Utilizing Tech - Season 7: AI Data Infrastructure Presented by Solidigm - 16: As AI is Transforming Our Tools, It’s Also Transforming Us with @ChrisGrundemann

Episode Date: December 8, 2020

In this episode, Stephen Foskett and Chris Grundemann discuss the impact of AI on the future of work. How will our everyday lives be transformed by the widespread application of AI? What about the datacenter? Do AI-enabled network management tools mean we lose jobs? We already rely on AI-enabled tools, from Siri to Marvis, and maybe this is the template for the future of work with AI. Not everyone needs to be a data scientist or programmer, we just need to see AI as a co-worker. Episode Hosts and Guests Stephen Foskett, Publisher of Gestalt IT and Organizer of Tech Field Day. Find Stephen’s writing at GestaltIT.com and on Twitter at @SFoskett. Chris Grundemann a Gigaom Analyst and VP of Client Success at Myriad360. Connect with Chris on ChrisGrundemann.com on Twitter at @ChrisGrundemann. Date: 12/8/2020 Tags: @SFoskett, @ChrisGrundemann

Welcome to Utilizing AI, the podcast about enterprise applications for machine learning, deep learning, and other artificial intelligence topics. Each episode brings together experts in enterprise infrastructure to discuss applications of AI in today's data center. Today we're discussing the future of work. If we take it as a given that AI is going to transform how enterprise applications look and feel and what they're able to do, then we're also taking it as a given that the people doing those tasks are going to be transformed by AI just as much.

So let's meet our guest. I'd like to welcome back my good friend, Chris Grundemann, who joined us previously on Utilizing AI and has quite a lot to say on this topic. Chris, go ahead and introduce yourself. Hi, thanks, Steven. Yeah, my name is Chris Grundemann. I work as the VP of client success for Myriad 360, which is a cybersecurity and data center infrastructure integrator out of New York City. I also do some other things all around the web and wear many hats. I'm very excited to jump into this conversation with you. Thanks a lot, Chris. And I'm Stephen Foskett, organizer of Tech Field Day and publisher of Gestalt IT. You can find me on Twitter at

S Foskett, and you can find a lot of my writing at gestaltit.com. So Chris, I think that that's really the way to think about this. I mean, if, you know, we've been talking about how AI transforms everything about IT. And last time you and I spoke, we specifically were going into the ways in which AI is changing enterprise networking, both, you know, LAN, WAN, data center networking, and also how networking has to change to support AI. But of course the people are the ingredient that we didn't really talk about. So this was your idea. Let me know, what do you think about

how AI is transforming work? Yeah, thanks. I do think it's impactful, right? And I was reading an article recently that I think the headline was something along the lines of, you know, in a world of AI, we need to change what we do, not just how we do it. And that really struck me and definitely resonates

with a lot of the conversations I'm having with my clients and colleagues and friends around AI. I think there's a little bit of an AI boogeyman of, you know, that this is something that's gonna come take away jobs or opportunity potentially. And which is interesting to me, because, you know, over the last several years looking into software defined networking

and the things around automation in the network specifically, we saw those same fears come up. And so to me, it's a little bit of an old saw of understanding how this kind of plays out when we see some form of automation, in this case, machine learning and broader artificial intelligence as a potential to take jobs.

And I think what ends up being true is that you're actually not looking at the problem correctly. And what really needs to happen is transformation. It's not to replace one thing or another. It's really looking at what are the new capabilities that are available? What are new ways we can work? And what are new experiences we can create for customers versus just, can I automate this function of this job, thus reducing my overall workforce? I think that's just a completely wrong way to look at it. Exactly. And I think that that's important to say that it's not, you know, I think that people immediately have this sort of robot overlords are going to eliminate work kind of idea. That's not what we're talking about here, right? I mean, we're talking about

transforming work and changing how work works. Absolutely. Yeah, I think that's exactly right. And so, that's where one of the things where, we're actually within my organization are kind of not struggling with, but looking at right now, which is, how much do we go in and optimize process ahead of time versus can we just use

some kind of robotic process automation to just automate that? And there's this interesting discourse where some of the folks who are really big RPA experts and consultants will tell you, hey, don't even worry about optimizing because it's also a waste of time because we're just going to go in and we're going to do exactly what you were doing before. We're just going to do it with a bot, which is potentially interesting in some ways. But again, I think it's missing the point that what we really should be doing is sitting back a little bit and trying to understand what is it that AI makes possible that couldn't be conceived of before. And so, yes, there will be some automation things, but I think it's also interesting to look at specifically how do you work with AI and what does that look like?

I know that there's an interesting idea that's been talked about quite a bit, which is in this world of AI bots, potentially a lot of the value comes from not necessarily what you're personally doing anymore, but how you're managing and designing those systems around you. And so the question of digital literacy really comes front and center to my mind around this idea of like, of course not everyone's gonna be a programmer, right? And again, this kind of hearkens back to the network automation talks we had where, you know, a lot of network engineers were like, well,

do I have to become a programmer? And, you know, I think that's also true for, you know, the modern office worker. No, I don't think everyone's going to be a coder. No one's gonna, you know, not everyone's gonna have to be a software developer, but I think understanding what the limitations are and the capabilities are of the software you're working with

whether it be, you know, just advanced heuristics or actual machine learning or some bigger AI is really, really important. Just like knowing your colleagues and your coworkers and what they're capable of is really important. And so I think we need to look at this as a really symbiotic relationship and a continuation of the co-evolution of technology and man

and if you look at it in that light, you can see a lot of opportunity and a lot less risk, I think. Yeah. And I think that that's the interesting aspect here is that we're not talking about how AI might, I don't know, destroy like work or how, you know, everybody needs to change immediately into like a data scientist or a coder or something. It's about reflecting on the fact that your tools are going to change. You know, you're, you're a veterinarian, right? Guess what, you know, there's going to be an AI application that's going to help you diagnose, you know, the animals in front of you. And it means that you'll be able to spend more time with the patients maybe in less time, you know, analyzing diagnostic tests or something. I mean, things like that, I think are the things that are going to be transformed for everyday people. But in the

enterprise, in IT, isn't it true that some people's jobs may be eliminated by AI? I think there's definitely a potential there, right? I know that I saw a case study about, I think it was UBS that recently did, I think they put in place like something like 2000 bots to look at their logs, basically, you know, the syslog view coming off. And not only, you know, dissect the logs and correlate them and look for root cause analysis, but also use natural language processing to pull that together and kind of provide an update to folks. Now, of course, that is something you could do with people, right? You could have just thrown people out. I think they estimated they would have needed like 10,000 people to do the same work. But I don't know that that's necessarily a really fulfilling role. Right. And so I think there is again, I think there is definitely a potential for some specific roles and functions to be replaced. If it wasn't, then there wouldn't be any advancement at all, I guess. But again, I do think it's more interesting to look at, okay, what are the parts of my role that can be subsumed by

automation or machine learning? And then what does that leave for me? And then what does that open me up to be allowed to do? It's kind of like thinking about you know and especially maybe you know an artist is one one way to look at right the starving artist i think is an interesting um model which is that you know the reason artists are starving is because in order to focus on creativity they kind of obscure any other work and they just dedicate themselves completely to that and so you know in that thought process i think it might be interesting to think about if if you didn't have to do those data entry parts or even some of the predictive parts or whatever it might be that becomes things that you can hand off to a bot, what are the other things that then become possible by partnering with that bot and by saving time by using that bot?

Yeah, absolutely. And, you know, it reminds me, you know, I guess we can go a little broader here into pop culture. One of the things that I keyed in on in the season three of Westworld was that the construction worker guy has a robot companion basically on the job with him. You know, so he's doing things and the robot is doing things and they're just partnered up. And that's not like weird or anything. It's just how work is in the future. And I, you know, even if it's not like an anthropomorphic robot that like lays cable for you, um, I think that it is very likely that we're going to be seeing, um, a, uh, a future where basically every one of us uses, uh, various AI enabled tools to augment ourselves and what we're doing. And in fact, we're already doing that. I mean, that's the cool thing. Like, think about, you know, Siri and, you know, Marvis and, I mean, all these, you know, things out there, you know, we're already

relying on, you know, AI-enabled tools to help us do our things, right? Absolutely. I think that's a really poignant point. And again, this is where I tend to think that AI is definitely interesting and is going to be very disruptive as we kind of gain more speed with it and more understanding of it. But I also think that it's just another color, maybe you call it, on the spectrum of our co evolution with technology. And as you said, right, this is something that's been going on for a long time, I'd be hard pressed, I think, to find a modern office worker who goes a day without, you know,

some kind of email application. On the sales side, maybe, you know, customer retention, or, you know, a CRM, there's all these tools that we use. We carry phones around with us 24-7 and feel isolated if we don't have it with us. So we've already got this really symbiotic relationship with technology. And this is just, I think, further augmenting what that technology can do for us. The technology that we're already living with day in, day out, I think. Yeah. So back to what we were talking about earlier as well, rather than saying, oh my gosh, in this AI driven future, everybody needs to be a data scientist or a programmer.

Maybe what we're talking about is, and it's because that's not really it. I mean, you know, in Charlie and the Chocolate Factory, remember, you know, he works at the toothpaste factory and then he gets a job at the toothpaste factory repairing the machines that make the toothpaste and replaced him. You know, I mean, that's, that's, that is also the nature of work. I mean, you know, you look at manufacturing today and there's almost nobody on the factory floor. You know, factories are basically just a forest of machines, but that's not really what we're talking about here. We're not saying that the AI is going to do all the stuff and that all of us are just going to like be

servants to the AI. We're talking about doing our jobs better. Yeah, absolutely. And like I said, I really think that we're going to continue to just see AI and this technology more and more as coworkers. And I don't think that it will supplant us at all. I think it's, like you said, it's expanding what we're capable of doing and changing the way we're able to do it by partnering with, you know, these potential technology co-workers. Right. And I think that that's the, you know, well, hell, it's a lot less scary to think about the fact that, you know, AI is my co-pilot than it is to think about AI is eliminating me. Because again, I mean,

look at a lot of the commercial applications. I mean, what do people say about autonomous driving? Autonomous driving is gonna eliminate truck driver jobs. And I'm guilty of that. When I first heard about all this autonomous driving technology, my first thought was, well, what about all those over-the-road truckers

who are gonna lose their jobs instantly? But now that we've got a better look at what it can do, specifically what machine learning can do, I think that all of us are being like, whoa, whoa, whoa. Yeah, that's not what's going to happen. It's just that the machine learning and all these tools are going to assist, it's going to be assisted driving, not autonomous driving. is that where you're going? I think so. And I think there's actually two sides of it. I think that in the one case

that we've been talking about a lot, it was where, you know, AI can assist us and help us, right, there's a kind of augmented intelligence, I think is the best term for that, where we're actually using AI to enhance what we're able to do through various different ways. I mean, and there's tons and tons of examples of this in farming, agriculture in general, in healthcare.

I mean, there's places where this is already happening quite a bit. Also within IT specifically, right? I mean, a lot of the new security tools are using some kind of AI to help with analysis and threat hunting and that kind of thing. So that's definitely happening and will continue to happen. There's another piece though,

which is that I think there's also new opportunities in where humans help the AI. And so we're seeing some of that today. Right. I mean, there's and not all of it's necessarily the best jobs at the moment. But as far as data entry goes, you know, a lot of there's a lot of new data entry jobs that are just identifying things and helping with machine learning, right? So a bit of a mechanical Turk behind the machine learning where you've actually got people and whether they're, you know, looking at someone's intestines and identifying polyps or looking at streets and, you know, boxing around stop signs and pedestrians. There's actually a lot of this going on across the world, not just in kind of where you would think that some of this would happen in the developing world where labor costs might be lower, but also here in the U as well. There are quite a bit of decent paying

entry-level jobs that allow you to work on training the AI. And I think that'll continue. And I think that's also was seen, as you said, with the Charlie and the Chocolate Factory metaphor, we saw that with the mechanization type automation, where building those machines and tending to those machines becomes a job that wasn't there before. And so I think you'll see both, right? The machine will help us, but then it'll still be us needing to help the machines as well. Yeah. So, indeed, I think that, you know, to kind of bring it back to IT here, what should people be doing right now? Okay, so you're a network engineer at a major company. You know, you spend

your day troubleshooting, you know, building things out. Well, actually we know what you spend your day on. You spend your day in meetings, but point is, you know, what, what should that person be worried about or what should that person be looking forward to? Cause I know this is kind of your background, right? Yeah, absolutely. And I think, you know think it's something where, again, I think the main thing is understanding kind of what's possible and what's happening, right? And I think that's true. Hopefully, that's easier for folks who are working in IT than maybe other roles, right? I mean, I think someone who's maybe working in an HR department who is just as likely

to be disrupted by some kind of artificial intelligence mechanism coming in, at least in the IT world, I think we're a little bit more equipped with understanding the pace of technological change and staying up to date with it. And this is an area where it's a little bit different in that if you're a network engineer or security analyst, you may actually need to branch out a little bit. And instead of learning more

about the specific tools you're using today and more about your specific methodologies, you need to learn about how AI might impact that and what AI is possibly capable of. And so it becomes just another tool in your tool belt versus something that I think could really replace you. Yeah, exactly. And, but let's say, well,

let's say somebody wants to get more involved in this. Can they become, do we need to have data scientists and programmers and developers and so on focused on AI in enterprise IT ops? Yeah, probably. I think definitely more and more, and this kind of harkens back to our overall ongoing conversation around digital transformation and what that means and what that is and i think as we turn more and more to you know again whether it's ai or some other technology helping us do our jobs and helping us you know as companies serve customers uh that's definitely something that's happening you know just like you may have gone out and learned some scripting or whatever to kind of augment your your abilities to work with the tools you use i think that as we expand that tool set inside of companies that are building their own IT

infrastructure still, absolutely understanding artificial intelligence is going to be very crucial to that. And then whether you're developing those products for internal consumption or external consumption, it is unique and different than a normal software development process, right? Because with AI, there's a number of things. One of them is unpredictability. When you're just running a product development effort for a standard software product, there's input and output. You code the thing up, you test it, and it works.

Whereas if you're building a machine learning-enabled product, there's a lot more uncertainty of the inputs and outputs intentionally. That's the value of the machine learning is that you're not just coding it to do specifically A and B and C. You're saying, hey, figure out whether you need to do A, B, or C, and then you've got to kind of see what it does at the end. So I do think that that is a unique skill set, and I think that it's a skill set that's more and more in demand for sure. Yeah. Yeah, it is an interesting thing. And this also comes back to some of the topics that we've been discussing on other episodes of this podcast,

alas, without you, but we've been talking about the comparison of DevOps and MLOps and talking about the, you know, the different requirements that an ML application has in enterprise. And again, not to get into that, but just to kind of continue this specific focus on work, one of the things that came up and continues to come up is that unlike most enterprise applications, AI applications have sort of the working set, they have the software, but then they also have this sort of the model. And there's

this whole aspect of making sure that the model is accurate, making sure the model continues to function. Is that an opportunity as well? So again, not as like a coder, like I'm going to develop the model, but sort of the care and feeding of AI applications? Is that a new job opportunity for people in the enterprise? I think so. And I think it's probably, at least for me, it's not very well defined yet because it sometimes can cross that boundary, I think, depending on the tool set you're using

and the products that you're using, right? There is this idea of potentially instead of, you know, going in to a CLI and again, hard coding exactly what you want the firewall to do or the router to do or the switch to do whatever it might be or the server, et cetera. There's a very strong potential that in the future,

if that is an ML enabled device or a system that's running it as ML enabled, what you're actually doing is teaching it to do better versus telling us specifically what to do. I think that again, harkens back to this idea that I really believe that we're all going to become managers of bots in a way. And so even if you're an individual contributor, you may want to brush up on some management skills as well.

Although the empathy piece might not be as important when you're managing bots, but definitely I think there's a lot of other skills in the management practice that a typical security analyst today doesn't really worry about, but may need to if their job becomes teaching the firewall how to protect your company versus specifically coding it to do that. And I do see that trend as well. It's kind of a shameless plug. I'm working on a research report for GigaOM right now on network observability. And one of the defining lines in the field came down to whether or not the tool included some form of AI ops. And the predictive power available there is pretty amazing. And so I'm definitely seeing that trend. And I do think we need to adapt to how do we work with AI versus

just telling things what to do. Well, that's amazing. Let me zoom in on that for a second. So AI has become like a checkbox in evaluating software? Almost. So I wouldn't quite call it a checkbox. So the way we do those reports is we look at kind of table stakes, which is you have to meet this bar to get on the report. And then- Must be at least this high. Right, exactly. And then the next set of things is key criteria. And so the key criteria are where folks are differentiating. And a couple of the key criteria we included within network observability was the ability for the tool to help you with

troubleshooting and the ability for the tool to help you with capacity planning or other forward-looking things. And it turns out that machine learning-based predictive AI models are really good at those kinds of things, right right about taking a data set and coming up with an estimator or an estimate. Michael Morehead, And then figuring out you know kind of predicting the future, a little bit looking at what the most likely cause is so you can do troubleshooting faster. Michael Morehead, When the most likely failure is going to be so you can do some you know capacity planning and and forward looking events so all we didn't actually include Ai as a checkbox at all, it definitely showed up as the way to meet and differentiate yourself in a lot of the key criteria we identified. So Chris, I think that it's pretty clear from my perspective that when we look at the future of AI and we look at how it's going to impact people and their daily jobs and their daily lives. It's not, you know, the

typical Hollywood view of, you know, the machines are going to come in and take all our jobs, or even the political view of, you know, y'all are going to lose out and there's nobody going to be behind the wheel anymore. It's more, I like what you said, AI is your co-worker, AI is your co-pilot. You know, that I think is much more likely to be the future of work than some sort of, I don't know, dystopia where there is no more work or no more workers. Would you say that that reflects your position as well,

specifically on the enterprise tech space? Yeah, absolutely. And like I said, especially in the enterprise tech space, where like I said, I think we've learned over years like how to adapt to technology. And this is just, this one is a little bit different because the technology you work on might not be changing, but the way you work on it is changing.

And so I think that's really interesting for folks who are, you know, getting deeper into their engineering careers, looking at, you know, how that path is gonna branch becomes interesting, right? And so if you're gonna go into, you know, technical marketing or product development or one of those things, I think that looking at, you know,

how AI is going to impact the customer experience is really, really important and what that means for your customers. Also, if you're gonna continue, if you're starting more early in your career or you're just starting out, you're gonna stay directly in the engineering field, I really think that looking at how you can use

machine learning and AI to augment your own intelligence, augment your own abilities, and understand how to manage those tools in new and different ways. I think it's gonna be really interesting in reinventing your work and reinventing what you're capable of doing. So yeah, absolutely. I think that augmented intelligence is really the thing we should be talking about. I think that's most important. And I think in enterprise IT, we're actually more equipped to handle that than some other fields, luckily. Yeah, I think I'm tending to agree. So thanks a lot. Well, Chris, thanks again for joining us

for this podcast. It's been great to have you on a second time. Those of you who listened to this, if you enjoyed what Chris had to say, maybe shoot back in the archives, you'll find another episode with him talking about specifically how AI is affecting networking and how networking is affecting AI. So where can other people connect with you and follow your thoughts on enterprise AI and other topics? Yeah, thanks Steven, at Chris Grundemann on Twitter or chrisgrundemann.com as on the website is the best place to kind of see what I'm doing

and then branch out from there. And if you need help with any enterprise IT tech at myriad underscore 360 is a Twitter handle for my company and we'd be happy to hear from you there as well. Great. Thanks a lot. And thank you listeners for joining us for the Utilizing AI podcast. If you enjoyed this discussion, please do rate and review the show in iTunes, since that really does help our visibility. And also please subscribe and please share the show with your friends. This podcast is brought to you by gestaltit.com, your home for IT coverage from across the enterprise.

For show notes and more episodes, go to utilizing-ai.com or find us on Twitter at utilizing underscore AI. Thanks, and we'll see you next time.