×
Evaluate Your Team's AI Readiness

Take our official assessment and explore training paths to build secure, compliant AI workflows.

Access Portal

AI at Work Podcast

Ep #8 AI Video Ads Are Beating Studio Shoots… and the Uncanny Valley Isn’t Even the Scary Part

AI at Work, AI Fundamentals, Featured Podcast

Audio episode

Listen to this episode

Summary

In this conversation, Elijah Szasz and Kevin Williams explore the intersection of AI and video technology, particularly in the context of advertising and marketing. They discuss the evolution of AI-generated content, the ethical implications of using AI in advertising, and the potential impact of government regulations on the industry. The conversation also touches on personal experiences with AI tools and the future outlook for AI in business and national security.

Takeaways

  • AI is transforming video technology and advertising.
  • The uncanny valley effect is still a challenge in AI-generated content.
  • Diverse media types are essential for effective advertising.
  • AI-generated ads are outperforming traditional studio-produced content.
  • Ethical considerations are crucial in AI advertising.
  • Government regulations are evolving to address AI’s impact.
  • AI can enhance organizational efficiency and creativity.
  • The intersection of AI and national security is a growing concern.
  • Personal experiences with AI tools can inform business strategies.
  • The future of AI in business is promising but requires careful navigation.

Chapters

00:00 Introduction to AI and Video Technology

02:44 The Evolution of Advertising with AI

05:28 The Impact of AI on Marketing Strategies

08:36 Ethics and Regulations in AI Advertising

11:40 The Future of Video in Business

14:40 AI’s Role in Creative Processes

17:33 The Intersection of AI and Traditional Media

20:35 Challenges and Opportunities in AI Adoption

23:28 Conclusion and Future Outlook

30:40 Navigating AI Adoption Challenges

34:37 The Impact of Executive Orders on AI

37:58 Understanding AI Regulations and Responsibilities

41:06 National Security and AI: A Global Perspective

48:36 Leveraging AI for Productivity

55:08 Exploring AI Tools and Innovations

 

 

 

Explore practical AI implementation: https://assessment.ascendlabs.ai/

Book a conversation with Kevin: tidycal.com/kevinwilliams

Show Notes & Links: https://www.spark6.com/podcast

Submit listener questions:

[email protected]

[email protected]

Check out Kevin’s stuff:

AscendAI: https://ascendlabs.ai/

Follow Kevin on LinkedIn: https://www.linkedin.com/in/kevinguywilliams/

Check out Eli’s Stuff:

SPARK6 Agency: https://www.spark6.com/

Sign up for FREE AI Framework Friday Newsletter: https://www.spark6.com/newsletter

Follow Elijah on LinkedIn: https://www.linkedin.com/in/elijahszasz/

Transcript and citation details

Transcript

00:00 Play from here I'm doing okay, buddy.

00:19 Play from here I'm doing okay.

00:21 Play from here We're live streaming, which is a first for us doing this and there's all sorts of tech

00:25 Play from here running in the background that I don't entirely understand, but we're just going to roll with

00:29 Play from here it, starting with that intro.

00:31 Play from here We're going to roll with that intro.

00:32 Play from here We're going to roll with that intro that was remarkably AI created.

00:36 Play from here I mean, really, it stripped at least 15 years off of me.

00:39 Play from here I like it.

00:41 Play from here I love it.

00:42 Play from here It wasn't me, but that dude's rocking something.

00:45 Play from here Yeah.

00:46 Play from here I'm feeling good about not having to go through a tattoo removal process because whatever

00:50 Play from here generated that just did it for me.

00:52 Play from here It just took the tattoos right off me.

00:55 Play from here I have gone through a tattoo removal process and I much prefer this version than the one

01:00 Play from here that costs many thousands of dollars and sheds blood all over the place.

01:06 Play from here Most people are just listening to this, but we just did a new intro video that was, I

01:12 Play from here think VO3 generated or was it Sora?

01:15 Play from here What is that?

01:16 Play from here I think it was VO3.

01:17 Play from here Yeah.

01:18 Play from here It was VO3.

01:19 Play from here That's Google's video engine and parts of it looked like better looking versions of

01:25 Play from here us if you squinted and other parts don't look like us at all.

01:31 Play from here Yeah.

01:32 Play from here Maybe somebody who met you once five years ago would say, oh yeah, that's Kevin.

01:35 Play from here But your wife would be like, that's not Kevin.

01:37 Play from here Right?

01:38 Play from here It was so cute.

01:39 Play from here It made you young.

01:42 Play from here There's the uncanny valley where something isn't just quite right.

01:45 Play from here Maybe it's a look in the eyes or just some kind of facial expression.

01:49 Play from here And then there's the, that's not me.

01:53 Play from here That's robot me from another dimension.

01:57 Play from here So for those of you experimenting with video out there, the uncanny valley is still real.

02:03 Play from here And we'll probably talk about some applications of a video, particularly in marketing or other

02:09 Play from here areas of business later, but it still has a little ways to go for realism.

02:14 Play from here It's just getting closer.

02:16 Play from here And you know, we'll say this all the time that if this is the worst it's ever going

02:20 Play from here to be, and it's just going to keep getting better.

02:23 Play from here We are still in a remarkable time.

02:25 Play from here Yeah.

02:26 Play from here Video is kind of a topic we're going to cover.

02:27 Play from here Do you want to talk a little bit about video?

02:30 Play from here Sure.

02:31 Play from here Sure.

02:32 Play from here Yeah.

02:33 Play from here Just jump around all over the place because I don't think we really have any order of

02:35 Play from here operation here.

02:37 Play from here Videos top of mind for me.

02:38 Play from here And I'm going to tell you why, as far as like an actual business use case.

02:44 Play from here So Meta, your biggest advertising platform for almost anybody doing any sort of DTC work

02:50 Play from here or on and on and on lots of brands.

02:53 Play from here Are you familiar with the Andromeda rollout?

02:57 Play from here Does that mean anything to you?

03:00 Play from here It's the new algorithm, but the details in how the algorithm is being applied, I have

03:07 Play from here to admit as somebody who was a digital marketer for a really long time, I'm a little bit departed

03:12 Play from here from it.

03:14 Play from here So roll into it.

03:15 Play from here Okay.

03:16 Play from here Okay.

03:17 Play from here So I think that's actually spot on as far as I don't know quite exactly how it works

03:20 Play from here because that's how almost all of these things go, right?

03:24 Play from here People try them.

03:25 Play from here They know there's a change.

03:26 Play from here Andromeda was this kind of skunkworks label that they put on it of using AI to actually

03:31 Play from here reach your audiences based on the type of creatives that you're pushing out there.

03:35 Play from here So what's ended up happening is that instead of you meticulously going in and creating

03:41 Play from here lookalike audiences and trying to find your ICP and all of that, it's largely looking

03:47 Play from here at your creative assets to be able to reach those people that you want to reach.

03:51 Play from here So what's ended up happening is that you can no longer have just one media type.

03:56 Play from here You can't just have video, can't just have a static.

03:59 Play from here It's really going to favor these advertisers that have a variety of different assets.

04:05 Play from here So I've started coming across these people who are now doing all of their ad creative

04:11 Play from here with generative AI, mostly VO3.

04:14 Play from here I just wanted to talk about it since we just saw one of the outputs of VO3 that may or

04:20 Play from here may not have been us.

04:22 Play from here But the stuff that I've seen, Kevin, it's mind blowing.

04:27 Play from here And all of the data that's been shared with me is that it's doing way better than stuff

04:33 Play from here that was shot in a studio or with a bunch of red cameras and fancy lights and hiring

04:39 Play from here actors. This is just good prompt engineering and the outputs are stunning and they're

04:45 Play from here performing really, really well.

04:46 Play from here So this ranges all the way from cinematic shots to UGC content, user generated content

04:53 Play from here where some of these shops will make these AI personas or even influencers build their

04:59 Play from here AI influencer Instagram accounts and then use them in ad creatives as well as statics.

05:05 Play from here So it's pretty interesting, like that's where I'm seeing video really shift the business

05:11 Play from here landscape the most is in that regard in advertising.

05:15 Play from here Yeah, so there are a bunch of different pieces there, the promise of Facebook and Meta for

05:20 Play from here advertisers for years was this idea that you could stick a dime in and get some multiples

05:26 Play from here of dimes out on the other end.

05:28 Play from here And for a long time, there was such an art of ad buying and being able to select different

05:34 Play from here audiences and being able to really get fine tuned on things.

05:39 Play from here And believe me, I know ad buyers who are still leaning into that.

05:42 Play from here And it is still definitely a thing, but it's become harder and harder and harder to pull

05:48 Play from here off. And the bigger the brand, the more obvious the value proposition that's being

05:55 Play from here presented, the better systems like this tend to work because they get just so much data

06:00 Play from here coming in. Smaller brands, I do think, are going to continue to struggle with algorithm

06:07 Play from here changes like this, and they're going to be fighting against just a sea of content that's

06:13 Play from here coming in. You know, one of the pieces here that you hear people grouse about online is,

06:20 Play from here oh, you know, oh, it's all AI generated slop.

06:23 Play from here Well, it was kind of human generated slop before.

06:27 Play from here There's like, you know, the movie Idiocracy in Action that that that, you know, people

06:33 Play from here like videos of people falling on their butt or cars crashing or whatever it is.

06:41 Play from here And it doesn't really matter if it's AI generated or not, if it's getting the consumer

06:47 Play from here that little dopamine hit that they want and recognizing that and not getting overly

06:53 Play from here precious about it for a lot of brands, I think, is is a bit of a leap of faith.

06:59 Play from here But the data is there that these these ads are are yielding success.

07:05 Play from here Now, it's important to note that that anything that's produced by Veo or by Sora 2 does

07:11 Play from here have watermarks in it. So the algorithm, air quote, knows that it's AI generated and

07:17 Play from here you're you needed to disclose an advertiser.

07:19 Play from here You can't ever take those off.

07:22 Play from here But technically, there are some there are some pretty sophisticated ways of doing it

07:27 Play from here that that of course you can, but it doesn't matter is the point that there was this idea

07:33 Play from here that you'd get penalized for it being AI.

07:36 Play from here Well, more content, more better for the engine because people will watch it longer and get

07:42 Play from here their laughs out of it or whatever it is.

07:44 Play from here And engagement is the name of the game for these platforms.

07:48 Play from here And if the engagement is coming from sloppy, stupid videos, fine.

07:56 Play from here That's what they're going to lean into.

07:58 Play from here What about real ads, though?

08:00 Play from here Right. So, for example, I saw an ad that featured this Grand Prix car.

08:05 Play from here It looked so real and it looked so good.

08:08 Play from here And the people looked so real and so good.

08:11 Play from here And so good. Right.

08:13 Play from here From an advertising perspective, does it really matter?

08:16 Play from here Because they're just looking at the outcome of performance, of reach, of conversions.

08:20 Play from here What's going to what's going to hold attention?

08:22 Play from here Right. So there's been a lot of discussion about labeling synthetic media.

08:29 Play from here And I see that coming into play more where you are shaping opinion, influence, political

08:38 Play from here discourse, things of that nature versus I'm an agency and I am producing creative to try to

08:45 Play from here sell this widget for this client.

08:46 Play from here Right. Which I think it's just fair to divide those two when we talk about all of this new

08:52 Play from here synthetic media coming out there.

08:54 Play from here Yeah, sure.

08:56 Play from here I think the other place where advertisers can get into trouble is drifting into just

09:01 Play from here falsehoods. I've been seeing these AI generated ads on like the New York Times, of all

09:07 Play from here things that are like best gadgets for Christmas 2025.

09:11 Play from here And it'll be something that's just completely made up.

09:15 Play from here But I'm enough of a gadget geek.

09:17 Play from here So it has me targeted that, you know, this left handed screwdriver device thing, whatever

09:25 Play from here it is. Completely made up like the product is made up or completely.

09:28 Play from here Yeah, like the product is made up.

09:30 Play from here Like it's not even a real product.

09:31 Play from here It's not even a real product.

09:32 Play from here Get out of here. Where are you seeing that?

09:35 Play from here The New York Times.

09:36 Play from here So, of course, it's not the Times itself.

09:38 Play from here It's going to be their ad server that has to be targeted in whatever way their content

09:42 Play from here distribution widget that's sitting on the bottom of the page.

09:44 Play from here Yeah, exactly. And it's like coolest gifts for 2025.

09:48 Play from here And yeah, it'll be something that's completely fabricated.

09:52 Play from here And it's basically it's a it's a hook.

09:55 Play from here It's an ad hook to get you to get the click that gets you into the advertorial that runs

10:02 Play from here it down, you know, top 10 gifts that have affiliate links or whatever it is.

10:07 Play from here But that thing that that like weird looking made up Star Trek phaser looking thing, it

10:13 Play from here just just doesn't exist.

10:15 Play from here So, wow, that's actually problematic as far as both some of these new laws, which may

10:21 Play from here or may not ever take effect given what's going on with the Trump administration, but

10:26 Play from here just existing truth in advertising laws.

10:29 Play from here So, you know, you think it's cool to just like make something up.

10:32 Play from here But once you promise that to a consumer, you I'm not a lawyer, but you might be following

10:37 Play from here a foul of of other laws that are floating around there.

10:40 Play from here But more from a nuanced perspective is are you presenting an AI version of your product

10:46 Play from here that can't exist in some way, shape or form?

10:49 Play from here So it's not if it's not labeled as AI and it's presenting, oh, I don't know, your Chevy

10:55 Play from here and it's a Camaro that has features that simply don't exist.

11:01 Play from here And you're not disclosing that it's AI and that those features don't exist.

11:06 Play from here You could be following a foul of a whole bunch of existing.

11:09 Play from here Right. That one seems really obvious.

11:11 Play from here This idea of generating a product as a hook to get you to jump into an affiliate page.

11:18 Play from here That one is wild.

11:20 Play from here That to me, that looks like a business opportunity.

11:23 Play from here Whatever that thing that got you to click, Kevin, I'm making it.

11:25 Play from here And then I'm running ads for it because it sounds like it might need to exist.

11:31 Play from here They got you to click on to some top 10 list of the best Christmas gifts, but the one that

11:36 Play from here you wanted, you couldn't have.

11:37 Play from here Now I'm like scrolling on my email to see if I could find one of these things or if I'm

11:43 Play from here just making it up.

11:44 Play from here But they were really common.

11:45 Play from here If anybody wants to start a business, what we'll do is we're going to put Kevin's feed of

11:50 Play from here everything he's clicked on to try to buy that doesn't exist.

11:52 Play from here And we'll put that show you can go start a product, a DTC product company for that.

11:58 Play from here I am a terrible consumer, so I would never buy anything anyway.

12:04 Play from here But apparently the ad feed thinks I will.

12:08 Play from here That is pretty wild.

12:10 Play from here That is pretty wild.

12:10 Play from here Well, OK, anything else you want to talk about video as it applies to actual business?

12:15 Play from here Because we're in a weird time with it, as you mentioned.

12:19 Play from here I mean, we had our intro to the show, which may or may not look anything like us.

12:24 Play from here I don't know how many prompts it took to get there.

12:26 Play from here I've had some middle of the line luck with generating videos, some stuff I've been really

12:31 Play from here impressed with, others not so much.

12:33 Play from here Ironically, the best looking stuff for Sora, which is just really pure entertainment at

12:39 Play from here this point, if you want to call it that, looks more like me in most cases or when we've gone

12:45 Play from here back and forth and put each other's cameos in our videos, probably shouldn't be doing

12:51 Play from here that. But it's just kind of it's going to escalate and get worse and worse.

12:55 Play from here But that with an unbelievably little amount of training is pretty mind boggling how close

13:03 Play from here it looks. And as far as realism goes, it it is really shocking.

13:08 Play from here I had a moment, so I'm totally being that dad.

13:11 Play from here My teenagers are just cringing that I'm like sending custom Sora to videos in response

13:18 Play from here to their teenage missives.

13:19 Play from here And they're just dying inside.

13:21 Play from here They're just like, what did I do to deserve?

13:23 Play from here What did I do? How did I get here?

13:24 Play from here Where did I go wrong?

13:26 Play from here But I totally had a moment yesterday where I created one of those.

13:29 Play from here And I was like, that is totally my house.

13:33 Play from here Like, I get that it has a cameo of me, but that is a potted plant that is actually in my

13:39 Play from here house and sitting behind my kitchen sink.

13:42 Play from here Like, how the heck did it figure that out?

13:44 Play from here And then I realized that when you set these things up, you do a cameo.

13:48 Play from here So you do just whatever the 10 second thing is.

13:51 Play from here It's actually contextualizing more than your face.

13:53 Play from here It's adding all the other pieces that are behind you.

13:56 Play from here And since I just was goofing around with it and haven't bothered to do another one, to

14:00 Play from here be honest, where I happen to be standing had a view of my kitchen.

14:04 Play from here So it has a good idea of what my cabinetry looks like, what paint is on my walls, like

14:11 Play from here that my fixtures are like black.

14:13 Play from here So it's carried that over into other videos.

14:16 Play from here And sometimes if it's placed in my kitchen, it's going to be my kitchen.

14:20 Play from here Other times it will take that aesthetic and it'll actually apply it to other videos,

14:25 Play from here which is really wild when you think about it.

14:27 Play from here Yeah, yeah.

14:29 Play from here So what are you what are you seeing when you kind of look ahead as far as is this just

14:35 Play from here the open AI answer to try to monetize a.

14:39 Play from here social network?

14:41 Play from here Is this going somewhere?

14:42 Play from here Like, are you connecting the dots

14:44 Play from here for a business application with video right now

14:47 Play from here outside of creators using it, ad agencies using it,

14:52 Play from here D2C companies using it?

14:53 Play from here Like, where do you see this kind of fitting

14:56 Play from here into the AI for work?

14:59 Play from here A couple of different things.

15:01 Play from here If anywhere.

15:01 Play from here That's the easy answer.

15:03 Play from here It's pretty much everywhere.

15:05 Play from here You know, you always set me up to be the fun killer here,

15:08 Play from here but the fun killer aspect of it is these whole,

15:12 Play from here like, nasty little copyright issues

15:13 Play from here that have been floating around about all this.

15:15 Play from here They haven't really gone away.

15:18 Play from here Well, I heard Disney did a little something too.

15:20 Play from here Well, Disney just invested a billion dollars into open AI,

15:24 Play from here which, you know, just a few years ago,

15:26 Play from here a billion dollars would have sounded like a lot of money,

15:28 Play from here but it's like, eh, a billion, that's cute.

15:30 Play from here We'll give it a nod.

15:32 Play from here You know, you reached into your couch cushions

15:33 Play from here and you found a little shekel.

15:35 Play from here Yeah.

15:36 Play from here But the underlying issues still haven't gone away.

15:41 Play from here Like, were these models actually trained

15:43 Play from here on copyrighted material?

15:46 Play from here I think at this point, people have decided

15:50 Play from here that it's unlikely to be show-ending,

15:53 Play from here but there are a couple of different points

15:55 Play from here that people do need to be aware of.

15:57 Play from here That if you're using this material

16:00 Play from here and it does yield copyrighted material,

16:02 Play from here you could be held to account for that.

16:05 Play from here It's not really happening in mass and small brands.

16:08 Play from here It's probably not a problem, but it's an actual risk.

16:12 Play from here And then for a while, you had bigger brands

16:14 Play from here were all leaning into Adobe and its Firefly products

16:18 Play from here because Adobe was saying,

16:19 Play from here hey, they don't really have great video products yet,

16:24 Play from here but they're saying we're going to indemnify you.

16:27 Play from here So if you're a big company

16:29 Play from here and you're concerned about this stuff,

16:30 Play from here you could lean into Adobe and be confident

16:33 Play from here that you were getting results that were solid.

16:35 Play from here Ha ha, but guess what?

16:37 Play from here Now in the Adobe suite, you have dropdowns

16:39 Play from here that allow you to bounce into Veo.

16:42 Play from here And now you're crossing out

16:45 Play from here of your cute little Adobe indemnification wall

16:49 Play from here and you're pulling things in from Veo

16:52 Play from here that probably, again, not a lawyer,

16:55 Play from here that's twice this episode,

16:56 Play from here but probably don't cover you as far as that indemnification.

17:00 Play from here So, you know, big caveat here,

17:02 Play from here like if you're using these things,

17:04 Play from here the wise move is to stay away from knocking things off,

17:07 Play from here but that's exactly how people are finding success

17:10 Play from here is they're finding ways of getting really, really close

17:14 Play from here to somebody else's intellectual property

17:16 Play from here and they're parodying it,

17:18 Play from here they're creating like movie trailers.

17:20 Play from here And a lot of this stuff is really funny and really creative.

17:24 Play from here The other quick fun killer part is,

17:27 Play from here remember that nothing that is generate,

17:30 Play from here what is settled law is that nothing that is generated

17:34 Play from here from AI is actually copywritable at the moment.

17:38 Play from here Proving that whether it was, you know, generated by AI,

17:42 Play from here pretty easy in the case of a video,

17:44 Play from here harder in the case of a static image,

17:47 Play from here but be aware of that.

17:49 Play from here That could be a real problem

17:50 Play from here for people down the line somewhere.

17:52 Play from here Yeah, yeah, I really couldn't say at this point,

17:56 Play from here just where video is gonna have a strong play

17:58 Play from here in the workplace aside from generating

18:00 Play from here these creatives for ads.

18:02 Play from here That's the obvious one that I'm seeing

18:04 Play from here in real-time play out right now.

18:05 Play from here And there's a lot of money behind that

18:07 Play from here just because of the favoritism

18:08 Play from here of the current algorithm of the platform.

18:11 Play from here But what I thought was more interesting

18:12 Play from here about the Disney deal wasn't the billion dollar nod,

18:15 Play from here but if you kind of zoom out

18:17 Play from here from the big business perspective, right?

18:19 Play from here You've got Disney now saying you can use these characters

18:23 Play from here inside of Sora.

18:24 Play from here And it almost feels like Disney saying,

18:28 Play from here we're not an AI company,

18:30 Play from here but we appreciate and recognize where this is all going.

18:33 Play from here And we wanna be a part of that movement

18:35 Play from here and to stay relevant with all of the impending change,

18:39 Play from here we're gonna do this deal.

18:41 Play from here That to me felt pretty big.

18:43 Play from here That felt pretty big from a media company

18:45 Play from here for them to make that move.

18:47 Play from here Because as we talk about all the time,

18:51 Play from here just because you're not an AI company

18:53 Play from here or you're not building an AI product,

18:54 Play from here it doesn't mean that you aren't gonna see

18:56 Play from here this ripple effect of it hit every type of company,

18:59 Play from here every type of business.

19:01 Play from here And that's gonna range from law firms

19:01 Play from here to entertainment companies.

19:03 Play from here It's gonna be everywhere.

19:04 Play from here So I thought that was really the standout,

19:07 Play from here like, ooh, okay, that's a signal.

19:09 Play from here Like that's a big signal in the noise

19:11 Play from here as far as what's coming down the road.

19:13 Play from here Yeah, there's a lot there.

19:15 Play from here From a practical application perspective though,

19:18 Play from here it'd be good to dive into some of the things

19:20 Play from here that we're seeing out there in the wild

19:24 Play from here as far as how people are actually using this.

19:26 Play from here On one side, there is just the basic technology.

19:29 Play from here Disney is one of the biggest media companies in the world.

19:32 Play from here If you think that they're not looking at

19:33 Play from here how they can apply some of these technologies

19:36 Play from here in terms of their own production benefits,

19:38 Play from here you're crazy, right?

19:41 Play from here I think that, we've talked about this before,

19:44 Play from here but one of the neat things to consider

19:47 Play from here in how you think about applying AI

19:52 Play from here are basically three different buckets.

19:53 Play from here AI can amplify you in the things

19:58 Play from here that you're already doing

19:59 Play from here in new ways that you couldn't do before.

20:01 Play from here It can replace things that you do

20:04 Play from here that you'd rather not do, which is great,

20:06 Play from here unless, of course, it's an entire role that gets replaced.

20:09 Play from here That's kind of not great for that person.

20:12 Play from here And then third is creating new capabilities.

20:15 Play from here And that creation of new capabilities

20:17 Play from here is one of the places where video falls,

20:19 Play from here that humans are naturally visual and auditory-based.

20:25 Play from here We see each other, we talk to each other, et cetera.

20:28 Play from here And the idea that you can now produce

20:32 Play from here single-purpose disposable content,

20:35 Play from here either for personal entertainment,

20:37 Play from here like me torturing my teenagers,

20:39 Play from here or for being able to address a specific issue

20:45 Play from here within content production or something like that,

20:48 Play from here that is a new capability for most organizations

20:51 Play from here and that they just have never been able to touch before.

20:55 Play from here And I think that thinking about the ways

20:57 Play from here that your organization can apply video

21:00 Play from here in new and exciting ways that would have been

21:03 Play from here either cost-prohibitive or time-prohibitive in the past

21:06 Play from here that could make a difference

21:07 Play from here to the way you operate with your clients

21:10 Play from here or the way that you operate with your employees

21:12 Play from here is where a lot of the applications should be focused,

21:15 Play from here outside of just the pure marketing.

21:16 Play from here I mean, we can also talk about just pure marketing

21:18 Play from here and all the cool stuff we're seeing.

21:20 Play from here Right, right.

21:20 Play from here I mean, it really depends what your organization is doing.

21:23 Play from here And if you're using video

21:25 Play from here or if you're thinking about using video,

21:27 Play from here now, neither of us really lean too heavy

21:30 Play from here as far as, oh, we need to generate all this video

21:33 Play from here for our marketing and this and that.

21:35 Play from here You a little bit more than me.

21:36 Play from here You're actually in front of the camera

21:37 Play from here pretty often inside of your lab.

21:40 Play from here But we have also dabbled in that world of video

21:44 Play from here in regards to the way a lot of other people

21:47 Play from here commonly use it for marketing.

21:49 Play from here And we collectively came back to each other and said,

21:51 Play from here wow, this is either incredibly time-consuming

21:55 Play from here or really, really expensive or both, right?

21:58 Play from here So when you look at this being another one of those things

22:02 Play from here where all of a sudden it can enable things

22:05 Play from here that you just couldn't do before

22:07 Play from here because there was just no way

22:08 Play from here that you could get it into the budget

22:10 Play from here or you have the time or the staff.

22:12 Play from here I mean, think about how expensive it is

22:14 Play from here to go out and shoot some of the video

22:16 Play from here that we are seeing generated by VO3 right now.

22:19 Play from here So that's a big unlock.

22:22 Play from here I really just think it depends

22:23 Play from here how your organization might be using it.

22:26 Play from here And, or if you haven't before

22:27 Play from here how it just opens so many doors, right?

22:29 Play from here You're seeing some really neat stuff

22:31 Play from here about product visualization.

22:33 Play from here I think it was Hasbro who put out a bunch of videos

22:36 Play from here about new toy designs.

22:38 Play from here And as opposed to having a storyboard,

22:41 Play from here which I think a lot of us can picture a storyboard

22:44 Play from here with a bunch of frames on it.

22:45 Play from here You can actually run 10 different prompts

22:49 Play from here in one of these systems

22:51 Play from here and come up with a really visual representation

22:55 Play from here of what you're talking about.

22:56 Play from here So do you want it to have a pink theme

22:58 Play from here or a blue theme or whatever it is?

23:00 Play from here You just visually can key off of it.

23:02 Play from here Now, a company like that,

23:04 Play from here well, I would say that they're not gonna go production

23:06 Play from here to production with that, but it's pretty close.

23:09 Play from here They're gonna be,

23:09 Play from here soon they will be going to production with that stuff.

23:13 Play from here But at the very least,

23:14 Play from here it gives this leg up to their production cycle.

23:17 Play from here And even since like the beginning of 2023,

23:20 Play from here I've been talking about something I call cycle skipping.

23:23 Play from here It's the idea of looking at your business

23:25 Play from here and looking at the things you're doing

23:27 Play from here and then questioning why things take so long.

23:30 Play from here And for those of us who have spent a lot of money

23:33 Play from here and a lot of time developing media spots,

23:35 Play from here it's always this like six month, eight month process

23:39 Play from here of meetings and back and forth.

23:41 Play from here And this ideation session

23:42 Play from here and this Pantone session or whatever.

23:44 Play from here And it's like,

23:45 Play from here okay, what this probably needs to look like

23:48 Play from here is all the smart people get into a room

23:50 Play from here and they talk about what the objectives are.

23:53 Play from here Okay, then they leave and they come back the next day,

23:57 Play from here everybody with 10 highly visual representations

24:01 Play from here of what the heck they're talking about.

24:03 Play from here And at this point,

24:04 Play from here they'd be video representations

24:06 Play from here of what they're talking about.

24:08 Play from here The creative director or CMO,

24:10 Play from here whoever it is, looks at it all and says,

24:11 Play from here wow, that's a really good direction.

24:13 Play from here We're gonna go in that direction.

24:14 Play from here You just cut three months off of your planning timeline

24:18 Play from here and all of these intermediate meetings

24:20 Play from here and all of this stuff.

24:21 Play from here And there are people who will cry foul to that.

24:23 Play from here Oh, the creative process takes a long time.

24:24 Play from here Okay, fine.

24:25 Play from here So punt it out a week.

24:27 Play from here So you have the first meeting,

24:29 Play from here let people chew on it a little bit

24:32 Play from here and then have a first production meeting.

24:34 Play from here And guess what?

24:35 Play from here Because you're not really filming anything,

24:37 Play from here you can tweak it all you want.

24:38 Play from here And when you wanna call in

24:40 Play from here that really expensive video crew to do it,

24:42 Play from here you have your aesthetic already defined.

24:44 Play from here You have your characters already defined.

24:46 Play from here Right.

24:47 Play from here And how many places in your business

24:49 Play from here can you be doing that?

24:50 Play from here How many places do you have?

24:53 Play from here You have schedules that are built

24:54 Play from here because the schedules were built there

24:56 Play from here because that was what was demanded

24:58 Play from here of humans operating the way humans do.

25:01 Play from here Legacy workflows.

25:02 Play from here Yeah, legacy workflows.

25:04 Play from here And we both see clients

25:06 Play from here who want to just like replicate,

25:08 Play from here you know, there are 72 steps in the process.

25:10 Play from here How can we automate steps 17, 19, and 24?

25:16 Play from here Okay, but why don't you look at the entire workflow

25:19 Play from here and figure out what's going in

25:20 Play from here and what's coming out

25:22 Play from here and how much time is involved.

25:24 Play from here Find the key points where humans need to be involved

25:26 Play from here and lean in there.

25:28 Play from here I think that's really the best way

25:30 Play from here to holistically look at

25:31 Play from here what's happening with video right now.

25:33 Play from here Because as you mentioned,

25:34 Play from here the amount of time and money it takes,

25:36 Play from here it would take you in the real world

25:38 Play from here to get some of these assets that VO3 is producing.

25:41 Play from here I mean, it's not even apples to apples anymore.

25:45 Play from here It's completely crazy.

25:46 Play from here But what you said about having spent time

25:50 Play from here in D2C and in marketing

25:52 Play from here and in the creative process and in ads

25:54 Play from here and how it really is both an art and a skill

25:58 Play from here to be able to reach the right audience

26:00 Play from here to get the right creative,

26:01 Play from here none of that goes away with this at all.

26:05 Play from here None of that goes away.

26:06 Play from here If anybody has ever tried to produce something with VO3

26:11 Play from here where you had some expectation

26:13 Play from here and you are maybe not inherently a creative

26:16 Play from here or a videographer or have ever done ads,

26:20 Play from here you will be really disappointed really quickly

26:22 Play from here and probably get frustrated

26:23 Play from here after five or 10 failed outputs

26:26 Play from here and then it says you're out of tokens

26:27 Play from here and you got to go buy a pro license or something.

26:30 Play from here We talk a lot about one thing AI hasn't done yet

26:33 Play from here is be an arbitrary of taste.

26:36 Play from here If you don't have the skillset,

26:37 Play from here if you don't have the experience,

26:38 Play from here you don't really know a good output from a bad output.

26:41 Play from here And that hasn't changed.

26:43 Play from here It hasn't changed with video.

26:45 Play from here It hasn't changed with a ton of the roles

26:46 Play from here that we talk about,

26:48 Play from here including the person who says

26:50 Play from here I want to take out steps 17 through 23.

26:53 Play from here It's like, well, if you didn't really understand

26:56 Play from here how this whole thing worked to begin with

27:00 Play from here and you don't understand how to really apply AI

27:02 Play from here from a high level,

27:03 Play from here then those pieces aren't going to fit together.

27:06 Play from here That's where we always talk about

27:09 Play from here not just a human in the loop,

27:10 Play from here but a human kind of sitting on top of this whole process.

27:14 Play from here Now, a hundred percent.

27:19 Play from here It's...

27:19 Play from here I'm Elijah Frost,

27:34 Play from here a serial entrepreneur with 28 years of experience in tech.

27:36 Play from here And now I run an agency that builds AI solutions.

27:39 Play from here Hey, I'm Kevin Williams.

27:41 Play from here I spend most of my time helping companies

27:43 Play from here figure out how to actually use AI.

27:46 Play from here Not in theory, but in the work itself.

27:49 Play from here This show is where we break down

27:51 Play from here how people are really using AI,

27:53 Play from here what's going right,

27:54 Play from here what's going wrong,

27:55 Play from here and how it's changing the way

27:56 Play from here that we all get work done.

28:01 Play from here Hey, Kevin, we're back.

28:02 Play from here We're back.

28:03 Play from here We're back for our very first ad interlude

28:07 Play from here that neither of us was particularly well prepared for.

28:09 Play from here Very meta in terms of ads.

28:11 Play from here You know, the irony is I had totally lost track.

28:14 Play from here I was so obsessed with thinking through

28:17 Play from here the creative, that creative input piece

28:21 Play from here that you were talking about

28:22 Play from here that I think I was about to have

28:23 Play from here one of those blank stare moments.

28:25 Play from here So thank you, ads, for saving me there.

28:30 Play from here I was busy doing the thing we do best,

28:32 Play from here which is just going down the rapid hole

28:34 Play from here on a single topic,

28:35 Play from here and then all of a sudden an hour and a half goes by.

28:38 Play from here Listen, okay, I'm going to stop.

28:39 Play from here I'm going to stop with the video right now

28:41 Play from here because this is how we roll.

28:42 Play from here We'll just keep going on this forever.

28:43 Play from here Yeah.

28:44 Play from here So, you know, let me say,

28:46 Play from here we always say, oh, we're not a news show,

28:49 Play from here and this is really about practical stuff at work.

28:51 Play from here And the last couple of weeks

28:53 Play from here have just been a flurry of news

28:55 Play from here around studies and regulatory.

28:58 Play from here And there's some stuff that we can't really ignore.

29:02 Play from here Before we get into Trump's executive order,

29:04 Play from here which I'm like super curious to get your take on this.

29:07 Play from here I've got a lot of conflicting feelings about it.

29:10 Play from here I'm backing up to that,

29:12 Play from here and I'm not going to go into some of the studies

29:15 Play from here that I've been reading lately,

29:17 Play from here but there's a lot of just.

29:19 Play from here opposing data out there as far as the future of work with AI, how many people are using it,

29:28 Play from here how many companies are laying people off or doing hiring freezes because of AI, and what's going to

29:35 Play from here happen in 2026. So my question for you, because we're both out there talking to a lot of different

29:42 Play from here businesses and a lot of different types of businesses all the time, anecdotally, okay,

29:47 Play from here what hits closer to you when you're talking to these different companies and people? Are you

29:55 Play from here surprised with how much they're using AI? That would be A. All right. Are you shocked with how

30:01 Play from here little they are using AI given the different tools out there and just obvious easy wins?

30:08 Play from here Or are your expectations pretty much level set? Is what you're hearing kind of what you'd expect?

30:17 Play from here Which one of those three kind of hits closer to you? Yes. All of them. All three. All of the

30:23 Play from here three, depending on the industry, depending on the company, depending on the size, depending on

30:27 Play from here how close they are to what I am right now going to call the existential horizon. So that's a great

30:33 Play from here term. I want to use that again. It's pretty close to the event horizon. But it's the existential

30:38 Play from here horizon of if you are a marketing agency, if you are a video production agency, if you were a copy

30:46 Play from here agency, because you're probably already toast, you are all over this. And you have to be because you

30:55 Play from here just have no choice if you're paying any attention to your business at all, if you care at all.

31:01 Play from here These are the very first people I was talking to all the way back at the end of 2022 into 2023.

31:07 Play from here They were on it and they still are on it. Now that pervading the entire organization is different.

31:14 Play from here It's still at the leadership ranks and then building a shared understanding and say a two

31:22 Play from here or 300 person company about what this means for the entire company versus just what it means for

31:27 Play from here the product. We have a long way to go there. I am still weekly, daily meeting with very large

31:36 Play from here companies that are tech forward in their identity that still haven't really gotten over the hump

31:45 Play from here as far as employing things in a really useful way. You hear a lot of news about cutting edge

31:54 Play from here Fortune 1000, Fortune 5000 type companies are getting over the pilot hump, but they're spending

32:00 Play from here a lot of time, a lot of money. They're spending effort to educate people. They're willing to learn

32:06 Play from here through failure in a lot of cases that is helping propel them forward. But most of the economy is

32:14 Play from here still just figuring it out. Right. Why do you think that is? What do you think the biggest barrier is

32:21 Play from here for this just adoption at a very base level that gets people to the next step? Like if you had a

32:30 Play from here name and again, you could probably say, well, in this industry, it's this or this business, it's

32:34 Play from here this. But is there any sort of just connective tissue that you're seeing across those different

32:40 Play from here companies where it stands out and you say, oh, this is either the mindset shift or the cultural

32:48 Play from here shift or the change management shift that really has to happen to start moving in the direction

32:53 Play from here that is inevitably going to sweep over everybody. So from a sales and marketing perspective,

32:58 Play from here it's been a bit painful because we keep thinking that we are going to end up the AI enablement

33:03 Play from here company for this vertical or that vertical or whatever. And we'll come in, we'll solve some big

33:08 Play from here juicy problem for somebody. We'll save them a ton of money. And then we'll waltz over to somebody

33:13 Play from here else in the industry and say, hey, we've solved this big juicy problem. And it doesn't usually

33:18 Play from here convert. And it doesn't usually convert because the issue is less of a demand or a need problem

33:25 Play from here and more of a leadership problem. And organizations that are ahead of this are organizations that have

33:31 Play from here taken a stand from the very top. So on the show, we've talked about AI manifestos before. So often

33:39 Play from here it's one of the first things that we do in an organization is we help literally the board,

33:45 Play from here the CEO or the founder, often who is a member of all of those. I really don't like it to be

33:51 Play from here punted to anybody else in the organization to take a stand on how they think about artificial

34:00 Play from here intelligence, how they think about the future of their business, the future of their industry,

34:05 Play from here the future of their people, and put forward a mandate that aligns all of those pieces

34:11 Play from here into capacity building. Because most organizations, and I do mean most, still treat it as like, oh,

34:18 Play from here there's this tool that does this thing. And then they wander around with a hammer looking for a

34:23 Play from here nail, as opposed to understanding that it's a pervasive technology. And you have to be opportunistic

34:30 Play from here about where you're applying it. And you can't be opportunistic about where you're applying it

34:34 Play from here unless you and your people understand how that's going to work out. So the very best organizations

34:42 Play from here lean into it from the top, develop capacity throughout their ranks, and then they start

34:49 Play from here uncovering those low hanging fruit sort of opportunities, high frequency, low risk, high

34:55 Play from here yield type automations that take a bunch of crud off of people's desks and make them happy.

35:02 Play from here If you want to build a culture of AI, solve a bunch of problems that people hate in the

35:09 Play from here organization. It's like the office space TPS reports. Nobody wants to do a TPS report.

35:14 Play from here The work you really don't want to do, the work you dread, right?

35:17 Play from here And all of a sudden, people are like, oh, I get it.

35:18 Play from here The tedious pixel pushing across different little Excel boxes, that kind of stuff. Yeah,

35:25 Play from here yeah, totally. Well, speaking of top down, let's talk about this. Let's talk about this

35:31 Play from here executive order with AI and Trump, because I think there is some real world impact with businesses.

35:36 Play from here And I know that you are pretty close to some government work that's going on locally here

35:41 Play from here in Utah. So what's your take on this? You want to unpack it?

35:46 Play from here Oh, boy. You're going to make me say not a lawyer again for the third time.

35:49 Play from here Listen, it's just an opinion. It's just an opinion. No, no, no, no, no.

35:53 Play from here It's really just what happened. Like, what happened?

35:55 Play from here I think everybody knows I'm not a lawyer, but what happened was that the administration

36:01 Play from here basically threw down with an executive order that said, directionally, if you do things

36:09 Play from here that we don't like, we're going to make your life super hard. It also had a bit in there about

36:16 Play from here ideological leanings of models. So this is the anti-wokeness thing. That is a whole rabbit

36:22 Play from here hole that we probably shouldn't go down in this case. But models are trained on big sets

36:27 Play from here of information. And whether those big sets of information are biased left, right, center,

36:31 Play from here or whatever, that's sort of the fault of the information, not the model. And when you start

36:36 Play from here trying to pull levers to rebalance it, you end up with things like Google's African-American

36:43 Play from here founding father type stuff from a few years ago. Anyway, so it's just not practical.

36:51 Play from here The other part is, I can't not say not a lawyer, but it's not an executive power to be able to say,

36:59 Play from here hey, states, don't pass laws. Congress can do that, but they haven't chosen to. So the reading

37:07 Play from here of it is states that are really leaning heavily into regulation. Colorado, believe it or not,

37:17 Play from here Texas, New York State, and New York City, Illinois. I mean, they're all 50 states.

37:22 Play from here Utah's got some stuff too, yeah.

37:24 Play from here Utah does, but I'm kind of leaving Utah out because it's unclear who's on the nice list

37:31 Play from here and who's on the naughty list as far as essentially what is impairing commerce.

37:38 Play from here And Colorado's law is the closest that you're going to come to some of the laws in the European

37:43 Play from here Union, which are sort of inarguably anti-development, that they will make it really

37:50 Play from here hard and really expensive to deploy systems in Colorado. So assumedly, Colorado's received a

37:56 Play from here nasty gram from the Department of Justice saying, we're going to remove your funding for, let's say,

38:03 Play from here broadband. So that's what they're, one of the big sticks they have is billions and billions of

38:07 Play from here dollars in broadband development, and they're going to pull that funding from you if you don't

38:11 Play from here listen to them. Obviously, there are going to be lots of lawsuits, but it's going to slow down the

38:17 Play from here pace of that sort of regulation. My bias is that we do need some common sense regulation that sits

38:23 Play from here at the top, that protects things like our own likenesses, that, oh my goodness, maybe protects

38:29 Play from here children from this stuff. That's a big one. Common sense laws that places like Texas are

38:35 Play from here trying to pass. Look at Australia, kids can't use social media if they're not 16 now.

38:41 Play from here Yeah, as of this week. Interesting lab there to see how that plays out over the next five,

38:46 Play from here ten years. Very curious. But it's not going to be effective to have 50 states' worth of laws.

38:52 Play from here So I'm sympathetic to the movement. Where I'm not sympathetic is that there's no real guidance

38:59 Play from here about what this is supposed to look like anyway. But from an effective perspective,

39:06 Play from here there's a lot of existing laws on the books that companies need to be aware of. And just because

39:11 Play from here it's AI doesn't obviate your responsibilities under those existing laws. Earlier, we were

39:17 Play from here talking about fairness in advertising, for example. But the EEOC, as far as equal opportunity

39:26 Play from here laws and hiring in housing, there's FERPA and finance. This stuff doesn't care that it's AI.

39:33 Play from here If you're injecting bias into a process. And it's all being impacted, right?

39:37 Play from here I mean, it's all being impacted. So you don't necessarily need these new laws. And believe me,

39:43 Play from here from a political focus perspective, I want the fewest number of effective laws and regulations

39:48 Play from here you have to accomplish the goal, right? And if you already have laws on the books that should

39:53 Play from here be being interpreted in this way, great. The problem is that there's so much of a gray area

40:00 Play from here in terms of what that interpretation means without regulations that define what that

40:07 Play from here interpretation means. And I don't want to be mealy-mouthed on it. But a lot of the issue

40:12 Play from here is with, say, measurement. And you'll hear me harp on this a lot. But the other thing that I'm

40:17 Play from here seeing, even in big companies, is no one spending the time on performance measurement or determining

40:24 Play from here whether or not the models are drifting or evolving over time. Just this week, we went from GPT 5.1

40:31 Play from here to 5.2. Well, if your models were connected or your apps were connected to 5.1, and now they're

40:36 Play from here connected to 5.2, are you checking the outputs? Are they drifting? Is it still performing the

40:42 Play from here way that it was before? If you are not and you can't demonstrate that, you might be following

40:48 Play from here a foul of who knows how many other regulations are out there. But forget that. Even if you don't

40:54 Play from here care about that and you believe that you're not going to get sued and whatever, because you

40:58 Play from here probably won't. At the end of the day, it's probably not going to surface anytime soon.

41:03 Play from here You need to care as an organization about the performance of the tools you've laboriously built,

41:09 Play from here because if they have drifted and if they're performing poorly, people will stop using them.

41:15 Play from here And all of the time and all of the effort that you've spent developing whatever that cool solution

41:20 Play from here is, is going to go to naught. I got a little passionate there. You did. You did. I don't know

41:25 Play from here about performance. You just kept going. I'm going to let Mealy Mouth slide on this. I was trying

41:30 Play from here to connect the dots and I think I heard it referenced in a movie I saw from filmed in like

41:34 Play from here 1948. I'm a boomer and a Midwestern. Neither of which is true. A movie that was shot in the

41:44 Play from here Midwest, by the way. So I think it's worth zooming out, though, for people who don't know and ask,

41:50 Play from here well, why the executive order? What's the intention behind it? We get into all the state

41:56 Play from here legislation. We get into all of the business impact. We get into all the punitive measures.

42:00 Play from here But if you zoom out, why? Why? Why? Why? Why do they care so much about making this a federal

42:08 Play from here mandate as opposed to letting states do their own thing? What is the reason? China, you know,

42:15 Play from here China, China, China, China. Yeah. So which which which you won't often hear. I mean,

42:24 Play from here you will hear kind of secondhand. You won't often hear the administration say this is all

42:31 Play from here because. And when you start then unpacking all of the potential things that could happen with

42:37 Play from here artificial general intelligence for a lot of people, including some people that might be

42:43 Play from here listening to this, it sounds very science fictiony, but it is real. Like what is what is

42:48 Play from here your take on that as far as just the Wild West and just cowboy this thing and build as fast as

42:53 Play from here possible and break things versus some of the potentially bad things that could happen by

42:59 Play from here doing that? Right. Because this is this is kind of a big deal when you do zoom out and look at

43:04 Play from here this holistically and where we might be headed. I tell you, this one, it is so hard. You know,

43:10 Play from here if I was David Sachs and like whispering in the president's ear, I would probably lean more into

43:17 Play from here some of the common sense guardrails and approaches than I would into anything that is restricting

43:25 Play from here the pace of development because we're not operating in a vacuum. And the fact is that

43:33 Play from here China really has many times our intellectual heft in terms of resources that they can throw

43:40 Play from here at a lot of these problems. They're behind in hardware, they're behind in certain aspects of

43:46 Play from here it, but they just have numbers on their side. And you don't want to turn this into like a new red

43:52 Play from here scare. But it kind of is in that these models, as they become more capable,

43:59 Play from here I mean, it's adorable that we can use them to create cute little videos that we send

44:04 Play from here to our teenagers, but it's less adorable to think about a broad scale development of cyber

44:13 Play from here penetration systems and new warfare systems, all kinds of things like that.

44:20 Play from here And it's not just the idea of prevention, like, oh, we have a bigger model than you

44:29 Play from here do so we can, you know, the whole arms race sort of deal.

44:34 Play from here It's that you're going to need a bigger toolkit to fend off the increasingly sophisticated

44:40 Play from here attacks that are heading your way.

44:43 Play from here And if your toolkit does not have the sharpest tools in it, you're going to, it's really

44:47 Play from here going to be a problem.

44:50 Play from here I really think that you kind of ain't seen nothing yet.

44:55 Play from here The capabilities of the models are already getting to a point that is super eye-opening

45:01 Play from here and we're not even seeing the best of.

45:04 Play from here We're not seeing whatever GPT-6 looks like.

45:07 Play from here We're not seeing whatever Palantir and crazy organizations like that are leaning into yet.

45:13 Play from here And it's only been a few years.

45:15 Play from here So yeah, I'm super taken with national security, long way of saying national security.

45:21 Play from here I mean, because that's, it's just, it's strange for some people to connect those dots of here

45:27 Play from here we are talking about ways that AI can make you more productive at work and do the things

45:32 Play from here that you hate doing and let you be more efficient or let you have new creative outputs.

45:38 Play from here But the very same base technology is also now being talked about as far as national

45:43 Play from here security.

45:44 Play from here That was a quick break.

45:48 Play from here We're back, Kevin.

45:49 Play from here We're back.

45:50 Play from here A little intermission, a little riff, a little riff.

45:53 Play from here Yeah, that was interesting.

45:56 Play from here Anyways, anyways, I think that depending on how much people know about this and where

46:03 Play from here it's going, you could be really knowledgeable about how this is making your job easier,

46:08 Play from here letting you do things that you couldn't do before.

46:11 Play from here But maybe it's not on your radar that, oh, wow, this very same tech that I'm using has

46:15 Play from here the potential to be a massive national security risk.

46:19 Play from here If put in the wrong hands and by wrong hands, I mean, the hands that aren't ours, really.

46:23 Play from here So I don't want to over what desk you're sitting on in what country, if you want to explore

46:28 Play from here this audience, go check out Project 2027.

46:32 Play from here And this came out about a year ago, as far as this, this, this hypothetical situation

46:37 Play from here of different types of models and arms races and how this plays out by all means wreck

46:42 Play from here your weekend by reading that.

46:44 Play from here But you should be aware of it because it's not impossible to know.

46:48 Play from here I mean, because when an order like this comes out, it's good to know why, right?

46:53 Play from here If your experience with it or your, your first hand applications are, oh, it helped me reply

46:58 Play from here to this email because I have Chepenhine in my Gmail.

47:02 Play from here What is this all about?

47:03 Play from here It is really good stuff just to be aware of because however these laws end up playing

47:09 Play from here out, they will affect all of us and the way we're able to use these tools.

47:13 Play from here Well, I don't want to, I really don't want to dive into any sort of political lens here,

47:18 Play from here but you see-

47:19 Play from here Sure you do.

47:20 Play from here But you're not going to.

47:21 Play from here You see the Gemini for government just launched in the Department of War.

47:26 Play from here It is pretty funny.

47:27 Play from here If you look at the Google press release on it versus the Pete Hesgeth, Secretary of War,

47:35 Play from here his statement on it, Hesgeth is all about rah-rah-ree, this is going to help us dominate

47:40 Play from here in the field.

47:41 Play from here And Google's like, yeah, this will help you with your procurement orders.

47:45 Play from here And it's like, eh, you know, all right, there's a, there's a reality somewhere in between.

47:50 Play from here But the, the, the fact that we're leaning into these technologies in even the military

47:55 Play from here structure is important, not because it's the military, but because I do believe it's, it's

48:00 Play from here likely to be first move of how the administration is going to deal with the rest of government.

48:07 Play from here So here they've taken a chainsaw to government through Doge, like dislodged all of these

48:13 Play from here people.

48:14 Play from here A lot of that work still needs to happen.

48:17 Play from here And you now have a clearing of the field to a degree of starting to apply these technologies.

48:23 Play from here And I'm sort of crossing my fingers that that's the plan, you know, not, not because I want

48:28 Play from here to begrudge federal employees or anything like that, but the work still needs to be

48:32 Play from here done.

48:33 Play from here And there's a lot of work out there.

48:35 Play from here So if we're leaning into the tooling from, from our own business, the business of doing

48:43 Play from here the people's business, basically.

48:46 Play from here And we have to clear the decks a little bit to happen, have, make that happen.

48:50 Play from here I think that's one lens that's out there.

48:53 Play from here You know, the other is just, it is just purely commercial that it's, this is just like, there's,

49:00 Play from here I mean, now our entire economy is basically being built around AI.

49:03 Play from here So it would be a bad thing for that to go poof anytime soon.

49:07 Play from here So let's make sure that we're not too much, too much fiscal momentum behind it for it

49:12 Play from here to go poof.

49:13 Play from here There's a little bit of fiscal momentum behind it right now, which kind of like keeps the

49:17 Play from here world turning in a certain direction.

49:19 Play from here You know, if we're going to be honest, like that's it, a lot of it comes down to the money

49:24 Play from here for sure.

49:25 Play from here Listen, Kevin, we're going to be running up on time soon.

49:30 Play from here And there was like all of these underused power tools you wanted to talk about that

49:35 Play from here are making you more effective in your organization and other people.

49:40 Play from here And we're not going to get to all of that.

49:41 Play from here So I want to leave it with this.

49:45 Play from here I have this idea.

49:46 Play from here Have you ever heard of a life ledger?

49:48 Play from here Have you heard of this?

49:49 Play from here No, I mean, I know.

49:51 Play from here Okay.

49:52 Play from here Yeah.

49:53 Play from here I made it up.

49:54 Play from here That's why you haven't heard of it.

49:55 Play from here There's this exercise I did where I wanted to see how my time was being spent and it

50:00 Play from here felt like the days were going by too quickly or I wasn't being productive enough or I wasn't

50:05 Play from here being good enough.

50:06 Play from here Dad or whatever the, you know, fleeting racing anxiety stuff going in my head was.

50:10 Play from here And so there's this exercise I read about where in 15 minute increments, you actually

50:15 Play from here just journal what you're doing throughout the day.

50:17 Play from here And then at the end of the day, you look back and you say, okay, this is how I spent my

50:21 Play from here time.

50:22 Play from here I want to do something similar with you with using AI, because I'd say we're both very

50:28 Play from here heavy users of AI in many different parts of our business.

50:31 Play from here I think it would be interesting to log a day.

50:36 Play from here All of the different things that you're using AI for, I've got like a zillion custom models

50:44 Play from here that I've built for all sorts of use cases that I'm sharing with team members that I

50:48 Play from here don't even think about, but I'm effectively using them because it's something my team

50:53 Play from here member is supposed to be doing and they're using the model that I made.

50:56 Play from here And we could look at some of those and I bet that there will be so many that we won't have

51:01 Play from here time to go through them in an hour, but we could compare some of those notes.

51:06 Play from here If you were to guess how many line items, like a hundred for a day?

51:11 Play from here Oh, it's probably ridiculous.

51:13 Play from here I'm looking at my monitor right now.

51:16 Play from here I think I see seven Gemini, four Clawed, six GPT that I have open right now.

51:25 Play from here And then that's not even Notebook LM.

51:27 Play from here All doing specific things.

51:28 Play from here All doing specific things.

51:30 Play from here Yeah.

51:30 Play from here I mean, some of them were like yesterday stuff that I moved on or whatever.

51:33 Play from here I sort of leave tabs open, but, and you know, the new one is definitely Notebook since we

51:40 Play from here spent like all last week talking about Notebook.

51:43 Play from here But I have six Notebook LMs open right now that I'm working on with different projects.

51:49 Play from here I'm still just getting into it.

51:51 Play from here We were sending some things to each other over the past week that we had put together

51:56 Play from here a Notebook LM and yeah, I'm finding more and more that I can do with it.

52:00 Play from here I'm really digging it.

52:02 Play from here Before we jump, you did mention GPT 5.2.

52:06 Play from here Any hot take on that?

52:07 Play from here Have you been beating on it much?

52:09 Play from here Any real notable things that you saw come out of that, that felt different in that model?

52:15 Play from here Boy, it's really hard for me right now.

52:17 Play from here I still, the pattern I'm trying to develop right now is in Gemini.

52:22 Play from here I'm trying to lean into Gemini because Gemini is integrated with my business documents and

52:29 Play from here everything else.

52:31 Play from here Like it is well positioned.

52:34 Play from here I have GPT 5.2 open.

52:37 Play from here I don't have any really firm opinions on it.

52:41 Play from here But you have a ton of custom GPTs, right?

52:45 Play from here You know, I have more projects than custom GPTs.

52:48 Play from here So because we don't have a ton of like little repeatable tasks like that, we're doing more

52:55 Play from here knowledge management across entire client accounts.

52:58 Play from here So we will have projects that contain many chats and that's starting to work its way

53:05 Play from here into our workflow more where it's normal for me to say, toss my business partner a chat

53:13 Play from here link that he can pick up and run with as opposed to some sort of a summary document there.

53:21 Play from here So, I mean, I do have a lot of custom GPTs.

53:24 Play from here We probably have 20 or so, but most of those are just little things I built for myself

53:28 Play from here to solve my own stupid problems.

53:30 Play from here OK, not really worker business related so much.

53:34 Play from here Well, it's your own little stupid business problems like that, too, right?

53:37 Play from here Yeah, but, you know, we're trying to be practical here.

53:40 Play from here Like, you know, here we're coming up at the end of the end of the year.

53:43 Play from here So I have this this bot that's designed around my small business accounting software that.

53:48 Play from here Yeah. Yeah.

53:49 Play from here I think I get really familiar with using those tools so much because a lot of them are daily

53:55 Play from here drivers of, OK, I have this custom GPT.

53:57 Play from here It does this thing. It gets used every single day.

54:00 Play from here And when a new model drops, you'll see the little flag up on top that says, oh, the publisher

54:07 Play from here made it with however old it is with with GPT-4.

54:12 Play from here Would you like to revert back to some old model?

54:14 Play from here And it's it's always interesting to see the difference in the outputs with those models

54:19 Play from here where I'm very familiar with the types of outputs I was getting and see how they evolve.

54:24 Play from here And very often it's just kind of vibes like, oh, this feels smarter or this wasn't as

54:32 Play from here verbose or this is more verbose or this isn't quite as sycophantic.

54:39 Play from here I would say overall, it feels like an improvement, like sometimes these things will drop.

54:45 Play from here And I think, oh, this is just an awful regression in terms of how effective this feels.

54:51 Play from here This this felt like an improvement to me.

54:53 Play from here But I, too, am also moving a lot over to the whole Gemini ecosystem for the same reasons

54:59 Play from here you mentioned. Really, I think innovation, just how it's already built into all of the

55:04 Play from here tools I'm already using.

55:06 Play from here And at the same time, you had said before that you were very heavily weighted towards

55:12 Play from here OpenAI, like as a business, it sounds different and feels different, too.

55:17 Play from here Like there is a weird kind of thing when you spend enough time in these tools that you

55:20 Play from here really do kind of feel like there's this personality behind it.

55:24 Play from here And the Gemini personality does feel very different than the GPT personality.

55:30 Play from here And it feels different from Claude.

55:32 Play from here And it feels different from Claude.

55:33 Play from here Claude is still my favorite model.

55:36 Play from here It really is.

55:37 Play from here If I am doing something that is that I really feel like I need a thought partner for in

55:42 Play from here the communication aspect, both to me and from me is vital.

55:47 Play from here I'm going to lean into Claude.

55:49 Play from here GPT still feels a little bit more sycophantic and sort of like trying to curry favor.

55:55 Play from here You're the smartest guy ever, Kevin.

55:57 Play from here I can't believe you thought of that.

55:58 Play from here And Gemini, Gemini, it feels like it's the workhorse.

56:02 Play from here Like I would have called GPT my workhorse before, but now it's kind of sticking to the

56:07 Play from here facts and the facts can be huge.

56:11 Play from here Just for curiosity, I brought up a Notebook LM and I think it has 65 different sources

56:16 Play from here in it. And now I'm really, really leaning in.

56:20 Play from here It's a different use case.

56:21 Play from here It's very, versus creating a gem or just one chat.

56:26 Play from here That many different sources, you're talking about different documents you've attached

56:29 Play from here or different documents you've attached.

56:31 Play from here Yeah. So, yeah, it's remarkable.

56:35 Play from here And it's levering what's a really underappreciated benefit of Gemini is it just has the biggest,

56:41 Play from here baddest context window out there.

56:43 Play from here Right. When you're hitting that number of documents, I don't have any that are that big.

56:47 Play from here When you're hitting something like 60, are you noticing any sort of latency issues or do you have to

56:52 Play from here specify routing as far as like what you want it to look like?

56:55 Play from here Because that's a lot to that's a lot to crawl through every time you say, oh, make me a

57:00 Play from here this. You know, are you being specific as far as, hey, there's I think these 10 documents

57:07 Play from here uploaded, they're going to be the most useful lean into those.

57:10 Play from here How are you how are you handling that when you have that much context in these attachments?

57:16 Play from here It's funny you say that because I tested this.

57:18 Play from here I was I was I was testing this out with that notebook LM right before we got on air.

57:23 Play from here And one, I just gave a basic prompt to like, give me an infographic that shows all of this

57:29 Play from here stuff. It was literally that simple.

57:32 Play from here And then the second one, I used Gemini itself.

57:37 Play from here So Gemini within the notebook, so the chat box in the notebook.

57:42 Play from here And I said, help me construct a prompt for an infographic that keeps X, Y and Z in mind

57:49 Play from here and pulls from these documents.

57:52 Play from here And then I worked with Gemini in notebook for a couple of rounds to make sure that the

57:59 Play from here prompt was covering the things I wanted it to cover.

58:02 Play from here And I was happy enough.

58:04 Play from here And then I just dumped that prompt into the infographic generator.

58:09 Play from here Oh, thingamajigger. And yeah, the difference is really stark.

58:13 Play from here So the one that was focused on tell me stuff about this, it went pretty superficial.

58:21 Play from here And the one that was even though all of the conversation today had been about some pretty

58:27 Play from here highly technical stuff as far as data feeds in agentic commerce and stuff like that, it

58:33 Play from here went high level because that was the sort of the base of the pyramid, I suppose.

58:38 Play from here And we were dealing with the.

58:39 Play from here of the pyramid, whereas the prompt was at the pointy end of the pyramid, and it yielded

58:45 Play from here something totally useful. Although I am seeing in the spelling, which at this point,

58:49 Play from here I'm disappointed, Gemini, I'm disappointed.

58:52 Play from here On the spelling and the infographic.

58:54 Play from here In one shot, massive infographic, and you know, we can't have misspellings like that.

58:59 Play from here I got one too yesterday.

59:02 Play from here Back to the glue factory.

59:03 Play from here I almost missed it too, because I just want to believe that visually it looks so

59:08 Play from here impressive and compelling that I'm not going to sit there and read all of the copy that

59:12 Play from here it's thrown all over it, but it is still doing it, right?

59:18 Play from here What do we say?

59:18 Play from here It's the worst it's ever going to be.

59:19 Play from here Worst it's ever going to be.

59:20 Play from here Spelling errors might be gone tomorrow.

59:22 Play from here And the AI hedonistic treadmill is the other one.

59:25 Play from here Oh, we got to talk about that in the future.

59:27 Play from here Can't talk about that today.

59:28 Play from here The hedonistic treadmill?

59:29 Play from here Yeah.

59:30 Play from here For another time.

59:30 Play from here The AI hedonistic treadmill.

59:31 Play from here The AI hedonistic treadmill.

59:33 Play from here And you're going to have to start your little journaling experiment with, I think when we

59:39 Play from here talk about where do you start and you're talking to these different management teams and you're

59:43 Play from here talking to different leaders, being able to come with real world examples of I do this

59:49 Play from here and it achieves this and I save this amount of time.

59:52 Play from here That, I think, if anything, just gets people thinking, right?

59:56 Play from here Like they start connecting dots and, oh, I could do that.

59:58 Play from here Oh, and because of my business, I could also do this.

01:00:01 Play from here Wow, that's really good.

01:00:03 Play from here I think that'd be fun to trade notes on how we're using this stuff and share it with listeners.

01:00:09 Play from here All right.

01:00:09 Play from here Great.

01:00:10 Play from here All right.

01:00:11 Play from here Excited for more.

01:00:12 Play from here Till next time, bud.

01:00:13 Play from here Have a good weekend.

01:00:14 Play from here You too.

01:00:15 Play from here We'll see you.

01:00:15 Play from here Bye.

01:00:21 Play from here Thanks for listening.

01:00:22 Play from here If you'd like to send us a question to answer on the show, email me at Elijah at Spark6.com.

01:00:28 Play from here That is Spark6.com or Kevin at Kevin at Ascendlabs.ai.

01:00:35 Play from here To learn more about AI strategy and training for your business,

01:00:38 Play from here visit Kevin's company, Ascend AI Labs at Ascendlabs.ai.

01:00:43 Play from here And to learn more about design and development of custom AI solutions, visit my agency at Spark6.com.

01:00:51 Play from here You can also sign up for my free newsletter,

01:00:53 Play from here AI Framework Friday, by visiting Spark6.com forward slash newsletter.

How to cite this episode

Credit the named speaker and AI at Work by AscendAI. Link to the episode or the passage you quoted.

“Ep #8 AI Video Ads Are Beating Studio Shoots… and the Uncanny Valley Isn’t Even the Scary Part.” AI at Work by AscendAI, 2025-12-16. https://ascendlabs.ai/podcast/ai-video-ads-are-beating-studio-shoots-and-the-uncanny-valley-isnt-even-the-scary-part/

Recent episodes

View all episodes
×
AI Solutions Graphic
Free Resource
Get Started Now