ChatGPT and the Rise of Generative AI – Techstrong Research Review EP 13

here Hey everybody, Mike Rothman here. GM of textural research Chief strategy officer of Textron group. It is another episode of the text wrong research review.

We missed last week. We had our Tech strong All Hands meetings. So Mitch and I were knee deep in Chef Baba and and boat trips and actually some meetings and we have some cool stuff planned for more productive work done, which I am principal with Texan research CTO with Techron group.

Yeah. It was it was just you know, you get a lot of stuff done when you're in person don't yeah, we do it was nice to see you know, that's I've been with the company since what July right? So there were a whole bunch of People actually have met in person.

So it was it was great, you know kind of just not just on the zoom box, but actually being in the same space as folks and it was a it was nice. It was it was it was a good event, but we didn't record the review last week. So we're back and I think this week we want to hit on something that everybody seems to be talking about because you know, if you got to jump on the bandwagon, you may still jump with both feet.

So let's talk about chat GPT Mitch. It is everywhere. Everybody's talking about it.

My kids are having you know it right their papers that they're actually not having it right there papers. I think it's only because they haven't figured out that catch you be can run right to papers and we'll figure it out and I'm sure they do before we did. Yeah.

Evolved. It's on tiktok. Then they must have known.

Yeah, that's the only way my kids know about anything on effort. But let's talk a little bit about you know, again generative AI in general and and really the impact it's gonna have on you know, operations and analytics and and really, you know, part of the challenge of whether you're trying to optimize your devops environment or you know, try to find and and isolate different malicious activities is making the connection between all these different, you know, data types and understanding that so so where does generative AI start to play into you know, aggregating and analyzing Telemetry? It's not about anomaly detection, right?

It's about you know, kind of trying to find the relationships between these, you know disparate, you know, data sources and data streams. So what do you thinking Mitch at this point? Well, yeah, I did a segment on one of our shows AI times and we talked a lot about this and A lot of the generative AI right now is about create this for me create an image that has this do a painting right this for me.

And by the way, I did a LinkedIn poll where I typed in a question and actually wasn't chat GPT. It was another one because it was it was busy. Right and I just pasted the text other than took the footnotes out of it, but where it got the text from and then said is this the answer to this question is that generative AI or not and 75, I'm sorry.

57% said it was 36% said it wasn't that it was a person that wrote it 7% said they're not sure. So this will Air after a polls close and so it actually was written wasn't they? You know, what's funny is that I got that wrong because it was it was recent data and I figured you to use checking GPT, and I know that hadn't been based upon, you know anything recent.

So that's kind of why I had that it wasn't but It didn't occur to me that you would you know hack the system and actually use a different generative AI algorithm. Well, I didn't have a choice. So other than if I wrote it myself and what's interesting is grammarly, you know, just GPT in these things are supposed to be perfect grammar.

And that's how you're supposed to be able to tell right grammarly said that it had an extra comment didn't need so I don't know. Yeah, but I don't know what I would have guessed. It looked pretty look pretty.

You know, it definitely was edited. I wouldn't have written something that clearly just on my own without some help, but for sure, so if you said, you know, if you said it wasn't It had would have had to have some help the point but the point being is generative AI right now is kind of pointed at this crate this for me kind of like one of the applications was in a recording like for a deaf person or in a live setting, you know pick out the voices what the different people who are saying because I can't lip read six people at once right and help me follow the conversation. So it's It's interesting.

It's not AI it's just a particular field of AI right and it kind of reminds me of you know, like AI has arrived like robotics arrived to win Factory started putting robotics in or blockchain become the Savior for security, but not really. It's kind of one of those moments. It seems like well and I think it is gonna have the staying power and it's going to change things but You know, it's not gonna wipe out every white collar job.

Is that kind of it's maybe Educators job more difficult, you know can't use calculators on the test. You know that kind of thing. Yeah, you know and and I really do have you know mixed feelings about this because you know clearly especially in a lot of the areas that we cover.

And we have skills problems, right? We have you know, we don't have enough resources. And yes, it's helping a little bit that you know kind of a whole bunch of big companies that you know kind of hoovered up.

A lot of the really talented folks are you know kind of right-sizing their employee base which is putting, you know, a decent amount of talent, you know back out there Mom, that's helpful. But again, it's kind of like a drop in a bucket right? I mean, they're just aren't enough people with these skills in order to do that.

So if there are tools that can help us be more productive or spend less time on things that don't really make a difference like documenting stuff you already, you know, if they were templates or forms or something that you can put in and you know, you press a button and and reasonable text comes out that then you can just spend time editing that really cuts down on the time to deliver stuff. I got the use case for you, right my compliance report for me. Every security profession would be that's right.

Right. So if you have a template and then you can you know kind of set one of these algorithms on sitting there saying yeah based upon this, you know, just pop it into into a format that makes sense. I mean that helps right especially since you've got information all over the place if you start to say, you know kind of what is that and they can you know, really intelligently look at it relative to the requirements and and the mandates, you know again, I think that that's a great use case, right?

Is it going to help you really find, you know an attacker not yet, right. You really undermines your environment not yet. Right.

And and so you again and part of what we're gonna be talking about a text from Conrad is AI in each one of these different, you know kind of areas and that that's certainly the area of focus that I'm gonna bring to, you know, the stuff that we're talking about. So and this is it's more of a Turing test, right? Can you tell if it's a computer on the other side?

It doesn't mean that it understands what it's writing. It's correlating like information and stumbling it together and there's actually intelligence in doing that artificial. By the way, all my intelligence is artificial, but that's all another topic.

So but it doesn't know what it all means. It couldn't write. You know, I'm yes, it could write some you know poem or whatever but is it really know what it means?

At least now, it does this not truly intelligent, you know to your point. Is it gonna go find The Hacker? That's right.

That's right. And and you know, I mean, yeah, maybe this is short episode. I know it could six months from now.

It could be something else it could you know, I think it's got staying power, but I don't think it you know, let's go Let's go find new jobs because our jobs are going away. That's right that rarely so let's talk a little bit. Right so, you know, you mentioned grammarly and you know, that's really the grammar checking, you know, kind of thing.

Obviously, we do a ton of writing so it's an indispensable tool. If only just to point out that I forgot, you know that I'm using and passive voice so that I forgot to put a comma you know in a certain place has that impacted our editing group not really right maybe a few brain cells from seeing stupid stuff, you know, just obviously Number but I think we're kind of at this weird inflection point where we're really trying to figure out what the best use cases are for the technology and and the thing that made you know, kind of chat GPT feel like a bit of an earthquake is just because it was able to put together stuff that we that we didn't know, you know could happen. I mean, I think people knew that you know, you could assemble, you know, kind of data and have it, you know generate stuff.

I think Dolly, you know kind of the the other, you know image Generations thing, you know, kind of said, oh, holy crap that's you know, that's amazing, you know from that statement. So so we're starting to see you know, what generated AI can do because we just hadn't been exposed to it before right. So that's what feels you know, so novel um, but I still think it's it's a it's a technology looking for a solution at this point or yeah, right Bitcoin kind of still do or about blockchain you what my guess is my This is that.

We might see things like instead of a chat chat GPT. We'll see a chat GPT for medical and a chat GPT for finance and I'm one for security and one for Tech in other words. The sort of general purpose AI kind of things those are really super hard to do and a lot of what we see in generative with these, you know Dolly and there's other image Art and things those are all sort of like a narrow slice of where we're playing a generative AI to it.

So I think the the curation of that information and getting even more refinement of getting the right stuff presented to you in a document, you know, like right this medical report and here's the data in here's the other 50. We've created like it and right do it right the 51st that seems to me a potential really good use of this and then doctors aren't s*** more. They probably don't write the reporting way but, you know sitting around writing reports and you know, a lot of that's done for us like checking Grammar with grammarly today.

Who knows? Yeah, what else I think that's you know, that's that's okay. So let's talk about stuff that is not generally right?

So, you know, we did research, you know earlier this year. At this year right last year about you know kind of AI and Ops, right? We've done AI, you know kind of in application testing.

So we've kind of danced around, you know, kind of what the impact is beforehand. We steamed any of this stuff really start to manifest. I mean you we talked last year about how you know, it was gonna help us refine what to you know, focus on from a code coverage standpoint.

If we looked at application, you know, kind of testing on that front you talk about observability and how you know, kind of AI is gonna help from that statement. So you give me a sense of what you're seeing in terms of, you know, kind of the real impact in that on operations today. So we did do some research or some sponsored research around observability and one element of it was looking at AI Ops right how much you're applying to AI in an operation setting and it was I think I don't remember the numbers right off top of my head, but it was you know, like in the 15 to 16% range people.

either experimenting with it working with it seeing some results for it, but it was things like correlation of alerts and reducing alert storms and sort of the manual reducing the noise kinds of AI things which Um, you know when machine learning took off, I went to one of the experts on the lab that I was working with. And so it's a wise machine learning suddenly taken up and his answer was because we had huge mounds of data that it can go search and then find patterns in and that's what machine learning one one flavor of it. Anyway, unsuperified unsupervised uses.

So guess what observability does to generate tons of data right from all Telemetry from all these devices and iot and network and software and everything else. So, yeah, I think we'll see it as an as something that is an enhancement. I mean right now every vendor technology vendor says and we use AI to fill in the blank.

Well, what does it mean probably machine learning in most cases if it's a lot of data? you know writing code for you is one of the things of course that that GitHub and Google also are doing some things with you know, my question about that is writing code is easy maintaining code is hard. There's gonna maintain what they wrote and and I say this from experience because Long ago, I was working with code generator products and the code.

It generates not what people would write. It didn't was anything close now. Maybe it's a lot better today, but it was yeah, it's not pretty let's put it that way.

So I I think it's going to be an element of a lot of things. and each vendor per inch vendor technology company will figure out where they can really leverage it and they have teams working on it and Enterprises will have models and some things that they can use Ai and machine learning against Will it will it be the next you know, I don't know doordash or whatever kind of change the model of the industry maybe but you know. Nobody's good at predicting those things.

Yeah, and and we don't really know when I mean, I think it's one of these things that we want to you know, kind of be excited about the potential but tap the brakes a little bit in terms of you know, kind of it's It's short-term, you know impact and and I do think the point about you know, kind of every company says they're doing AI ML and some way shape or form and I think they are right. I mean, I think it may help, you know, again, I look at things typically through a security prison and it helps when you're looking at, you know, kind of all this different Telemetry having, you know, some algorithms look and try to refine and fine, you know, unexpected activity or different patterns that you hadn't seen before that assists in terms of you know, kind of your detection capability. So that's goodness.

Right? So we're inching along right and then every so often we'll get this, you know, discontinuity at least in perception and I think that's kind of where we are now and that's why everybody's going bonkers over, you know tragic Boutique and I think it just to be It's really cool. It is really cool.

But is it gonna help you do your job better, you know and then you know in the very near term probably not. I wouldn't change my major right right yet if I was in college, yeah, maybe well, I would to go work on it but not thinking I'm this one's gonna go away. If you've got any kind of math capabilities and understand that, you know, definitely continue to do that.

That is gonna be yeah, but data science has been really hot for a long time. So that's not you know, like I keep pushing my steps on in that direction because he's kind of a mathy guy and you know, it's I'm inching him along. We'll see we'll see but it's again, it's it's really where the future is.

There's no question about that. It's just how quickly we'll get together. It's hard to answer for your kids.

How am I going to use anything from algebra on right now you trying to Draw some common ground with your kids and get them excited about this stuff. Just tell them that's what tiktok uses to figure out what you know keeps you on that freaking app for hours ago. Right?

It's an algorithm based on what you've seen in the past and what you've you know tend to favor and where you go from there, you know from a navigation standpoint. So, you know, oh, okay. So, all right.

We have anything else any anything else Mitch on your you know in your crawl that you need to get off your chest. You know, I'm I'm looking forward to rsac, you know, kind of thinking about what that's like and we'll be there together this year because you are walking around looking good from your from your knee. Yeah, you're whatever we go that they can progress for sure doing great.

So that'll be it'll be great to do that with you and we'll be doing a bunch of interviews and we're starting that process that planning as well, but with you know, all the travel and we still have kind of weird covid or no covid or just happening in that happening. I've trying to figure out if we emerged out of The people don't go to things or they only go to the events. They know or we are we still sort of stuck in this I just go to rsac because that's the one to go to and I don't go to the thing in downtown Denver or downtown Atlanta and I don't know that it's covid related.

I think it says much work remote work related and that way yeah kind of when folks had to go out and they were going to the office every day if they're gonna go to a conference center or whatever. It is like that. I mean that's easy, right, you know kind of fine.

Yeah. I'm just gonna go somewhere different but you know kind of the whole process of actually getting out of the house and going to do something and and I think that remains a pretty heavy lift. So it seems to me that you have kind of the upper Cross of the conferences and and they're starting to rebound I think, you know kind of reasonably nicely.

We'll see right. I mean as our say gonna be what it used to be Pride on right but you know, is it gonna be better than it was last year probably. Um, so I think we're making some improvement but I think a lot of the conferences that I bet the lower end of that, you know smaller, you know community events, you know kind of these Road shows that go to you know, all these different cities.

I think that's tough. I just I think that's tough nowadays. So I do have something on my crowd.

I remember to bring up and we were talking about sort of the hype of Robotics coming in and now yeah, I did generative AI. Yeah, what happened to 5G? I thought that was going to transform the world and the applications and I wouldn't we wouldn't need internet into our homes and all this.

Yes on my phone. I have 5G but I can't tell the difference. Can you tell that suddenly you're like I'm here speed test tells me it's faster.

It does that's probably got a little lower latency. But yeah, if there's killer apps for 5G, I don't think right you don't read about it. I mean certainly it's being deployed everywhere.

It's just to me it's seems like a better 4G. Maybe maybe we just I will tell you it is it does feel kind of magical to me. If I'm you know, standing online somewhere or doing, you know, whatever I'm doing, you know, not in my house that I can pull up, you know kind of TV wherever I want, right and I have sufficient man with to do that.

I don't have to be on Wi-Fi. I don't you know that feels, you know impactful to me but only from a personal and it could be Important it doesn't really help me do anything better. I'm not a better person because I can check out, you know, kind of ESPN, you know clip, you know what it aligned, you know for my groceries.

If anything it makes me less, you know kind of in the moment and and less aware of what's going on around me. But to me that that's how I kind of perceive, you know 5G is it's really taking a lot of the issue out of you know, kind of do video wherever it is that I am but in terms of new applications in terms of anything along those lines, not so much well, and I don't know where you live in relation to Atlanta City proper, but I'm kind of been the edge of the suburbs. We still go in the grocery store.

You're going to Hobby Lobby you're going to wear a lot of storage, you know, you can't get any Wireless any Yeah cellular, right? It's to it's still a big issue. At least my coverage around me is is pretty good in my specific house.

I don't have strong enough coverage to want to get rid of my, you know, one G, you know kind of fiber connection, you know from from the carrier, but it's around around Atlanta. It's it's pretty good. So that's you know, that's been helpful.

But again, I don't I don't perceive any of those things. But again, it's one of these things where you know, the height machine has got a hype. So now that machine is AI and you know, will I for a long time before, you know, one of the gym, you know, sometimes there's stuff that comes out and and you know that you're you're just jealous you didn't come up with so, you know gardeners hype cycle.

I love right. I just because it's so apropod everything that we've just seen and you know what 5G is down, you know kind of I'm not sure if it's the the trophy the solution yet, but it's probably pretty close where generative Another AI things or you know close to the top of the hype cycle now and we're inevitably gonna, you know ride it down when you know folks are really trying to figure out what to do and as we've talked about before things like devops, right? They're starting to go up towards, you know, the plateau of productivity and that we're not talking about it anymore folks are just doing it.

Right and we're kind of focused on the edges about how do we refine it? How do we you know kind of make it a little bit better? So so I think that's a perfect model for how technology markets and hot environments, you know, really kind of play out and you know, we're just we're just riding.

We're right. We're riding the roller coaster again with AI and some point it will transform things in 5G, you know kind of down there. I think it does provide some pretty cool, you know kind of use cases but limited right not Universal and that's I think that's where the hype machine goes wrong.

Yeah. I love their cycle too. And before they came out with it.

I had this mental model of well things go for about a year and a half and then the next thing takes over. Right now, I think it's a lot quicker than that for many things but you can see the you could see the hype you could see the everybody's doing it. You could see the down then that's not it.

Let's go after the next, you know, hot potato or whatever. Yes. All right.

So like we got we got a generative AI we got yes, you are coverage exactly. We've got all that conferences. Once you connect 5G to chat GPT, then magic it's gonna happen, right?

That's kind of that's right throw a little bit of devops in there and and we're all good mentioned, you know RSC so we will be there. So if you are going to the RSA conference, we're having sex ops days again. I know somebody maybe on this show is gonna be keynoting the event.

So there's a something I can't remember his name. Yeah. Yeah.

That's something that will be the keynote speaker there. So come It'll be great to see everybody and really I missed it last year. I'm looking forward to that.

But before that, you know shorter term issue if you happen to be in the Singapore area, we will be there in I think it's two weeks from today Mitchell crap two weeks from today two weeks from day. You're right is the devops Asia Summit in Singapore. So come see us there and we've got Tech strong con, right?

So so March 16th Tech strong con sign up for it's gonna be you know, really a great show. We're gonna talk about a lot of this stuff a lot deeper. So, you know, come come see us there virtually.

Anyway coupon is the week before in in Amsterdam is these will be there right now and my bizarre will be there for for coup calm a lot of things going on those airline miles wrecking up exactly exactly. All right. So with that we will sign off from this week's Tech strong research review.

We'll be back next week to talk about something. That is of critical importance. As well.

It's at least us anyway and everybody.