Techstrong Gang – July 9, 2024

Hi, everyone. Happy Tuesday. We've got an exciting program for you here today.

We're gonna talk hurricanes, myths, and do you know where your barcode is? It's 50 years old. All that And more on Textron Gang.

Hey everyone, welcome to Textron Gang today for Tuesday. I hope you had a great holiday. I know yesterday most of us were still in sort of that July 4th hangover here in the US anyway, where we're easing back into work.

Well, for me, I'm, I'm easing back out 'cause uh, after Wednesday I'm actually on vacation, but the text on gang will go on with or without me. Uh, speaking of the gang, let me, let me introduce you to today's cast of, of gang members. First of all, it's Monday and he's joining us, well, he's still up in Ohio, but he's also headed to vacation.

He is the founder and and host of the Tech Field Day, as well as chief editor and, and founder of Gestalt It. And he actually has a past as a, uh, it person and, and emergency backup. Well, he's gonna tell us where it's out.

Fred, Stephen FoST. Hey, Stephen. Welcome.

It's good to be here. Good to be out of the sights of a hurricane, that's for sure. We don't have so many hurricanes in Ohio.

No, that's one thing about Ohio. You're relatively hurricane proof there. Um, but not, you know, everything has a price.

You have twisters, um, or tornadoes or however we call them. Sure do sometimes. And who knows, even earthquakes.

Speaking of earthquake, let's go high to the Rocky Mountains and, uh, where we're joined by our CTO and, uh, CTA. Mitch Ashley. Hey Mitch.

Welcome. Good to be here from the 100 degree, degree high yesterday on Monday. Um, yeah, it's nice and hot up here in the Rockies.

So It is, it's go higher To get some better temperatures, I guess. Yeah. Or, well, you know, when people in Colorado talk higher, it's a whole different thing.

But Steven, is it a hundred degrees by you as well, or? No, it's actually pretty beautiful here in Ohio. So, sorry.

Good for you guys. Join and then joining us. I think he's still up in his native New York licking his wounds after a spanking by the Red Sox this weekend.

It's our Chief Content officer, Mike Ard. Hey, Mike. Hey.

How you doing? And as far as I know, hurricanes are things you carry around in open containers in New Orleans. Fair enough.

And then joining me here in studio, it's our Echo Insights editor and analyst about all things green and sustainable. Bonnie Schneider, welcome Bonnie. Thanks, Alan.

Great to be here. Great to have you here. Alright, um, well, let's kick right into it.

It is hurricane season down here in South Florida. Everyone's, you know, it's supposed to be making preparations. We've had our first, well, our second name storm, but our first hurricane hit twice a little rebound shot.

Bonnie, why don't you talk to us about it here. What, what's going on? Well, Hurricane Barrel made landfall Monday morning around 3:30 AM central time, uh, near Matagorda, Texas.

That's about 90 miles southwest of Galveston. And it came in as a category one with winds at 80 miles per hour. But this storm has been around for a while.

It started off as a tropical wave on June 24th off the coast of Africa, and did a lot of damage in the Windward Islands where it was impacted as a Cat four. And what's interesting to note is that this is the first storm to strike the US so far this year, so early in the season. Typically we see peak activity in September.

Um, but one of the points I wanna make is that this is going to be a very strong hurricane season. I know meteorologists used to say that every year, but 85% chance of a very active hurricane season, meaning we're going to see between eight and 13 hurricanes. And several of those will be major meaning category three or higher.

So we're coming off of some very warm water temperatures. And also, um, LA Nina is going to be developing, and that means more relaxed winds. And usually in order to decrease hurricanes, what we need is to have strong winds because they will topple down the, the high, the high points of the thunderstorms.

So unfortunately, with the warm water, LA Nina, um, we are in a, a right position to see more hurricanes and to see a very active season for 2024, which makes us think about, okay, you mentioned people are preparing and, and we all do that, you know, for people that live in the coastline. But what about for businesses disaster? Um, how do you prepare a data center?

What do you do if you have, uh, data damage? And also, unfortunately, the times when we face these storms and these, these, um, disasters, bad guys come out, you know, for cybersecurity. So there's a lot to prepare for.

Um, Steven, I know this is a, an area of expertise that you're familiar with. And as we get started with such a really ramped up start to the hurricane season, what comes to mind? Well, The first thing that comes to mind is, uh, I lived in Houston, Texas, uh, when a hurricane struck.

And, um, you know, hurricanes, uh, can be really, really damaging to, of course, um, uh, people's lives, um, infrastructure, everything. But as you mentioned as well, also it, especially in Texas, uh, where there's very poor, uh, drainage, uh, you know, very little to block the winds, uh, you know, even after the hurricane is gone, uh, storm surge and flooding can cause all sorts of trouble for people as well as infrastructure. So we have to think about things like that as well.

Um, when I was there, um, my job was actually disaster recovery planning for a big, uh, household name Texas company. And it was a pretty big challenge for us. It was something that we were worried about constantly.

Uh, how are we gonna manage to survive a hurricane hit? What's gonna happen to the data centers? Uh, what's gonna happen, uh, in terms of, of, uh, disaster recovery, uh, preparation, planning, you know, is the plan, are the plans gonna work?

Uh, you know, one of the things as well that we were very concerned about was whether our, uh, disaster sites would be far enough away. Uh, because in the case of, uh, the company that I worked for, the initial disaster recovery site was located so close to the data center that we used to joke that if a hurricane hit the building might actually fall on the disaster recovery site. Um, that being said, part of what I did was, uh, we moved the d dr further away.

Now, one of the things that's really helped in terms of disaster recovery for, uh, it is, uh, virtualization. Uh, once we managed to virtualize all of our systems, that was actually one of the big reasons that we invested in VMware to begin with, was that we could, uh, build sites that actually could fail over and, um, and fail back in the event of a disaster because we could move, uh, virtual machines to different hardware running in different places. That was always a big challenge before virtualization.

Now with cloud as well, we have the opportunity to have, uh, workloads running, uh, for example, in virtual machines in the cloud. And this opens up a whole new world of possibilities. Frankly, uh, the technical solutions for disaster recovery have never been better.

But as you point out, one of the challenges with, uh, disaster recovery as well is that, uh, there's a confounding factor right now of ransomware and, uh, ransomware bad actors are attacking, uh, DR sites. They're attacking backups. And, um, that could really, really cause uh, kind of a one-two punch for a company that's really struggling in the event of a hurricane related disaster.

So it, it really is a lot of different things. Now, I know that Textron as well, has recently written on this subject. Um, uh, what, what do you guys think?

What are the steps that people should take to prepare for dr? Yeah, there's an article on Techstrong ITSM where they're outlining the seven steps. And a lot of it has to do with just fundamentals.

To your point, it's 3, 2, 1, you need to have three copies of somewhere remote and two local. Um, but part of the issue in my mind, and I'd love to get Mitch's thought here, is we've been talking about this issue for as long as I can remember. Disaster recovery shows up on the top 10 concerns for CIOs for the last three decades.

And I wonder if we're actually making any progress despite, you know, the advances that in technology that Steven outlined. So, Mitch, where are we? What makes recovery hard?

Well, I think it's, if the, it is like a moving bla right? That you're, you're trying to build while the plane is moving. 'cause it's constantly changing.

I think one of the hardest things is just keeping your plan up to date. Um, both with changes in your environment and what you're working with, but also technology changes too, right? Um, you know, using Veeam to copy your data off to multiple offsite, you know, centers for Recovery was a great strategy 10 years ago.

That's a still a good element of a strategy, but it's not a full, full blown one. Um, I think a lot of it is in my experience is having somebody like Steven who is dedicated to working the disaster recovery plan. Because if you don't, that's tough for a smaller organization to do that.

But if you don't have someone or a small team focused on it, um, it's one of those that never gets done. It's like, you know, cleaning out the garage and organizing your, you know, tool cabinet. I'll get to that someday, right?

It's one of those kinds of things until, oops, when it's too late, it's way too late. So I, I think it's on the, it's, it's on the mind of the CIO 'cause that's the person who's going like, yeah, but what happens when it does happen? Right?

Uh, we can't drive into the site to open up the doors to cool the data center if the power goes off or whatever, if we have a failure in our, in our hvac. So I, I think that's number one is having some people who are well qualified to put this together. And second, it's just making it very practical.

It's really easy to make these plans super detailed and super complex, um, and kind of get lost in that detail. At least my experience. I, I had to go through a refresh.

Not anything is significant as I think what Steve's describing, but it was really simplifying and modernizing it, right? We're not gonna send out text, we're not gonna send out pager messages, right? That was back when it was written five, 10 years ago.

Let's update that to modern communications. What is the real process? And getting real clear on that so people know where is it?

What do we do? We don't have to find it 'cause we can't find it. 'cause the dust is so thick on the plan.

Let's make sure we're ready for action and do some tabletop top exercises. So I, I think we're avoiding the elephant in the room and that elephant's called climate change. Okay?

When you have 75% or more of your population living with, what is it within 50 or a hundred miles of the shoreline, right? In this country as well as in many other parts of the world, when you are building data centers where the population is next to the coastline or in extreme places to take advantage of environmental factors, hydroelectric power, uh, colder temperatures or what have you. And then you bundle that with climate change, pressuring, creating warmer waters, more extreme storms, ice glacier melting and, and seas rising and so forth.

What do you expect's gonna happen? I mean, this, this is this, all of the disaster planning in the world. And, and to Steven's point, we have better DR tools now than we ever had, and we have much more capable of hot swapping and, and stuff like that.

But the fact of the matter is, this is about climate change. And if, you know, I would hope everyone out here acknowledges climate change and that we have to take steps, either climate proof, our data centers and, and disaster proof our data centers, you know, build them. I am reminded of, uh, the, uh, the old, if you've ever been to old IBM data centers back in the day, um, these things were built to withstand nuclear war, right?

I, I remember we an inter reliant in the late nineties. We were looking at buying an IBM data center upstate New York. This thing was like built into the mountain and no amount of storms or, or wars were going to affect it had its own internal generators and everything.

But, you know, climate change is real. And this is just another way of, of that it's affecting us, right? I, I live somewhere close on the water and, and I see it when we get king tides and stuff like that.

That water's coming over. You're gonna get storm surge. We have great building codes down here in Florida.

All the buildings post Hurricane Andrew are built with, you know, uh, rebar reinforce or steel bar reinforced cinder block. So even 150, a CAT five hurricane is not necessarily gonna knock the buildings over, but that doesn't stop the water from flooding. Yeah.

And that's over time, you're exactly right, because we get sunny weather flooding, like you were mentioning. Mm-Hmm. And then there's the cumulative effect of that, of just the constant wind and not just, you know, from a hurricane, it could be a nor'easter up in, in New England, So, yeah, no, well, they get, yeah.

'cause they don't have the building codes up there, right? So a good nor'easter does wreck wheat on. Well, to that point, Alan and Monte, um, we had, you know, some significant fires in Colorado a few years ago.

Actually, the business that I was running it for, I had left by then. But, you know, fire literally went up to that building built, burnt down several homes and buildings around that, that was due to high winds, you know, another climate change, climate issue, right? Whether you can say is climate change or not, but certainly the dryness and all of that is a part of climate change.

So it's something that affects everybody. You know, the closest coastline to Colorado is what California. So we're a little ways away from, from something hitting that way, but it still comes up the, you know, up the continent.

And we have other issues that happen too. So, you know, and in response to this then is do we have to rethink disaster recovery in a, in a, you know, a climate changing world like this? What does, what does that mean for your disaster recovery plans?

Yeah. And it's not like a once in a million, uh, lifetime thing. At least things are happening more often, Steven, When I talk to people, I get this spot that says, yeah, we understand VMs, we got that.

It's in the cloud, but they can't figure out how to actually rehydrate the application stack that sits on the VMs. And that stuff is overly complicated. There's too many dependencies and nobody actually documented anything.

So, Yeah. Yeah. That's one of the things that we learned when we were trying to do dr.

Um, the only way to make sure that your DR plan is gonna work is that if it's not a DR plan, um, essentially if you have a continuously available application that can run in multiple locations and that you're active, active in multiple locations, that's the only way to make sure that it worked. The company that I worked for, uh, again, this was, uh, 20 years ago, we decided to, uh, initiate an active, active inside, uh, Texas and outside Texas so that the system was constantly running in multiple locations. And we knew that if the, if the system that I was responsible for went down, it would continue to run because we did it literally every single day we would do failovers.

So it wasn't something where we were making, um, emergency plans and, and, oh, no, pagers don't work anymore, or, you know, we can't drive to the data center because it's flooded, because we literally did this every single day. And that reminds me of modern application design, where, um, most modern web applications, I mean, remember in, in the web, uh, we have, um, availability zones and, uh, geographic dispersion of applications. Most modern web apps are gonna be a lot more resilient to, uh, interruption in a specific location.

But of course, as we saw, um, uh, sometimes, uh, for example, I'm thinking about fire in France a few years ago. Uh, sometimes the web can prove remarkably not resilient in the event of a disruption because it turns out that, uh, there's unexpected, uh, locations of, uh, of critical infrastructure. Yeah.

To your point, uh, we, we need disaster recovery is a 70-year-old concept, right? This is a sixties data center idea that we created and have lived with. And so you said the right word, which is resilience.

We need to stop thinking recovery and think resilience across our application stack, across our infrastructure, across geo, geo diversity, whatever that might be. It's a great place to apply. Kinda what you're talking about is, it's almost a form of chaos engineering, Steven, where you're like, failure is gonna happen.

Disasters are going to happen. So it's not about how long is it gonna take us to recovery? How many days are we gonna be out of business?

It's like, no, you know, you know what happens. So let's, let's, let's, uh, tabletop it, let's experince with it. Let's chaos engineering the crap out of it, because that's, that's our world.

That's the world we live in. And that backs up your point, Alan, of it's, it's not the elephant in the room, it's the elephant, right? Yeah.

Agreed. Hey, we're gonna need to take a little break here, not for disaster recovery or anything, but it's time. We're gonna come back and we're gonna talk about ML Myth.

Stay tuned. You're watching Touch Strong Gang. All right, folks, and we're back.

And we're talking about ML Lops, otherwise known as Machine Learning Operations. And I'm just gonna go on the record right now and say, you know, we're gonna talk about a subject that we've talked about before, but I have to tip my hat. Alan might've been right about this one.

All right, here we go, Mitch, there's a survey out and it says that the number of organizations that have a dedicated ML ops team is relatively small. And basically, it kind of tells us that everybody's pressing anybody who can do anything related to AI in the service. And so I sense is that there's a boatload of on the job training going on, and is there such a thing as a ML ops team, or did we just get everybody who knows anything about a model and put 'em into service?

Well, we, we do a bit of disservice to it by calling it machine learning algorithms, because that's only one third of the equation, right? You have definitely developing algorithms. You have also data preparation and data engineering, because you don't just stick a bunch of data into a, into some data store that you can access.

It's gotta be actually in a format that that particular model can access. It kinda like when we think of prompt engineering and how we, how we set things up in a, in a gen AI world, you, you do some other different things, actually call 'em feature sets, which is the combination of the code, the algorithms, the data that's been engineered for that application. And the third is domain expertise, right?

If, if you're talking about, um, something for, you know, oil exp exploration or analysis of weather or, um, uh, what's happening in our development cycle and how to create a product that makes that more efficient, you have to have to have that domain expertise. So throwing it over the fence at the developers and say, here's a new shiny object, go learn, uh, PyTorch, and you'll, you'll, you'll, we all have, we'll have, uh, machine learning. So I think if you step back and just say, okay, wait a minute, we have this whole new engineering process, um, of how we create models and the data and the domain expertise that goes into it.

How does that fit into our development cycles? How does that fit into our workflows? DevOps are not, but how do we do that?

Because data has a different center of gravity than code does. It gets changed differently and managed differently and maybe even released in, in different ways or in coordination with other parts of the system. So that's where ML ops and tools can come into play.

Now, not everyone ha knows enough or has been down that road far enough to say, you know, look, this is different. This is hard. We need to kind of think a little bit outside of the box of what we were doing and adapt to it.

I think that's what most, most organizations I'll bet they didn't start with an ML ops group. They learned the hard way that yeah, we got, this is a different challenge. This is a data management problem.

It's an AI problem, it's a domain expertise problem. I don't disagree. I'm sorry.

Go ahead, Mike. Steven, is this all coming back to centralized it and will there be a separations maybe of inference engines are managed by one team and training is done by somebody else? And are we gonna, 'cause we just don't have enough people, it seems like.

Yeah, I think it comes down to, uh, absolutely. Uh, everybody's gonna have to be an AI engineer in the future because everything is AI right now, isn't it? Uh, you know, that's the world we live in.

I, I should be point out though, I, I'm not surprised to see that only half of organizations have an ML ops group since, uh, the whole idea of ML ops is not even 10 years old. And the idea that, uh, somehow organizations would've a whole bunch of people that were trained and ready to go in this new group. Um, well, that's pretty new concept.

Uh, you know, I'm not super surprised by that. I'm also not super surprised that to see that this, uh, lack of machine learning expertise is causing companies to bring in people who don't have deep machine learning, understanding to build and, and maintain and operate these applications. Because, again, those people are hard to find.

Uh, there's, people are, you know, many people are still learning, uh, this stuff still coming up to speed. Uh, that's just the nature of, of it. But to your point, I, I do think that, um, over time we're gonna get to a point where, uh, machine learning, uh, operations is different from, uh, application development is different from building, uh, and training and so on.

And, and this is all gonna get a lot more normalized. Um, I think that we will have, um, m ML ops specialists, but I don't think that that's going to be a entirely unique silo. I think it's gonna be part of the overall operations and DevOps world.

And I think that the ones, the companies that are gonna be the most successful at deploying machine learning, uh, applications are the ones that take this, uh, seriously and take this the same way that they do all other IT applications and, and, and have people who specialize in application development on one hand and have people who specialize in, in, uh, ongoing operations on the other hand and have, uh, folks who can build and maintain platforms and, and, and kind of link those groups, uh, as a third community. So I, I, I think that that's sort of the nature of it, that we, you know, we, we kind of go off in these frontiers and then we build a more stable and sustainable, uh, system to manage them. So, you know, I, I think when you look at ML ops, though, to me ML ops is a way station on the way to ai.

You know, what, you know, what's gonna fix our ML ops problem? ai, right? I think we'll have maybe less ML ops engineers, but smarter ways of dealing with the data.

You know, you think back to when this whole big data issue first came out and Hadoop broke off. I think there was a bunch of guys from Yahoo went out and started Hadoop. And, you know, getting ahead around that much data was, was kind of mind boggling.

And, and today, the, the data lakes and data sets are so much bigger even than back that, and we've gotten better at it. But really, I, I think a lot of, and whether you call it ML ops or AI ops, I'm not a hundred percent sure what the difference is, but whether you call it ML ops or AI ops, to me, they're precursors to, okay, then can I have generative AI somehow work on that and, and not manage it per se, but give me actionable intelligence based upon what's going on with my ML ops and really have less people involved in the, the, the, you know, the guts of it, if you will, right? People should be the benefactors of it, not the, the, the worker bees of it.

So I, I think we are in a sort of very transitional, um, stage of how total AI solutions deal with our big data possibilities. 'cause it's not a problem. It's great to have all this data as long as we don't drown in it.

And that's what I was thinking in terms of ML ops. It's really about, um, the quality of the data, maintaining the data, cleaning the data, making sure that that data isn't biased one way or the other. So ML ops isn't just from the starting point of building the, the model, but it's the maintaining it through throughout and adjusting the algorithm accordingly.

Yeah. So I'll give you an example, exactly what you're talking about in the words of Mike Baard. I think you had it right, Alan.

Um, Google, uh, vertex AI does exactly this. It's a platform. They don't call it an, they may call it ML lops, but essentially what it's doing, it's, yeah, it helps with the collaboration on the front end, but it's doing the analytics on the back end of what are you learning through monitoring the model, learning diagnostics, you know, actionable explanations of what you could do with that information.

Uh, we're still probably at the beginning of this, but that's exactly what we should be doing, is leveraging AI to help us build better ai. Why, why would we throw people at the problem? We can't hire enough people that understand ml, uh, more or less enough to really kind of feed into the process, to, to support all the efforts we wanna take on.

You might not agree with me on this point, but I feel like specialization has been bad for it. And the reason I say that is, yes, it's great. Somebody decided back in the day, I don't know, I'm gonna become a VMware administrator.

And they got a raise 'cause they weren't your traditional administrator. And then over the years, we've just created all these silos around specializations. It's just made things more complex and more challenging.

And I know that people need skills to do that, but I don't know, Steven, what's your thought here? I feel like we're too far down the path of, you know, let me have a, a title for every silly little function we need to do, We need to get back to the world of, uh, um, uh, storage specialization. That's my, my suggestion as Mr.

Storage here. Uh, we need, uh, store ops specialists like me. Um, no, I it sounds like a campaign kind of.

I you heard it here. So Yeah. I, I think it's, it's the push and pull, you know, between specialization and, uh, and generalization.

Um, I, I wonder if machine learning operations is specialized enough to need specialists that are different from other application management long term, I suppose we'll see. But I suspect that machine learning may turn out not to be all that different, uh, from other types of modern application management and, and could be handled by DevOps people. Once every DevOps person is an ML specialist, maybe Mike, I would say as long as we have job titles, we will have specialization.

'cause that drives salary. So get rid of job titles. I think you'll lose at least some of the specialization.

You know, back in, uh, DevOps, when we were running, you know, co-founded DevOps Institute, we were doing the, uh, the, uh, DevOps, uh, employment job market skill. Excuse me. It was the DevOps skills, uh, report.

This was a, a, a very common thing, right? Do you want what we call t-shaped individuals, eye shaped individuals, or kind of broom shaped individuals, right? And, and clearly, you know, you want a certain amount of broom shaped individuals like out of your Phoenix project kind of learning.

You want some broom shaped individuals will be able to deal with a variety of different issues, bottlenecks, you know, areas. But as our, it becomes more complex, the level of specialization needed to do some of these things just lends itself to those t-shaped people, right? Because it's very hard for a generalist, you know, jack of all trades, master of not, and sometimes you need a master of something.

I wonder, there's a, there's a joke going around that says, you know, um, the trouble with AI and data scientists is they're all academics, right? So they go build an AI model for a company and they come back and report that, you know, every seventh day there's a drop in revenue and the business looks at them and says, of course, we're closed on Sundays. So, um, I feel like some of this specialization is so removed from the rest of the organization that it only contributes to this it versus the business divide we've been talking about for years.

And I'm not sure it's getting any better, Steven. Yeah, I got one more thing that stood out to me about this study, by the way. And that was buried in the study.

There's a mention that many businesses are rolling out, uh, machine learning and AI applications just because they can. And there was this, this suggestion that many of these applications aren't actually serving the business. They're just, you know, we gotta, we gotta get us some of that ml, Hey, how do we get that, that AI in here?

And so they're trying to roll something out that doesn't do anything. And, um, and, and that to me is really emblematic of where we are in 2024 when it comes to AI that, um, it shows up in a survey that businesses are just sort of running off half cocked saying, yeah, put AI in it. Put AI in it, that'll fix it.

Uh, spot on. I saw that too. You know, I think it's, it, it, we're talking about this as a binary choice, right?

Do we have specialization? Don't we have specialization? The fact of the matter is, as new technology comes on board or comes along, it may come, you know, heavily out of academia like we see with ai, but other things, just like moving to the cloud, right?

Not everybody had cloud skills or understood what might be different about it, how to secure it as it's the same or different. What do we do? There's an adoption curve, and I'm not talking about just the hype cycle curve, but there's an adoption curve of where it may reach a reach, a point where it is a general skill across a large number of people.

In the beginning, it's gonna be a few folks that figure it out. And you'll have a team that, you know, tries some of the first projects, and it's your job to figure out how you take those learnings and, uh, begin helping others come up to speed if it's something that's gonna impact a lot of people. So I think, I think the idea that AI is gonna make everything so general, nobody has to know anything in terms of specialized knowledge, is, is, I don't believe that's gonna happen.

'cause I think our world changes too fast and well, AI just keep up for us, you know, ahead of the curve for us. Well, maybe it's, you know, general intelligence and it can do that, but you know, we're, we're certainly not there yet. So I think, I think it's just be smart about it.

This is an adoption curve. It's not Do we need specialists? Yeah, we need specialists.

When we don't have that expertise, that's when we need specialists. I think it folks wanna stay current and I understand why, but yes, there was a significant amount of resume padding going on, and there always is. Every time there's a new technology, Like Mitchell says, as long as that's driving salaries, that's not gonna change.

Not gonna change. Anyway, I think we're about ready here for a break on Textron gang. We're gonna come back and talk about your trusty little barcodes.

They're 50 years old barcodes. Are they, are they still viable? Are they secure?

What are some of the implications? Barcode QR code is one better than the other from a security point of view? I don't know.

Stay tuned. You're watching Textron Gang Cloud native now is the web's leading resource for the growing cloud native ecosystem. com is your destination for news, thought leadership, features and webinars on cloud native architecture, Kubernetes, serverless, cloud native application development, microservices, service mesh, cloud native security, and more.

Stay on the cutting edge of modern application development at cloud native now. All right, folks, we're back and we're into our next segment. And yes, it's amazing that barcode that we all take for granted is 50 years old, and some of us remember when it first came out, so we're all feeling our age all of a sudden.

But Steven, what is your sense of the barcode? It's ubiquitous, but is it gonna be with us forever or is other technologies gonna start to supersede it? What's going on in the world from where you sit?

Well, first off, um, I think it's important to point out that barcodes in, uh, retail are 50 years old, uh, 1974, but barcode technology is older than that 1952 baby. Um, and, uh, and, and it's, it's so fun. Um, we're all nerds here, right?

Nerds about one thing or another, right? Well, there's a lot of crossover between, um, it nerds and radio nerds, uh, and trains. I mean, who doesn't love a good train?

Right? Well, that's where barcodes came from. Um, you've got, uh, you've got train cars, you've got morse code, uh, and that's what barcodes originally represented in the fifties.

Uh, they tried to have, uh, train cars, read information, uh, while they're rolling along the line. I mean, what could be more nerdy than that? And, uh, yeah, so as you mentioned, uh, one of the goals of, uh, barcodes, one of the reasons that they were finally commercialized was, uh, because, uh, when you go to retail, it's, uh, often, uh, a challenge to put prices on things.

Um, back where I come from in, uh, in, uh, Connecticut and Massachusetts, the law actually said that everything in the store had to have a pricing sticker on it. Barcodes were invented to basically address that challenge that you needed to have prices on everything you needed to be able to, uh, to scan things quickly. And, um, and barcodes met that.

But of course, uh, nowadays we're all used to seeing QR codes, which of course are, uh, something, uh, somewhat newer, but maybe not as new as we think. Uh, QR codes came into, uh, into production in, uh, the nineties about, uh, 20 years after the barcode. But, you know, barcodes are still here, and I think that they're going to be here.

There's a lot of, uh, inertia when it comes to technology, especially for the uses that we put barcodes to. I mean, thinking back to that original train car, the goal there was basically you had a linear system, a train traveling along a track, and you needed a linear code that could easily and reliably be read as it's, as it's going. Same thing is true of, uh, scanning barcodes at retail.

I mean, you look at that, that stupid self scan thing that they should have a person working at. And, um, and it's a little spinny thing that, uh, that reads the code that way. The same is true of, uh, many cases, uh, that you see barcodes in use in industrial and so on.

When it comes to QR codes, uh, there are two dimensional scan, um, basically a photo, uh, we'll scan it. Um, qr I would love to nerd out about how cool the technology is behind that thing too. But, um, the bottom line is a QR is fundamentally a different kind of code, and it's not useful in some of the situations where a barcode is useful and vice versa.

So I do think that we're gonna be stuck with barcodes here for a long, long time. You're making me nostalgic because 40 years ago I was the price changer in a supermarket. You know, I was one of those guys walking around with a little kaching thing, and you could always tell us, because we all put those things in our back pocket.

And then when you were walking around the street, we were the guys with the ink stain butts. So, um, you knew exactly who you were. But it's become interesting in, in the sense that I feel like it was easier to change the prices back then.

Now today, it's a whole computer exercise that has to be tied back to a barcode scanner. Um, Mitch, we're at this for 50 years plus, then we're just getting around to like, to the point where we think it's ubiquitous. Sure.

It takes us a long time to wrap our heads around innovation, doesn't it? Well, it's, it's one of those that, you know, if you, barcode is two things, right? There's a technology for analyzing what the barcode represents and then what the, what it represents is the universal pricing code, right?

So it's giving every product or every, every variation of a product, um, for whatever reason, it can be the same product that's in Costco that shows up in Walmart. You know, they put their own bin, uh, their own pricing codes on it, UPC codes on it. So that's, that's what it's really about is how do I translate very easily?

What is this product and then what's the pricing? And is there, you know, discounting and all those kind of things with it. We have this in other areas, VIN codes for your car, for your automobile, right?

I don't know when they were invented, but those things have been with us for, you know, decades and probably will be for a long time. I think it's one of those, if it don't, if it ain't broke, don't fix it. Well, maybe you improve it with QR codes 'cause UPCs you're gonna have to pay and register and all that kind of thing.

Anybody can create a QR code with a website that does that in a, in a URL or would you, whatever you want to translate to a, you know, business card address, all that kind of thing. So I, I, I think it's something to celebrate. You know, we, it, we, it was a right thing at the right time.

It fed a need. Think about how much of our economy is driven around barcodes and our ability to, whether it's us shopping for ourselves or, you know, really getting the right price for something does totally solve that, but at least hopefully we have a better chance of getting the right price for product. Is this the same product as something else?

So will, will there be something better to come along? I don't think it's QR codes, and why do we need something better at the moment? Let's solve another problem until that one isn't working for us anymore, and then somebody will come along with a better idea.

I, I think there's another aspect here that we haven't touched on, and spark codes are not just for pricing. You think like what drove, what's the biggest employer in America today? Walmart, right?

What made Walmart great? Well, they had cheap prices, you know, and they worked on a very, like that low model of, uh, margin. But what really helped Walmart is they had the best inventory control of any retailer ever.

And a lot of it was driven by those lowly barcodes. They knew from there using their barcodes exactly how much stock they had on the shelf in the store, what was sold, how much stock was back in the warehouse based upon when that barcode got scanned at the register. They had a system, okay, we only have three units left, we need to send some from warehouse to store, 86 and, and so forth.

It really drove, enabled this whole inventory control just in time kind of, uh, system that made Walmart and, and subsequently most other retailers. Great. Amazon uses it too.

Um, so it, it's more than just making it easy to get price or to do the self-checkout and, and make cashiers scarce. Um, it, it's, it's really an inventory control system, but there is an, an implication we haven't touched on, and that is security, right? It's not hard for me to go take a barcode from a, let's say a cheap price product.

Maybe I I print it on a little portable printer and, and stick that barcode on a higher price product and just scan that barcode, add a at a self checkout, what stops someone from, and I hope I didn't give anyone any ideas. Oh, but honesty. Yeah.

What stops people from doing that, right? What, and there, I mean, that's an obvious, to me anyway, an obvious security implication. But, you know, what are some of the other security implications?

It's not exactly encrypted, Right? Some people were talking about going back to, uh, cashiers because a lot of people are quote unquote scanning stuff at the counter there when they're like missing every fourth one and they did the math out, and then it's cheaper to have the cashier based on what we're loosened on wasted. That's really the human factor.

I'm glad you mentioned that, Mike. And, and, and also Walmart, because Walmart's also known for their greeters and their interaction with customers. So maybe they went a little too far with having the self-checkout, because I've read that they, they're pulling back on that people wanna have that interac.

Well, people don't. Yeah. And the same thing also during, um, I was in New York during Covid and every menu, you know, you had a scan, right?

And that was not well received. And now, uh, not even before the pandemic was over, I can recall going restaurants and they said, no, we're giving menus. People wanna hold the menu.

They don't wanna scan anymore. So it's interesting how valuable the barcodes and, and the QR codes are, but there is that human factor to see how much people will take of it and when they're gonna push back on it In, in defense of the QR code, the, it's not the code, it's the problem. It's the application written around the code.

That's the problem. I was just, you know, using one at the movies this weekend where I went to one of these theaters that has, uh, you can order dinner and you're presented with a QR code. So I got to the app, but the app was such a nightmare to navigate that.

I think I spent 10 minutes trying to figure out how to order, you know, two hot dogs. It's just crazy. So true.

Fair enough, fair enough. Hey, I think we're going to wrap up our, uh, barcode segment here on Textron Gang. Guys, I hope you've enjoyed this Tuesday, uh, show as usual, we have a full day of Textron TV behind this.

We've also got some exciting stuff I think tomorrow, Steven, you, we have a, uh, tech Field Day event, don't we? Yeah. So, uh, Wednesday and Thursday of this week, uh, we're gonna be, uh, doing networking field day, which is, uh, really, uh, a lot of fun.

Um, we've got presentations from a number of different, uh, uh, companies in the enterprise networking space. Uh, we've got, uh, acus Hedgehog, Intel Selector, AI and C Packet presenting. And those are all gonna be live streaming right here on, uh, tech Strong tv.

com. But, uh, you'll be seeing these, uh, tech Field day presentations all over the Techron network and, um, they've been really well received. I'm really glad to have so many great people watching.

Uh, thanks for being part of it. Yep. And if you're interested in Tech Field Day, you can watch, uh, past Tech Field days on YouTube, on the Tech Field Day channel.

I think they might be on Textron TV, or at least some of the recent ones we've been involved in. Uh, but they're really a great event, you know, the delegates and the companies presenting. I highly, highly recommend that that's Wednesday and Thursday.

Don't miss it. Um, I don't think Barney, anything else, Um, for Textron tv? Yeah, well I have some stuff coming on too.

Well, Well, let's hear About it. Yeah, we have some, uh, good, uh, interviews and some analysis. I'm gonna be doing something this week on the right to repair, which is, uh, interesting because it's coming up more and more where people wanna have the ability, ability to prepare their own devices.

And I talk about different companies, the tech giants, how some of their takes on that whole process has changed. Excellent. Alright, Mike, Mitch, Steven and Bonnie, thank you for joining me.

I'm Alan Shiel, you've just watched Techron Gang. I'm Bonnie Schneider, sustainability contributor to the Techstrong Group. I'm excited to introduce you to a groundbreaking new initiative from Techstrong Research, the sustainability pulse meter.

The pulse meter offers valuable insights into how environmental responsibility factors into tech purchasing decisions for key players in the industry. Position your company as a leader in the industry and differentiate from your competitors with a sustainability pulse meter offered exclusively from Techstrong research.