're having to take all that new information in. You're having to do distributed computing and decisioning through algorithmic approaches and select members to accomplish all the things as part of that overall mission that you're doing. And it's quite complicated. I mean it's a very complicated thing. As we sit here in 2025, I think we'll nail it gradually, I think we'll nail it in iterations and I think it won't, I think somebody a century from now will be like, yeah, well of course we did that, you know, but today it's, it's a tough problem to solve.

C (58:54)

So at what level do the humans interact? So let's imagine that you've created a swarm of vehicles and it's, it's, it's a legit swarm. It's not a, not a swarm. Swarm.

B (59:05)

A legit swarm.

C (59:06)

Yeah, I was thinking about that as you said that, you remember the old Jeff Foxworthy thing, like you might be a redneck. We could do a whole line of like, might not be a swarm. Like if you got a boss, you might not be a swarm. You know, if you've got a path they gave you to fly, you might not be a swarm, but let's say you have one. And this is like, you know, Chris approves and it's a bunch of drones. Let's just do that. At what level does the drones receive their mission from the humans? Like, is it very generic? Or is it very specific?

B (59:35)

It can be either. It depends on what you're, what you're building toward. And swarms have different purposes. So remember, swarm is not a generic thing. They're purpose built, you know, for certain capabilities.

C (59:45)

Okay.

B (59:46)

And so you, and you do have that C3, which is command, control and communications. That's inherent to that. And one of the other phrases I use, which people outside of the military context may not be as familiar with, is human on the loop, not in.

C (60:01)

The loop, but on the loop.

B (60:02)

Not in the loop, but on the loop are two different things. So. And in the loop is where you're. A human is controlling a technology directly or. And they are, they're making it so like a human in the loop may say, make a choice for a task. So they may say, yes, I'm going to now have you drop that package on that person's front door. And you know, yeah, it's clear, we've looked at it, it's safe, there's nobody in the way. And we're going to have you put the package on the front door because it's safe. And we did not want the drone to do that until me as a human verified that that was okay for us to do so that we didn't hit people or hit things on the loop. You are essentially tasking that. It's kind of a, the human has a supervisory role and maybe a mission giving role, like your mission swarm is to deliver the package to that or maybe more. It might be, here's a bunch of packages and to the swarm and I want you to go to this neighborhood and deliver all these packages to the right houses. And that is the mission. And then the swarm understands that geographic layout. It understands the, the real world environment it's in, and it figures out which member. They each pick up a package and it figures out how are they going to do that. Some of the packages are more than the eight pounds that Adam talked about. Some of them are 60 pounds. And it takes multiple swarm members to get that package airborne and to collaborate. And so as they go, as they go into that environment and they're looking, I got to get this package to that address. And oh, by the way, that address might have been reachable by a four pound package on one swarm member acting alone. But we, it's now 60 pounds. We have multiple sworn members. And even with all those swarm members, it's outside of our range. So how do we address getting it outside the range given the fact that we have other concerns that may be limiting that the swarm would work that out through its distributed computing and, and collaboration that we just talked about that, you know, where it, where it kind of comes to that consensus on how it's going to collectively solve the problem. Does that make sense?

C (62:12)

Yes, I think that it does. I'm wondering if maybe I'm sniffing danger eventually because.

B (62:20)

Oh, go for it.

C (62:21)

Well, because at a certain point you give a directive and maybe that directive is completely benign. Like you, you have a swarm of cleaning bots, you know, in your house and you say, okay, bots, you know, clean the bathroom. And that's as far as you get into it. You're on the loop, but you're not in the loop. And so they go about doing that and we've accomplished Chris Benson level swarming. So, okay, I now have a, I now have numerous independent, fully autonomous, multi agentic platforms in my bathroom exhibiting highly coordinated locomotive and emergent behaviors with agency and self governance. Right. So at a certain point, couldn't they just say, this toilet's really dirty. What if we just removed it? Wouldn't that be. The bathroom would be even cleaner. And then they all decide that, yes, that's a great idea. I come back, I don't have a toilet.

B (63:15)

That, that is. Adam had a great moment a moment ago and you just had a great moment there.

A (63:20)

Jared.

B (63:21)

That was.

C (63:22)

Most of my great moments involve toilets.

B (63:27)

You're cleaning up your act, man. But yeah, so like that, that's a great thing. And that comes down to you're not giving. What you're really telling about in swarms is when you get down to the task level, then you're talking maybe not about the whole swarm making a decision. It might be a few of them that are, that are addressing a task and figuring out at a more logistical level, like how am I going to, you know, operational level, how am I going to do this? And, and, and that is one of the things is that when you're doing, you know, we're back to AI safety and AI training on this.

C (64:01)

Yeah.

B (64:01)

Is that, is that maybe removing the toilet in most cases is not an acceptable thing. So, so we need, we need some technology based guardrails there. But that's also where, depending on the circumstance that you're looking at, that human on the loop needs to be able to go, no. Yeah, there's kind of a kill button, if you will. You know, figuratively speaking, meaning killing of the swarm, not killing a person. Just to be very clear, so that nobody misunderstands me.

D (64:29)

Super clear. No killing here.

B (64:31)

No, no that's why we're not talking about killing people here.

C (64:33)

That's why I use a bathroom and a toilet as my example.

B (64:36)

In this context it might be don't kill the toilet.

C (64:38)

Right.

B (64:39)

You know, kind of thing. And that's where the human on the loop as, as an oversight where we still have these amazing capable human brains that have, they, they can't do everything that digital technology can do. But you know, digital still hasn't yet arrived. It will, but it hasn't yet arrived at what our capabilities are. And we can look at and go, taking the toilet out is not acceptable to the homeowner. We're not doing that. Maybe if you're Jeff Bezos, maybe instead of cleaning the toilet, you just remove it every time. Maybe Jeff Bezos will have the toilet removal every day. The, the drone swarm goes into Jeff Bezos bathroom and it just takes the toilet out and it puts the new in the back.

D (65:20)

There's just a sea of toilets back there. It's like a pile of them.

B (65:24)

There you go. Well, it cleans it all up. Because you know what, when you can see throw, when you can throw $8 billion at something you haven't really identified yet, you can probably afford to have your toilet.

D (65:34)

6.2.

B (65:35)

Oh, 6.2. That made a good, that made such a big difference in my know.

D (65:40)

Sorry about that. 8 billion is close though. Let's just round up to 8.

B (65:43)

Let's just round down to okay, we're, we're there. But yes. So other than Jeff, I don't want the drone swarm taking my toilet away. That would get rather.

C (65:51)

I'm with you.

B (65:52)

Yeah, it's a, it's a little bit too.

D (66:12)

Well friends. When you're building and shipping AI products at scale, there's one constant complexity.

C (66:17)

Yes.

D (66:18)

You're wrangling models, data pipelines, deployment, infrastructure, and then someone says, let's turn this into a business. Cue the chaos. That's where Shopify steps in. Whether you're spinning up a storefront for your AI powered app or launching a brand around the tools you built, Shopify is the commerce platform trusted by millions of businesses and 10% of all US E commerce. From names like Mattel Gymshark to founders just like you with literally hundreds of ready to use templates, powerful built in marketing tools and AI that writes product descriptions for you, headlines, even polishes your product photography. Shopify doesn't just get you selling, it makes you look good doing it. And we love it. We use it here at Changelog. Check us out. Merch.changelog.com. that's our storefront. And it handles the heavy lifting, too. Payments, inventory, returns, shipping, even global logistics. It's like having an ops team built into your stack to help you sell. So if you're ready to sell, you are ready for Shopify. Sign up now for your $1 per month trial and start selling today at shopify.com PracticalAI Again, that is shopify.com PracticalAI. Can we get into this scenario where it's like Aladdin and the genie and he's like, hey, you know, I want make me a prince, right? I think was the first. No, the first one was to. He tricked the genie to get him out of the cave. We'll skip that one. And then the second one, technically, the second wish was make me a prince. And he didn't really make him a prince.

B (68:01)

Great memory. Gosh. Like, I've seen the movie, but, like, you're really bringing it back to me.

D (68:05)

I'm. Well, I got a good brain, you know, over here. My brain.

C (68:10)

His. That's his second great moment right here.

D (68:11)

That's right.

B (68:12)

Two senior moments here with me. So. Okay, keep going.

D (68:16)

He doesn't really make him a prince. He just clothes him as a prince. He mimics a prince. He doesn't really give him royalty. He doesn't really give him lineage. And I guess I'm sort of sidetracking to some degree just to be accurate about my Aladdin reference, but point is, is there's times when he asked, or I guess in. In all of the lore around the Aladdin figure and a genie figure where you ask the genie for something, but you have to be careful. That's where this term comes from. Be careful what you wish for, because you wish for something without the full awareness of the agency behind the genes, behind the swarm. And so you might get your toilet removed. Is that a concern?

B (68:52)

Like, how Totally.

D (68:53)

How are you guarding against that? How do you guard against that without the human loop or the kill switch? Is there an os? Like, I don't know. How do you. How do you guard against this genie issue?

B (69:03)

No, I think there's a lot, and I think it's at many different levels. And it's a real thing that we talk about in real life today without having achieved full Chris Benson level drone swarming. In that. We talk about that in terms of AI safety all the time now. That's a huge part of the AI world, is, what is AI safety? How do you keep unintended consequences from coming to pass? I think anyone who's reasonable Recognize that some of those will still come to pass out there. You can put guardrails around things, but, and you can even ask AI to put guardrails around other AIs, as we're doing, you know, because we're using the tool to build the tool. But we will have bad outcomes across the board, just as we always have with software and always will. So I, I don't have the magic bullet on that, but there, there is, there is training the distributed swarm brain, you know, this abstraction of computing, of grid computing, where they're all doing this and using their algorithms. You'll, and that will, where it goes wrong may happen different places. You know, we often talk about today's LLMs coming out with, with inferences that are, that are suboptimal. Sometimes quite funny, sometimes quite tragic, actually, but that will continue to happen. We have, we have software issues. We're also moving into the physical world where, you know, you know, if you have these physical agents that are imbued with a whole bunch of AI agents that are doing stuff and they're acting as a member of a larger swarm, there's a lot of places where things can go wrong. So it's going to be, it's, there's going to be a learning curve on that and we're going to have, we're going to have problems along the way. So I don't want to, I certainly wouldn't want. I know for a lot of listeners and viewers, they probably think of a little, you know, a little bit pie in the sky. Not everyone's going to believe that this is probably sooner than they would otherwise expect. But we'll get through it and stuff, and we'll, we'll try it, we'll do the best we can, and the, the responsible people will put a lot of safety around it in, in the best they can, but we will make mistakes.

C (71:13)

Where do we stand? Where are we in this initiative to create this thing or these things?

B (71:20)

So I think like many things, you'll see it coming from specialists. You know, there's, there's a whole area of expertise, you know, that, that you develop around trying to solve these problems. And some companies are specializing that. And just like other things, you'll see that. But I think over time, especially given the fact that it's not one industry, it's many industries, there'll be many players. I think one of the things to make this happen isn't just can we get there? Because if you think about it, once you can get there, almost everybody kind of does A close, close copies that, you know, once we had our first, you know, chat GPT, it wasn't long before that we had competitors and other models that were, that were nipping at its heels and I think you'll see that here as well. But it'll draw. I think it really comes down to getting organizations and, and motivated individuals into it so that they are producing some level of whatever is productive in what they're doing in their industry and they're in their world what's productive. And costs will drive down. And I think as those costs drive down, that's where you see it really pushing out into lots of different places in life. So a lot of it isn't just a technology question, it's an economics question. But I think the pervasiveness of it will drive that.

D (72:37)

Let's get. Since we helped create a show called Practical AI, let's get Practical. Yes, you'd mentioned this is obviously burgeoning. You're coining this. I kind of feel like swarming is the protocol. Maybe there's a specification there somewhere and the implementation is more of a product potentially. But take us into the practical nature of let's just say over the next three years. Well, we see swarming of any sorts in a consumer level home lab, put it my home level. And if we do like be realistic, practical, if you can, like, what will it be?

B (73:10)

At the level of the definition that I provided, which is a very high bar, I think you'll see lots of things that are calling themselves swarming things developing within that two to three year horizon, I, I don't think many of them will rise to that level. They'll be, there'll be kind of quasi swarming capabilities that you're starting to see in consumer and commercial products and stuff. I do think, however, there are so many really smart minds around the world working on swarming because by opening up an entire new category of capabilities that don't exist today that people already have productive use cases in mind for, there's a lot of money to be made there. So you have not only commercial entities and motivated makers, but you have nation states that are highly motivated to do that. And it's a, it's a big scientific topic of research. I think you'll see it probably first in areas where people can throw lots of money at it. And so you know, if we do talk about in the commercial space, our 800 pound gorillas, you're more likely to see it in a narrower case of use cases there. I think in the, in the Military space, an intelligence space. You're likely to see it there because you have the, you know, the economies of nation states that are, that don't want to be left behind, you know, that it's, it's. If we don't, if we're not able to produce a swarm first or are very closely following whoever is first, then we have a national security issue here in terms of what's possible. And so I think you'll see nation states prioritizing that probably in very close collaboration with commercial entities, which is really common today. I mean, if you look at certainly how both the US government and most of our allies, as well as, you know, the Chinese government, you know, there's a lot of, of overlap between nation state resources and commercial entities that have special knowledge and skills working together to produce that stuff. So I think that's, I think those types of collaborations are likely to be the first ones largely because they can throw resources at the problem until you get there. I think the key is it's thinking about the problem the right way. And I think that's where, where people struggle is breaking down that complexity that we were talking about earlier, you know, that Jared pointed out and saying how can we discreetly address those points of complexity in a way that you can then pull those, those, those many solutions together to achieve the grandiosity of the definition that I provided.

D (75:48)

Let me see if I can not predict, but this is where I would.

B (75:53)

Because I'm not going to predict anything. We know.

D (75:58)

And I think the two of you, I think, will agree with what I'm going to say here. I think the area where I'd like to see this type of swarming is in energy conservation. And so I think there's multiple devices in my house that consumes energy from a H vac system above me that both heats and cools my home, to the lights that power my house, to let's say a kettle that is electrified. All the things I want to give my home the task of being energy conservative.

B (76:28)

Right.

D (76:29)

This, this swarm. I want to have a swarm of devices that help me be that and it can it. Hey, Adam and the family are not here. It makes sense as an agency to be, to be conservative with our energy use because there's no one here to do it. And that's where the, you can do like individual device level smart home automation, which is here today. Yeah, but it's not matter.

B (76:52)

Matter supports that.

D (76:54)

It's not a swarm though, right? It's not a swarm.

B (76:56)

That's right.

D (76:56)

So I would like energy conservation to be my first swarm tactic. The next would be, I live in Dripping Springs, Texas, just outside of Austin, Texas. And we always have water challenges right now. We're always in some version of a drought. There's actually a big bet on the wall on Wall street against Texas running out of water. Like there's a bet essentially shorting Texas running out of water. At some point I just heard this headline. That's a headline, only I don't even know what the truth is behind that, but I heard it, so it must be true. Okay, so the next thing is water conservation. Help me as a household, maybe even help me as a neighborhood, a swarm neighborhood, be conservative when it comes to water conservation. So my child goes to flush the toilet or, I don't know, some sort of action tries to take place, but the swarm is like, like, hang on a second. We're in a conservative nature. We're going to use the 1 or the 0.5 gallon version flush versus the 1.2 because it's a, you know, it's number two, you know, some reason.

B (77:57)

Right.

D (77:58)

But for whatever reason, like we now have new tech in my household that gives me things that really matter. Energy, water. And I think the last one for me is food. There is so much food waste in America. Tremendous amount. I know I for sure buy some chicken once, twice a month and I'm killing chickens constantly because I'm wasting my chicken, not making it. So I don't know if that's a problem. That's me. But at some point, my tech, my swarm tech can help me solve those three key things. Energy, water, and food. And I think you start there because that's what matters. My laundry, kind of a me problem. Maybe my washer can say, hey, you put a white in with a darks, probably not smart, eject it or alert me, you know what I'm saying? But like I don't need help with laundry. I don't need. I mean, I like my iRobot and vacuuming. That's cool. But I think that the thing I would want to conserve on is those three things that'd be helpful.

B (78:48)

I. And I think, I think you'll see that. I don't think it'll just be the swarm doing that though, because like, even today, you know, if you start with where we're at right now and talk about the fact that energy monitoring is really common within a lot of these existing.

D (79:01)

Not a swarm though, not a swarm.

B (79:03)

Not swarm yet, but we're getting there.

D (79:05)

We're getting.

C (79:06)

Keep saying It.

B (79:06)

Sorry, yeah, bear with me for a second. So we have what we can already do at the individual device level. And then as we really start imbuing our homes with AI agents, which is going to happen even before the swarms are hitting so soon.

D (79:21)

Yeah, that's next.

B (79:22)

You're going to have AI agents doing lots of different things, including the monitoring. And those AI agents will be monitoring your matter driven devices and thinking, oh, we need to make some adjustments. They'll communicating with the devices that, that are being governed by that. And so they're able to get you a great deal of the way down that use case that you just talked about. But there are also going to be things in your home, you know, where things like for energy, the energy conservation thing you mentioned, things like, you know, airflow and temperature, where it's not an explicit device that's matter enabled and has the, you know, energy monitoring built in. But it may be like that corner of the room is cold and in that case that swarm, it's monitoring the house. And maybe it has other functions that aren't just monitoring. Maybe it's doing a cleanup, you know, it's doing the cleaning job. But it also notes that, hey, this corner is not getting good airflow. It's the temperature's changing to your vision, Adam, that you just talked about. That's where the swarming capabilities of having different devices work together will do it. Now an individual robot could also detect that device. It doesn't have to be a swarm. So you really get it for a swarm to be effective there you're really going to be looking for how does a cluster of members working dynamically together get me something I don't already have. And I think that's the question to answer in that use case if you're actually wanting to introduce the swarm to it.

C (80:52)

Well, we humans have our own form of swarming. It's called open source software. And I'm curious if there's a place where people who are as passionate or maybe even just potentially interested in this initiative, this movement, this, I don't know, this next big thing of swarming tech. Is there a place they can gather? Is there like a, is there a framework, is there a conversation, is there anything in the world of open that people could gather around?

B (81:18)

There are and I probably should have brought a list. Maybe in the show notes we can add some stuff in. Some of the things that I often tell people to start off on is, you know, robotics has been a big part of this kind of robotics role, you know, being Part of developing to the swarm is ROS2 exists. ROS stands for ROS2. ROS is the robotic operating system which is open source and it is the most widely used robotic software technology out there. It's not the only one. There are many and some of them are closed and some of them are open. But there's tons of books now on ros and so I often when people are interested in this and they're like, but how do you do like, like aside from the swarm, I can't even make a single robot work. Like what do I do? Well, there's tons of information about that. Start off maybe not solving the overall swarming problem that we, that we were describing as being as remaining a hard challenge, but start with something more accessible you can buy, you can get on to. You know, you know, we mentioned Bezos so much Amazon and others and there are a lot of maker kits that you can get that are open maker kits. You have ros, they're very similar in terms of. But if you want to not do robots and you want to do drones, there's a whole bunch of open source drone stuff. And then the thing that I love doing, I do this all the time is diving in on GitHub at different software communities that support, you know, open specs and stuff. There's tons of repositories on GitHub that are designed to do this that just interested people said I want to go scratch an itch, I want to solve a problem and I go there and I'll then also go to hugging face and look for small models that may, if I need AI in the mix that that can contribute because really, you know, small models are where the future is. You know, it's not. We talked at the very beginning of the conversation about the giant versus the small. Go for small stuff. You have, you, there's very likely that you have a GPU and at home it may be in your laptop or something that you can buy for a couple hundred bucks that, that can do all sorts of cool inferencing with an existing model that you can then go do some of this stuff with. So, but with open source that's the place to go. That's where I think, that's where I think the majority of innovation is really driving from. And it's a good place to start and figure out what is interesting to you and, and even that area I'm, I'm really into, I, I'm gonna, I'm gonna also pitch a language that I'm into which Is Rust. I mentioned go at the beginning of the show, love go. And I use that in a lot of environments. But I've been using Rust as a replacement for C C because. And it's great for embedded. You can use it with no operating system at all and it's fast as can be. And so I've been. That's been like when I go play on my own, aside from like work work stuff in this area, I'm always. I mean I'm every day I'm looking at all the innovation in the Rust community to do small little projects that I can do for fun that drives my own passion forward. So it doesn't have to be a giant 800 pound gorilla or defense industry or whatever kind of thing it can be. It can be something that you're that the kid in you or maybe the kid in your house can go do on their own.

D (84:42)

Give some shouts to I guess some crates or some projects out there in the rest world. I think probably Tokyo or Tokyo probably is one of them. Saturday is probably one of them.

B (84:50)

Yep.

D (84:50)

What else you you playing with?

B (84:52)

Tokyo is really good because it, it allows you to. To you know, kind of that the multi threaded things, many things happening at once, which is really important in robotics. And so that's really taken off there is. I'm trying to remember the Embassy is the name of it. I was trying to remember for embedded. It is a runtime in Rust that allows you to do a whole bunch of embedded capabilities without writing everything from scratch. It kind of gives you this framework and so you can go get a Raspberry PI, even one of the small ones I think in the Nano and stuff that doesn't have that isn't there supporting the OS and use Embassy to create an executable that runs on something that's too small for an os. And so I like exploring all these different possibilities in terms of how you're gonna. And when I said Tokyo being multi threaded, it wasn't multi threaded, it was concurrency. I said the wrong thing. So I just want to correct that before we got too far. But being able to do highly performant concurrent things on very small pieces of hardware out on the edge is a real thing. Like 5 years ago it just wasn't possible to do anything like what we're doing now. But we're. You know this. At the beginning of the show we talked about the revolution of all these different areas coming. Now anybody can go use several different languages, but in my case Rust and find small bits that cost me 10 bucks out there and put some unique software and do something that scratches my itch that no one in the world has done. And it's no longer out in the cloud or out on some computer. It can be something that I'm carrying around on my body or is literally a robot. This is all reachable now. And so that would, that's really what I would encourage people to do is the future with people. I get asked all the time about the future of AI and I really think the next big revolution in AI is going to be physical. AI is AI imbued in all these things in our life that we've been talking about that we refer to as on the edge in the software world. But that's going to be the new normal. And now you can do that without any real budget, on your own, anytime, from any place in the world. So this, if you want to go create the future, and I said, this is the coolest time we've ever lived in, well, you can go create that right now, no matter where you're living and no matter what your budget is. So that's what go do it.

D (87:23)

If you're tinkering with Russ right now. So let's say you're done with this podcast, you're off for the day. Let's just say magically you have nothing to do. You're going to go pick up your next or your current Rust project. Maybe you've got a new model you want to play with. Where, where are the places you're going? You mentioned hugging face. What are some of the stack that you're. That you're tapping into?

B (87:43)

So there's, there's the swarming stuff that we've talked about and trying to, and trying to figure out robotics and all that. And we've talked about home automation and I think that that feels for answering this question. That's an accessible thing that I like to do now. So as I've picked up this kind of home automation stuff, I'm trying to figure out, like, what can I do? I go get some Raspberry PI's or I can use a slightly larger, you know, like a mini PC to do something in the house. None of this costs much and I'm now on my day to day when I'm just at home and I'm not thinking about the day job, if you will. I'm looking at all the things that I do with my family and thinking, wow, you know, like, I can go pick something to handle that. So, like, almost all the lights in our house are Automated. A lot of the appliances are automated. We have voice command from anywhere in the house where we can tell a particular assistant, go do this. And it happens. I've been starting to integrate AI agents into that workflow now that that is becoming super accessible with all the. There's so much open source that have made agents very easy to do. And you can get small models off hugging face and run it off compute that you have in your house already. And so that's. That's the kind of thing that I like to do. And I think it's amazing because it's gotten people in my family who are like, oh, my God, Chris is doing technology again. You know, like, if the family members, they're like, yeah, yeah, I don't want to hear it because you're talking about that with everyone else all the time. But now they're like, they're using that and they're getting interested in like, yeah.

D (89:18)

They'Re like, tell me more, Chris.

B (89:19)

Yeah, they'll start like, not a swarm, though.

D (89:22)

Tell me more.

B (89:22)

Not a swarm. But my wife will say, you know, how could we make this. This. You know, how can we automate this to make it better? And like, I couldn't get her to.

D (89:30)

To care.

B (89:31)

Yeah, she didn't want to care. Like, that was my thing. And just stop talking about it, Chris. So, yeah, and.

D (89:38)

And.

B (89:38)

And my daughter is starting to get really. She's 13 and she's really starting to think about what could we do? And like, it just sparks the imagination because it's real and it's tangible. And so that's like, that's why I get to, like, go do something. Just decide, today, you're a maker, go get some cheap stuff, have a vision, recognize that every part of it is either free or only a few bucks, and just go do something in your imagination. If you can't think of anything, there's tons of websites with maker projects out there and find something that you go, oh, God, that's cool, and just go do it. And like, even if it. It doesn't have to be the greatest thing in the world, just go do it. And then you're helping push all this stuff forward. You are diving into the future and making this stuff happen. And that's why this is the greatest moment in the history of the world. It really is.

D (90:24)

I mean, we went from photography or from painting photos to photography in a blink of an eye, and now we're thinking, gosh, I just wouldn't, like, paint the picture that way ever again. I would just take the photo because that's the way. It's a cool moment in life. I'm super curious about one, one particular area that you mentioned. You mentioned voice. Are you leveraging Alexa or leveraging the behemoths or are you home assisting in it and you're doing something with home assistance?

B (90:52)

So I am moving. We have been, we have been for a while. Alexa all over the place. And given the fact that I am, I am increasingly concerned about privacy just in, you know, in terms of it like surveillance is so built into everything now that I am generally moving from cloud based systems into more private systems that are completely under my control and local and stuff. And I realize that may not be for everybody. I think part of that is because I work in a world that is obviously touching on intelligence and I'm more aware of what's possible from a surveillance standpoint than probably most people are and how pervasive it is. And that makes me obviously wanting to kind of protect our own privacy a little bit. So I'm, I'm keenly interested in automation that's not specifically commercial cloud dependent.

D (91:50)

We should circle back in the new year for a deeper conversation because I'm sure you'll have some time away, maybe new progress, new projects and new insights. Because these are things I'm about to go into in my curiosity is I haven't automated anything in my house. They're like, Adam, you're such a nerd. You care about home lab. I'm like, yeah, I don't care about that part of the home lab. It's, it's a different area of the home lab that I'm trying to conquer.

B (92:14)

I didn't either. It really took it for me, the kick in the butt was moving into a hat, buying a house that already came with a lot of automation in it.

D (92:23)

Yeah.

B (92:23)

And, and it's, it's not just catching up on that and learning. There was a certain like ramp like I had to level up. But then there was also the. It starts getting your imagination going like you didn't. Like you knew in the back of your mind you could do this. But now you're like, you're living it and then you're thinking about the next five things after that. And I think that's it. Once you do a little bit, you, it wets your appetite and you start seeing all the possibilities. And that's what it took for me, you know, professional technologist. But I wasn't really doing it until a year ago. And now this last year is just.

D (92:54)

Take off Being able to host models locally have that privacy. The fact that Home Assistant is so pervasive and so massive as an open source project that they have a, you can tap into via the API, you know, whatever local, you know, models you have running for inference, they have voice capabilities. There's just so much happening there. Why give that data to, you know, to Amazon? It's not that they're bad, it's just that I have preferences and the preferences don't involve me telling you what I want. And then now I get hit with ads for X, Y and Z as I scroll the Internet.

B (93:29)

People often complain about how creepy it is that you're almost just thinking about something and then it shows up in your Amazon cart kind of thing, you know, or Google or whatever. And like, but you're doing, you're doing that, you're giving them that power over you. And so to some degree, and it's not happened all at once, but I'm taking responsibility for the fact that that's been my choice because it was the easy way to go because they were, they were providing this ecosystem. I didn't have to do much. It just happened. All I had to do was let them was say yes every time they send the updated terms and conditions and, and they would take my data and do whatever they wanted and there they are. And I've kind of gotten to that point where I'm done with that and to some degree and, and turning around.

D (94:14)

Just gave me an idea. Chris. They, you know, somebody should. I don't know if this is actually a good thing or not, but like AI is great at scanning an entire document, like in terms of conditions. There was a documentary, I think, on Netflix about this that if you tried to read all the terms and conditions you would agree to in modern society, you would spend more than your entire life just reading terms and conditions. So to keep up with the updates and, or literally scrolling them to say yes, I accept is not, it's not possible, it's not realistic.

B (94:46)

Yeah.

D (94:46)

Of a request from the people. So we're agreeing to a lot of things just out of the nature that we don't have the time to do it.

B (94:53)

And you're not going to. If you're trying to get something done and now you have to do through terms and conditions to get something done. They do it at that moment because they have you. They know you have to get something done. And what are you gonna do? Go. Well, I had to do that thing. It was really important, but now I can't do it. Because I'm not going to do terms and conditions correct.

D (95:10)

You're bricked now.

B (95:11)

Yeah. So I'm. I'm starting to invent my own.

D (95:14)

Whereas the kids say cooked. You're cooked.

B (95:15)

Yeah, I'm starting to invent my own world where I. I'm not bound in that little prison, if you will.

D (95:23)

Well, that was cool. Thanks for deep diving on the swarm. Not a swarm, rust. All the things. Make sure, if you don't mind, some of the things that you can link us to in the. In the show notes. I'm sure you got lots of links. Just spam us with all your links. We'll put them in the show notes for everybody.

B (95:40)

Fantastic. And thanks for. Thanks for having me in, guys. It's been great catching up with you and fun conversation.

C (95:46)

Tons of fun. Go listen to Practical AI PracticalAI FM if you want more, Chris, that's where you find it.

B (95:51)

Thank you, Jared. I'm so glad you did that because I. We would love for people to join the conversation and we all. It's one big happy family, as people can see here. And that I love changelog and I hope some. Some of the Changelog people who haven't given us a shot will. Will give us a shot and join our conversation.

D (96:08)

So Practical AI practical.

B (96:11)

That's it.

D (96:12)

Go there and be square. As they would say in the 80s or 90s, which is cool now. It is cool now. Yeah, the 80s and 90s are cool again.

C (96:21)

Good stuff, Chris. Bye, friends.

B (96:22)

Bye, Chris.

D (96:23)

Bye, friends.

B (96:23)

Thanks, guys.

A (96:31)

All right, that's our show for this week. If you haven't checked out our website, head to PracticalAI FM and be sure to connect with us on LinkedIn X or BlueSky. You'll see us posting insights related to the latest AI developments and we would love for you to join the conversation. Thanks to our partner, Prediction Guard for providing operational support for the show. Check them out@prictionsguard.com also thanks to Breakmaster's Cylinder for the Beats and to you for listening. That's all for now, but you'll hear from us again next week.