Techstrong Gang – August 13, 2024

Hey, everyone. Happy Tuesday. We've got a whole bunch of great stuff to cover today, and we've got a great gang to do it.

We're looking at trans digital transformation, cloud security, and even a little ai, believe it or not. What's a show without ai? You're watching Textron Gang.

Hey everyone, it's Alan Shimmel here for Textron. Welcome to, uh, Tuesday's edition of Textron Gang. We've got a great gang light up today and some juicy, juicy topics to go over.

Let me introduce you to our gang members for today. First of all, joining us fresh off his vacation Fresh is a daisy. I dig his shirt.

Looks a little cuddle fishy, I don't know. But anyway, it's our good friend, the CEO and founder of Tech Field Day, as well as member of the FU group, Steven fst. Hey, Steven, you look great.

Well, thank you. I've been in, uh, Denver and, uh, and, and beautiful Kansas City. Uh, last week I got to go to the, the share, the longest running tech conference in the universe.

I, I've been to my shares over the years. That's a great show. How was it this year?

Um, it was a little smaller in this time than it was in March, but even so, it was, it's, it's always really cool because as I like to say, it, it's like going to any tech conference except all of the nouns are different because it's all mainframe stuff instead of open systems and cloud stuff, but they're talking about all the same stuff. Digital transformation and AI and DevOps and, I mean, everything will be familiar to the audience here on Textron. Absolutely.

Well, and we also know you were away on vacation before that, so welcome back and happy to have you on today. Those, Those, those main framers, like the French, they got a different word for everything We well did, but maybe they put on a good Olympics. Um, joining us also on the Gang today is our CTO and futur of CTA.

Mitch Ashley. Hey Mitchell, welcome back. Hey, welcome.

From Denver. Still decompressing from Black Hat, but you know, well, I'll I'll be back to reality Soon. Yeah, well that is black hat for you.

Alright, joining us here live in our, uh, studio at our Booker Ratone headquarters is our sustainability, echo Insights analyst and editor Bonnie Schneider. Hey, Bonnie, how are you doing Great, Alan. Welcome.

Thanks for being here. And then to Bonnie's left my far left, still a resident of Florida, maybe at least for just a little while. More our chief content officer, Mike Ard.

Happy to be back. Always happy to be in Florida, but you never know, One never knows. Sometimes home you can't go home again.

There you go. Um, anyway, we've got some great topics too. I wanna kick things off with an article, well, coming out of an article over in our di digital CXO site.

com and Security Boulevard and Cloud native now and Techstrong TV tech, strong guy. But you know, for digital transformation for leadership, our digital CXO site has been laying down the wood now for probably two to three years. Great site.

If you get a chance, check it out. We're going to take a look at a recent story there on digital transformation revisiting digital business transformation. The in the notes will be the link to the article, but Steven, you're you fresh off Cher, you were talking about how even the main framers are in a transformation.

What do you have on this? Yeah, I think that sometimes people give, uh, they, they, you know, things like digital transformation can seem almost like, uh, like a just buzzword bingo, you know, and people are like, ah, you know, what is this? You know, but the truth is that it makes a lot of sense.

And, and you know, you mentioned share. It's funny, uh, we had Broadcom talking about this just like in one of the articles that, uh, that you guys have provided here, but we also had BMC, uh, talking about metrics. And really it comes down to the question of, um, how can you prove that you're doing what you're doing?

And I know that it's a little unsexy, you know, especially for techies to talk about metrics and to talk about value in business transformation. But it actually makes a lot of sense. Um, you know, if you look at this digital CXO article, um, this is really actually a nice primer on what to look for in terms of metrics, in terms of measuring the value of all the things you're doing.

And yes, some of it is a little bit eyerolling to the techies, but bear with me, techies, bear with me. Because ultimately we have to prove that there is value to the things that we're doing. And the way to prove that there's value to business people is to prove value using metrics that business people understand.

And so we have to look at things like customer satisfaction surveys. We have to look at things like, um, performance, uh, net promoter score. I know, I know, bear with me now.

Resource utilization, quality, financial performance. Because ultimately everything we're doing in business is supposed to be related to the financial performance of the business. And it's easy for us to sometimes take our eyes off that, but we have to think about it in these, in these terms.

Honestly, the best thing that I ever did career-wise was take a bunch of courses in business school that helped me to understand more about what the rest of the business was talking about. And it let me translate my, IT work into words that the business people can understand. And I think if more of us, um, thought like A CXO instead of like an IT person, that we would be more effective at our jobs and, uh, we would have a better career path.

And so, so I definitely think that that, you know, take, take a look at this article, take these metrics to heart, uh, don't be too cynical about them because really there's actually some really valuable nuggets in here in terms of proving out the value of the things that we're doing already. I think we're kind of gone through a cycle, and I'm not sure if we're at the end of the second cycle or starting a third cycle here, but the first cycle with digital transformation was all about, you know, new use cases, emerging technologies are gonna drive additional revenue. And then Covid hit, and then it became all about optimization, uh, business processes to drive more value out of it.

And kind of going back in and re-engineering what we already had. Now we're on the cusp of a bunch of emerging technologies, including your favorite new one, ai Mm-Hmm. And, um, it feels like we're on that might be driving a whole set of new use cases that are gonna feed back into these digital process transformation initiatives.

And to Steven's point, if you don't know what the business side is talking about, you're never gonna get the budget for the IT thing that you wanna do in the first place. Yeah. Just to note we made it 10 minutes.

There you go. Before you mention that word, but, um, two words. So I, I do agree.

Yeah. Well, yeah, I do agree though that there, we do seem to be, have moved into a next generation of digital transformation. I think originally digital transformation actually started kind of with the internet, if you will, right?

We started moving everything onto a online connected world. I think during Covid was really the second, uh, generation where it wasn't just a nice to have or something I'm going to do someday. It was, if you don't do it now, you're going outta business.

And, and things that we didn't think we're gonna be going digital went digital a lot faster than we could have imagined. But I think what's happened now, and I was reading an another article about this on LinkedIn this morning, is we're not, you know, in the famous words of Tony Montana, we're not the only dope dealers on the block. Our competitors, our customers, our neighbors, friends enemies have also undergone their own digital transformation.

And now just doing that kind of perfunctory digital transformation, moving stuff onto the web and enabling, you know, transactions and and business to be conducted digitally, like that's not quite enough because other people have done it too. And now we could go faster, we could go smarter, hopefully we could go more secure and now add that bed word AI into the mix and it, and boom, it's, it's like a catalyst for even more. So I, I think we're on the cusp of, of yet a third wave of, of digital transformation.

He Don't say Mitch, I love to get I think so Too. Go ahead, Mitch. I'd love to get your opinion on this other thing that Broadcom is up to with, um, value stream management because it seems like we've had all these islands of automation over the years and now we're trying to stitch it together for some sort of value prop.

And we, I know value stream leans back to, uh, lean manufacturing as it were, but not a lot of folks are always familiar with that, especially digital cxo. So, you know, can you kind of, how do you measure all these investments? 'cause I know you talk about this all the time.

Yeah, it's, I was gonna say, I've been talking about this for a long time. You know, value stream management coming on a lean has having a little trouble getting, getting foothold. There's been a lot of companies start down that path and not necessarily succeed at it.

And, and it's because it's hard to fund itself, right? And trying to look at the whole sys part of the system and what, what things are adding value and what do we mean by defining, adding value? You know, I I think to make it a little more simply as it organizations, technology organizations, we tend to be guilty of creating metrics that were meaningful to us, but are in nothing to the business.

Like, how many deploys per hour can I do? Well, who cares? Why do we need to do you know, how many, what is the right number?

What, what do we, what do you want? Well, if what I preach, what I talk about is always create metrics that are line of sight to a business metric. And that's what's fantastic about this article.

This is tons of business metrics, customer engagement, satisfaction productivity, operational efficiency. There's a bunch of 'em. And you'll hear, you'll see buzzwords in there.

You might have heard before that were business metrics, but as technology organizations, it's a is too easy to say, well, everything in debt we do helps the business. Well, maybe not. Maybe some of it is just inwardly focused and we don't know.

The business doesn't know why that really helps us. So as I think as tech leaders, not only setting the group goals saying our objective is we need to be able to deploy twice a week because we have a product coming out that they need to experiment and market and they need to make updates to that product every, every week, maybe twice a week so that they can quickly react to market conditions. 'cause it's a very competitive or very new market.

Okay? That's the reason why I need to do multiple deploys per week or whatever it is. But just doing it because velocity sounds good.

It's not a great reason. Yeah. And to the point of this article with, uh, Broadcom that you mentioned there, Mike, um, with their value stream management platform, um, you know, this is, this is, uh, what we saw at share.

This is the kind of things that we're hearing from companies like Broadcom and BMC, especially now that, uh, you know, VMware is in the mix there. Uh, companies are trying to come up with platforms that they can get objective metrics out of, uh, this digital world and prove the value of what these applications are worth for the business. And I think this is something we're gonna go into later with Bonnie and, and the, the whole, uh, AI world.

I mean, how do we show real value from the all the, the, the time and effort and energy and, and literally, you know, atoms that we're putting into the compute infrastructure. How do we, how do we make sure that we're delivering value from that? How do we show that value on an ongoing basis?

And how do we communicate that value to the business to make sure that the business is good with us investing all these resources in these areas. Um, the Broadcom, you know, this, this, uh, value ops, insights, uh, platform that they're working on, uh, I'm not sure I haven't used it. Um, I read the article about it just like anybody else could, looks good, uh, looks like a good idea.

Uh, but you know, ultimately the proof is in the pudding. The challenge for a lot of this stuff, what I found in my consulting career was that it's, um, it's, it's hard to come up with a model and a metric that is acceptable to the business. That doesn't sound like you're just pulling things, let's say, out of the air.

Uh, and to make sure that it's really a metric and a model that everybody can buy into so that they're on board with the results that you're trying to demonstrate. In other words, uh, you know, if this value ops insights platform is accepted by the business as a source of truth or a source of information at least, uh, then, then it could provide real business value in terms of demonstrating that we in it are actually doing something that's useful to the business. But first I think that you have to get buy-in on the model, and then you need to continually develop these reports and kind of show this value.

I am optimistic about one thing. It may not always require a big giant platform to track this. We're seeing, uh, project management app vendors are embedding vs capabilities or the tracking capabilities directly into the software itself.

So I don't have to go get this whole separate platform thing. I can now kind of manage that directly alongside the project itself. So I kind of like that idea, but we'll see how things align.

Yeah. Yeah. I, Oh good, Mitch.

I think there's one key ingredient that if you miss it, it's all for not, and that is if you don't understand what the business goals are, but the business metrics are all the metrics in the world, in the technology organizations, we don't know whether they're the right ones. I have this test and it's the, and because test and because we did this, this business outcome happened, right? This impact for our customers happened, this result happened.

And then the business has to agree with that. They have to either, it has to be that metric that you can say that's the business metric right there and give we an increased gauge engagement by 10%. Or here's the line of sight to that engagement metrics and how we're contributing to it.

Even though we say, here's the value we offered as a technology organization, if the business doesn't agree, it's, it's all for Naugh. I think we have to have alignment and agreement when we do demonstrate that we're contributing to it. And because this happened, You know, if you're gonna measure flow, you know, value stream management, you've gotta, everybody has to agree on on what the, the key metrics and how you measure it and, and the numbers in there.

Um, Well, I mean, Bonnie, you know, I I love asking you this question though, but does it not occur to you sometimes that, you know, we've been investing in it for 20, 30 years and here we're talking about getting the value prop out of it. Doesn't that seem like we're a little kind of horse cart before the horse kind of thing? Yeah, I'm not, I'm surprised that, hasn't that come up before?

Well, no, it has, right? I mean, historically it was looked at as a money. Oh, right.

You, you invest in it, you don't, it is not a profit. It's an expense. It's an expense.

And, and I think, you know, putting it on the right side of the balance sheet is, is relatively recent anyway. Business trans digital transformation is real. And do check out digital CXO, it's a great resource.

We're gonna take a break here on tech strung rank. We're coming back to the trouble with cloud security. Sounds like a Star Trek episode Mitchell, doesn't it?

I do. I think they had that problem. The Trouble with Chiles.

You're watching Tech strung gang. All right, folks, we're back. The Black Hat conference was last week and a lot of things got overshadowed.

And IR would argue one of the things that we didn't talk enough about was cloud security and Cloud Security Alliance folks have a survey out that kind of shows that while in between the platforms, the people and the processes, the problem is the people. And there was also a report outta Black Hat at the end of the week where Aqua Security was talking about how they discovered all these flaws in the AWS cloud that were just traced back to, not technology, but just the way things were configured in the cloud. And AWS fixed them.

And probably other cloud service providers have this issue as well. Um, we've been talking about this for a while. DevOps folks have a role in this.

'cause sometimes the developers are frequently misconfigure the cloud services 'cause they don't have the right level of training. How come we're still talking about cloud security 10 years on now? Well, black Hat was last week.

Um, you know, there's always, that's the time we always see a good breach right before Black Hat. And we see every company, you know, who's there, has a solution, has some new solution. You're not accusing people of holding back in Info.

Yeah, no, no, never. Never, never. You know, so, so here's the deal.

Cloud security, the biggest inhibitor to cloud adoption. Mitchell, how long did you and I hear that right? Security from day one?

From day one. That was the issue. And, and the fact of the matter is, I would say stuff is probably more secure in the cloud than it is not the cloud.

I think the AWSs and Googles and Microsoft of the world have put more resources and making the cloud infrastructure more secure than any single entity could, right? Any company going by itself isn't gonna be perfect. No, it's never perfect.

Um, but, you know, we have all of these security boulevards, black hat, uh, notes in, in, in the notes here, articles that came out last week. And as I said, every company adds their little piece to the pie. And, and together, you hope it's a symphony and not noise.

And that, I think for folks watching out here, that's your assignment. Is it a symphony? Do you, do you see a melody here or are you just hearing noise?

Uh, I'll, I'll give you my takes on some of these clouds are the cloud security alliance. You know what our good friends, Mike Rothman, rich Mo, rolled out the next version of the cloud security, uh, training classes that they've been writing for years. And Mike and Rich actually taught it there.

That a great response. And it is updated with a lot of these new things. Uh, our friends at Aqua are crowing that they're part of, uh, a named vendor in Gartner's Cnap Quadrant or something.

And you know, and they, they have a great research team. How critical were these flaws that they found? I, I don't know what, you know, I'm assuming they did Responsible disclosure.

AWS fixed them. Um, RAD security is great. Cloud security company have been for a long time.

They fly under a radar a lot. And you know, to me this was, guess what we're adding ai, congratulations to RAD security. Um, our good friends at sig, they look, they've been doing great stuff at identity with identity and cloud security for years, and they're continuing along their way.

Uh, you know, or another one, our friends at Orca, Harvey Schuler and them great cloud security. So these are sort of the incremental evolutionary kind of watching the paint dry. I am, I'm optimistic about one point on this.

So if people are the problem, and we do have AI and we have algorithms crawling around, we should be able to find the mistakes that people made faster. We should be able to identify the configuration issues. And to me, much of the problem has historically just been this kind of shared responsibility model.

Sounds good on paper, but in reality, people who aren't security experts, it's hard for them to assume that responsibility without some help. So I think AI is gonna give them some help and maybe even that out a little bit in a way that could be interesting. But Mitch, I don't know, what do you think?

Yeah, I think AI will definitely help. I think platform engineering is a help. And one of the things that we, we, we suffer from is we still use the same kind of model for how we put together infrastructure and technology.

It's like an erector set. You take all the pieces and you start to build the platform that you want, right? And that's where a lot of the human error comes in this misconfigurations, because okay, I need an S3 bucket, great.

Did I do all the security on it? I don't know, but I got an S3 bucket. I'm moving along, right?

And if we have predefined, we know our secure platforms, configurations that we use, et cetera, that's one good step to, to helping. Also, it makes it easier to update those things if the, we have 25 of those instances in use. We know how to update 'em, roll out that update versus 25 different ways of doing it.

It's interesting, we're, I'm doing a, a study on application security right now. That'll be out shortly, probably about a month. Um, and, and, and the no two biggest, uh, security concerns in the cloud are security depth, all the vulnerabilities that haven't, we haven't got to fixing yet.

And, and misconfigurations human error. Uh, a Alan and I had a great conversation, as we always do with Brian Cox from Fox Type, Fox, Fox, sorry, it's, it's, it's Morning Brian Fox. Anyway, he's huge in the open source world Maven, uh, artifact directory for, um, Java.

And we were talking about how how much log for J is still out there and what was his estimates, like a quarter 20, 25% of are still using vulnerable log for J. Yeah, it's still out there. So some of it isn't even in the misconfiguration itself, it's in the human processes of keeping it secure and keeping it updated.

So again, if we could turn AI loose and finding those problems, Mike, you're talking about, I think if we could even turn AI loose on our technical debt, going out and patching and testing and doing some of that, I hope that's the real cringe work that we don't like to do that would help us. Well, Steven, you're an expert in hu human error. So, um, you know, what's your take try Just because of his shirt As aer, but you know, you talk it, I can't compliment this.

Steven, You, you talk to IT people all the time and, um, you know, to Mitch's point, are they just trying to, you know, happy just to get the thing up and not worried about the security? Or are we just winging it too much? I think people are winging it a lot.

Um, I, you know, I think that the dirty secret that a lot of us at techies know is that there's a lot of copy stuff out of, uh, stack Overflow. And you know, there, there, there's a, you know, there, there's, uh, a lot of things that are, are, are just sort of, you know, ah, it looks like it works, let's go with it. And so I'm, I'm with, uh, I'm with Mitch here thinking that, um, you know, as, uh, products can become more automatic, that's gonna be better for everyone.

Uh, one of the things that I'm gonna call out here from the, the links provided, number one, um, the, uh, cloud Security Alliance, uh, analysis, uh, one of the things that really stood out to me was that, um, the things that people used to be worried about when it came to cloud security, like, uh, DDoS like, uh, vulnerabilities because people are sharing the same infrastructure, um, like data loss. Those are kind of fading into the background to the extent that the CSA isn't even tracking those in this study because they found that those aren't even, uh, top of mind, they're not even like middle of mind now for companies that are using the cloud. That really stood out to me.

And I think that that shows, not that those are solved problems, but those, those are problems we're all working on problems we're all aware of, and problems that are increasingly fading to the background as we're concerned more about things like ai. And of course, you know, the ever present threat of buffer overflows and so on. The other one that I wanted to call out here from this, uh, that I thought was really interesting was from Orca, as Alan mentioned, you know, this is a really clever group.

Uh, the fact that they're specifically saying that the attackers are using DevOps methodology, and that's one of the core problems here. I mean, why did CrowdStrike happen? It happened because we need to be remediating in real time or else the, the bad guys are gonna beat us because they're out there, they're deploying multiple times a day with new attacking tools.

They've got a, a, you know, a dispersed organization that's working on different angles. You know, basically they're doing DevOps against us. And so we have to be responding quickly.

And unfortunately that really bit us on the butt when it came to CrowdStrike. But the, the bottom line is we have to be there ready to accept DevOps threats. And then another one that I want to call out here that's not part of this, uh, actually a couple things, uh, that really struck me from Black Hat number one, um, we had a one password vulnerability that was responsibly disclosed, fixed, and then properly announced at, uh, black Hat.

Awesome, awesome work. Uh, everybody involved there. And then the other one that I really love is CloudFlare, that's done basically trying to do exactly what we're saying here.

They're trying to make more of what their security efforts automatic. So basically it's on by default and it automatically configures itself and it automatically does the thing That's, I think ultimately what's gonna happen here is we're gonna get more and more remediations that just sort of happen. And essentially we're gonna let these companies, to Alan's point, we're gonna let these companies master their domain and do their job on our behalf instead of constantly making it something that the customer has to turn on or opt in.

com. So there, oh no. Well, Readers are readers.

Eyeballs or eyeballs, but you know, Steve, let me riff on that, Steven, because I I couldn't agree more with you and I want to, and on, on some of the points you made there. First of all, I've always said, and for a long time now that the killer security app or cloud security is IAM identity and access management, that's where the, that's the main battlefield today in cloud security. Yes, we have the things Steven mentioned that we had, you know, from the beginning, but they have faded in the background.

We've gotten better at them. IAM is is where the seems to be the, the battleground there. Um, also, uh, uh, on our friends from, from Orca and, and what they've done, visibility's important.

But Steven, you mentioned something else and that was that one password incident. And you're right, it was a picture perfect response to a breach or to a vulnerability maybe before it became a breach. And you know, Mitch and I, and some of, a lot of you out there have, we have a good friend Chris Hoff.

Hoff is the, uh, big F at ciso or chief security person at one of one password's, biggest competitors. And, uh, you know, even Chris came out and said they handled that picture perfectly. Right?

And that is a model. Another person who was at Black Hat was our friend George Kurtz, right? CEO, founder of CrowdStrike.

And you know what George got up there on the CIO panel, and before he even said anything, or the Cecil Pal, before he even said anything, he said, look, I wanna apologize. Did it like lurk in the background, did it, you know, held his head up, came out there and said, Hey, we messed up this on us and we're we're gonna fix it. These kinds, this is what the security industry needs.

Yeah. And this is what makes it better. Mitch, Um, and Steven kind of alluded to this, but I'd love to get your take on this.

'cause I feel like the whole game is changing. It is now, uh, instantaneous near real time, uh, detect and response. And if you don't, you know, the bad guys are moving in real time.

So if you're not, you're gonna get creamed. So has the whole game changed? It?

It's flipped to 180 in some ways. We, we've always invested in defense, but now we realize the response is equally, if not more important. So we're investing greater amounts into response.

And what I mean by in, in an incident, not only what our process is, but uh, uh, table topping gaming and making sure that when something does happen, because it has and will happen again, uh, we can respond to it quickly. You know, in our previous segment we were talking about, um, metrics and one of my arguments for multiple deploys per day, per hour, whatever it might be, it's not that you wanna do it every day, it's this incident. It's a security incident that I need to get something into production quickly.

And I don't want the one time that I need to do it, be the one time I've ever done it. I need to know that I can reliably do it without making everything fall down. Exactly.

So there are things we've changed about and, and just the attitude of it isn't if I will get hacked, we will all, and we have all had security instance and we'll continue to. So we have to be really good at response to it. And the transparency, I think is, is something that we're applauding when we see it done really well, which is great.

'cause then if, when we need to do something along those lines, um, we don't do what OpenAI did, which is wait six months to tell people we had a a breach. Exactly. So Being more accountable, like You're Saying with CrowdStrike accountability with CrowdStrike, I'm one password that is the key.

You can't hide your head or, or think you're going to get away with it or whatever. But Mike, to your point, I think what you asked Mitchell is yeah, we need to move cloud security at the speed of cloud business. And and to do so, you probably do need DevOps and AI and, and some of these technologies, but we are, and we will.

So I'm optimistic. We're gonna take a break on that here on Techstrong Gang, we're coming back. Bonnie has a nice, a nice, uh, piece of work she's been working on.

You're watching Textron Gang, Discover Textron Group, the epicenter of tech innovation. We are your go-to for reaching IT leaders and practitioners worldwide. Our secret impactful content that sparks awareness, engagement, and top quality leads with us.

You'll access editorial websites, streaming videos, virtual events, custom content analyst research, and more. Join our satisfied clients. Let's revolutionize your tech journey.

Contact us today and tell your story to the world in the most powerful way with Textron Group. While we may be used to using AI for Siri or Alexa, but the open AI has something different now with chat, GBT, it's a brand new voice feature that's only available to some select users, but I decided to take a deeper dive into it and show you not only what it is, how it works, but also the energy that it might consume. Hi, I'm Bonnie Schneider with your Ecotech Analyst Insights.

Imagine talking to an AI model that feels like you're having a conversation with a real live person. That's the essence of open AI's advanced voice model for chat CBT. This innovative feature is available to a select group of chat GBT plus users with a gradual rollout anticipated in the fall.

Envision spontaneous natural conversations where you can interject in the middle of a sentence. And here's the standout feature. The AI has the ability to detect emotional nuances as it listens and interacts with the user.

Open AI's advanced voice model speaks in four different voices. Let's take a listen. Hey, it's great to meet you.

I've got a really great feeling about us teaming up. If There's anything you'd like me to focus on first, just let me know. This technology seamlessly combines speech to text and text to speech, minimizing delays and creating a smooth conversation.

However, this significant breakthrough racist questions about sustainability, intricate algorithms for speech recognition, language processing, and voice synthesis consume a lot of energy. For example, generating authentic sounding speech even for a 15 second sample power-wise. Well, that's comparable to running your dishwasher for several minutes.

Open AI addresses environmental concerns by leveraging Microsoft's carbon neutral Azure infrastructure. But challenges do remain across the greater supply chain. And of course, the energy used by data centers.

Researchers are actively seeking ways to improve energy efficiency, and here is an interesting prospect. Could advanced voice AI actually help to mitigate energy use? Well, potentially That's because its ability to understand context and detect emotional nuances opens new possibilities.

This AI might detect stress in a user's voice and adjust the room temperature and lighting accordingly. This could fine tune energy consumption based on a user's emotional state rather than a direct command. Open AI's advanced voice mode brings the potential to transform communication from industries like customer service to healthcare and beyond.

It's important to note the energy demand of these new voice technologies will only grow as systems become more widespread. This underscores the urgent need for a comprehensive approach to environmentally responsible AI development. So it's pretty exciting to think that this voice in can actually detect emotional nuances as it's talking to you.

Uh, I'd be curious to see, um, you know, how, how it really works in, in the real world, but then of course, as we've been talking throughout the show brings up a lot of security issues, which have already been, uh, talked about with AI and voice and what's real and what isn't. So since we're filled with, uh, security experts, I'd be curious to hear your take on that. You know, I, I guess the, the, if it works, it's great.

If it doesn't, it's a something show. Um, I, uh, well, I'm not sure that I have all the members of my family can detect emotional nuance either. So kinda like, you know, you're Right, It's true.

Isn't This eq? Is that what we're measuring AI by? Emotional?

Yeah. You know, it's interesting intelligence. I, I watched a show on, uh, I think it was Netflix, mind of a Dog, or it's called The Mind.

It was a documentary because I'm weird like that. And it was the mind of a dog. And it, it examines how the human dog interactions are.

And what makes what has made dogs so successful is they are experts at picking up emotional nuance, nuance, nuance, nuance, nuances, nuances from our voice, inflections, mannerisms, tone, everything. They're really good at Intuitive about it, right? But this is going way beyond just the machine, right?

I mean, 60 Minutes had a whole thing where they were talking to robots and the robots were coming back and forth. So eventually the robots gonna pick up some emotional nuance, and you Think they're gonna be smarter than the average dog or bear or Go. You better perhaps you, You, you might come home and, you know, kick the robot, I guess instead of the dog.

You never kick my dog. You kick the dog Mike, kick The dog with you, man. You don't have, Hey, let me just make clear.

I have nothing to do with that. I always said he was a little weird. There you go.

Project, There it is. It is a proverbial saying. So I'm just co co-opting it.

No dogs have been harmed in the production of this show you. Yeah. Um, but no, I, you know, to me like, like ai, this is a great, you know, new, uh, paradigm for all of ai.

If it works, like we think it could work one day, wow, this is, this is gonna change civilization, right? When, when 80, 98, 90 8% of the code it generates is free of bugs and good and working man, look out, boy, you know, Katie bar the, the door. But if it, if it's only 20% or 30%, it might wind up doing more harm than good.

So to the energy side of this thing, does it consume more energy with more nuance? So no By nuance, so funny. No, it goes, uh, just, just by generating, as I pointed in the piece, like 15 second sample of voice is like running your dishwasher or another appliance in your home for several minutes.

So that adds up because now imagine there's, you know, millions of people with these prompts going at the same time. So this is gonna be another feature where people that are watching AI and energy are gonna come down on, as you know, as as exciting. It is.

It as it is. It's, it's an energy producer for sure. Um, that, that's gonna have to be offset because I think once it is right now, it's only available to select chat GBT plus users.

And then in the fall, they're gonna unload it for, for all the chat GBT plus users. I think it's gonna be popular and they are gonna have, uh, a lot more to answer for with energy. You know, Mike, I think it's only been proven in human males that it consumes more energy to, uh, yeah, that's true.

I understand nuance. So other than that, we don't know about AI Yets walking around. Go in there either who Yeahs will be walking around the house going turn off the damn lights and stop talking to the chat.

E yeah, That there'll be some. Well, so, but so to my point earlier, and I hate to be the ai, you know, skeptic and on the panel, but a a, all of these great AI innovations are only gonna be made, possibly possible at scale across humanity by doing them in a more energy efficient way. It, it's not chat GPT-4, oh that's gonna jive it.

We need, we need GPUs that are not power hards. We need more, better, cheaper, more re uh, renewable energy sourcers to power these things. You know, this is a huge, as you as Bonnie reports all the time, and if you don't follow her on text, on TV and echo insights, you should, uh, these, this is a whole area that we need to solve.

We enable all these great things we're talking about. I'm not entirely sure it's a hardware issue. I think it's maybe a software issue and maybe we need to figure out how to train these models in a way that doesn't consume so much resources.

'cause it feels like it's kind of piggish at the moment. But, so I don't know, Steven, is this a software problem? Uh, no.

And, and in fact, um, I'm actually pretty bullish on what, uh, Bonnie was talking about here in the video from a, uh, uh, a res or a resources perspective because what I'm seeing is increasingly we're kind of shifting. We're turning the corner here from investing in basically giant supercomputers to train models that, you know, burn the rainforest down and, and, and, and getting these things into production and into people's hands. And increasingly the things that are gonna be in people's hands are gonna use the resources that can fit in one of these.

In other words, if you look at what Apple is doing, certainly, uh, they're, uh, making all of their AI features run on device or if not on device, in a very, uh, energy efficient from the looks of it cloud. Uh, apple this week published a research paper detailing how they're doing ai and the way that Apple is doing it, man, it, it gives you some hope, uh, to the point. Uh, like you said in the video there, you know, you know, a lot of this stuff is being run on device.

Um, a lot of it is being done in a very lightweight fashion. All of their models were trained, it looks like using, um, uh, Google's, uh, public cloud. Uh, so they didn't end up having to build their own giant infrastructure of things, and they're building a very energy efficient cloud to offload some of those operations on their own, um, processors, right?

I feel like that's gonna be the, the direction that the whole industry is gonna happen. And we're gonna use a lot less energy to do a lot of this stuff than we would've suspected simply because you can make these models more efficient. I agree.

I We start out, I was gonna say, we start out with the brute force method, right? Let's do a general life, well then Refine it, Open AI implementation. What where it's really valuable is it's gotta be personalized and it's gotta be done in a much smaller, efficient way.

You know, Alan's an east coaster, Mike's an east coaster. They have a different way of emoting than, you know, us people from the middle of the country. And so, you know, even just levels of voice.

Yeah, we do have a motions. I had one once, um, yeah, I'll have to look back when that was, but it's gotta be personalized. Otherwise you get kind of a generic, okay, that works for, you know, 40% of the population, but not the rest of us.

So that personalization is what I'm saying is a part of it. And that's gotta be local. That's gotta be things can be done to, to Steven's point efficiently on your, on your desktop or your mobile phone or things that can travel with you.

I'm, I'm deeply concerned that I'm reading the same white papers as Steven. So, um, the thing about the Google thing is that they were using their TPU instead of GPUs right? To training and now is the most energy efficient part in this thing.

So we could get to the point now where maybe we don't need, you know, these big Nvidia GPUs for every model. Look, there's gonna be a time in the future where they're gonna look back at this age, you know, the dawn of ai like this, like we looked, remember though the IBM computers of the fifties and sixties took up a whole room with tubes and everything else, and we were punching cards and so forth. Good.

All this, Yeah. Um, and, and now, you know, we, we have it on our phones more powerful than what the computers that took us to the moon. So that's just the way of it, right?

We, we refine, we, we get better at it. We, you know, and make it more efficient. There's not a doubt as skeptic as I am, and as you know, bearish as I am sometimes on ai, not a doubt in my mind that we're gonna make those kinds of improvements here.

Um, it's just a question of what timeframe. Um, I just wanna mention something Steven mentioned about the, uh, Mac and it's, it's interesting and this is a little teaser for next Monday, uh, next Tuesday show. I'm gonna be talking about the rumors about the Mac Mini and that they're gonna have a redesign, um, coming up this fall.

And a lot of that is what both of you're saying, local device, um, less energy. So interest, there's so maybe a preview that that's coming up, uh, that we'll be talking about next week. But it seems to be happening across the board with having, having better trained models using less energy on locally.

You know, what, put your order in now every day and save shipping and handling, uh, every day is just chock full of information here on TRO Gang. But we need to wrap this version of the Gang Up. It looks like we're a little bit in overtime.

Steven, thank you for joining us today. As always, you add so much to our conversations, tech Field Day, check it out. When is the next Tech Field Day, Steven?

Well, we just posted the videos from Share, uh, where we had some, uh, our first ever Mainframe Focus Tech Field day event. And, uh, but we'll be back September 11th and 12th and 13th with AI Field Day. Uh, check that one out.

Uh, we've got a bunch of companies coming in and, uh, a lot of these topics are gonna come up there. Fantastic. Look forward to it, and of course we'll see you next Tuesday on next Strong gang.

Mitch, thank you very much. We'll be in touch, Bonnie, Michael, and most of all, you out there. Thanks for joining us today.

As usual, we have a full Textron TV LightUp following this, do check it out. But until tomorrow, this is Alan Hummel. Have a great day for Textron 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.