Democratizing App Development, GenAI and Lawsuits – Techstrong AI Podcast EP15

Hello and welcome to the Techstrong AI podcast. I'm Amanda Ani and with me today is Mike Baard. How are you doing?

I'm doing great. Back from, uh, yet one more trip to Las Vegas. Yes.

Well you have lots to share with us from that trip actually. So we'll start with, um, ServiceNow has, uh, democratized app development and you wrote a couple of articles, um, about different things they shared at the event from leveraging genai to democratize app development to partnering with IBM to extend generative AI reach. So can you share a little bit more?

Yeah, there's a lot going on here with these folks. So, um, at the core of the democratization effort is what they're saying is a no code tool and this no code tool. We will enable anyone to kinda type in using prompts to say build me an app that does the following, and the tool will build that app and then connect it up to the ServiceNow automation framework to execute those tasks.

And as we go along, and it'll be interesting to see how things develop because I can use LLMs to automate some tasks, right? There's a reasoning engine in there. And uh, as those things get smarter, I can just rely on the LLM to a degree.

But then for more complex things, I'm probably still gonna need a core automation framework. And how these things kinda interact with each other going forward will be the next big challenge. And then their overall strategy is, it's not dissimilar from what Salesforce is up to in the sense that they are supporting multiple LLMs.

You can either use the ones that they build, you can use open ai, you can use uh, Microsoft, and then they partner it up with IBM now. So you can use theirs and I think there's a variety of open source ones through hugging face that you can use or you can bring your own. And I think what we're gonna see is that all these things are gonna get leveraged in concert with one another.

And you know, that's kind of gonna be an interesting development in terms of how we operationalize AI because, well there's probably some work to be done on that side of the equation, but um, it was interesting to have all this stuff there at the same time. And I'm just curious. I mean they're saying that, you know, everybody and anybody can develop apps that might create some governance challenges.

We talked to the folks from NASCAR about that and they said, you know, it was awesome that we could get rid of our application backlog, but we underestimated the governance challenges and so that might need be the next big thing we need to solve. And do you think, it sounds like more companies are gonna work to collaborate together, be just to make it easier and more efficient. Nobody's gonna have to be locked into one particular thing or another.

They can use whatever they choose and it will all collaborate with each other. Yeah. I don't know how seamless that's gonna be, but theoretically yes.

I mean there's still probably folks who might find the Microsoft approach where it's all on one cloud and integrated by them a little easier to work with. Uh, but ultimately I don't think that, you know, Microsoft is gonna have access to all the data that's required to automate everything using their LLM. So I think there's certain data sets that will be exposed to other LLMs where it makes sense to use that LLM versus the one that Microsoft has.

Um, which kind of flows into our next set of stories. So maybe we should jump there. Yes.

So next you have an article and this is over on our brand new site. If you haven't checked it out, it's Textron, ITSM. So go check that out.

Lots of great information surrounding all things IT service management. So to that note, you just recently shared, um, an article over on that site about ServiceNow aligning with Microsoft to foster AI assistant interoperability. Yeah.

This is an example of what we were just talking about where Microsoft is saying that their copilot will be integrated with whatever AI assistant ServiceNow comes up with it. And there's gonna be a, a certain hierarchical relationship here, right? Microsoft is probably gonna have one of the dominant AI systems out there.

They have a copilot for just about everything these days. And, um, what's gonna happen is you'll be able to ask on Microsoft copilot, for example, to perform some sort of IT task, like, I don't know, provision a new laptop for me or go buy new one and it will hand off that task to the ServiceNow assistant, which is more of an ITSM type of platform and is optimized to manage those kinds of tasks. And they'll be a little more, uh, I guess hand in glove relationship as that evolves.

I think this is the gonna be the case for just about everything out there. I think there's gonna be some major AI assistance that are dominant in the, in the first case, and then there'll be handoffs to the second case. And um, it was also interesting to watch, there's a story that'll be on digital CXO talking about how, uh, NVIDIA and ServiceNow were showing digital avatars that were built with ai.

And you could actually engage with something that has a human-like experience where it's appears to be a person, it has, uh, attributes that you would recognize. It smiles, it tells jokes, and we will, uh, you know, do basically anything you ask it to do and might even give it a name someday. Um, so it's pretty clear that these AI assistance are gonna rapidly evolve into avatars.

It might take another year or so before we can actually afford to do that. 'cause the processing power is not cheap, but it will come along in its own way. And I'm sure by now everybody's kind of asking, well, what does that do for my job?

Well, this is the interesting debate, so there's no doubt it's gonna change jobs. The question is is will it add more jobs or will it eliminate jobs? And I guess, you know, there was a survey that ServiceNow put together with Pearson and they were just pointing out that, you know, for countries around the world to maintain their economic growth projections, they need to hire millions of more workers that they don't have.

So the only way to keep the growth going is to basically augment those folks with more ai. So we increase productivity otherwise countries won't be able to maintain their current level of economic growth and things might start to, uh, collapse inwards in a way that would be bad for everybody. So I think all our jobs are gonna change, there's no question about that, but I'm not quite to the point where maybe we're all sitting at home all day watching the AI do all the work.

Yeah, absolutely. The human element's still very important, but I, I think it'll be great when we can just kind of interact with it, uh, almost as though it's another fellow human and be able to just have that natural language conversation to get things done. It will certainly make it easier.

Yeah, I was reading this article, I think it was the times where somebody did an experiment when they had a dozen AI assistance that they were using in their life and they were using to automate all kinds of interesting tasks. And uh, you know, you can't help but wonder if, uh, there's a lot of things that we should be doing that, you know, fall by the wayside and including, you know, we're at work but we're supposed to set up a dental appointment. Something as simple as that.

Well, maybe that AI assistant can take care of that for us and then they want to show up. And so, um, you know, there's a lot of doom and gloom in the air, but I gotta say there's a lot of little tiny things that conspire to make life harder than it needs to be that maybe this AI thing can take care of. Yeah, I think certainly, you know, doom and gloom, yeah, maybe some of those jobs, you know, assistant jobs might be eliminated, but overall that's gonna be really great for businesses to be able to just use an AI rather than looking for, you know, virtual assistants for example.

Alright, well we can hope for the best change is coming whether we like it or not. So as, uh, Jensen Wong said, I might as well get on the train now For sure. So moving on now we're gonna discuss a little bit about Oracle and, um, they are committing to generative ai.

com and it's about Oracle planning to apply generative AI to Java and sql. So can you share a little bit about that? This should not come as much as a surprise for folks because, well, everybody who does anything with code these days is embracing gen ai.

So the fact that Oracle is essentially demoing a tool that they're using internally and that at some unspecified data that will make available to the rest of the world, that will, uh, enable you to take old Java and they're hoping you'll use that to migrate to a newer version in Java. Same thing with sql. You can generate SQL code, you can even generate infrastructure as code for their Oracle cloud infrastructure platform.

This is not much different than what anybody else is doing in this space, except that it's more specific than Java and SQL and they vetted the code that they're using to train it. So we should have less hallucinations because we won't have just a bunch of stuff that we're pulling from across the web and hoping for the best. But, um, I'm not quite clear that everybody's gonna use it in a way that Oracle intents, right?

I may use a gen AI tool to take a Java application and convert it into Python, convert it into rust or some other language. So it made me that, uh, this is gonna enable a grade switch because the cost of switching from one platform to another is gonna be sharply reduced as we go along here. There's a lot of, uh, for example, proprietary code used around all these automation frameworks that could be converted into standard languages that everybody does.

So I think we're on the cusp of something bigger than people realize here, because you can swap out a platform now in a matter of weeks rather than months. And that's always been the barrier for entry and it's always been the reasoning incumbent at such a strong position, which I might argue is not gonna be as strong as it used to be. Yeah, and I think it's much like with, uh, communication and languages, the fact that we can be able to translate all the code, all the code into any kind of language is amazing.

Yeah. It's a similar idea whether you're speaking, uh, Farsi to English or uh, Spanish to Italian programming language is say mm-Hmm. And next up we're gonna talk about more lawsuits.

We have several law, you know, they seem to be coming every day almost. Um, but you know, some of the big players that are continuously getting hit with these lawsuits would, it would be Google, Microsoft Open ai, and this usually is pertaining to how they're training their models. So most recently we have, um, a photographer and some cartoonist that filed against Google's parent company Alphabet for use of their, um, their work for training the model.

And then we have Media News Group, which filed against OpenAI and Microsoft for similar reasons, and we seem to see more and more of this. So, um, do we think this is gonna stop anytime sooner or how are they gonna resolve this problem? Well, I kind of wish they would just fast track the whole thing to the Supreme Court, right?

But it seems like we're gonna go through, I don't know, multiple appeals and this will drag on for years at the, at the core of this thing is a debate. Does the using content taken off the web constitute fair use to train something? Because I didn't quite steal everything.

I just kind of lifted it and used a, you know, a couple of graphs or this, that and the other, but I did it at such a scale that I got the benefits of the training or, um, isn't really copyright infringement and did I steal that content? And, you know, that seems to be, uh, a matter of, uh, opinion at this point. So somebody needs to settle that.

And the issue is, in my mind, is gonna be, it's a little cynical, right? I think that the Microsofts and the Googles and all these folks are betting that even if they lose, they're gonna make enough money off of Gen AI to pay for whatever costs that they eventually might have to pay. But in the meantime, they're like, let's just keep going full steam ahead and we will, you know, sort it out later.

Yeah, it's a difficult question to answer, I feel like, because you know, when content is online and say you're writing a report or you're quoting some source, you know, that's all you have to do. Just quote the source, put where you got this information from. So, um, one could say, well, they could do that with the large language models, but that would be impossible.

How would they do that? How would they quote every, you know, source? So therein lies the problem is just how it's being used, and there's really no way to, to say everything that was scraped and used and, and give back, you know, credit to those sources, Right?

And what will happen in the future when AI starts generating content that another AI engine uses to train itself. So, you know, is Google gonna suit Microsoft because they stole their content? I don't know.

Yeah. It, it, it's a weird scenario. It'll be interesting to see how it plays out.

I don't, I don't know if any of them have actually come to a conclusion on any of the lawsuits, so, um, I guess we'll just have to wait and see. Yep. Lawyers are gonna make a lot of money For sure.

Well, that brings us to the end of today's Techstrong AI podcast. Mike, do you have any last thoughts for today? No, as always, AI every week is something new and different, so stay tuned as always.

ai. com, textron, itsm, cloud native. Now, uh, security Boulevard by all means.

And of course, don't forget texture on tv, but one thing we can promise you right now is, hey, all that content generated by us humans. Yes, that's right. Alright, well thank you and go back and check any of those articles that we were talking about today across those platforms, and feel free to share in the comments what you're most interested in hearing about and let us know.

Bye for now. Bye.