What happens when everyone can create AI videos with just a text prompt?

We’re entering the “watch bait era” where AI-generated short clips may start dominating social media feeds, making it harder to distinguish between AI content, misinformation, and reality.

The technology is improving but still has real limitations. While AI avatars and lip-sync are getting better, we’re still dealing with 8-second clip restrictions, consistency issues (your character might lose their glasses between scenes), and the ongoing challenge of writing effective prompts.

Joining us is Myra Roldan, Founder and Chief AI Officer at UnDesto AI, who’s been working in the AI space for 12 years. She shares practical insights on what AI video can actually do today, why you need authorization before using these tools with company data, and why storyboarding remains essential.

Learning points from the episode include:

• 00:00 - 02:45 Introduction to Myra Roldan
• 02:45 - 03:41 12 years in AI: pushing limits and finding considerations
• 03:41 - 05:18 Current state: better avatars, but still limited
• 05:18 - 06:58 Understanding the "watch bait era"
• 06:58 - 08:52 Corporate AI use: authorization and strategy first
• 08:52 - 10:11 Cheap, fast or good, pick two
• 10:11 - 12:38 Consistency challenges with AI characters
• 12:38 - 16:00 Writing effective short prompts and accepting waste
• 16:00 - 17:05 Generation costs and platform caps
• 17:05 - 18:08 The "drunk intern" analogy
• 18:08 - 20:14 Moving from experimentation to everyday use
• 20:14 - 23:45 Why video editing skills still matter
• 23:45 - 29:12 Speed round questions
• 29:12 - 29:19 Myra's final take
• 29:19 - 31:04 Outro

Important links and mentions:

• Connect with Myra on LinkedIn: https://www.linkedin.com/in/myraroldan/
• Find out more about UnDesto AI: https://intelligence.undesto.ai/
• Watch Hey It’s Myra on YouTube: https://www.youtube.com/@learnwithmyra
• Camtasia: https://www.techsmith.com/camtasia/  

AI Video is here to stay. And I think

that we should really start to think

about how we can use it

to drive our business goals in a business setting,

in our personal goals. If, you know, if you're, you want to,

you know, get bigger on social media, how can you use AI to help

you boost your social media presence,

right? Or even in advertising your business, how can you use it to

like, you know, you no longer need a big studio to hire

talent. So how can you use AI to

boost your company's social presence also?

Good morning, good evening, good afternoon wherever you are and wherever you're watching from. My

name is Matt Pearce, host of Visual Lounge and today we're going to be talking

about the thing that is prevalent everywhere. Another day,

another AI, Right? AI is taking over, but it is having a

serious impact. It is having an impact on images and videos in particular, which

at the Visual Lounge we are sincerely interested in.

It is changing the way we start, are starting to see image creation,

video creation and even the outputs from like tools that

what can they do to help you get faster, get smarter? And now of course

we've talked about our human framework and how we want to keep humans involved

in that and have that kind of oversight. But we've got an

expert today who is going to help us understand how

this world is changing and maybe some of the things that we can do to

maybe live better in it. So, so let's introduce our guest today,

Myra Roldan. Myra is a digital transformation strategist

with 20 plus years in technology and 12 years specializing in

artificial intelligence. She has trained over 2,500

professionals across AI, cloud and quantum technologies and

led global initiatives at Amazon that reach more than 1,000

employees. Her work spans Fortune 100 and 500 companies

as well as US government entities including the army, Navy and

state leadership. Myra's cross sector impact continues to shape

how organizations build AI ready talent and

drive enterprise wide change. And with that said, please help me welcome Myra

to the Visual Lounge. Hey Mayra. Hey. Good day,

Matt. Good day. It's good, good to have you here.

I, I'm such a fan of, of you and your work and it's about

time that we had you on the show to talk about this really pressing topic.

AI is kind of everywhere, isn't it? Yeah,

it's everything and anything people can talk about these days.

So. Okay, so I set us up here a little bit at the beginning that

obviously AI is changing a lot of things across a lot of

industries, a lot of kind of workflows,

particularly I'm curious, you know, you've been doing AI for a long time, which is.

I was going to ask you about 12 years. That's a long time for AI,

considering I think most of us have been doing it for like a year or

two. Yep. Yeah. Actually,

I've been in the AI space for 12 and so I've been

really playing around with video and just media in general

putting out, you know, trying to push the limits to see, you know,

what we can create and what are the considerations that we need to have

in place right now and the things that we should be watching

for. So. Yeah, so,

so foundationally then, you know, we've had

like someone come on and they talked a little bit about their experience with VO3,

you know, and obviously it creates, it takes a lot of churn

to create something that works well. And you can edit, but you can edit it

together. So what's the current kind of current landscape then of

AI video that we should be aware of? What's, what's maybe

front of mind, what's kind of now old passe stuff.

Yeah. So I would say that right now it's getting better. Right? Like

slowly but surely you're starting to get

at least AR AI

avatars or people in, in the videos that look like real

people. Right. Where the lip sync motion is getting better.

But there's still limitations, right, because you're still limited to 8 minutes

of video at the most with. With

Gemini. Right. If you create it using

their tool. And then there's Sora. Right. Which

also has the ability to create these

videos. However, I think that there's still a

component in here where we still need to get the prompts right,

to be able to get like the output that we're looking for.

And the videos aren't going to always be consistent when you have

to like piece together these little, you know, eight

seconds. It's not eight minutes. I said eight minutes. That was wrong. It's eight

seconds, eight second videos and you have a. Your character

changes. You know, either they, they lost their glasses or.

And so you have to really still have a strategy. You have to

have like, what's your design

element going to be like, how are you storyboarding this? Right. Storyboarding is

still a step you should not skip.

I, I truly believe in that. And I think that

until these tools get better, we really need to be

careful because we're entering what's called the Watch

Bait era. So have you heard of that? The

Watch Bait? No. In fact, I was just gonna ask you what is Watch Bait

era. Yeah. So Watch Bait is

where AI created short clips are going to start

dominating social media feeds. And so we're going to start

seeing, and you might already see it OpenAI sora to being

integrated into like social media platforms and into platforms like

Synthesia and other

video generation platforms. And it's really going to

signal a shift towards rapid automation of content.

So people are going to start producing this AI content at scale

and, and sharing it on, on social media. And so

now us as viewers, we as viewers have to like be able

to disseminate. Is it AI or is it not AI? Right. Is it

misinformation? Is it not misinform information?

What's, what's reality? Right. I think that's,

that's what we're entering. So clickbait. We

all know about clickbait. Now we have to enter Watch Bait.

I. Now that's the connection right there. Yeah.

Okay. So obviously it sounds like, it feels like. And I,

and I'm, I'm out there trying to generate stuff and do stuff and play with

it as well, but it feels like we're, we're, we're not there yet. We're

on the cusp of. There are good things. I think there are, as you mentioned,

I think avatars have come a long way even in the last year. I mean,

voice generation, which obviously is part of the video, has come.

Listen to like 11 labs. Holy cow. Amazing. Kind of just quality

and sounding. So from a practical

standpoint, then if, if I'm, you know, in a department

L&D or something else at, at an organization and I'm

hearing, hey, you should be using AI, right.

What's the reality here? Like what I can really get done

is it, are we functionally at a point where we can do stuff or is

it really, gosh, we, if we could wait a little bit longer, we might be.

Feel a little bit better about it. Yeah. So I would say first

of all, you know, if you're, if you're an L and D and a

corporation in a corporate setting or in a business setting and you're being told

to use AI for like this video generation, you need to make

sure that whatever tools you're using are like

authorized to be used within your organization.

Because I have seen some incidences where people are

using tools that are not authorized and they're putting

company data into these video platforms and there's

no delete button. Once you put that information in there, it's like there.

Right. And so I think that the first

thing would Be like, what's the goal? Like, why are we using

AI? Why do we want to use AI? Is it a. Is it a

budget thing? Is it a speed thing? Do they think

it's going to be better quality. Quality than a person?

Like, why are we using it? Right. And then if you do have an approved

platform, what's your strategy for using it? Right.

And how are you going to track the ROI on it

to be able to say, oh, this was better and faster,

or maybe it wasn't better, but it was

faster. You know, it's that. It's that cheap, fast,

good paradigm, you know, where you can only have it if it's.

If it's cheap, it won't be fast, right? Yep. And

it won't be like, you can't have cheap and fast and have it be good.

If it's cheap and good, it won't be fast kind of deal.

So you need to be able to weigh that out. And then there's the ethical

part of it too. Right. Like with these

platforms, like I said, there's no delete button, so are

you infringing on someone's likeness?

Mm. Right.

And you just have to watch out for that. Yeah. So it does sound

like. And, you know, actually I just went through our AI. We had an AI

training that we had to go through, and I thought about you the whole, the

whole time, because we've had. You and I, I know, have had conversations about these

things. Right. About that. We. We have to have those things in place before

we're doing it. So let's say, let's say we're in our organization, we

do have those in place. We've got our safeguards. We're not, we know not to

put our personal identifying information of our customers or anything else in.

Into the platform. Um.

It. It is kind of that practical thing. I, I love your. The. The triangle's

perfect for this because I do think I can get fast, but

it's maybe not good. And I can get. Or I can get good,

but I've spent a long time on it, but it wasn't cheap to maybe pay

for the models or whatever. Um,

so is there, are there other things like, like, are there practical things we

could be doing that allows us to get better outcome? It's not. I know you're

playing with a lot of different tools and are there things that you're seeing that,

like, have changed? Whether it's that maybe thinking about prompts or is

it. Is there structurally things? Are, are you feeding it models of,

like, I Say models like imagery, which again, ethically,

you gotta get on that, make sure it's not some model that didn't agree to

it. Yeah, yeah. So you know, when it comes to like,

image creation. So like, I've authored some

books and I've created images

using AI, and one of the things that it

struggles with is this idea of consistency, right? Like image

consistency, whether it's video or images.

And so I've had some luck with

ChatGPT on image consistency, where I

have fed it and you created an image and then fed it

that same image instead. Just put this image in different poses, right.

So that I can do with using ChatGPT. I

always use illustrations. I never use like real people.

And then for video, that's been a, a little

bit more of a struggle, you know, because

I, I've put. Been putting out this series of Abuela and

tech and AI where she explains different

AI concepts in like eight seconds, which is a feat. Right?

That's awesome. And. But I haven't been able to get a consistent

character. She's close, right? With every video,

but she's not consistent. And so I would say you need to be

okay with that. So you need to, when you're developing like your content

strategy, your video strategy, your image strategy, you need to put

a threshold for like, how much of a difference is okay,

most people don't notice. They'll. They'll let it fly, right?

But also like being careful about using

these avatars to represent like an entire community that you're

presenting to. Right? That's, that can be

tricky. And avoiding, like

trying to avoid stereotypes, trying to avoid,

you know, misrepresentation of groups, we run into some, like,

tricky areas where we have to be careful. And so I would

say when you are

planning to use these AI platforms for either image or video

creation, make sure that one, your

prompts are strong. So that's going to be the first thing

you want your prompts to have

a good foundation of what you

are trying to do.

You know, when you're using these tools, your prompts,

you might notice that if you try to put in a

prompt that's too long, you're not going to get the best

output because it's not going to know how

to process your prompt

correctly and you might end up with garbage.

And if you use Tool, for example, and just give you like Google Flow,

where you can then go in and create videos, right,

Using the Google tool,

what you'll find is that it will

only take a certain amount of characters

for you to be able to effectively generate a video.

And so like, if you want it to have a script, that

script has to be super short. And if you, you know, you

can and you should define your character

down to like the T, like if you were to describe a person

as far as look, what they're wearing, what their hair looks like, what their eyes

look like, what are they doing, what setting are they in. Right. That's the

most important part. Important part. But that script has to be like

super tight. And so that's a consideration that you have to have

when you're designing. Because eight seconds is

not a long time. No, it is not a long time.

And it's. It's interesting just because, you know, I'm. I'm

hearing what you're saying. Obviously there's a lot of con consideration to put into this.

Right. Because it's gotta be thoughtful about. Not only was what that

person looks like, which it is. It is a challenge to describe a person. I

find. Right. To get the kind of details because you inadvertently could say,

describe it one way and you end up with something that you didn't want. And.

But that also goes to like, if you're trying to represent someone in a particular

community, are you giving that a proper kind of

description and not making sure it's stereotypical and all those things.

So I love the considerations though, because I do think this is a

challenge. We're in this phase where I hope it all gets easier,

but it's not right now. Right. Like it is. I'm

finding that AI is a good partner for a lot of

things, particularly text. I find it's a good thought

partner. Ask. Help me think about questions. Maybe I'm not

wouldn't have thought about from a media creation. There are things

I know feel I feel pretty consistent about again, voices, avatars, pretty

consistent. I can get some pretty good images depending on

what kind of information I provide. But there's

a lot of stuff I'm also like, I'm throwing out a lot. I feel like

I'm doing stuff throwing out. It's a lot of still iterative testing

versus like I get one prompt and I. I've

nailed it. Yeah. And I think that's something that

people have to be okay with right now, is that you're going to generate a

lot of things that you won't be able to use. Right. And it

does. There's a cost to generating video and

images right. When you use any of these platforms.

So one of my favorite platforms to generate images is Ideogram

and So with ideogram, though, you get capped. Even

when you pay, you know, there's a cap. So, like, if you don't get it

right, like the first few times, it'll say, like, you're

done. Right. Come back.

And so that can be frustrating. Right. Because it's, it's.

We have to hone in our ability to take what's in our head because we

always can picture, like, this is the specific thing I want.

Right. This is what I want. But then we go to write a prompt on

it and it's garbage. Yep. Right. Or

it gets misinterpreted by the AI. Yeah, I've

heard a couple times the, the analogy one is

more funny. Mike Parkinson, I don't know if you know Mike, he's. He calls it

the drunk intern. But I've heard the intern analogy. Right. Like I, and

I work with actual human interns and, and they're brilliant and they

have their own skills, but it is often without. They are without

context of a career at an organization

that has. Of someone who's been there a long time. And so I often find

I have to be much more descriptive to an intern because they

just, you know, they might not know what outlook is. Right. And, and that's not.

Because they're not capable of knowing what outlook is. They just haven't experienced

outlook. And so I do, I do appreciate that, that sentiment. Right.

Like, you gotta. Yeah. Be willing to express more.

Yeah. And it's just, it's hard. Right? Yeah. Without

expressing too much is what I wanted to say. Like, because too much and then

it gets confused because you've added too many details and now it's like I'm overwhelmed,

just like an intern. I've said too much. Yeah. Yeah. Like

you gave me way too much information. Now I don't know what to do with

it. Yes. So,

Maya, as you are often on like, the cutting edge,

you are always looking at, I know from knowing you

for a while now that you're always looking ahead at kind of what's next.

And I'm guessing that gives you some pretty good vision to

look ahead here. Short term, where do you think

we're going in terms of AI images and

videos in the next. Let's say. I know it's hard to predict. I'm not asking

to necessarily put a flag in the sand, but like six

months, a year, what do we think will be true? That's not true today.

I think that we're going to move from like this

experimentation phase and to everyday use.

Right. So more People are going to start to use it for different purposes. We'll

see an increase of use for social media. I don't know if we're going to

see an increased use of AI in corporate just

yet because I think we're still feeling it out, right. So corporate

usually lags behind a little bit,

but we will start to see

consumer apps start to make its way

into the business setting. And

I think what we need to

really watch out for is whether it's in the corporate setting

or not, is the, we're going to see an increase in synthetic media

that's amplifying misinformation, wrong

information, or emphasizing, you know, the, the need for stronger

providence provenance and like digital liter

literacy standards across the board.

Because anyone can create a video, an AI video right now.

Right. And you have,

I think it's having guardrails around how

we create and what those tools look

like and how they evolve into like the tools that we use every day. So

I think we'll see some video generation and copilot we might see

some video generation integrated into

other platforms that we use in business, you know, and we may

start seeing the ability. Through.

Middlemen like a Zapier or through

APIs application interfaces

to these, like commercial consumer products to

generate video. And you know, there's still going to

be a need to be able to edit those videos

to make them make sense. And so I think we need to get good at

our video editing skills so we can start

cutting out like the bad stuff. So instead of throwing the clips away,

right. Keeping what we can use and figuring out like, how do we splice

this together so it doesn't, so it doesn't look bad.

Yeah, well, if you need a tool, I've got one.

Oh, I, I now I use it every day. Well, thank you.

Well, I, you know, I think I, I, I see what you're saying too, right.

That social media, obviously there's lots of, of things that are

happening and it's happening fast and it usually is a precursor to coming

into the workplace. Right. We saw this with,

gosh, I'm gonna date myself. But social media, Twitter became like yammer, right?

Twitter was happening and people mic y and then Microsoft

bought yammer and brought it in. Now I think yammer might be dead at this

point, but it's a thing. And so I

definitely see that. You know, I have a

prediction that, and this is what I hope to

be true. Whether it is true or not, I don't know because I'm not very

good at predicting Things. But the ability to have more and

more control in what the AI generates, I think is going to be really important

to, to people. And so, you know, giving,

giving, maybe more finite control over, you know, backgrounds,

characters, like letting it layer things rather than just kind of whole. Right

now, it's just a wholesale. I prompt it. If I'm good at describing things,

I might get what I want. If not, you know, it puts

in a different scene or whatever. I think that's going to be

tricky because we're working with generative AI, and the whole point of

generative AI is that it generates new content based

on historical data. So it's making new content every

single time. And so it's not persistent, meaning it

doesn't. I think what we need is more persistence and

character consistency. So, like, if I create a character, I want to. I want it

to be persistent. I want it to. To look and feel the same, no matter.

No matter where I put it. I would love to have the ability to

have character consistence where the background is

transparent and I can dump that character anywhere. Right. That would be.

That'd be a beauty. Right? To be able to do that.

Yes. Yeah. And I mean, I think the reality

is images have gotten not perfect, but it's better.

And so I imagine video is. It's just. It seems video is lagging

about a year to year and a half, kind of where images were.

I mean, we're getting to the point where our images now no

longer have six fingers or 20 fingers. Right. So, like,

that's a good thing.

Yep. But there's. It still generates some weird images occasionally, depending on what you prompt.

It's. Yeah, yeah, yeah. But it's

getting better, right? And it's going to get better over time. Yeah.

It's the worst it will ever be. So. Well, Meyer, this has been. I love

the take on it. I love the thoughts about ethics. I love the thinking about,

like, you know, how to. How we can apply video. Video today, obviously, it's not

a. It's not a one and done. There's work that needs to be done. And

that's a good news for us as humans who still want to do jobs and

still want to work. Right. There's lots of room for us to be involved. But

I do want to move us along here into our speed round, which are quick,

fast questions with quick, hopefully quick, fast answers

determined by the role of an eye. So here we go.

Okay, so here is our dice tower. We have a die to tell us which

questions to ask. Myra. Here we Go. First roll is number

12. So question number 12. Oh, what's your

one? Go to tool that helps you get your job done on a

regular basis. And when I say tool, it could be software, physical

gear, whatever, something that you use often to

help you get your job done. Yeah, I would say my phone.

I love it. Yeah. My smartphone, because,

you know, sometimes I'm not near my laptop and I need to pull information

up or I need to get to my files, and my smartphone is connected

to, like, all my external drives, so I can just pull a file up and

send it. Like, I can. I can run my business

off of my smartphone.

Yes, well, and we know how people feel when they get detached from their

smartphone, and, yeah, you know, it feels a little bit. Whoa, what's going

on? So. Love it. Well, are you ready for your next question?

Yeah, let's do it. Here we go. Question number two. Oh, we can't

have two 12s. I'm rolling really consistently. All right, the dot

is at the bottom. So that is a. That is a. Actually

going to be a nine. So question number nine here. What's the

one thing that you're most proud of in your career?

I would say the thing that I'm most proud of has to be

the people that I have been able to interact with and who I've had

an impact on and who come back and share

their, like, the things they've done with, you

know, either the con. The consultation or mentorship that I gave them.

I have so many people that I engage with and meet

on a regular basis, and that it just. It

makes me feel good that I'm able to, you know, share the knowledge,

help others grow. And I. And I love cheering

on for everyone. Like, you know, if you're going after a goal and I.

And I've helped you in some way, or I gave you an introduction or I

gave you some information or you went through training or whatever it is, right. You

come back and you tell me, like, hey, like, this went great or this. You

know, I love that. I also love when they come back to me and they're

struggling. Right. And they're like, I still need help getting there. So I'm

always. I'm happy that I can help.

Well, I know when I want someone in my corner, you're one of the people

I want there. Cheering for sure. Thanks, Matt. You

bet. Okay. Cheering you on for a long time. I know, and I appreciate it.

So we got one more question here. Here we go. Okay,

question number 10. Ooh. Okay, this one's

interesting. If you had to shift careers, they said, no more

of what you're doing today. No more, no more AI business. No more of what

you're doing. And you could do anything that you wanted,

you know, out of the world. AI, what would you do? What was the thing

that you would move to? I would be a,

a travel advisor. Like, I

love traveling, so I would want to like, like a travel

influencer almost, but not an influencer, but like helping people, like figure out their

travel. I want to go to the hotels and see which hotels are good.

I want to go to different countries and, and then be able to buy. Create

like travel experiences for people that I think that would be

fun. I get to travel and I get to help people like, you know,

have a good trip. Yeah. Well, I'll have to ask you next time. Hey,

Myra, what should I do on my next. Actually, you know what, you can help

me out when we're at Dev Learn. I'll be Myra, what should I go do?

DevLearn. That's what I'll do. That's all I have time for. It's the conference.

Yep. I'll tell you. You should go do your session. That's right.

That's right. Do go do your session. I love it. Well, Myra,

I always enjoy talking to you. It's always so much fun and always so

insightful. I feel like I just, I learned so much from you. If

people want to connect with you, they want to learn more from you. Where should

they go? Where should they turn? Yeah, so

connect with me on LinkedIn. I'm under Myra Worldan. I'm pretty easy to

find. If you look for Matt, you can find me also under

Matt's profile, my website. So I

ondesto AI, but we have an intelligence suite which I

think is a better. Is more useful than my website.

So it's Intelligence ondesto AI. We have resources

in our intelligence suite to help, whether it's

regular users or businesses, figure out how to get started

with AI and integrating it. Because personal and business use

of AI are two different things. And so,

yes, that's where the two places I would say intelligence suite

and LinkedIn. Feel free to connect with

me on LinkedIn. I love connecting with people. I love. I like. I look at

my feed every morning to see what everyone's up to. So I

like to doom scroll on LinkedIn and see what everyone's doing before

I start doom scrolling on TikTok,

which. You can also find you over there, I'm sure as well. Yeah, you can

Find me on Talk under Learn with Myra. Yeah, and I'm on

Instagram under Learn with Myra also, and I post,

you know, educational content. That's.

Hey, good, good stuff. So, well, Myra, as we wrap up the show,

we always like to ask our guest, what is your final

take? Yeah, my final take is

AI Video is here to stay. And I

think that we should really start to think

about how we can use it

to drive our business goals in a business setting,

in our personal goals. If, you know, if you're. You want to,

you know, get bigger on social media, how can you use AI to help

you boost your social media presence?

Right. Or even in advertising your business, how can you use it to,

like, you know, you no longer need a big studio to hire

talent, so how can you use AI to

boost your company's social presence also?

All right, well, Myra, thank you so much for joining me here in the Visual

Lounge. Yeah, thanks for the invite, Matt. This was

fun. You bet. So, all right, everybody, real quick as we wrap

things up. Uh, look, AI, we know it's here. We keep talking about it. We're

going to keep talking about it because it is so prevalent in the things that

we're doing with images and video. If you're looking for some AI, you want to

dive in a little bit, you want to try something, I go recommend, you know,

lots of tools out there. Go give Camtasia AI a try.

It's. It's free, it's a labs, so it's beta, but it's got

some really cool stuff. You upload some images and watch it, create a video from

images. And the cool thing is you can take it into Camtasia and you can

manipulate and change and move things around. So give it a chance. Go

try it out. We got some other great things that we've been working on that

we announced at our October event. If you didn't get to see that, but with

that said, you know, there's so much going on, so much change. You got to

make sure that you're staying up on top of these things and all the things

that will help you feel like, able to do the job that you need to

do. So I encourage you, take a little time and to

spend and level up for yourself. We'll see you next time, everybody.