Today’s episode features an interview between Matt Trifiro and Stacey Higginbotham, Founder and Editor of Stacey on IoT. In this interview, Stacey explains the intersection between IoT and edge, how to think about IoT from the bottom-up perspective of the devices and manufacturers, and shares some interesting use cases and innovative companies tackling the most challenging problems in the space.
Today’s episode features an interview between Matt Trifiro and Stacey Higginbotham, Founder and Editor of Stacey on IoT.
Stacey has been covering technology for major publications for almost two decades and now produces a weekly podcast, newsletter, and daily articles covering all aspects of the internet of things under the Stacey on IoT brand.
In this interview, Stacey explains the intersection between IoT and edge, how to think about IoT from the bottom-up perspective of the devices and manufacturers, and shares some interesting use cases and innovative companies tackling the most challenging problems in the space.
Key Quotes
Sponsors
Over the Edge is brought to you by the generous sponsorship of Catchpoint, NetFoundry, Ori Industries, Packet, Seagate, Vapor IO, and Zenlayer.
The featured sponsor of this episode of Over the Edge is Vapor IO, the leader in edge computing. We want to be your solution partner for the New Internet. Learn more at Vapor.io
Links
[00:00:00] Matt: [00:00:00] Hi, this is Matt Trifiro, CMO of Edge infrastructure company, vapor IO, and co-chair of the Linux foundation state of the edge project. Today. I'm here with Stacey Higginbotham. Stacey is as a technology journalist analyst and prolific podcaster. She is the founder and editor of Stacey on IOT, where she writes about the internet of things and is also host of the internet of things podcast, which has almost 300 episodes strong.
[00:00:21] We're going to talk to Stacey about your career in technology, her work as a journalist covering the tech industry and everything, IOT and edge. I Stacey, how are you doing today?
[00:00:29] Stacey: [00:00:29] Doing so well, Matt, thanks for having me on the show.
[00:00:33] Matt: [00:00:33] Yeah, you bet. Before we dig into the minutia of IOT and edge, I just, how did you get started in technology and technology journalism?
[00:00:41] Stacey: [00:00:41] Oh, it is so random. And basically I wanted to be a genetic engineer when I was in college. And I was like, yeah, biology is awesome. But my college, which was UT Austin at the time that I chose, they were like, you can only do that as a graduate degree. So your [00:01:00] undergrad has to be. What was it? Chemical engineering and they were huge in petroleum and all this.
[00:01:04] And I was like, that sounds horrifying. And they said, okay, we'll pick a major. And I was like, I like to read, so I'll pick magazine journalism, because I think there's a lot of reading in that there wasn't. And I actually hate to write, but I love finding things out and I love learning stuff. So journalism actually fit me really well.
[00:01:24] And. Because I love science. I don't really love math and statistics, but I kind of fell into business journalism and then tech journalism, because it's a nice, it's a nice gig because it's usually nine to five ish.
[00:01:39] Matt: [00:01:39] How do you reconcile being a professional writer and. I mean, maybe not so much now because you're running a company, but certainly back when you worked at fortune and gig, with deadlines, doing something that you don't really love to do, I mean, in terms of writing and I get all the other part
[00:01:54] Stacey: [00:01:54] when you're writing about technology writing is really only like 25% of your job.
[00:01:59] The [00:02:00] rest is trying to understand how the stuff works, how it fits into everything else, and then thinking about it and talking to people in like 50% of your job as a journalist, it's really learning.
[00:02:09] Matt: [00:02:09] Yeah, that's really true.
[00:02:10] Stacey: [00:02:10] I guess, you know, everybody has stuff they hate, right? Nobody likes filing expense reports.
[00:02:15] This is writing the stuff down is just the payment for getting to do the cool stuff at my job.
[00:02:20] Matt: [00:02:20] Yeah, that's actually, I mean, I think of myself as a writer, it's certainly how I began my career as a technical writer, but I don't really like writing either. I like the fruits of writing and I certainly like teaching which writing is a good form of teaching.
[00:02:33] And I love the understanding and synthesizing. I think that's why journalists makes such good marketers when I've been able to learn them over to the dark side, because it's that ability to make something understandable that good journalists tend to have
[00:02:46] Stacey: [00:02:46] has still working on that man. I keep thinking if I just find that magic editor, they're going to like, get everyone to care as much about whatever minutia I'm excited about today, you know, I can make everyone see the light.
[00:02:59] It's so [00:03:00] important.
[00:03:01] Matt: [00:03:01] Yeah. Well, and so you've been on the internet of things beat quite a long time. I mean, I did some quick Googling and it looks like the term was invented, in the late nineties,
[00:03:11] Stacey: [00:03:11] Kevin, Ashton.
[00:03:13]Matt: [00:03:13] yeah. To describe RFID tags, which isn't really how we think of it today,
[00:03:18] Stacey: [00:03:18] which kind of is right.
[00:03:20] His original thought is, was that he was at Revlon max factor, some makeup company. Yeah. I don't remember who owned the makeup company at that time, but it was
[00:03:31] Matt: [00:03:31] doctrine gamble. Is it okay? I don't remember.
[00:03:32] Stacey: [00:03:32] Yeah, there you go. And Kevin came on the show at some point in time and he told me the story. So, but yeah, he was like, yeah.
[00:03:40] He would go to the store and he would see that he was in the product he was in charge of, was out of stock and he had no idea. Right. And he was like, if I could just figure out a way to have it, tell me when it was out of stock. This was the beginning of the internet. It was RFID only because that was what was most logical at the time.
[00:03:58] But yeah, I mean, the idea here is [00:04:00] that. We have with computing, right? We have cheap computing. Now we have cheap sensors. We can start pulling off all this information and synthesizing it at a very cost effective and efficient way. And we can know I'm going to say everything about anything, but most things about everything.
[00:04:19] And that's, it's beautiful.
[00:04:21] Matt: [00:04:21] Yeah, well, I mean, we have the internet and then we have things, and we connect the two, I mean, w why isn't the internet of things? Just the internet.
[00:04:29] Stacey: [00:04:29] Okay. So this is a thing, this is a whole thing that people talk about and the internet. Right. Now, when we're talking about the internet, people think of the internet as like, Ooh, it's Facebook it's tik tok but the internet is just this underlying medium, right?
[00:04:43] It's the physical infrastructure. It's pipes. Yeah. It's so series of tubes, Ted Stevens was so excited about, but with the internet of things, What you're really thinking about, think of that as the physical infrastructure that ties the internet with the physical infrastructure needed on the things themselves.
[00:05:00] [00:05:00] So the sensors, the wireless connectivity, energy harvesting, or some sort of super efficient battery, right. We need all of that. You combine that's your internet of things, that's physical infrastructure. And then you add to it AI, which is the Ana analysis and the synthesis of that data that's going to be coming in.
[00:05:18] And then that's the soul. So the AI is the soul IOT is the body. And then you have this wonderful, like cloud of information. Right. And I don't know how else I'm just, it's so transformative.
[00:05:33] Matt: [00:05:33] I love how the writer goes to the metaphor. That's awesome.
[00:05:36] Stacey: [00:05:36] I want to make it easy for people to get excited about this.
[00:05:39] Cause this is really a big deal. It's as big as the invention of broadband, right? It's as big as bringing mobile connectivity to everybody. So you get all these new applications and we're just scratching the surface and it's going to be awesome unless we utterly screw it up, which we totally are.
[00:05:56] Matt: [00:05:56] So I interviewed Joe Zhu, who's the founder and CEO of Zenlayer.
[00:05:59] And [00:06:00] we were talking about his definition of edge, which I tend to not go into because you know, alive because there's like 180 different definitions. But his definition, I think resonates with your definition. He said that the edge is where the internet meets the physical world. And I thought that was a really great way of describing it because right now, for the most part, those two worlds are largely separate except for these intermediating devices that we hold in our hands, right.
[00:06:27]the iPhones or the devices or the keyboards and the browsers and stuff. But there are a lot of other things out there that don't have screens and keyboards. And in fact, you know, when I talk about the 20, 20 state of the edge report, One of the concepts we introduced was this idea of the three acts of the internet.
[00:06:46] And, you know, the first act was like, it emerges, like, it's the amazing thing that you can connect to any computer anywhere in the world from wherever you are. And it works. Now it may take a while to get a video, or it may take a while to get a [00:07:00] rich webpage, but it was amazing. And then we started to, you know, consume videos and things and the CDN was invented.
[00:07:05] And that was really the first. Edge computing. Right? We said, okay. The internet through the tubes from point a to point B is too slow for certain applications or it's too congested. So let's cache things. Let's put some storage devices out into the field. In fact, when I did research on the origin of the phrase edge computing, the first.
[00:07:23] Reference that I could find was the paper that the Akamai founders wrote when they're at MIT and they called it edge computing. So anyways, the CDN is the original edge computing, but now we're entering this third phase and the primary driver of the phase in my mind. And the way that I think I like to explain it, I think you'll like, this is we're moving from a world of primarily humans talking to machines, or maybe humans talking to humans.
[00:07:45] Like we are now. To a world of machines talking to machines. And in that world machines are on 24 seven. There are hundreds of billions of them, not ones of billions, and they generate copious amounts of data. [00:08:00] All the time and the value of that data decays. So you need to analyze it faster. And so that's creating a lot of pressure to make at least some portion of the analysis, which would involve the internet, hopefully happen a lot faster, which is driving a lot of this edge computing.
[00:08:17] And I'm wondering if that jives with your
[00:08:20] Stacey: [00:08:20] view. Some of it does. And I will say, I think you're coming at this like a cloud guy.
[00:08:27] Oh, definitely.
[00:08:28] Matt: [00:08:28] So, so here's
[00:08:30]yeah. So give me the other perspective.
[00:08:31] Stacey: [00:08:31] Here's what's crazy. I came, obviously I came from the cloud world. Right. So yay. But when you come at it like a cloud guy, You think in terms of think about the history and the arc of computing.
[00:08:44] Right. So back when I was just starting out in tech, people used to show me their server closets, you know, they'd be like, Oh, I'm so proud. This sun machines, the blue lights, so pretty
[00:08:55] Matt: [00:08:55] gave names to the machine.
[00:08:56] Stacey: [00:08:56] Yeah. So, and then there was this shift over to [00:09:00] time to one Sys admin doing like five servers.
[00:09:03] That was amazing. And then you get to the cloudification. Throwing weird words out there where you have one guy can manage, I don't know, a thousand servers, you know, Google and Facebook. They're just like, I don't know if they have a robot probably doing it now, but the point is you treat it instead of like a pet, you went to treat it like cattle, right?
[00:09:24] So the cloud world, we're like, yes, this is the way this needs to go. And that's how we think about edge. We think about it from the standpoint of how can one person manage all of these devices and. Where does it make sense to put the computing right from a, maybe it's a store, sorry, bandwidth perspective.
[00:09:43] Maybe it's legal. Maybe it's all of these things. When you go to the people who are industrial office managers, when you start talking to people who deal primarily with things, they have a lot of concerns about that, and they're not cloud people. They don't, they're thinking about this a lot of [00:10:00] times from a life safety perspective, for example, It's not cows that you're managing.
[00:10:05] It's a bunch of rabbit wolves that will rip the arms off of somebody who, if something goes wrong. Right? Think about it that way. So they take this and they think, okay. My challenge here is I have to manage these rabbit wolves and they focus more on the education of the people. They focus more on. They ask a lot more questions about security.
[00:10:26] They focus more on latency because of a rabbit Wolf is going to go crazy. You know, you need to quickly stop. It is a terrible metaphor. I don't know what I'm doing here. so. I don't know if we're ever going to get to. The cloud version of the edge, or it may be, I don't disagree with your characterization.
[00:10:48] Maybe I think you're missing some of the really essential elements tied to what you need to think about. When you're thinking about computing at the edge, you can't expect the wolves to all behave gracefully. So we have [00:11:00] to build more safeguards in which right now is going to be less efficient, which the cloud people hate.
[00:11:05] And the people talking to the cloud, people are like, It's really important. And I like we got this, you know, so I guess there's that dichotomy there that I think that is under appreciated and that's not just edge.
[00:11:19] Matt: [00:11:19] Yeah, I think that's true that it's under appreciated. I would argue if I'd gone out of the five minutes, I would have gone all the way down to the embedded systems on the device and the safety control loops and things like that.
[00:11:29] But you're right. I mean, I definitely come from the top down and I mean, that is why it is the internet of things. You know, this idea. I remember when IPV six was first coming out and said, we're going to cave. So many IP addresses. every grain of sand can have an IP address. And I think that is to me.
[00:11:44] The most transformative opportunity because, you know, we've arguably had some form of device, edge computing for a very long time. I had a TV that wasn't a connected TV and it had a microprocessor and didn't have a control loop and it did things [00:12:00] on the edge. Right. But what really. Changes the possibilities is connecting that back to the internet.
[00:12:07] And it's not just about where's the compute, right? Because that's part of the equation, but it's solving, it's like, where's the business problems. How much battery do I have? What's my transmission bandwidth. And I think that maybe the way to think about it is that the internet is a continuum.
[00:12:24] From the edge of the sensor, the thing that's picking up the data or generating the data all the way back to centralized cloud and compute and storage and networking are going to exist.
[00:12:34] Stacey: [00:12:34] I agree with you. And, Oh, that was the other point I wanted to make. You talked about the data generated by these devices being copious and actually.
[00:12:43] In a lot of cases, it's not. And the reason why is yeah. The battery life. And so like when you're talking to these people, they're like, that is actually a consideration they have of, you know, I've got a plant, I've got whatever, I need this to last. Five years before I send somebody out to change [00:13:00] all 800 of these sensors.
[00:13:01] So they're like I can send data once a day and usually it's time series data, and that is not copious amounts of data. Sure.
[00:13:08] Matt: [00:13:08] I agree with that.
[00:13:09] Stacey: [00:13:09] So I can talk to, you know, if I talk to you about a car it's like 25 gigabytes of data at an hour is the most recent stat I pulled from somebody. And I was like, ah, so much, but if you're talking about like a centrifuge or something that could be like, A gigabyte a year, if that, because it's
[00:13:24] Matt: [00:13:24] or a stoplight where every light has a bit.
[00:13:28] Stacey: [00:13:28] Right. But the stoplight bits are only good for like, however, the long the cycle of the stoplight is. Right. So you only care about that stuff like data for like 45 seconds, right?
[00:13:37] Matt: [00:13:37] Right, but you might want to be combining it with all the other stuff, like data and the motion data of all the cars and pedestrians and those sorts of things.
[00:13:44] And that's where you need some sort of a network and some sort of a, you need to have compute at a nexus where those things meet
[00:13:50] The
[00:13:50] Stacey: [00:13:50] cloud thinking is how can we form a monolith, right? How can we build an application that can run across, you know, a bunch of these servers right. In the [00:14:00] IoT? I think we're going to come into this realization that what we really need are a bunch of individualized, highly.
[00:14:07] Personalized customized applications. So the opposite of the monolith it's back to, and I think about like right now, what everybody's doing is building these very siloed, deeply focused on an industry, vertical kind of. Efforts. And we're getting back to this day of like the nineties when you had pre Oracle and SAP buying up all these companies.
[00:14:28] Right? So we're back to the silos in IoT and all the cloud guys want to make the silos into this horizontal platform. And we're just not there yet. And I don't know if we're going to convince everybody and I don't know how to get there and what we're going to do the trade off to get there. If that makes sense.
[00:14:44] Matt: [00:14:44] Yeah. So let's spend some more time coming up from your world, from the devices at what are the big trends? I mean, what are you saying? That's exciting.
[00:14:51] Stacey: [00:14:51] you say trends. are you talking about like use cases? I mean, there's so many things. What do you want?
[00:14:56] Matt: [00:14:56] Yeah, there really are.
[00:14:57] There really are so well, let's start with use cases. I [00:15:00] think that's a, that's an interesting topic.
[00:15:02] Stacey: [00:15:02] Probably number one use case in industrial IoT is maintenance, right? So. Predictive maintenance tracking your machines. It's not on the production line. This is separate from all the control loop stuff that we're talking about.
[00:15:15] This is just like, Hey, what's this doing? And it's an add on thing. It's totally ingrained. It's been there for a while. Like a couple of years now, the other is asset tracking and asset tracking is really a cool use case because it is essential and it requires like crazy different wireless networks that are like uncontrolled.
[00:15:37] Matt: [00:15:37] So what for our listeners who may not be familiar with the world of asset tracking, what is it?
[00:15:42] Stacey: [00:15:42] Oh, it's knowing where your stuff is.
[00:15:44] Matt: [00:15:44] What's give us, give me an example. Like what kind of stuff would I need to track
[00:15:47] Stacey: [00:15:47] that stuff could be a fleet of cars, but it could also be like, if you're in a hospital setting every single little, like.
[00:15:54] Ivy thing or
[00:15:56] Matt: [00:15:56] EKG machine. Yeah. Right.
[00:15:59] Stacey: [00:15:59] EKG [00:16:00] machine, or even, you know, that $500 package of bandaids. That's not actually being tracked yet, but one day it will be, but
[00:16:07] Matt: [00:16:07] back to the RFID tags.
[00:16:08] Stacey: [00:16:08] Yeah. So that's asset tracking and it's really important. And they're even doing asset tracking on these big cleaners.
[00:16:14] Right. You know, those big, like Zamboni, like machines that clean office buildings, because they're like. $300,000 and the people driving them will drive them into a place. And they're like, shift's over, I'm stopping. And then the next night, when it's time to clean someone wastes like hours going through these floors, looking for the Zamboni cleaner
[00:16:31] Matt: [00:16:31] and it's on the bottom level of the parking garage or something.
[00:16:34] And nobody told them,
[00:16:34] Stacey: [00:16:34] yeah, like. So that's like wasted time and productivity. So that's why asset tracking is essential for that. It's also essential for things like security, because when you know where something is, you can hopefully help secure it. And it's important for, you know, so many use cases like cold chain management, like think about all your seafood or milk or.
[00:16:55] Precious drugs, knowing where those are and that they stayed at the right temperature. That is like one of [00:17:00] the top use cases right now. So, so many reasons to track your assets that are maybe obvious and not so obvious. So those are probably the biggest use cases. When we
[00:17:10] Matt: [00:17:10] talk about the internet things, we tend to think of sensors and to some extent you think about, okay, what are you, what do you want to know?
[00:17:19] You want to know. You want to know the current state of the device you want to know potentially what's around it or what it's sensing, right. Is it sensing vibrations on a engine or a motor or something to do predictive maintenance and you want it to where it
[00:17:32] Stacey: [00:17:32] is? Okay.
[00:17:33] Matt: [00:17:33] Those seem like the three most important things, but there's another part of this, which I also think it was a United things and that is making actions.
[00:17:40] So, you know, robotics, how does sensors and robotics merge in this world?
[00:17:47] Stacey: [00:17:47] Probably any way they want to. So right now where we are with IoT, 90% of the use cases are I am, it's a sensor triggering some sort of notification on a dashboard and [00:18:00] then a person. Solves it right.
[00:18:02] Matt: [00:18:02] It's an indicator light. Yeah.
[00:18:04] Stacey: [00:18:04] It's a fancy indicator light. Eventually we will get to that close loop. I've sent something. I notify someone, but I also know that. Thing I need to do. And then I will take the action that doesn't actually have to be a robot. It could, I mean, it depends on what you think about robots and it could be like a servo.
[00:18:24] That's just like a tiny little motor. I mean, you could think,
[00:18:27] Matt: [00:18:27] well, it could be,
[00:18:28] Stacey: [00:18:28] yeah. I was like, I have little robots all over my house that are not at all robotic in my mind. So in that case, we're already there in some stuff we're going to get further along and others, if we're talking about robots, like.
[00:18:42] There's robots and factories like pick and place robotic arms for pick and place machines. Are we talking about like robots, like pepper that are driving around and greeting people in Japanese hotels or, you know, are we talking about nurse, like robots, like diligent robotics makes that go [00:19:00] through and like stack sheets and towels and we'll actually bring patients things they need.
[00:19:04] So I guess the answer is, I think sensing and robotics will come together. Right now if you are a robot, you usually have your own sensors. You're not reliant on any other sensors. Like if you're a robot that is a autonomous robot, that's moving around on your own. You're your own separate world for control purposes.
[00:19:24] And if you are like a, I'm a servo, or I am turning on a light, you probably can be part of this world. So I guess that's where I'm thinking things are right now.
[00:19:32] Matt: [00:19:32] Yeah. So, so now I do want to ask you about edge computing. So again, from this device up from the thing up to the internet, like where do you see edge computing playing the role?
[00:19:42] Well, where do you see it playing a role today, if at all? And w where do you see the biggest opportunities. So
[00:19:48] Stacey: [00:19:48] what I think about edge, I spent so much time, like two or three years ago, trying to define the edge because I got so frustrated with
[00:19:55]Matt: [00:19:55] yeah, I did too. I wrote, I started an organization to try to define it.
[00:19:59] Stacey: [00:19:59] Yeah. [00:20:00] Well, and I was thinking, so the Linux foundation, you're the state you're doing state of the edge. That's correct. Yeah. But they also have LF edge. Is that part of the state of the edge? Right. So,
[00:20:12] Matt: [00:20:12] so state of the edge is actually a project of an LF edge now. And it was, I contributed it to the Linux foundation and we've just finished harmonizing the taxonomy of edge.
[00:20:21] And it actually is starting to make sense.
[00:20:23] Stacey: [00:20:23] Glossary guys, you and Jason. Okay, lovely. Sorry. I'm just relegated you right there to the glossary guys. There you
[00:20:31] Matt: [00:20:31] go. another connection we have that's right. That's us.
[00:20:33] Stacey: [00:20:33] Yes. Three years ago. Tried to define the edge, you know, there's the. Telco edge. There's people calling it the dumb edge or the farthest edge or the tiniest edge.
[00:20:43] I don't care when you're thinking about the edge, you've got to think about it, what you're trying to do. So I think in terms of time, so latency, like how latency sensitive is it. If it's super latency sensitive, it needs to happen right there. Right? Power. Do you have enough power to do the computing you need at that place?
[00:20:59] Like, [00:21:00] is it plugged in? Is it. You know, this tiny solar cell, do you want to change the batteries every day? Probably not. Security is this data you want leaving your factory, your plant, your whatever, right? So that's a consideration. And then bandwidth, which do you have capacity on the network to transfer that?
[00:21:17] Is it running off a Bluetooth that it only goes this far? I don't know. So, and all of that. Cost is the other concern, right? And cost can come from the standpoint of like your components in there. It can also be the cost of like someone who has to maintain this sensor. Right? So all of that is how you have to think about how you compute at the edge and where you put it in.
[00:21:39] Most jobs are probably less latency sensitive. I know that we'd love to talk about latency, but I was actually blown away. Cause I talked to Matthew Prince who is the CEO of CloudFlare. And I was talking to him about workers because it's a really cool platform for IOT. And two years ago, when they launched it, they thought latency was going to be the big, [00:22:00] like driving demand for that lowering latency.
[00:22:03] Nobody cared, no one cared about it. So that was, I was kind of surprised by that, but I guess I shouldn't be because the people who are most concerned about latency, those are the Wolf herders. They are not trying the new stuff because they're like, what if my wolves break free? So.
[00:22:19] Matt: [00:22:19] There's one exception to that.
[00:22:20] Well, and really profound extension of that. And that is the virtualized network functions for a 5G network.
[00:22:25] Stacey: [00:22:25] Yes.
[00:22:26] Matt: [00:22:26] Very latency sense of running on white box servers, definitely edge computing and a lot of people care about it. Yeah.
[00:22:32] Stacey: [00:22:32] I agree. those guys are cloud guys. they're it people they're not the.
[00:22:37] Matt: [00:22:37] The emedded system people, right
[00:22:39] Stacey: [00:22:39] so like you're right. that is totally true. And your edge is way more expensive than my edge, because I look at that and I'm like, is that really edge computing? I mean, is it a thing.
[00:22:51] Matt: [00:22:51] Well, I mean, so there's, I mean, that's the back to my original question, which is why aren't we just calling it all the internet?
[00:22:57] Because I think eventually it is, I think this idea of edge computing is going [00:23:00] to seem kind of silly because obviously you want to put the compute in the ideal location for all the factors that you described, whether it's bandwidth or cost or, you know, radio transmissions with security, there's probably an optimal place to
[00:23:11] Stacey: [00:23:11] put.
[00:23:12] Yes.
[00:23:13] Matt: [00:23:13] And it could be anywhere depending on the application and all the constraints.
[00:23:18] Stacey: [00:23:18] And that's why I'm like, we don't have to talk about the edge as much. What I think is more important, right? let's assume that we go this way. What's more important is things like. Containers or some sort of way to run an application throughout that continuum without having to manage it at each point in that continuum, by a different person, with a different set of expertise.
[00:23:40] Right? Sure. And so like, I get that, we all have to talk about the edge for now because we're all still figuring that out. But if I look forward five years, And this, I started looking at this like two years ago. This is where I think we need to go. This is where it gets exciting is how do you manage things like that?
[00:23:59] How do [00:24:00] you build processes? How do you train your developers? How do you think about, you know, sectioning off computer science even, right. Because we have historically had these big divides between embedded and, you know, traditional 86 or risk-based processing. Right. You know, and at some point in time, you're like, maybe you need to be able to develop for.
[00:24:25] All of that. I don't know. I'm not a computer scientist, so this is where my brain is just like, but that's what needs to happen.
[00:24:32] Matt: [00:24:32] Right, right. That makes a lot of sense. so, so who's doing the most interesting work in solving those problems?
[00:24:40] Stacey: [00:24:40] Are you familiar with, Balena IO, okay. So
[00:24:44] Matt: [00:24:44] no, tell me about it.
[00:24:45] I'm
[00:24:45] Stacey: [00:24:45] going to get in so much trouble. this is a company that is, it's trying to build basically like I'm just going to call it container orchestration. I don't th it's not, but they're trying to build a platform where [00:25:00] you can run. You can run your application on ever smaller. Devices with ever smaller memory requirements, for example.
[00:25:08] And I actually think Microsoft has seen this and is working hard on this. Obviously, if you're dealing with anybody with who's playing with Kubernetes, it's interesting. CloudFlare has decided screw containers. They haven't decided screw containers, but they're doing their whole isolates concept with.
[00:25:25] Their workers stuff. That's interesting because you can run an isolate in a lot of different places. And then there's other companies who are building small bits of this that are like thinking about how you carry your operating system and keep it current all the way through something like this.
[00:25:41] So there's pieces of this everywhere.
[00:25:45] Matt: [00:25:45] So in some future world, we have, software developers who. In today's world in the cloud, they don't really even have to think about servers necessarily. Right? They are building their, their code. It's [00:26:00] going into a continuous integration pipeline. It's getting orchestrated onto some Kubernetes cluster somewhere that they don't know where it's running.
[00:26:06] And if a machine goes down it restarts and the workload goes back and you can see that. Extending potentially, you know, again, I'm not a computer scientist anymore, so I don't remember. I don't know. I don't know all the complexities in this, but you can see that extending all the way down to the device where you hand your workload to an orchestrator that is.
[00:26:25] Aware of probably cause it has a lot of sensors feeding into it. Interestingly enough, it's aware of network congestion. It's aware of cost. If you have to like buy a worker from CloudFlare or a, you know, a NBC two instance from Amazon or some other equivalent and it understands latency and the SLA of the workload, it may even understand like whether it's life critical or life safety, so it needs to run or maybe.
[00:26:49] It influences life safety. Cause you're probably never going to take the life safety off the embedded system. but yeah, but it can provide decision support. I mean, an example is autonomous cars. Like they can't see around corners. I don't care how good [00:27:00] your LIDAR is. So the network and the compute in that network can provide you information about what's happening elsewhere to make better decisions, but you're never going to have the break or the airbag deployed by the cloud.
[00:27:14] Stacey: [00:27:14] Probably not, but it's exactly actually, you know how I was debating with you about 5g and what's happening with network function, virtualization and all the back and forth that's happening there. That is a precursor to this in the sense, like that's still happening on a big white box, but eventually you're going to want to, like, you know, if you talk to at T and T or Verizon, when they start talking about like, Managing like the different classes of spectrum they'll have for 5g.
[00:27:44] Right. They're going to have the device come into play saying, I'm here. I have this capabilities. This is how many DBS I'm shutting. and then that's going to come back into where they're going to land on the network. Right. So you're going to start seeing feedback. From devices. And I [00:28:00] love when people talk about edge because they have to talk about autonomous cars.
[00:28:03] And if it were not nine in the morning, I would drink because that is the game. when you mentioned autonomous cars and you're talking about the edge, we must drink. So feel free, anyone listening to this after, you know, whatever time you'd feel, but. That's, it's not as real of a use case yet because we don't actually, I mean, yes, they're out there, but they're not actually talking over the cloud very much,
[00:28:26] Matt: [00:28:26] although it's a super relevant use case that is autonomous vehicle or drones.
[00:28:30] And what's interesting about drones is you can't put the compute cause it's got to fly and carry it and has battery life. And so you've got to have a whole different strategy to controlling the drone and a lot has to come from the cloud.
[00:28:41] Stacey: [00:28:41] Yeah. And that has to, that gets us back to that whole idea of.
[00:28:44] Such individualized use cases. We're not going to have these monolithic apps or even architectures possibly for these things. I don't know. Maybe we'll have big classes.
[00:28:54] Matt: [00:28:54] Yeah. that's interesting. Yeah. I mean, it's, although I think if you look at the history of, and I'll just broadly say [00:29:00] computer science, you have this push and pull where things are first built in this kind of bespoke solve a single use case maybe with a single device.
[00:29:09] And then. As we learn more and we learn where the points of leverage our, in terms of creating abstractions, we create this abstractions and, you know, a simple example is we create an extraction of a computer called a virtual machine. And we created a simple version of it called a container. I mean, I'm oversimplifying, but you know, there's the joke that all problems in computer science are solved by another layer of extraction.
[00:29:31] And to some extent that's true. And what's really interesting about this world of billions of devices and potentially. Millions or hundreds of millions of machines, servers, everywhere that you could run workloads on. you need a later abstraction and you need these orchestration layers and you probably need artificial intelligence to make the right decisions.
[00:29:52] How are you going to figure out where to run that workload? You know, in the old days, like I talk about Amazon. I don't pick on Amazon in particular, but it's a good example. Everybody understands. So like [00:30:00] today, if you want to provision an ECE, two instance, you know, ignoring the tests, going with Verizon and wavelength now, but.
[00:30:07] You want to create a traditional ECT incidents on Amazon? You have two choices. United States, you are requesting USCS and that's it. A human can make that decision. Right. But when you've got Chicago, West and Chicago, East Atlanta, West Atlanta, East, and maybe, you know, a much more granular neighbor neighborhood one and neighborhood two, or device one, or device two, you can't have, as you said, the example of cattle versus pets, you can't treat these as pets anymore.
[00:30:33]and I think. You know, I'm optimistic. So I think these two worlds, eventually do merge, you know, where these embedded, so you look at like the smartphone, right? I mean, there were lots of devices, you know, the Newton even, or the Palm pilot, or, you know, we had this idea of a device that supported our lives with applications.
[00:30:52]but it didn't really explode until. Apple will built an amazing device. but then put, also put an abstraction layer and an app store on top of it. [00:31:00] I wonder if you see that trend line,
[00:31:02] Stacey: [00:31:02] I think we will get there in part, I think what will be. It's totally trying to go that way in, even on the embedded side and the safety security side, they want it to go that way.
[00:31:16] Right. And you're going to see more and more things get eaten up by the cloud. And I, for anyone listening, I am using my little cookie monster. Pac-Man hands to like the cloud is coming to eat it,
[00:31:26] Matt: [00:31:26] eat the CloudApp. Yeah.
[00:31:28] Stacey: [00:31:28] There probably will ensure, yeah. Always be devices that. Are away from that. And the reason I say that, and you can think of it like the nuclear codes, right?
[00:31:38] Like, or what did I just discover about the Boeing? I'm not an airplane person, so I'm going to mess this up. But there are some Boeing planes that are their navigation, software is still updated via floppiness three and a half inch or three and a quarter inch floppies every day. Someone pops that in or once, sorry, it's once a month, they pop that in there and they update the navigation.
[00:31:58] Right. To me, that [00:32:00] could be over the air, but that's okay. But there are things that won't be part of this and the reason will be safety and kind of five, nine reliability at least. And yeah. I don't think that's a bad thing. I mean, think about what you lost. remember landline telephones.
[00:32:16] Matt: [00:32:16] I do actually.
[00:32:17] I've never rotary telephone.
[00:32:18] Stacey: [00:32:18] They worked right. Like you never had that last call. Sorry. But now we tolerate that because our cell phone networks, you know, the compromises you no longer have to have literally wait by your phone. Right. You just drive around talking to people and yeah. Okay. Sometimes you lose people.
[00:32:34] Sometimes the connection gets lost, whatever, but there are always going to be things as computing eats that world that are still need the five nights for more than five nights. And I think the cloud people are underestimating. How many of those there are, and I don't know how we bring that in.
[00:32:55] Matt: [00:32:55] I think that's a fair, I think that's a fair criticism.
[00:32:58] I think that's a fair criticism.
[00:32:59] Stacey: [00:32:59] I talked [00:33:00] to a guy who runs it. It's at Schneider electric and he runs a plant that makes those breaker boxes. And he was like, Stacy, if I have five, nine reliabilities and. Five nine reliability in terms of manufacturing, those boxes like 1600 people a year would die because they would explode.
[00:33:20] Right? He's like, I can't have that. He's like, it needs to be super reliable. I have to have, you know, everything there. So I'm not saying it will never go that way. There will be some that will never go that way. I think we need to, as computer science people, we need to be thinking about how to boost reliability.
[00:33:38] And make that more of a priority than we possibly have in the past, because traditionally with service, you're like, Oh, like, think about like, Ooh, when hot swappable servers came around, that was so amazing. Right. We're like, woo. And you think about ECE two instances. It's like, Oh, it's bins down. It spins back up.
[00:33:53] Right. But we have to think about moving applications. That can never go down. Is that a [00:34:00] right way to think about it? I don't know.
[00:34:03] Matt: [00:34:03] Well, that's certainly our applications that you never want to go down because they're like flying a rocket or something. And I know a lot of people are trying to figure out how to.
[00:34:14] Introduce elements into the compute stack that maybe you don't have the same level of liability as the embedded systems, but work in conjunction with the embedded systems to when they're all working, deliver a better experience. You know, I mean, you think about, I'd be interested in, if you look at the space X rockets, you know, what percentage of what that rocket does at any given time happens.
[00:34:36] From a control loop running on the rocket versus something that's coming from mission control. I mean, that's a really interesting question and it's a version of what you're just
[00:34:43] Stacey: [00:34:43] helping a lot cause they go to space.
[00:34:45] Matt: [00:34:45] Yeah. I mean, that's an example of, you know, in some ways that just could be a super high latency.
[00:34:50] Is that how high you go? Speed of light is really fast and there's no network routers and BGP routing tables in your way, between yours, your rock, [00:35:00] whatever. You know, maybe we'll come back in five years and we'll see where the intersection of the cloud and the devices are. Let's go someplace completely different.
[00:35:08] Cause I love the fact that you're in this world. Like what is the craziest thing you've heard of somebody putting a sensor on or like connecting to the internet?
[00:35:16] Stacey: [00:35:16] Oh, well personally it's a toss up between, I had some connected socks and a connected egg carton. So that's just me as an individual in that.
[00:35:28] Matt: [00:35:28] Can you describe both of those, like what they do and how they benefit you and like how the H
[00:35:32] Stacey: [00:35:32] how this was the early days. Is it the internet of things? There was a company called quirky, went through many owners and it made one of the first products it made was, first of all, it was the quirky pivot thing.
[00:35:43] It was a power strip that actually moved around. It was super useful. So we were like quirky. It's pretty cool. And then they came out with a connected egg holder that held for some reason, 14 eggs. Not a dozen. So you put the eggs in there and it would literally just count. It was like you have [00:36:00] five eggs.
[00:36:01]
[00:36:01]Matt: [00:36:01] it would tell you when you're out of eggs so you can reorder.
[00:36:03] Stacey: [00:36:03] Yeah. and that came on the market for $50 and surprisingly, very few people bought it. I only did it because it was my job. I'll be honest. If it had been $10, I totally would have bought it and give it to everyone as Christmas presents.
[00:36:17] Cause it'd be like welcome to the internet of things. It's.
[00:36:19] Matt: [00:36:19] Well, I wouldn't be surprised if the Samsung refrigerators sold in South Korea have that built in.
[00:36:24] Stacey: [00:36:24] Oh, I'm sure they do. Now. Why not? So that's was that, and it was pointless. The other thing was my connected socks. And those are actually surprisingly not pointless.
[00:36:34] Those are socks with sensors in them, and they have a little Bluetooth module that you clip onto the socket, and then it talks to your phone and it tells you how you're running. So if you are a runner, this is really interesting. It tells you if you're pronating or. Sup and whatever the opposite of pronating is
[00:36:51] Matt: [00:36:51] super needing.
[00:36:51] Yeah.
[00:36:52] Stacey: [00:36:52] Yeah, that's right. That's it. and it gives you like information about your stride and then you can like take that information down, visit the [00:37:00] service, did not do this. It just gave you that information. You're like, well, that is, that has information. The cool thing would be the next step, which one day could happen, but probably will not.
[00:37:10]Where it says, Hey, these are the shoes that would work best for you. And here's how they're running down. Right. I think you're going to see that
[00:37:17] Matt: [00:37:17] and press this button. And Amazon will send you a pair that you can try on and sent back if they don't work.
[00:37:21] Stacey: [00:37:21] And I think we'll see that. I think we'll see Nike is already embedding stuff or UA is also embedding stuff in their shoes.
[00:37:27]we'll see that. So that's kind of the crazy use cases where. I think they're going to be they're crazy today, but they're going to get pulled into our lives in ways that are really interesting on the I'm trying to, I think on the industrial side or the enterprise side, probably the craziest thing I've seen is probably Microsoft has some really whacked out fun ideas about building smarter buildings and facilities managements with digital twins.
[00:37:56] So they are, they have this [00:38:00] concept of. Productivity rate three 65. so this idea that you could tie your productivity software to your physical infrastructure, how COVID changes this? I don't know, but think about if you schedule a meeting in outlook. If you schedule a meeting in outlook and it automatically reserves the conference room for you.
[00:38:19] Right. And the conference room then gets updated and you see that on the little screens in some places.
[00:38:25] Matt: [00:38:25] Yeah. I mean, why can't I order a conference room? Like I order an Uber, right.
[00:38:28] Stacey: [00:38:28] That happens. And then, you know, if no one's in it, motions, if no one goes to the conference room though, cause you like, we're being all like.
[00:38:35] Sly and tried to get a conference room you may or may not need. Then after 10 minutes of no one being there, it would actually free that up. So that's the idea. The other cool stuff tied with that is once you start doing that, then you can do things like now let's bring COVID into the picture because it's here.
[00:38:52] Imagine if you have more people in a conference room, then you're 25% occupants see can support. You could get an alert. [00:39:00] Facilities or security might get an alert. So you start seeing ways that your world interacts with you. That feels really crazy, but actually could be kind of good. The question that we need to ask though, when we start doing this is do we do it with privacy in mind?
[00:39:17] Like anonymity in mind, probably not in your office building, but you don't really expect that. and what kind of rules we have around keeping that data and sharing that data and that's gets to, and we're not here to talk about that, cause it's the edge focus podcast, but that is why we need to answer these questions because otherwise we're not going to get anywhere with this because people are going to rebel and say, this is stupid.
[00:39:39] I actually want to
[00:39:39] Matt: [00:39:39] talk about this.
[00:39:40] Stacey: [00:39:40] Oh, great. Okay.
[00:39:41] Matt: [00:39:41] Yeah. I mean, it's about the edge and the future of the internet, which I think this fiscal funds
[00:39:47] Stacey: [00:39:47] that's great. I have a whole column on that.
[00:39:49] Matt: [00:39:49] Yeah. So please, well, we'll put that in the show notes. So educate us in the audience and me about like some of these issues and how people are approaching them, security [00:40:00] and safety and you know, sharing information.
[00:40:02] Stacey: [00:40:02] My God, that's like a forever conversation. so think about, I always tell people. IOT makes the invisible visible. Right. So you can see things that you haven't seen before. Well, it's a state of a
[00:40:15] Matt: [00:40:15] conference room or the speed of tractor.
[00:40:17] Stacey: [00:40:17] Yeah. Or pollution at, you know, mouth level in your town. Right. So all of that's available to you, but it also means.
[00:40:27] That people and the story, the cheap storage and cheap analytics also mean that as you move through your day as an individual, it is possible to track you. I'm not saying people are doing it right now at every point in time. Right? We're not there yet in China. They're getting there. People can track you, whether you're not, it's like an unhinged individual who has access to systems or a state.
[00:40:52] Right. And that becomes very scary, especially in a place like America, where individual rights, really Trump collective good in a lot of ways. Right. [00:41:00] So people are inclined to distress that. And I don't think that's the wrong way to think. So I think right now, especially when we talk about privacy and security, we're kind of having this debate over convenience versus art.
[00:41:14] Potential loss of our individual liberties. Right. And I don't think people, many people are not thinking of it that way, but they probably should be. And so very broadly, that's the big overarching thing. And we see this in tiny ways. Like we see Google getting warrants from police departments saying, Hey, there was a crime here.
[00:41:35] Could you tell me any passers by carrying your phone that you know, we're here at this time of day and. That's scary. Right. We don't have a legal infrastructure for that. We're relying on Google to like, sort of help us here because they don't want to become an arm of the state. But again, that's the big picture.
[00:41:56] Matt: [00:41:56] But they're the ones driving cars down every street and nearly [00:42:00] every country taking photos of nearly everything.
[00:42:02] Stacey: [00:42:02] Oh. But they blurt, they blur out all the weird bits. Come on, man.
[00:42:05] Matt: [00:42:05] Right? Right. Well, where's the unredacted version of that image who has access to that.
[00:42:09] Stacey: [00:42:09] Yeah. So I think if you're going to do this, you have to establish trust with your users.
[00:42:14] You have to establish trust by establishing robust privacy policies and security policies. You have to secure everything. At rest, right? Are the device in the cloud. You have to know who your providers are and understand how they're securing things. And then you have to secure things in transmission, but you also have to build out new data retention policies.
[00:42:34] And maybe it's not as crazy as GDPR. Maybe it's as simple as like after three months we'd dump everything. I have lots of companies in the consumer IOT who are like, they save your wifi password. When you log on, most of them will acknowledge. They don't have to do that. But they do it anyway because why not?
[00:42:51] Data's data. It's the new oil. and I think we need to get away from that sort of practice. You need to decide what you need and keep it for only the time that [00:43:00] you need it and then dump it. But it here comes the, but. What about AI? What are we going to train this stuff on that if we don't have all this access to free information, and the answer is, I don't know, a more proactive opt-ins getting users who like it, and this is a question.
[00:43:18] I mean, I can't argue about privacy and anonymity by design and all of this without acknowledging that there's going to be a cost down the road and there is a cost, but I think we aren't having those discussions now that was so much. That was awesome. That was so much I'm so sorry.
[00:43:35] Matt: [00:43:35] Well, no, you're right.
[00:43:36] We do need it. Do you think you need to be thinking about these issues and, you know, to bring this back to the show? I think that edge computing can both be. the cause and potentially part of the solution, but you're right. A lot depends on how we approach it.
[00:43:51] Stacey: [00:43:51] I mean, storing things locally, you know, makes a lot of people feel really good because it's under their control.
[00:43:58] It's kind of the same [00:44:00] argument though that we had at the beginning of the cloud and everyone was like, I will send my test, you know, test data to the cloud, but I won't do anything crucial there because you know, it's.
[00:44:10] Matt: [00:44:10] Although there is a very strong argument that the cloud providers are way better security than individually it teams, which is fascinate
[00:44:17] Stacey: [00:44:17] until your partner that you're working with, like leaves an unsecured S3 instance run in your,
[00:44:24] Matt: [00:44:24] or your employee that you're working with
[00:44:26] Stacey: [00:44:26] your production data.
[00:44:27] You're like, Oh, it tells you that
[00:44:28] Matt: [00:44:28] isn't it. Isn't it something like, you know, there's some outrageous statistics in the high nineties that when they look at security things, it's human error.
[00:44:34] Stacey: [00:44:34] Oh yeah. it's us. And that's actually, Oh, that's another point where the embedded guys. Totally have it over on the cloud guys, right?
[00:44:42] The embedded guys, they have safety protocols and training that focuses on people as the, just the dumbest, the weakest link. That is
[00:44:54] Matt: [00:44:54] a really important point. Consider that it's just the mindset and how we're taught and how we're taught
[00:44:58] Stacey: [00:44:58] to think. [00:45:00] I go to the Emerson user group or the Honeywell user group meetings, right?
[00:45:04] The beginning of every presentation. I think it's Emerson does this. They're like, because our motto is safety first. You know, they show a slide that said, that does like literally a safety checklist. And that's because these people are dealing with toxic chemicals and giant machines that again could slice their arms off.
[00:45:19] Right. But every presentation, they point they pick out something like recently with COVID they're like, does everyone have their mask on? Right. And they ask it every single time and they all have their checklists for things they have, like, where did this USB stick come from? Before you, like before you plug in a USB stick, it wants to know where you got
[00:45:41] Matt: [00:45:41] it from.
[00:45:42] Yeah. Where's the chain of custody on this
[00:45:45] Stacey: [00:45:45] stuff. We don't really have. We have like, like phishing training. That's pretty lame. we don't have anything like that. And I think. We probably should. It's the same thing you do with pilots, right? They have those checklists are in the operating rooms. Remember how, [00:46:00] like, I think it was like a Malcolm Gladwell story where like, when they gave them a list of surgeons, a list of things to check before each surgery.
[00:46:07] Matt: [00:46:07] Yeah. It's that, the checklist manifesto, the book that talks about that in particular how yes. Yes. Just by making the red Sharpie Mark on the limb, you're supposed to amputate. Re reduces the incidence of wrong limbs, being amputated, dictated by, you know, 80%
[00:46:26] Stacey: [00:46:26] and when you're moving and when we're moving from software and things that only happen on the web to things that happen in the real world, we need to bring that mindset with us because otherwise we're like, I mean, what's the equivalent of the wrong limb in a factory setting.
[00:46:42] Maybe it's a toxic gas gets released, maybe, you know, or in a hospital setting. Maybe it's your patient data goes. You know, on the dark web and people get, or maybe it's like your IV machine distributes the wrong dose. I mean, these are really important.
[00:47:00] [00:47:00] Matt: [00:47:00] Yeah. and I completely agree with you that there's a type of person who is in that world, lives that world and thinks that world, I mean, the dirty secret about cloud developers is that we don't actually know how our software works.
[00:47:12] We don't, it's too complicated. We know that it works enough for the time to be confident that we do it. But I mean, I worked for a very, shall remain nameless, credible software company, that had a bug that they could not figure out what it was. And the solution was to, basically restart the machines, once a day.
[00:47:32] And so they would go through and restart each of the ECE two instances once a day, because the bug would only show up after 48 hours. Now, if all you're doing is providing, you know, a news feed on Facebook, that's probably fine.
[00:47:46] Stacey: [00:47:46] But if you're in the real world, an example of that is I talked to a hospital CSO who said that one of their Ivy machines that was hooked up to a patient at the time, Had a bug and they found it or it had a vulnerability and they found it after it was already [00:48:00] hooked up to someone.
[00:48:00] They had to pay a nurse 85 or a hundred thousand dollars for a full day that patient was stuck on that machine to sit there 24 seven to make sure that person didn't die. Right. That's their restart option. It's expensive. It's scary. And you can't actually do that at
[00:48:22] Matt: [00:48:22] scale. That is fascinating.
[00:48:25] Stacey: [00:48:25] These guys are called Doug's by the way, this is my fit, these, the embedded we'll call them the embedded guys, my friend, who was in it, went to work for a utility and he calls them Doug's dumb old utility guys. And the reason it's really mean it's like
[00:48:42] Matt: [00:48:42] the plain old telephone.
[00:48:43] Stacey: [00:48:43] Yeah. And he's like the Doug stand in our way.
[00:48:46] And the reason they do it is because. They're working in the real world and they know how something works and they don't want to risk breaking it. And he's like, and I have to respect it, but I also hate it because I am an it guy and I want to tweak it.
[00:48:59] Matt: [00:48:59] Yeah. Yeah. [00:49:00] You can't move fast and break things. And when lives are dependent on it, that's really interesting.
[00:49:05] So I've got a couple last questions. So the first one is, if you look at how IOT has changed since you've been covering it and. You look forward into the future
[00:49:20] Stacey: [00:49:20] and you
[00:49:20] Matt: [00:49:20] think okay. To get to this imagine world where a lot more things are connected to the internet and we're providing extreme benefits.
[00:49:32] Exceptional benefits and we're doing it with the proper levels of security and things like that. when do you think that's going to start to emerge in reality? And what are the, if there's a set of dominoes that need to topple for those things to come, what are the ones that if you were godlike, you would go and like push on.
[00:49:50] What are the things that you would push on to make go faster?
[00:49:53] Stacey: [00:49:53] I think I would focus on. Trust things that build trust, and that is everything [00:50:00] from establishing reasonable data retention policies, good security, and then also a legal framework for both companies and individuals when things go awry, I guess, because I think it's important to develop towards, I know that innovation, you know, the Silicon Valley folks hate.
[00:50:23] Regulation, because it interferes with innovation. I think actually developing your solutions with this sort of regulation in mind is probably more important when you're dealing with the real world for all the reasons we discussed. So I would say I would like us to have those conversations because I think developing towards those, you have to have guide rails here.
[00:50:47] This isn't an area where it. I don't think a wild West is really what we want here. I really don't. And so that's that on the technical side, I'm not a computer scientist. [00:51:00] I don't know what we need on like a software side. I will say energy harvesting, I think is really important. We need to have better systems for generating power for these devices.
[00:51:12] I think. I think machine learning at the edge, both training and inference are going to be essential because I think that will help bring computing at low power to the edge. Right. And I think I thinking about networks and connectivity and I don't know what I want there, but I want more reliable connectivity that, and maybe I just want it to go.
[00:51:37] Back up a layer to the internet itself, right. to communicate. I'm like, I hate that we have such a patchwork of systems and that's frustrating and it's also causing delays. And I also think we need way more rugged computers that can be run by idiots and implemented by idiots. And I don't mean idiots in the sense.
[00:51:58] I just mean people who don't know what they're [00:52:00] doing because that's, what's going to make this. Faster. So I can deploy an air pollution detection network with my kid, because I'm curious about my local air quality. Have you done this? I have not. There are actually people who can do this. Like there are,
[00:52:15] Matt: [00:52:15] I've got a little device in my room that, measures that are
[00:52:18] Stacey: [00:52:18] closed.
[00:52:19] Oh, I have it. Yeah. I have local air quality, but I'm talking about putting it around my neighborhood, right?
[00:52:23] Matt: [00:52:23] Oh, gotcha. Yes.
[00:52:24] Stacey: [00:52:24] So I want to start taking action as a citizen, right? Because I think that's going to be a really powerful element here. I think those are important, but it goes back again to that trust thing.
[00:52:35] I need to be able to authenticate my data. I need to have frameworks where I can take that data and make it, send it someplace in bulk. I think it's reliable and trustworthy. So there's a lot of, there's a lot of things that need to happen. We're just at the very beginning.
[00:52:51] Matt: [00:52:51] Yeah. So, so when do you predict, and I won't hold you to it, but when do you predict that will, well, that we'll start seeing the internet of things [00:53:00] is actually part of the internet.
[00:53:01] Stacey: [00:53:01] Oh, I already see it as part of the internet when everybody will. I don't, I, I think it's going to vary based on like, Where you are like, is your home on the internet? A lot of people might start saying yes. And the three to five years, especially if we get standardization with like the big smart home players, right.
[00:53:19]in your workplace, it might already feel kind of like it is right. in your cities.
[00:53:26] Matt: [00:53:26] I mean, my children's sorta feel that way. They sort of feel like everything in their life is on the internet. I mean, maybe not the refrigerator or the eggs, but you know, their watches, their phones, their computers, their tablets,
[00:53:39] Stacey: [00:53:39] their toys in 20 years.
[00:53:42] Yeah. In 20 years, I think you're going to have. Something that helps you make decisions every day of your life, right? Like, do I need to stop at the store and get milk, or maybe it'll deliver the milk to you? I don't know how far that's going to go. And people might be able to dial that wherever they want.
[00:53:56]so I think you'll have technology [00:54:00] mediating, your individual decisions and life choices by then, based on all of the information that it's gathering around you. And. Who you are, but I think to get there again, we'll have to have trust in. So yeah.
[00:54:15] Matt: [00:54:15] Dear Silicon Valley, please listen to the dogs.
[00:54:18] Stacey: [00:54:18] Yeah.
[00:54:19] Listen to the Doug's and consider like what the value is that you're trying to offer people and maybe consider asking them to pay for it. I don't know. Yeah,
[00:54:29] Matt: [00:54:29] Stacy, thank you so much for this wonderful pan optic conversation. it's nice to actually, get a little away from edge. This is all relevant.
[00:54:38]before we go, how can people find you online? How can they find your podcast? How can we follow you?
[00:54:44] Stacey: [00:54:44] Sure. You can find me at www dot Stacey on IOT, and you can find me at the internet of things podcast, or I think it's the IOT podcast.com or you can find me on Twitter at giga Stacy. [00:55:00] So those are all places.
[00:55:01] Oh, and I write a column for I Tripoli spectrum called the future of the internet.
[00:55:06] Matt: [00:55:06] Excellent. Well, let's face it very much. Thank you very much for joining us here today. And I appreciate you. You think a part of our,
[00:55:14] Stacey: [00:55:14] I was so glad to be here. Thank you.