The Media Odyssey

TMO - David Rudnick (audio)
===

Evan Shapiro: [00:00:00] I think there are a lot of people who work in our industry or any industry right now, they hear someone like you in your French calculating accent say We're going to cut the fat. And I think that they fear that means they're gonna lose their jobs. And I think that is the fear.

Marion Ranchet: You make me sound like Ratatouille. It's awful, but I love it.

Evan Shapiro: Or Pepe Le Pew. One of the two.

Marion Ranchet: Welcome. Welcome to the Media Odyssey Podcast. This is Evan Shapiro,

Evan Shapiro: and that is Marion Ranchet. And what are we talking about this week?

Marion Ranchet: So this week we're talking about AI.

Evan Shapiro: AI.

Marion Ranchet: Yeah.

Evan Shapiro: Everyone's favorite or least favorite topic, I feel.

Marion Ranchet: Yeah, it's very much a love and hate relationship and I [00:01:00] think that's why it was long overdue, yeah?

Because this is gonna be Episode 23, I think of The Media Odyssey Podcast. And we haven't spoken about AI that much.

Evan Shapiro: I feel like everybody has a love hate relationship with AI or a fear-love relationship. So we're gonna talk about why people feel this way, what to do about that emotion, either emotion if you're having it, but we'll also have a really great guest, the CTO of LG ads, Dave Rudnick is on the pod, and he is an AI expert, especially as it applies to connected television, which I think something that intersection of AI and CTV seems to be something you are very focused on as well.

Marion Ranchet: I think what's interesting with David Stake and what LG Ads does, I think it's a good way of overcoming that fear and speaking to companies like that who understand exactly how AI fits in their day-to-day, in their workflows.

'Cause we'll speak about that at length, but, essentially, I think it's a way to augment ourselves, [00:02:00] right? And in CTV and advertising I think there's a lot of opportunities to cut a lot of the fat. Let's put it that way. The complexity so exciting to have him on the pod to discuss that further.

And again, overcoming that fear by having more of that test and learn approach.

Evan Shapiro: Well, I think you talk about cutting the fat, right? And I think that's a big part of the fear around AI is. I think there are a lot of people who work in our industry or any industry right now, they hear someone like you in your French calculating accent say We're going to cut the fat. And I think that they fear that means they're gonna lose their jobs. And I think that is the fear.

Marion Ranchet: You make me sound like Ratatouille. It's awful, but I love it.

Evan Shapiro: Or Pepe Le Pew. One of the two.

Marion Ranchet: No, but you, I think you're right. Okay let's back up for a second. So you have a lot of data, of course you do, around that love and hate, that fear. [00:03:00] But what are we actually doing with AI? Are we actually using it or is it just a lot of talk for everyone, and especially from people who are trying to sell us to use AI in some form or another?

Evan Shapiro: It's, I think the National Bureau of Economic Research here in the United States did a massive survey around AI usage.

Who's using it? Why they're using it, who's not using it, and why they may not be using it. And what I found interesting in the data is that the younger you are, the more likely you are to have tried or experiment with AI on a regular basis, the older you are, the least likely. And that's the norm. The older, the olds tend to adapt to new technology last.

But what I found fascinating was that there was a huge, a 10 point differentiation between men and women. Men seem to be much more likely to experiment with AI in their workplace than women are. And the big differentiator, whether it was age or [00:04:00] gender centered actually on a lack of trust. It wasn't necessarily a fear of being replaced, it was, I don't trust the ethics of the companies, or I feel there's a privacy risk involved in me giving information or involving my work with AI.

And given the fact that we are different genders, I'm curious, and I'm much older than you are.

Marion Ranchet: You tick all the boxes.

Evan Shapiro: Yeah. I'm curious as to your thoughts on that.

Marion Ranchet: So I will say that from a personal standpoint, I went, my relationship to AI has evolved a lot in the last 18 months.

I think there was a bit of excitement at first, then all of a sudden I felt overwhelmed. Overwhelmed because anything new, you have a lot of people who are jumping at the opportunity of teaching others, which is great, but I don't know if you remember, but last year you had prompt engineers and this would become the new job to be a prompt engineer.

And you had thousands of [00:05:00] people sharing do this or that. Ultimately, I never had time to do any of this. So then I moved onto a moment where I was like, damn I'm missing out on this thing. I just simply don't have the time. So I went in this year with really a mindset of, okay, I'm gonna try little things.

And it is very limited what I can do. And I still, I think I'm still at that moment where, the output is not what I would would, what I would like, and a lot of people are still telling me this is because the input is bad, which, I take offense with that, but it's okay. And so it's changed a lot.

And honestly, me right now, I'd love to find good use cases where I can augment myself. I'm a solopreneur. I have a very small team. We do so much. I do so much. If I could find a few things that, again, when I say cut the fat, for me it's I could almost say cut the boring. The tedious would be maybe a better expression.

Evan Shapiro: Yeah, I think it's, I think it's,

Marion Ranchet: And then focus on stuff I [00:06:00] love. You know what I mean?

Evan Shapiro: Yeah. I think I agree. I think the job enhancement is substantially a better use case now than job replacement, although there are jobs being replaced in finance. And in other areas. I did this weird Algo Wars game show thing with one of my friends, Jonathan Burke.

And we demonstrated how an editor augmented by their AI platform, Social Department, produced 45x the number of five second promos for social media over the course of an hour. Than did the editor by themselves with Adobe's normal suite of services. And the reason for that was, and it wasn't an hour was actually four hours.

It was that the editor who used the AI platform didn't have to watch the shows that they were editing. They basically prompted, watch these three episodes of Yellowstone and find me the scenes where this character is centered in frame. So that right there is a great case [00:07:00] study on what humans shouldn't be doing.

And what, how they can augment their work with a good piece of AI. Like it's a waste of time to ask an editor sit down and watch three hours of a program when they could be writing or editing. And when you look at this piece of tape, and we'll put a, we'll put a link to this thing that we did in the episode description, she gave great prompts.

She said, be funny. She said this, said, then do this. And it just saved her a tremendous amount of time, and I think that's the best piece of advice I can give to someone who is curious about AI or is afraid of AI is try it out. And keep trying it out.

This weekend I was on a train with my mother-in-law and she, we talked about ChatGPT. We were asking about life expectancy and she said, do you think a piece of AI can predict how long you're gonna live? And I said, it really depends on the question you ask it.

So I just asked, how long do you think I'll [00:08:00] live? And ChatGPT said I can't possibly tell based on the amount of information you gave. However, I then said, this is how old I am. I'm a cancer survivor. I do not smoke. I exercise every day, blah, blah, blah, blah. And it then gave me a very specific life expectancy, which was 81 and a half years old.

Marion Ranchet: Oh. That gives us a lot of seasons of The Media Odyssey.

Evan Shapiro: Yeah. It's only like three years left. I'm 78 now, so.

Marion Ranchet: I think that's what's interesting though is 'cause I watched this video you did. And yeah, the editor was blown away. And I think you should, you guys should do a second episode and actually take this guy to do a mix of both because essentially I remember what she did and the output was great in terms of quality and quantity.

Having said that, there's no one in the background to check to cross check. If there's no mistakes, et cetera. And again, perhaps this guy could have taken it even further given the level of experience that he has. So the video was great, but [00:09:00] if I were this guy I'd come out be completely depressed and I think

Evan Shapiro: He knew going in, he was probably gonna lose. It's like you have a car, you have a bicycle, who's gonna win?

Marion Ranchet: Yeah. Yeah. But I think, yeah, there's value in showcasing maybe exactly how it could be augmented, right? Because it really felt from this video, that he could be replaced. So there's one thing about the question about job replacement, enhancement.

There's actually job depression. Let me explain. I spoke to a friend of mine this week who's a, he's a developer. He codes, all day long. I spoke to him and I said, you must be happy, right? A lot of the stuff that you were doing before, you don't need to do anymore. He said, actually, right now I feel down because I love the searching parts. I love the figuring out parts. I don't get to do that anymore. Or if I do, it's depressing me to think that AI could do it in five minutes and I'm taking hours to do it. [00:10:00]

He spoke to a few folks within graphic design, jobs like this, and a lot of folks felt the same, right? So there's that in-between right now where people need to make that move. Stop being afraid, embrace it, and don't get depressed over it, right?

But it's tough.

Evan Shapiro: Yeah. And I think when we were in Spain recently for Gemma we saw this guy get a lifetime achievement award. He's a calligrapher and a designer Seb. And he talked about inventing a new typeface and he said it took 3000 hours of drawing to create this typeface. And I think there is no replacement for the artistry.

I get asked all the time, can't you find some AI to automate what you do with your map? I'm like, I could but I won't learn as much if I do that. On the other hand, yesterday the new gauge report for April came out and just a quick side note, YouTube distanced itself even further from the rest of television in the gauge report.

[00:11:00] And so I took screenshots of the gauge from Nielsen's website and dropped it into ChatGPT and said, create an editable PowerPoint slide with an Excel spreadsheet on the backend for me, and make sure it can be downloaded. 25 seconds later, I had slides that I could manipulate, which I released today on LinkedIn.

Now, I didn't ask it to design the slides. I didn't ask it to have a point of view. I asked it to just convert these screenshots into editable PowerPoint slides. It saved me at least a half an hour to an hour's worth of work. I also didn't have to ask, Jesse and Paul are sitting behind the scenes on this podcast.

They're normally the benefactors or the actually the victims of me saying, can you convert this to a PowerPoint slide very quickly for me so I can print it out? Yeah. It saved all of us that work time, that frankly, I'll use another case study and then I'll shut up.

Last year [00:12:00] Amazon converted all of the systems to a new programming language across their entire enterprise. Think about that.

Marion Ranchet: We spoke about this, I remember.

Evan Shapiro: Yeah. And they have this AI assistant, coding assistant called Q Developer, and they estimate that in asking Q Developer, now the coders came up with the code and the language, but then they asked Q developer to reprogram the site, all of their sites, all of their enterprises for them, and they estimate that it saved them 4,500 years of human time, so it's not even in that case, a question of whether humans could do the task. They couldn't. We wouldn't be alive to see the results.

Marion Ranchet: But, so, I've been talking to a lot of companies and what I think is interesting and why right now we're closer to people figuring out than just being afraid and not doing anything is you are seeing a lot of AI committees in corporate, well corporate loves committees, but it's interesting that they're putting a bit of a task force.

[00:13:00] A few folks who are tasked to get an understanding of what AI could bring to that given company. They're testing tools, et cetera. They're also ensuring that they're educating and teaching the workforce because you've said right this gap. From one industry to the other, one gender, to each, to another extent, it's problematic.

So I thought that was an interesting piece of, okay let's get cracking on this. Having said that, I think there's no doubt that some jobs are gonna be lost, and I sent it to you, but the CEO of Shopify sent a memo.

Evan Shapiro: Wait, what did he say? Because he took down the link. So what did he say?

Marion Ranchet: Yeah, he took it, yeah. Yeah, he took it down. He sent a memo and essentially in there he's saying, adapt and use AI or get the door, right? It, if I have to be simplistic. And so he's saying you need to use AI every day. You are not, you need to justify why you're using, you are not [00:14:00] using AI.

It's, yeah. It was rough.

Evan Shapiro: Kind of draconian but yeah, that's, that checks out for him generally. He's not the nicest person in the world.

Marion Ranchet: No. But what he is right is that any executive, we gotta overcome that stress and that fear, and finally

Evan Shapiro: Absolutely.

Marion Ranchet: Somehow, to use a bit of AI.

Evan Shapiro: A hundred percent.

Marion Ranchet: Or you're left on the side of the roads because others will be good at it.

Evan Shapiro: By the way, this is not a new conversation. Every industrial revolution comes with this fear and hyperbole, and every, in every industrial revolution, jobs are lost. But jobs are gained on the other side of it there is this great piece of research from McKinsey that came out I think either late last year or early this year, that showed that around 14% of the Globe's workforce is gonna be either replaced or dramatically, their careers will be dramatically changed as a result of AI.

So that's 14% in the United States of around [00:15:00] 22-23 million people. In France, you can do the math yourself, but I think the people who will be influenced the most need to get ahead of it. They need to understand the dangers and the benefits to AI.

And if you don't, you will be overwhelmed by it. You, the odds increase that you will be replaced if you're not experimenting with it now. And what I found interesting for you and I, and a lot of the constituents we talk to here, is that when the National Bureau of Economic Research asked various folks who is using AI now and who isn't, computer and math folks,

Marion Ranchet: Of course.

Evan Shapiro: were the number one. That makes sense. They're the ones who invented the technology in the first place, but number two was entertainment.

Marion Ranchet: Yeah.

Evan Shapiro: And right behind it. Over 50% percent of people in entertainment are now saying that they're experimenting or they're using AI on a regular basis in the workplace.

And it was a huge drop off to finance, 20 point differential. And I'll [00:16:00] put all of this in the in my newsletter and you can check it out. But like to me, that makes sense. If you look at the people who predicted, I love to talk about Fritz Lang who did Women in the Moon movie and a big director of the 1920s.

He created the countdown clock for rockets. They didn't exist before his movie, to the point where NASA brought him to the first launch where they were using it. Comic book writers, science fiction writers, movie directors and writers. They're the ones who usually predict new technologies and their usage long before the practical application of it comes along.

So I feel like it's incumbent upon us, the matrix first predicted

Marion Ranchet: Yeah.

Evan Shapiro: The dangers of AI. But I think we can also be the ones who can help show what the benefits are gonna be.

Marion Ranchet: I a hundred percent. And I think we also need to be ethical about the way we use it. Yeah. And we need [00:17:00] to, so one thing that is important is, because we're grownups, our heads, it's done.

Evan Shapiro: You more than me, but yeah.

Marion Ranchet: It's done. Yeah. It's done, actually learning is tougher for us, but there's a set of skills that we have and we will have those for life. And so I will say that I'm gonna be very careful to make sure that, again, I don't delegate the things that I love to do or the thing that challenge myself.

And I think that's one of the risk, and especially for younger generation and I see my kids, et cetera. How do you make sure that they just don't, always keep that process, that thought process and just get to the end game, right? And don't learn anything on the way. If you look at math like, my kids are really good at math and I'm telling my - I'm really bad - but I'm telling my kids you need to, it's not so much the result that matters is the thought process that will matter to the teachers. So I think that's one of the concerns.

Evan Shapiro: That's true. Show your work. Yeah. That's what, that's that cliche. And there's a reason for that. It's not that they wanna [00:18:00] see the work being done. They wanna see your mind processing the information as you go along.

And that, that goes back to something I, I use as a catchphrase all the time, which is wake up stupid every day. Yeah. If you are forcing yourself to learn these things on an ongoing basis, the chances are you're not gonna be made obsolete.

Marion Ranchet: I don't know if you saw, but this week the UAE announced that they were giving ChatGPT Plus for free to everyone.

I think that's fascinating.

Evan Shapiro: Smart. Very smart. That's smart, right? Everyone has it. Will everyone use it? Maybe not, but at least there's not that barrier of having to pay 20 bucks a month in Europe.

Evan Shapiro: That's that, that's a really important thing. And I would imagine in Europe this will, because you guys are socialists in a good way, supplying that technology to literally every school student in the United States would give the, give us or any country who does that substantially greater advantage over everybody else in the rest of the world. It's the broadband gap here in the United States, where there are [00:19:00] neighborhoods that don't have broadband access or computers in schools, and the learning gap that's created there is tremendous.

It's the same thing. We should be, if Sam Altman was a good guy, actually, he would be issuing free licenses for ChatGPT to every public school in the United States. He's not a good guy, so he won't do that.

Marion Ranchet: No.

Evan Shapiro: But here I publicly challenge him. Prove that your foundation is for the ethical use of this stuff.

Allow everybody to have access to it.

Marion Ranchet: So another thing, 'cause you and I, we talk about bundles a lot, but we talk about utility and utility status a lot. And one thing that will also actually disrupt our industry, and I'm not sure people really realize that, but in Europe, more and more telecom operators are actually bundling either Perplexity, Claude, ChatGPT as part of, so it's a hard bundle.

So like I got, a [00:20:00] Perplexity for free for a year. I didn't do it in the end. Not yet, but free did that in France with the French equivalent of ChatGPT, et cetera. And so one thing where B2B businesses will be disrupted, streamers, is that right?

They're gonna have to fight to still be a bundle candidate for telcos. Because they found one of the biggest threat, and this is this thing that anyone can use, anyone has an interest potentially in using it. Whereas not everyone wants to watch Netflix. Not everyone wants to watch Disney plus, et cetera.

So that's another thing more from a B2B perspective of disruption coming and the last one I will say is the fact that, and that's one where I'm very concerned because the same way that Google came in and disrupted the media and the access to news in particular, the same thing is happening right now.

Did you notice on Google now this thing at the top?

Evan Shapiro: Gemini, yeah.

Marion Ranchet: Curating AI results. And the same challenge is that there's attribution to news [00:21:00] publishers and others. If we're staying on this page or within the,

Evan Shapiro: They're not going to the, yeah. They're seeing traffic diminished to the source.

Marion Ranchet: Yeah. Traffic is gonna go down. There's no way to monetize. And so that's one thing where right now I'm full of fear about how, a lot of the media companies are gonna ensure that they get the prominence and the distribution that they deserve.

Evan Shapiro: I think sort

Marion Ranchet: According to gatekeepers, I feel.

Evan Shapiro: Yeah. And I think I tell this to students of mine all the time, if you want an area that is never gonna be short of intellectual, people needing jobs filled and smart people, intellectual property law.

Marion Ranchet: Yeah. That's what I did.

Evan Shapiro: Is only gonna get more important over the next 10 years than it's been. It's been very important up until this point, but now that you've got these large language models out there, scraping literally every word ever spoken or written. Then used in curated, AI results at the top of a Google page, or Apple is desperate to catch up in AI. And Mark [00:22:00] Zuckerberg reorganized his AI department just yesterday, again, because they're falling behind.

That the idea that the world's intellectual property and the world's AI platforms somehow have to ethically coexist. That is a major challenge and a major opportunity though. That's, I guess the thing is yes, it's very, there's a lot to be afraid of. Also, the, just the evaluation of Open AI itself, it shows that there's a tremendous amount of hyperbole and bubblicious stuff going on there, but somewhere in the middle, there's a massive amount of opportunity for us all to advance, each of us in our careers. I'm faster at PowerPoint than I used to be, and I'm a top 10 PowerPoint person globally.

I would, I brag of myself, proven with no metrics whatsoever.

Marion Ranchet: Says the old guy who still uses PowerPoint in 2025.

Evan Shapiro: I love PowerPoint. It's great. I wrote a love letter to PowerPoint. How many people do that? But I've gotten

Marion Ranchet: Oh yeah, I remember that.

Evan Shapiro: But I've gotten speedier at it as a [00:23:00] result of ChatGPT.

Marion Ranchet: Yeah.

Evan Shapiro: And that's a massive advantage for me over other people out there.

Marion Ranchet: Yeah, so I think that's a great segue to our guests, right? 'cause this is a company who's not afraid of playing with AI and he's found very different use cases apply to CTV advertising.

Without further ado let's welcome David Rudnick CTO at LG Ads.

Evan Shapiro: We love to get big brains around the CTV universe right, Marion?

Marion Ranchet: Yeah, absolutely.

Evan Shapiro: And so we'd love to chew your brain up a little bit about topics around tech in the CTV universe. Marion, you wanna kick it off?

Marion Ranchet: Absolutely. I think the biggest the biggest one we have, and also because we're not so AI, knowledgeable, I would say at least I consider myself not so knowledgeable.

Evan Shapiro: AI adjacent maybe.

Marion Ranchet: Yeah, exactly. And I'm keen to understand, because we spoke to Tony in a prior episode and he said that AI is transforming every aspect of [00:24:00] advertising. And so I wanna understand, what you guys do at the AI Innovation Lab. Can you tell us a bit more about how you see AI shaping the CTV landscape?

David Rudnick: Yeah, so at LG Ad Solutions we kicked off an innovation lab last year, and the purpose of that was to be able to experiment with some of these newer technologies to see how they would impact the lifecycle of the CTV ecosystem. And from my perspective the things that are happening are moving so fast.

Every month there are is a whole other generation of change and new things that are available capabilities that we couldn't even possibly contemplate.

And if you the way I like to, I'm a visual person, so the way I take step back and think about our ecosystem is there are all these stops along an ad in its lifecycle through a TV ecosystem.

You know how it's thought about? [00:25:00] Where are we gonna put it? How often do we wanna run it? Who do we want to target it to? What happens when it runs? Can we optimize it during its runtime? Do we wanna modify it during its runtime? When it's finished, what happened? How do we learn from that? If we're gonna do this in fast cycles, can we iterate on the fly every microsecond or millisecond?

And so what we're starting to see is that at these handoff points, or decision making points, machine learning was getting really good at helping us make decisions, but with AI it's the next level because it's almost like a little person who's doing a bunch of stuff and computational and making adjustments. And so it's really, it's making it superpowers within each of these stops along the ad exchange.

What we're doing in the innovation labs is we're taking a look at what can be affected, what can we do above and beyond that we haven't been able to [00:26:00] do? And part of it too has to do with on the technical side, compute and capacity of just being able to run very big models.

You're talking billions and trillions of computational things that are going on. And in the old days you couldn't do it. You didn't have the horsepower to do it. Now we're opening up the capabilities for us to begin to do that. And so with that we think about adaptive creative. We think about how ads are placed on a contextual basis as they're running optimization layers doing predictions during all of this stuff.

So even if you're throwing that dart, that darts guiding itself all the way as it gets to its target. And then when it's done, what can we do better? How can we make this more relevant the next time? And could we find audiences that we didn't even know existed? So all of that gets sewn together in a really cool ecosystem and it's, there's a lot there, right?

There's a lot to unpack and so we have to look at it and in certain cases we may want to proceed with certain sub components and [00:27:00] certain cases of vendor is an extending a new functionality to us and certain cases we don't want to do anything with it. And so as we think about our future it's very interesting and it's gonna save, I believe it'll save a lot of time, make things more efficient.

And really the most important thing is it'll be awesome for the advertiser and amazing to the customer, right? Because they're gonna see things that are more relevant, more interesting. And if we can eliminate waste, or we don't have to serve things that aren't relevant in a given moment, we're in a better place.

Evan Shapiro: Let's talk about the, those last two constituencies there that you mentioned. You have the advertisers and you have the consumers, right? And then you talked about, I think part of the friction that both of them are experiencing in the connected television world here. I, one of the theses I have, I don't know if Marion shares it, is that connected television's disadvantage to pure play digital is the lack of personalization. Is the lack of real time micro-targeting [00:28:00] that pure play digital and specifically the walled gardens promise. And for the most part after fraud, deliver. Doesn't the fragmentation and this real friction of over frequency and lack of personalization that consumers have had and advertisers have had.

Doesn't that require us to use AI to improve that, improve, create infinite personalization like the internet has on the connected television?

David Rudnick: Yeah. Yeah. And I think what the, especially in social media did really is I have a profile, right? I live in the mountains of Colorado. I'm a rancher. I like to fish. I'm a western guy. Those are fixed things. Those are guardrails, but then there's interests.

Interests are moments and interests are shifting with social media with me or with you at any given time. And what, the way we like to compartmentalize is we have baseline information about a consumer, as does everybody who's doing [00:29:00] advertising work.

And then as we have an interest that begins to peak, right? What social media has been doing really well is catching onto that interest. So we have the capability of doing the same thing, right? Using ACR technologies of, hey, these people are really into this type of content right now. And we know there's a window, there's a moment in time.

And so if I think about it and I'm gonna use it, I'll use an LG supported thought process. If I was running a, if I had an LG washer and dryer, and I had a box of 60 count Tide pods. And I knew that I was running that washing machine 50 times. The opportunity for a purchase intent after 50 loads is if somebody's not going to get laundry detergent, somebody's gonna get yelled at.

But once you buy it again, you don't have any more interest. Or if I was going on, if I was gonna go on a trip to Europe, I might be watching outdoor pro- or programming around traveling to Europe. But once that interest wanes and leaves, you need to move to the next thing of [00:30:00] interest, right?

And I think television is catching up to that whole scene of how do I get more relevant? How do I catch you in a moment of time or with imminent purchase intent. And activate that.

Evan Shapiro: But don't you think that television needs to become better at it faster? Don't,

David Rudnick: Absolutely. That's what

Evan Shapiro: You believe we're in a, we're in a race here between TV and other devices or other platforms to perform at the expectations of both the audience and the advertiser as well.

David Rudnick: All day long. Yeah. And I think we're getting, I think we're absolutely getting there with these AI systems, right? Is to understand that moment and activate that moment.

Evan Shapiro: So what do you think is the chief aspect of that AI will add to the connected television that will accelerate this rate of change and personalization and actualization for the advertiser?

What do you think are the individual components that most add value to the ecosystem?

David Rudnick: Sure. So number one, for me, it's about relevancy. So if I increase relevancy through an [00:31:00] ad, if I increase relevancy through content recommendations, the consumers are gonna spend more time dwelling in that area and we can measure that, right?

So the goal is if we increase relevancy, we increase dwell time, and then we wanna make sure that we also know when to stop. So we don't have saturation of messaging, right? And then we switch to the next thing that's relevant because what we don't wanna do is saturate you. We want to catch you in the moments and then leave you, it's like a good waiter at a restaurant.

They know when to come in and hook you up with whatever you need. And then boom, they're out. And our goal behind the scenes with our technology is really to try to think that way. Is how do we step in when we need to and provide something so important that they can activate on it, and then boom, we're out of the way again.

Marion Ranchet: Can I ask you, how does a company like yours embrace AI? Are you buying out technology and doing that integration work? How is that [00:32:00] working?

Evan Shapiro: That's a great question.

David Rudnick: Yeah. So I'll lean back into kind of how we're thinking about AI. There's certain types of AI where you go to the internet and you try to scrape everything, and that's not really our game.

Our game is staying within our, within the things that we know. We are, there are a host of different AI applications out there and standardized systems that are running, and we are creating our own individual agents in certain places. And in other cases we're partnering with other companies to look at what they're doing and if it's relevant, it makes sense, we can fit it into our system, or if it's an extension of our tools, we will activate them.

I will say from a technologist standpoint, we have had it do a tremendous amount of infrastructure upgrading so that we're in a position to take this on. And that has to do with being able to consolidate thousands of different data systems that we utilize in our ecosystem into a centralized location to be able to have cloud-based compute [00:33:00] where we can scale horizontally and vertically to support what we're gonna be doing with either delivery or just running models, right? These are all high compute things.

So we as companies, have to shift in order to support these things. And we've spent a lot of money investing into our infrastructure. And I always think about, I always wanna be able to play offense, right?

Which is, how do we attack, how do we find out? But we also have to do it in a fast fulfillment mentality, which is partner with people or try things. If it's working and it makes sense, adopt it. If it doesn't make sense, then we take, we don't take those programs out of their cycle and move on.

Evan Shapiro: It's interesting, you see the test and learn fail fast model in tech, which is, really not only are you a CTO, but you work for a technology company for the most part, as opposed to a program, a programmer or a content based company. But I think one of the things that interested me what you said there specifically around the use of data, how do [00:34:00] you see AI improving the measurement or the attribution aspects within the ecosystem. I think this is, again, this is another friction point for ad buyers, is the clarity around reporting on certain aspects of the streaming universe in CTV. How do you see AI improving that ecosystem or making it better for everyone to use?

David Rudnick: So the, I think I can answer that a couple of different ways. One is speed and efficiency and accuracy. If I'm running billions of lines of information on compute models, the more I can, the more models I can run, the more accurate I can become. If I run one model, I'm not gonna be as efficient as if I've run 10 models to find the most optimized one.

So that's, that I think is one piece of it. Speed is another, just within the infrastructure. If I can speed things up on decisioning, I can get more relevant things in more and in the proper destination quicker. And so that also helps. And then when I look [00:35:00] at the data that we use whether this is generic, household type of data, or I'm using ACR type of data. We have the ability to look at things that we may not have been ever been able to look at before.

And as these are looking at, let's say, targeted segments, we have the ability to run those segments in a different models to actually find relevancy between things that we may have not have thought of in a legacy type of model.

And so all of those things add up to be to help us become a better system.

Evan Shapiro: To better predictive models that you're running on the front end.

David Rudnick: All day. All day long.

Evan Shapiro And by the way, better attribution models on the back end as well.

David Rudnick: A hundred percent. And so when we go back at it we're smarter. That's the beauty of this whole concept around AI is it's improving on itself, right? And it will always take, these aren't easy.

Evan Shapiro: This is by itself. If the models are built correctly, each buy will improve itself during the buy, using the intelligence, it's gathering from each impression that's being served.

David Rudnick: Yeah. And right [00:36:00] now they're like the old school, I think about like old school F1 cars where you need a mechanic living on them because you're constantly tuning these things. Some people, I think that you just turn it on, like a light switch and it'll always be a light switch. These require attention, these require us to continue to make sure that we're doing the right things within those ecosystems, and that we, as these models are improving themselves, that they're stay within.

The capabilities of what we're looking to do. And I think that's part of, it's exciting and fun for the engineering teams and the data scientists and the researchers. But it is, it's a new world for us as we embrace this stuff and go,

Marion Ranchet: We wanted to ask you what you were, to do a bit of, future looking but what you were saying, it's changing so fast that I think your potentially your answer will be, moot in a few weeks time.

But let's still have a go like, what

Evan Shapiro: But answer it anyway.

Marion Ranchet: Yeah, answer it anyway and we'll put you on the spot in six months. No, but I'm kidding. But,

David Rudnick: I could go way [00:37:00] Orwellian on you and say that AI could be taking over what I do and everything in between.

Marion Ranchet: I don't buy that, not just yet. I don't buy it.

Evan Shapiro: What do you think is that, what do you think? In a utopian world? Not a dystopian world. You seem like a utopian kind of person.

David Rudnick: Oh, totally.

Evan Shapiro: You seem like you're on a ranch. And what do you see in your optimistic view, the most positive change AI will bring to the television experience? On any end of it?

David Rudnick: Yeah. I think for the, if I was the agency or the advertiser lens. I would say that I'm gonna get the best bang for my buck. I'm gonna reach the people that I need to get to. I will have tremendous efficiencies. I'm gonna find people I never knew existed. That is their lens. I think for operators like us, I think our systems will run a lot smoother.

Our world is an, is amazingly complicated and we are commodity brokers and we're yield management organizations, and I think it'll help the companies operate more [00:38:00] efficiently. So I see nothing but positive elements out of this. And I also think the last lens, which is probably the most important one, is the consumer.

If the consumer embraces this as part of their life, where it's actually beneficial and it's not seen as a barrier, that to me would be the biggest win of all. And I think that's where you find Utopia is the relationship of a perfect world between an agency or an advertiser. A brand, right? And how do they get to the consumer? And whatever their goal is, they can achieve it in a really thoughtful manner. I think that would be the best of the best.

Evan Shapiro: If I had an LG agent curating my content, where do I find this film? Where do I find that, that I would find useful?

A friend helping me curate my own TV.

David Rudnick: And you trust it, right? It's like the water cooler effect. And that's why, if you look at some of the stuff that social media's had really, why is it so viral and why is it so sticky? Is in certain cases it takes you down some pretty wild rabbit holes as you find different associations.

But [00:39:00] they're leading you and I think that there's good and bad with that 'cause you find yourself in rabbit holes that you probably didn't wanna start into. Like me, I get strange bull riding fails and things like that. But the, I think the important part is, making it, making that relationship really solid where it's trust.

And I think that is, that's where we want to get to.

Marion Ranchet Awesome. Thank you so much. I feel smarter. We need to spend more time with Chief Technology

Evan Shapiro: Professor Rud.

Marion Ranchet: Yeah, exactly.

David Rudnick: Come on out to the ranch. We'll do silly things out here and learn more about technology.

Marion Ranchet: Oh, wow.

Evan Shapiro: Field trip.

Marion Ranchet: We'll you up on that. Careful.

Evan Shapiro: Thanks so much.

David Rudnick: Thank you guys. Appreciate it.

Marion Ranchet: Thank you so much for being on the pod.

David Rudnick: Thank you.

Evan Shapiro: That was a great interview. He's got a big brain thanks to LG ads for sponsoring. Actually, they were our first sponsor, so thanks to them and thanks to Dave Rudnick for joining us here.

So what's your takeaway on AI? How do you leave this episode and your approach to AI? Do you, has your mind shifted at all or you a little less [00:40:00] fearful, a little less anxious, or.

Marion Ranchet: No, I'm excited. I think I'm at a point where I'm excited. You've said it at the top of the episode. It's all about learning, well, waking up stupid, learning something new. I think that's the exciting part. I love blank pages. I love the fact that for a lot of us, it means that we need to figure out how to to put AI within our day to day, whether that's personal or business. So no. Super excited.

He was a big brain indeed. That's how we roll, right? We bring smart people to the put.

Evan Shapiro: That's right. And if you're curious about whether AI is something that you can play around with, come up with a problem that you would normally spend a lot of time on and give it a challenge. And if it doesn't work, challenge it again, reword it and see what you can do.

I didn't think it was gonna be able to convert screenshots to PowerPoint slides, and it did. And when I told my class.

Marion Ranchet: It's funny, it's my same case. I have the same one. But let's find let's figure one more. Is there anything else we could do? Let's think about the pod. Is there anything that we could be using for for the pod with [00:41:00] the ChatGPT?

Evan Shapiro: I think let's challenge it to figure out how to get more people to click subscribe to our podcast, either on YouTube or on Apple Podcast or on Spotify. Let's give it a challenge and see what it says.

Marion Ranchet: I love that. I love that. And also subscribe to your newsletter, which is

Evan Shapiro: Media War and Peace on Substack. And yours is

Marion Ranchet: Streaming Made Easy on Substack as well. Substack is is hot right now.

Evan Shapiro: Yeah.

Marion Ranchet: Many new people coming, joining us.

Evan Shapiro: We started it.

Marion Ranchet: Yeah, of course.

Evan Shapiro: Thank you very much for listening once again. If you have any suggestions for the pod, we have been factoring your suggestions into our creative.

So let us know what you think and share the links with friends. We will talk to you very soon.