Welcome back to Data Driven. This is episode 400, and

no matter what Andy's weather station tells you, it's always sunny in

farmville. Today we're talking AI vibe,

coding, building real systems with small teams, and how the last

few years have completely changed how we ship software and podcasts.

Let's do this with some different music this time.

Hello, and welcome back to Data Driven, the podcast podcast where we explore the emerging

field of data, AI, and, of course, data engineering.

With me today on a special most auspicious day. I

think I use that word auspicious, right? I'm not sure. Is my

favorite data engineer in the world. How's it going, Eddie?

It's going well, Frank. How are you doing? I'm doing

fantastic. We're recording this on January

7, 2026. Can you

believe it? My goodness. 26. This

is the future, man. I'm waiting for my flying car, but

until then, any day. Any day now.

There actually is something at CES that they were showing was, like a personal, like,

drone that you can, like, stand in and, like, fly around.

No way. Way. I saw it and I was like. But now we're at the

point. Now I'm like, could that have been an AI video?

Never really can't tell anymore. Yeah, that's true.

They are getting better, aren't they? They really are.

You've seen some of the experiments I've been doing on Frank's world

on YouTube. Oh, yeah. See, like, you know,

a lot of that. You know, I'm not going to tell you which ones, but

you. You know, some of them are AI, so it's. It's.

It's pretty amazing. Plus, I can generate

videos longer than 8 seconds on my Spark.

Nice. My DGX Spark. So. Which I don't think we've

talked much about on the show is kind of like the. This 2025 was the

year for me of the home lab. Right. Where. Yeah.

And you, too, as well. Right? Like, it was an interesting. It was an interesting

year. So I, in October, convinced my wife

to let me get the DGX Spark. And if you're

watching this on video, there's the box right there. I

leave the box out because I know more than anything, the DTX

Spark is as much of a status symbol as

anything else. And

seeing you cough is making me cough. I know I choked you up there.

I know. Scrambling for the mute. I think the mouse was over

on the third, like, three screens over. That's funny.

Trying to get to the mute button. So,

yeah, so I got DGX Spark. And it

basically, it's. You know, Nvidia sells it as a personal

AI supercomputer. It basically has

shared. A shared memory model, so it has 128 gigs of

RAM, which means that, you know, I have about.

Depending on how much the system is using, when you're using it, I have about

120 gigs of video RAM,

which, if you were to replicate that with a traditional kind of desktop

PC, would be very pricey. I mean, you'd have to get

three or four of the 5070s, like, working in unison to get that much

VRAM. Right. So that is. And it also

has a Grace Blackwell chip in it, which is one of their, you know, one

of their more professional ones. I know we're recording this while CES

is going on, so there's. They've announced the new Vera Rubin stuff and

all that. So there's definitely a lot more hardware innovation

happening that I can't keep

up with it, man. Like, it's just, you know, you can either pick

the hardware side or the software side and, you know,

but. And then also this year, I picked up

a. I think I did talk about this on the show where I picked up

the an i9 with a 4070 in it.

That's a gaming PC, but it's a mini PC, so it's, you know,

probably about the size of a cable. Cable box. Nice.

Which. This will make listeners. I know it made me feel old, but

I. I told my. My teenager, he's like, you know,

what's it look like? And I'm like, well, it's about the size of a cable

box. You know what he said? What's a

cable box? So I'm not

surprised. Yeah. So

that. That is the world we live in today, I suppose.

Frank, you mentioned we were the OGs. We really. We really are.

So you were just on another podcast. So tell them. Tell. Tell our good

listeners why this one's special. Although I think Bailey may have

ruined it by now. Well, probably. But that's okay.

Bailey's doing her job. And. And thank you, by the way,

for helping Bailey do her job. You are the man, the

creator of Bailey. This is our

400th episode of the Data

Driven Podcast. Can you believe that? 400?

No, I can't, man. It's. It's wild.

And it was about, what, nine years ago, we started,

you know, figuring out the logistics of the recording and things like that

and working through it, and it. It's been.

It's been an interesting ride. That's for sure. Definitely.

It definitely has. And yeah, show has evolved.

You and I have grown and we've,

we're 10 years older and. Yep. You know, and

hopefully 10 years wiser. Hopefully, hopefully

our families have grown. Yep, your family

has grown. I. And, and you've added to your

family in the last 10 years. I, I have not.

But I did reach a milestone last August.

My old, my youngest son turned,

turned 18. And I realized

in, in thinking about that, I realized that

for 43 consecutive years

I was the father of a minor, at least one

minor child. Now in order to, to kind of

span a gap there in the middle,

I have to count time that Stephen was in the

womb, but I knew he was there. When my younger daughter

from my first marriage turned 18,

we, we knew, we knew Christy was pregnant with Stephen. But

43 consecutive years, Frank, that's a.

Solid lifetime that is of being a dad.

So it's different. It's, you know, it's, I don't like

regret or, you know, have regrets or anything about. Well, I do

have regrets, obviously, but I don't. It's not like a big loss.

It's like I changed, I shifted gears. I'm now in this

new phase of it. But, you know, it's not just me shifting

gears, Frank. You've done a lot of gear shifting. The market has done

a lot of. You mentioned keeping up. And I'll say this and then I'll shut

up. The way that I've found to keep up with both,

at least the kind of hitting the high spots, both software technology

and even physics when it comes to quantum, is

when GROK enabled the ability to create

something called Grok tasks. And I want to say it was in

2025, early in 2025, I set one up

for technology news,

really. And I haven't really played with Grok that much.

So at 5:40am every every

weekday, actually it's every day at 5:40

an announcement shows up. That's when I scheduled it. It pops

up and, you know, I get a little thing on my phone and I'll hit

it. Or if I happen to be on Grock, and I am, sometimes

I'll see the task gets populated there. Just click on it and

it gives me six or eight, sometimes 12

paragraphs on how things are going. I don't

know if it's limited to Super Grock

or not. I got in on Super Grok when it was

basically an add on. If you went from

paying x 8 bucks a month

to 16 bucks a month, you could, you Got

a subscription to Super Crock. So I think now if you buy it

out of the box, it's like 30 bucks a month, but I'm still

paying 16 and I get to use it. It is 30 bucks a month. Yeah.

Yeah. Apparently I'm on the free plan. I didn't know that.

I thought I was, I thought I was paying for this, but

now I don't feel so bad about not using it as much.

Will they let you do a task on a free plan? I don't

know what's available. It says. It does. It says the task.

Yeah. I have to take a look at that because that looks interesting.

So to find it in the ui,

but interesting. I've got the client on my phone. That's one of the few.

I've got one of the few AI clients I've got on the phone. I know

you use some too. Yeah, I have. So one of the things I did was

I don't know if this deal is still available. If you sign up for

perplexity using PayPal, you get a year of Perplexity for free.

Oh, neat. Yeah, I don't know if that's still the deal, but that's. I

have Perplexity, so I propel. I have Perplexity,

I have Gemini And I have ChatGPT and,

and the free version of Grok. I also have

Claude because I love Claude code and I thank

you for turning me on to that because it is, it has been

transformative. Right. Like so, you know, you did mention I have a

three year old and I have a teenager and I have a tween,

so I don't get a lot of focus time.

So I

have to be very judicious with what I choose to focus on.

And the big

advancement is cloud code. Right.

Because there's a lot of things I want to code up. I just don't have

time for. Right. I don't have time to do that. So this one project I'm

working on, I'll kind of give a preview of it. It's called podz

and Podsy is Ponzi, is, is. Is meant

to solve a problem that we're having with the creation of the podcast.

Because we have this show and we have the Impact Quantum show

which is doing gangbusters. If you're not already subscribed,

check it out. Impact quantum.com. but you know, we

won, it's won awards being the top quantum, you know, computing

podcast by Feed Spot. It's, it's, it's on a

really solid growth trajectory. Because again, we really,

we've been podcasting now for nine years, right?

And that makes us kind of OGs in this game, right? This

is, you know, absolutely. And you know, I was talking to somebody and I was

like, you know, like, I'm good at this and I don't say that at arrogance

because I've done it about four or five hundred times, right?

You know, if you something four or five hundred times, you're not, you don't get

improved at it. You know, you really have to be trying not to improve.

So, you know, and you know, we have enormous, we

have fantastic numbers on this show. But like, you know, if you look at the

growth trajectory, I mean, the potential is enormous. But one

of the big challenges is, is that how do you track

all of the assets that are related to a particular episode?

And that's the problem that PODSI is meant to, to address, right? So

every podcast is going to have a transcript, an audio

file or video file, maybe both. And it's also going to have

a, you know, a thumbnail image, right? So every show

has at least four assets, right?

How do you deal with that? Right? Because then at this point we're looking at,

you know, 1600 assets for all our shows, right? That's

not a trivial data problem to master or

to manage. So that's what podzi is meant to do. So

the reason why I say this the way, the reason why I think this is

an interesting use case for Claude Code is because

on December 5, I had a car accident and, you know, we're all okay.

Car was totaled. The airbag did, though, leave

a Honda shaped imprint in my hand

because it was a Honda and the logo went smack into my hand.

But aside from the hand injury, you know, we're all okay. And,

but that night I just was like, I had

this idea because I was getting frustrated because if I want to add a third

show and that's, you know, we've talked about adding a third show, maybe, you

know, more shows. Beyond that, the logistics of managing

all of that, you know, content becomes

an issue, right? And that's, you

know, that's what POSI was and how that ties is the accident is

that I, I was feeling frustrated with the lack of progress of

Ponzi. I, I, I had thought I had this grand vision that we don't in

a week, right. Even with,

you know, vibe coding it or whatever,

you know, when it got to like Christmas, I was like, all right, what the

heck has taken so long here? Right? And yep,

I basically looked at it,

and it turns out that by that point I'd written about 30,000 lines

of code, or it had written 30,000 lines of code.

Right. And it does about 90% of what I wanted to

do. Now think about that.

Two of my boys have birthdays in December. I had a car accident.

I had to close out a lot of things for my day job. And you

know, you have the general, you know, run up to Christmas. There's

no way in hell I would have been able to do that

normally. 30,000 lines code, right? Yeah, yeah. Even if I

was doing it full time, no job, no children, no responsibilities. That's

a stretch, you know, in 20ish days to hit

30,000 lines of code. Absolutely. That is,

I'd have to, you know, I mean, that's a full time, maybe two full time

coders. Right. Again, I'm not going to

shell out the money to do that. Right. So in a sense,

I'm not so much taking a job away.

I'm adding to my productivity as an

individual. Does that make sense? Yep. And

that's exactly, you know, that's all of it. And I mentioned

to you that I was listening, I am listening still to an audio book called

Vibe Coding. And it occurred to me, and I, I think I posted this

on social media yesterday or the day before. It occurs to me

in this that there's more than one way

to change the ratio for return on

investment. And the role AI plays in that

is exactly what you just described. And what it

does is it takes one person, you or me,

and it makes us ten times more productive.

Right. And it turns out that, you know, we don't need to jack up

the price, which would be one way to do it. Kind of a, you know,

maybe not in a very effective way to do it. Increasing the value

would certainly support jacking up the price. But to do,

to increase value, you need to add features or functionality

or some, some kind of way improve the code. Well, because,

you know, ostensibly using an AI like Claude code

to do that, you, you nailed it. You don't have the

inclination to spend the money to hire a couple of Vibe coders, or, sorry, a

couple of coders to, you know, knock out 30,000 lines of

code. And it costs you the cost of a

cloud subscription to do that in the

middle of all the rest of your life. And what's happening is

all across the board, all of these projects that have been on our

list of when I get some time, I'll do that. Right. Or

it's, you know, it's I joke about if it's not in the top

three on my to do list, it doesn't get done. That's

changed. It's now probably the top 20. Because

a couple of things that were hanging out there that A couple of weeks ago

I spent some. I spent a day and a half revamping

all of my Andy weather stuff surfacing. The data that's

collected from my weather station sitting out here on the deck and it took a

day and a half. I had two applications that I rewrote.

One of them was the one that posts on X. That's running again.

It had run since 2018,

rewriting the one that transfers the CSV blob

from the old Emachine sitting over here up to

Azure Blob storage container.

Finally, the website, you know, it looked like an engineer built it because an

engineer built it. And that's not a compliment

and I'm the engineer that built it. But I had, I had

Claude rewrite it. Now it's got a little bit of savvy to it and in

addition to that, Frank, it's got two APIs running behind it,

so. Interesting. Why do you need a weather station? Isn't it always sunny in

Farmville? It is always sunny in Farm,

except when it's not. You know, I. I noticed

you're wearing a jacket, but it's January and I know you're in a basement and

it's probably chilly down there. Yeah. It is actually

74 degrees as we record

on January 7th. It is 3:46pm

in Farmville, Virginia. 74 degrees is what my instruments are showing

me right now. Yeah, that's crazy hot.

It went from being like really cold. I mean, I am in the

basement and you can always tell if I'm running an AI workload or not based

on, you know, and I'm not running a workload

currently, but when I have the machines going, it does get toasty in here.

So. Nice. We're in an

abnormal heat wave for January in the mid Atlantic. And I know you

guys are warm up there too. We're like at 50 today.

Wow. And like a few days ago it was, you know,

1720 degrees Fahrenheit. Right. So it's

not. Yeah, it's kind of unusual, which is going to have. Yeah,

amazing impact on my sinuses. So

that'll be fun. That'll be fun to experience. So it is always funny in

Farmville. You're right. I got these glasses. I love those

glasses. These, these are blue blockers. And so if you're not

watching. I kind of look. I don't know what I look like. He says I

look cool. I look like a cool. So.

No, like, one of the things that some people think could be

triggering the migraines is. And headaches is kind of like too much blue

light. So I figured, I'll give this a shot. I'll give it a shot. Plus,

I. I feel like I look like AJ from the Y Files because. Like, you'll

see, like, he wears glasses like that. He wears glass like that. And then

I saw that and I. I looked up and I was like, what are those?

And I find out. And I think the exact pair he wears

is like $200. But we don't sell

merch, like hacklefish and things like that, so.

And if you don't know what I'm talking about. Well, the Y Files is an

interesting YouTube channel. It's also a podcast, but it's very, very well

done. I think when I first discovered it, I binge watched it. Then I shared

it to you, and I think you binge watched it. I did, too, yeah. Lot

of good stuff on that channel. Well, and he has a lot of

throwbacks to some other stuff. And for listeners that, that are

thinking to type the. And the letter Y and files. It's a

W, H, Y. The Y Files. Yeah.

But the throwbacks go back for me to

start. So I used to work third shift as a plant

electrician back in the old days and when the years began with A one.

And it wasn't uncommon for, you know, me being on

maintenance to have plenty of time to sit around after I'd done my,

you know, my scheduled maintenance stuff and fix whatever may have broken.

I was often idle. It was not always, but,

you know, probably greater. Only so much happened. Almost only so

Much happens at 3am. But I had the radio on and I would

listen to. To WRVA 11:40am in

Richmond, and the Art Bell show would play overnight

on that. And he's a. AJ Gentile, the

host of the show. Huge Art Bell fan.

Oh, yeah. If you don't know who we're talking about, it's this whole. So

welcome to conspiracy theory world here, right? Yeah. He was

Alex Jones. He was Alex Jones before there was Alex Jones.

Art. And he covered all of them. So Alex is, I think, a

little more political. Yeah. You know, cultural art was

more like sci fi, UFOs. Yeah. Really

X Files type stuff, right? Exactly. Mel's Hole.

And you'll see, you'll see stuff about Mel's

Hole, even on Y Files merch. They've got some much formula sold out

there so. But it's a fantastic podcast. Just

the art of it. And AJ has

is a career in. In entertainment and I didn't

realize this particularly definitely radio. But he was also one of the

producers or associate producers on Family Jewels. The

Gene Simmons. Oh, I didn't know that. I

didn't either. But there's somewhere I want to say in

social media and stuff. He posted some pictures a few years ago about

him and Gene, you know, collaborating on his head. Gene Simmons, if you don't know.

Not everybody knows. Bass player and leader of the band

Kiss, huge rock band in the

70s and 80s and their.

Their lead guitarist passed away a few weeks ago. Ace Fraley

passed away and they actually did something for him at the White House.

I believe there was a. I think so. Yeah. Some sort of ceremony there.

So yeah. Yep. We're telling on ourselves

Frank, but we're. Going, we're going on our off road thing which I think

that's. And 400 shows we

managed to. To miss this. 399 shows in a

row. And here we are going off on a tangent. First time ever.

I don't know about the first time. First time in the four hundreds. But

the. No, I mean it was. It's. It's a cool. It's a. It's an interesting

podcast and I think what, what I like about it is

he's advancing the art of podcasting. Right. With Heckle Fish and like as a

character and things like that.

A lot of innovation. You know. I think one of the

things that you know, also is, you know, we've had Bailey. Bailey is

no longer in the Quantum show because some of our list. She

doesn't really apply there. I think so. But we

know we're doing. We're all going to do fun stuff like that.

I'm toying with the idea of like a Professor Cubitt

kind of like a thing that could work. Wasn't really ready in

time but you know, just conceptually. But again like I

think people focus on the jobs that AI will take away.

But I also would, would, would pull from our own experience. When we hired the

voice actor to do the first, you know,

couple of seasons of, you know, the, the intro,

we eventually replaced her with, you know, AI Once

synthetic voices got good enough and somebody's like, well,

you know, you really. You took away a job from a

voice actor. I was like, well no, because getting a custom

recording for every show would have cost. Would have been cost prohibitive would have been

somewhere between 75 and $100. Right. Yeah. And

you know, that's not really feasible

and I didn't think it was, that. That was going away no matter what.

But I wasn't going to do that anyway. Right. Like, so I think also with

this, with this idea of, you know, most of us, you know,

every developer I know, every data engineer I know always has like a back,

like a side project in the back of their mind they want to build. But

life, responsibility. Now that excuse is

really a lot lower. Right. Like,

I mean, now the question I have,

now the stuff, now the question I have is, you know, how many machines can

I have running code at the same time? Right. Because, like, I have

all these other ideas. Like, you know, I have Dingo, which is, you know, right

now is just a command line tool. I want to convert that

to a web application. Sure. Like podzi.

And you know, I don't know, like, podzi is amazing.

Cool. Dingo is amazing. Thank you.

It's good work, Frank. Frank does all the work. I've said this before, but

I never say it enough. Frank does all the hard work here on the show.

And it's as our schedules have just gotten crazy,

you show up here less and less. Well, I show up

and not even, I don't even show up that, that often anymore. That's so

sad. But it's, you know, I, all I have is

excuses, but it just, you know, it's a good

thing that work is as busy as it is. Yeah, yeah.

You know, the consulting and I, I, I'm in a,

I've been in an interesting situation for about the past six

months. I don't know if it's going to last. We, but there's been enough work

out there that the teams have been doing for,

for me to not, I can, I can pay the bills and

I don't have to do consulting billing and

that's unusual. It's the first time ever in my life

that's been true. So I had time and money

and I invested that money pretty heavily

in. Oh, sorry. I invested the time pretty heavily

and some stuff that's coming out of data integration, lifecycle management

suite stuff. But the big boost was

exactly what you were describing earlier.

Applying Claude code specifically to these

coding tasks has made it possible for me to

finish up stuff that I've been working on for 10 years.

Yeah, I know. You were showing me some stuff that you were building and some

cool code names that you have as well as you have some acronyms.

I Like to think that I inspired you to come up with the cooler code

names, but we don't know. Absolutely. Well,

I think the coolest code names actually Claude code

suggested. Oh, really? It's been cool. Yeah.

The one I think the one you reacted to most, the

new big project I shared with you, that's the biggest thing I've ever even conceived.

It picked the name and the name of the roles and so it's

first, it's very applicable what it, what it shows the analogy holds

and, and then the name names of the roles

that are going to be parts, kind of like sub parts of that

project. Those are also very apropos.

Oh, very cool. A Ponzi I got name. I got. I asked

ChatGPT to come up with a list of interesting names describing what

it's going to do and Ponzi was one of them. Nice. So, yeah,

great minds, Frank. Great minds. There you go. There you go. I'm actually, I'm

actually talking about Podsy because I think I'm building it from the get go

with the idea that it could be a SaaS for other, it could be a

service for other podcasters. Absolutely. Yeah. We're having

this problem. We can't be the only ones. Right.

And if nobody buys it, at least I solve the problem for myself. Right.

You know, and for you too. Like, I mean, one of the, you know, one

of the things this does leans pretty heavily into the AI side of

how do you generate content from a podcast that you already made?

Right. How do you track, how do you track the content that you've, you've

created as a result of, you know, other tools like Opus

or. And if you've seen, if you've seen us on LinkedIn,

like a lot of those short clips are generated by Opus where it'll, it'll

show, you know, kind of us talking. I know you use it

and, you know, it does the captioning, does the

slick editing. You could have it at B roll. I mean, it's just amazing what's

possible, right? Like. Yep. And I

mean, people are amazed to find out how small of a team we are

between this podcast and the other podcast. Right. Like, yeah,

I mean, it's a lot of content if you think about it. If you're doing

content per headcount, we're up there. Oh, yeah. And I

say we. And it's mostly you. Again, I would say you're

doing all of the, you're doing the, you're doing the lion's share of it. You

know, north of 90%. Not so toot my own horn.

But yeah, I mean, you started this with, with an eye

towards automation and you kept looking for

shortcuts and. And shortcuts for the shortcuts, and that's what's

grown these tools. And I mean, it was. Frank, we had only

been doing this a couple of years when you started coming up with, you know,

scripts and stuff like that, that would. That would, you know,

grab a transcript or, you know, parse

the URLs and, you know, post things for

us. It wasn't long at all, and it's just. It was what was available

at the time. And I think you should do a demo of Potsy

just to show people. How cool it is, like right

now. Yeah, I'm gonna need a

second to. And you got to keep in mind that Frank's been.

You've only been working on us, what, a couple months? A month. Maybe

a month today. Yeah, yeah.

And you started, you know, you started with the idea.

Yep. And. And, well, I was actually. I was

thinking about this for a while, like, how to do it. Yeah.

The real challenge was the real

thing that broke the cat, the straw that broke the camel's back, so to speak,

was the Advent calendar that we did for Impact

Quantum. Okay. Which if you go. We'll make sure it's in the show

notes. Right. But if you go to impactquantum.com advent calendar,

filling in all of that content into one place.

Was. That was vive

coded, actually. Right. That whole HTML experience. But the issue

I had was the effort to collect all the

content. And the data that I had

was. I'm typing, I'm telling Claude to start the dev server.

I'm so lazy now. I don't have to type in.

But that's efficient. It's efficient operational

efficiency. It's not lazy. So

ultimately I realized, like, you know,

we have this, we have the transcripts, we have all this, but, like, collecting it

in the one place is just way harder than it needed to

be. And I kind of had this in

the back of my mind, like, what is this going to look like? And things

like that, and what am I going to do with it? And then really

kind of the accident, actually, like, I had a lot of free time to kind

of think, you know, later that day

and, you know, nothing will jar you out of

anything faster than like an accident,

you know. And so this, ladies and gents,

and AIs of varying levels of sentiency.

This is the current homepage for Pozzi. I actually do have A domain name that's

registered. Podzi Studio, I believe. Okay.

And you know there's going to be more marketing material

here. Right. But I'm going to sign in.

These are all test accounts, so you're not gonna.

I'll sign in. So right now we have two shows.

There's Data Driven and Impact Quantum. And so I can

see from my dashboard

what I. What I have. Right? This tells me I have

97 episodes in the system, which I think is about.

Right. Based on what the feed would be pulling because we don't share every

episode on one of the feeds because it would make the feed file super long.

Right. But anyway, so you'll see my

4 to 1 ratio is pretty spot on. Right. There's

401 assets here and 97

episodes, right. So it's about a 4 to 1 ratio. So let's

just go to. From Molecules to

Medicine. This was an episode

36 of

Impact Quantum. And you will see that

I have these tracked assets now where it

pulls. Most of these tracked assets are from.

Are from the RSS feed, right? So every RSS feed. Yeah. So,

like, and I think you and I were talking and was like, we use

Captivate, so there's other fields we could also capture too. But I want to start

with the vanilla. If you have a podcast, you have an RSS feed,

full stop. Yeah, right. Good call. Because I think

that's. I could always add on later.

And so basically you have media, which is audio or

this. You have video, you also have the audio and you have the

transcript. So what I have here is I

have the web page that's associated with it. So I can go here. I click

on this. This takes me to the webpage for it,

that episode. Right. Now, some of the stuff you do have to add manually,

but that's fine. You can just add an asset and tag it to this. I'm

also working on the ability to add a person,

right? So we would add the name of the person, the guest, what organization they

work for. My ultimate goal is to get a map of a graph

of how many people work in this industry. How many people work in this industry,

and kind of see that. And that could tell a story too,

right? Because at my heart, I am a data visualization nerd.

So that's kind of the thing. But as you can see, I have the thumbnail,

I have the transcript. So if I click on this, this is the.

The transcript from that episode. And then what I can do from

there is I've using the power, the magic of

AI Right. So if I want to create an

infographic or pull quotes, right? Let's

just say let's create a LinkedIn post. I know if you follow me on LinkedIn,

this may be like telling you about

the Easter Bunny, right? So I'm going to click on this

and I'm going to click generate. So what this is doing

here is this is saying, Please write a LinkedIn post about

this episode, right? And I will open,

I'll open a Grok because I just added this today based on a conversation you

and I had. Cool. So I click

on that. I want you to notice two things. One, that prompt

was automatically put into that text box.

Nice. And Grok is working on it right now.

I could do the same thing in Claude, Gemini, et cetera,

et cetera. But while that cooks, I will show you that prompts.

I have a template, little template language you could do for a prompt.

Right. So if I go here, I start

off the prompt. This is actually based on the. The

guest that we had on the

first show of the season, Jeff woods, was it? Yes.

Great show. That was a great show. And it was basically

how his book, the AI Driven

Leader and his prompt framework. This is really kind of based largely

on that conversation, right? So I have a thing here. This

is what I type. And I say based on the following data about this

podcast episode, provide a list. Oh, that's a typo.

Right? Can you give me a list of

comma separated tags for WordPress? Right? So podcast

name, I can just inject that as part of this script kind

of code here, right? So podcast name.com,

podcast name, episode title,

episode description, here's the URL. So transcript, etc, etc,

etc. I'll save changes. And so if I go back.

Let's go back to here.

Okay,

so if I go here now, I will see, I

will go tags,

generate and we'll do this one in Claude just for grins

and you'll see that it's there.

I click the button now it's going to generate the comma

separated tags. So this is useful for things like YouTube, where they want

this type of metadata for, to help us with SEO and things like that. Sure.

And then while that's cooking, we can see, look at, this is the finished product.

Nice here. Now there's obviously things I need to do to

clean it up, right? Because it says host Frank Lavinia. I wouldn't post this on

my LinkedIn, like talk about myself in the third person, but

just think about what's possible with this, right? And

all you have to do is load your RSS feed into the system

and all of this stuff becomes available. I can also add,

I can add more metadata. What's really cool about this is

that I picked this one because I'll put fun facts. So that way when we

do like a retrospective show or like someone out wants an

anecdote about a particular episode we can talk about,

I can say like, well, I was actually in a lot of pain recording this

and you can actually see it in my face. And I was attending

ignite and you can see the hotel room in the background.

And the reason why I was in pain was because we had a hot water

leak like literally 36 hours before I had

to fly cross country. So I had to clean up the basement,

put everything in a dumpster, and then get on a cross

country flight, which I don't recommend it at any age, but I can tell you

anything that bent or moved hurt. Oh,

Frank. So that's. And then the other thing I'm going to add is like, you

know, add a sponsors, affiliates like you mentioned. Yeah.

You mentioned an audiobook called Vibe Coding. Fun fact. Did

you know that we have a sponsor? That sponsor is

Audible and if you go to thedatadrivenbook.com

you will be routed to Audible and you will get a free audiobook on

us and you can have it be Andy's book

or that he mentioned called Vibe Coding or Jeff

woods book called the AI Driven Leader.

So it's really kind of taking this to the next level,

right? So the advantage of, the advantage of this is that

realistically, if you're going to launch a podcast, and I say this to any

podcaster or budding podcaster out there, right?

Obviously the microphone's important, the camera's important, all that stuff.

But there's a lot of other things to think about, right? Not just the mark,

a lot of it is the marketing of it. Right? Who's your target audience? You

know, what's your. The cool kids call that your avatar, right? Your typical

thing. What's your audience? What are you trying to do? But the other thing,

increasingly, in a noisier and noisier world, how are you going to

market this show across various social media networks? Right?

Right. And we've solved that problem

with things like dingo and buying

opus and things like that. But the

next problem you have is

creating content, right? Like, not just creating content, but

managing it. Right? So generally speaking, opus is the

clip that'll take a video and kind of cut it into like little

short form videos, you get about 1 of those per minute,

roughly, right? So let's just say, so a, a 60 minute

podcast or. I already mentioned that you already have four bits of content,

right? You have the actual audio of the show, the transcript, the

thumbnail, possibly video as well, like the full on

full length video. But then now you add,

if it's 60 minutes, you're going to have roughly 60 short

videos. So now you go from, you know, now

you have 64, 65, 64 items

now to track, right? And it doesn't sound like a big

deal, but when you're trying to organize things, it becomes a very,

very tedious work right away, right? And

what if you make an. What if you write a blog post about, you know,

the particular episode? Well, that's another asset. You have the track. Be nice to know

you. Because up until then I was googling it

basically against my own site. Like, where did I write that? Where did I put

that? So now with this ability, you have the ability to track

all that metadata in one place. And

as it goes further along, I'm going to add the graph feature

where I'll be able to upload metadata about each guest. Like

this person works here, this is what they do. And then I want to be

able to kind of track that. And I think that that will also, aside

from satiating my inner data visualization nerd, I think it

also could help people figure out where the next

opportunity is, where the next opportunity to

find guests. Maybe I'm over covering one type of industry,

maybe I'm not covering this.

Eventually I'd like to tie it into

download statistics so I can say like, hey, every time I talk about

bioinformatics, like this happens, right?

That sort of thing. And also help manage kind of, you know, all these

affiliate programs that we have, right? We have a program for Audible, we

have a program for Opus. It'd be nice to kind of have that in one

place as opposed to going around. What I do now

is I go around different websites and find it. Candace, to her

credit, shout out to you, Candace. She has them all in one spreadsheet. But that's

still, I. It's still a cognitive kind of

switch of, oh, I got to go to a spreadsheet. It'd be nice to have

everything, one place to rule them all.

So that, that's my stump speech. I know I got to work on

shortening my elevator pitch for it, but you're

muted, Andy, so we can't hear you. I'm sorry, I. I

didn't Want to cough last time. It's a great stump speech, Frank. Thank

you. So if folks are interested, let me know. Not that hard to

find LinkedIn or whatever if you reach out to me.

And we do go into wider beta testing unless you have it.

And obviously if you're a longtime listener or even a short time listener,

whatever. I tried to decide to charge, you'll get a solid discount from

Gotcha but cool, man.

Testing the data driven book.com

and I believe the,

that that was first off, I know the link has changed.

It's been. I don't, I don't want to use the word hijacked.

When did it work? Did it work the data

driven dot com? It doesn't.

No. That works for me.

Does it? Yeah, it works on my machine.

You know what, we'll include willing in the show notes. We'll include the actual

affiliate link.

I'm, I'm not sure what's going on.

I've got us. I've got a sneaking suspicion I do know what's going on.

All right. So in any case, we'll make sure we put in the proper link.

But you know, there are other things too that like you said, like I've been

working on optimizing things and things like that. It's more about,

you know, it's a testament to 1%

improvement every day will compound into a

ridiculous amount of optimization. So, I mean, like I can turn one bit of

content into, no exaggeration,

like I said, like 60, 70 bits of content. Oh,

goodness. Yeah, Easy. Opus is a big driver. Opus is going

to be a big driver of that. But there's other, there's other secret sauce I

haven't shared yet. Like, like shared with you, but I'm not shared publicly yet.

Right. So there's definitely things you can do and if you use your imagination, it's

a lot hard to figure out. Right. With all the other tool, generative AI tools

that we have, you know, whether it's

infographics, Notebook, LM like that sort of thing, you can

get kind of, you know, one of the things I discovered is you can

get kind of orthogonal views across different AI models

off of one transcript and compare them

and see what resonates in one engine and what

resonates versus another. There's a lot of opportunity there. Right.

And I certainly, I think if nothing else,

transcripts are incredibly important, not just for

accessibility but like just for the ability for

AI to ingest the content and it

becomes very malleable. And it's,

you know, I think That I don't think people in Mass have realized

that just yet. Hopefully. I totally agree.

Hopefully Podzi will change that. Hopefully Podzi will be a big driver of that. But

we shall see. And if you think about, you know, what LLMs

actually do, it's right, they're. They're all about the words

Brother. Right? 100%.

And so transcripts filled with words.

I like that. That should be the. That should be the tagline

filled with words. You heard it here first.

And I know you've been doing some exciting things with Claude.

It's just. It's such a. It's very easy to

get into the Debbie Downer mode of oh, my God, AI is

going to take over. But I see opportunity here

left and right. You have to.

There's a lot of opportunity here when it comes to

how you can use AI to be more efficient. Right. The

idea of us having a podcast. Doing a podcast is one thing, but

doing a podcast and

having it appear everywhere. One of the best compliments I got was when I was

in. When I was at Ignite, somebody knows me from LinkedIn,

they're like, oh, my God, you. You're everywhere.

You know, and it's funny. And, you know, we get emails

and. And it's like, you know, this is a ghost for you and your team.

And I'm like, I don't think they. The team's not as large as they think.

I take it as a compliment. It is, you know. Yeah.

So. And I'm excited because, you know, again,

we're going to. Sometime in the spring, we'll will have season 10 of

Data Driven, if you can believe that. Wow,

is that crazy? I know.

And that's like nine more seasons than I

thought we'd have.

And I didn't expect us to get nearly to this many episodes. I mean, we

were excited, don't get me wrong. And we, you know, we got out there and

did the best that we could at the time. And I thought we had a

fair start. You know, we. We lined up some

excellent guests. I think we nailed that to start with.

And that may be the secret sauce that propelled the podcast

through to episode 400 here and beyond.

So, yeah, 101 of the best.

One of the next logistic challenges, logistical challenges I gotta address

is scheduling. Right.

Calendaring. We. We did use Microsoft bookings.

I want to keep our clean language rating, so I won't give people my opinion

of there. There's two versions of bookings.

There's the version that, the way you think it's going to work or the way

the advertiser works and the way it actually works. Yeah.

And so I actually switched back to calendarly,

so as soon as I get that fixed up, that'll be on the site. I

know we have a lot of folks reaching out to be in the show. Don't

take it personally. It's just. It's been. I was talking

to somebody this the other day, like, between, like mid November when

the water heater broke till basically this week. I feel like I've been,

like, running behind, you know, trying to catch up, but.

But Again, thanks to AI, I was able to crank out over 30,000

lines of code in the spite of this. Right. That,

you know, and I. I think we all have had those projects that have been

back burner ideas that you think I'll get to it someday. And

someday never comes. Right, Right, right.

You know, if you look at a calendar, there's seven days of the week. None

of them are someday. I mean.

Very true. So thank you.

I want to say thank you to everyone who helped us get to 400 episodes.

It's pretty wild. Like, I remember listening to

podcasts just, you know, when I was just a young lad living in

Richmond, I was listening to.net rocks and they

were already at, like, episode. I think by then they were like, at episode

three, 400 they were up there. And I was

like, wow, man, that must be some achievement to get to that many episodes.

Not just, you know, having a podcast and doing that, but getting to that

point. And here we are. I

know, it just seems surreal that we, you know, that we got

here and we definitely

could not have made it here this far without our

audience. Y' all rock. And we do really

appreciate y' all hanging with us through all this. We

had, I'd say we had a couple of challenges with.

With, like, scheduling. Certainly you mentioned that, the schedule with the

calendaring part. But even there was.

There were a couple of times where we had trouble getting guests lined up

and the number of shows being put out lapsed.

And part of that was just due to stuff

going on in my friend's life that happens.

It certainly wasn't intentional. And

one thing I took away from it, even then, I was surprised by how

many people would reach out to me and say, when's the next episode of Data

Driven coming out? You know, it weren't mean or anything about it. It

wasn't accusations. It was all right, right. Lines of gosh,

you know, we missed the show. We miss hearing from you guys, and that's

Good. That's nice. That's nice. It was a compliment. Yeah, yeah, yeah. We

are going to work on that. Like even if it means that we have. We

had Candace as sub in for Andy. I don't want

people to think that, you know, if Andy doesn't appear for a while, there's nothing

personal. It's just that he couldn't make it. No, right. Like I don't want to

start that. I don't want. Candace is a lot more strict with

the scheduling, so she's effectively

joined our team. So she's a lot stricter with the stress. You know,

she's awesome. She's

joined our team and you know, so if Andy can't make it, she'll show up

and maybe we'll have a few more surprise like stand in guests.

Johnny Carson did it. Yeah. Work for him.

Yeah, he did pretty well. Most kids today won't know who Johnny. Carson

is, but he was the host

before Jay Leonard. And you know what? Some kids won't know who Jay Le. I

was gonna say. Was it Jimmy Kimmel?

No, it was. Who. Who has guest hosted,

I think on the Tonight show for a while. He did. He did.

And then. Yeah. Who got the Tonight Show? Was it. It was David Letterman? Was

it Conan? No, it was Conan and then Leno took it

back and then now it's somebody else. I think it's.

He used to be on Saturday Night Live. Yeah. I don't know who's doing

it now. It's been so long since I. I don't stay awake that late at

night anymore, Frank. And you. Only I know you don't either because

there's been plenty of times when I've been here, been in, in the office

here at like four in the morning. And

I'm texting with you. I'm gonna check. Yeah. Get a chat, a

text. This is Frank. Who

has the Tonight Show. Now I gotta know now who runs.

Yeah, it's. I can see the guy's face. I

can't think of it. It's Jimmy Fallon.

Jimmy Fallon, that's right.

So apparently it used to be here. The. The former host was Jimmy

Fallon. Johnny Carson, Jay Leno, Conan o', Brien, Steve Allen, Jack Par.

Wow. Yeah, that goes. Jack Parr.

Steve Allen. Jack Parr. Like our parents. Generation would

know them. Yeah. Yeah. So

that's wild. I don't know how we got on that tangent.

It's kind of what we do. I really would like to start making money with

it so we could have an off road race, rally, sponsorship Car.

Imagine like this data driven on it. It's like, why do you do that? Because

we're always going off track. But dumb, dumb.

That was Steve Gainsworth's idea actually. I think he said that.

Yes. Yeah, yeah. Stu and I are in a

very similar place because he posted recently, recently like

four to six weeks ago, about how he's largely

doing work outside the Microsoft data ecosystem.

I think you and I were talking about that and I'm largely, I'm

largely outside the Microsoft ecosystem these days.

I'm lagging behind you guys, but I'm moving that direction.

I'm telling you, man, I know you finally installed Linux

and you did that earlier on when you got your new laptop.

True. I'm, you know, I'm

not going to bet use this space to bash Windows 11. I already have done

that on my LinkedIn newsletter multiple times.

But our next episode, which is set to air, recorded with Andrew

Brust, who is a rd, which is like a

Microsoft mvp. And he, we, we

have an interesting chat about fabric and kind of all of

that. So. Okay. Yeah, I'm really sorry I

missed that recording. Yeah. He says to say hello. He says that at the end

of the episode. Tell Andy I said hello. But we definitely, we got to have

him back on. Right. Because there's only so much you could talk about fabric. Right.

There are definitely a lot of things I wanted to say about fabric that. But

anyway. But I think fabric is

in the right direction in terms of how

you do that. But I, I think though that the world at large,

and this could be my bias is

it's becoming more and more about private AI, local AI,

I think in a very real sense. Right. You know,

because that, that was my number one goal was to be able to run an

LLM locally. Right. And then after that

I want to learn more about how I could train

that LLM, you know, kind of shift its focus

even. And you know, I totally get it.

We, we have some clients at Enterprise Data and Analytics

and some, I guess

I'll use the word partners, businesses that we're

engaged with or communicating about engaging with.

And I learned from one group that is in Europe

that, you know, there are a number of countries where part of the

culture, the technology culture in that country is very

much anti cloud. And it's, it's not paranoia,

it's just a lack of trust. I don't think

the tech industry, big tech industry has done a lot to

engender trust. If you look at. Totally agree, you know, some of

the, the privacy violations that have happened across different

social networks and things like that. I can totally understand it.

And you also remember, like, if you're in the United States,

you have your choice of multiple cloud providers across multiple

time zones. There are not

Microsoft or AWS or Google data centers in every country in the world,

which means your data has to leave your country. Which it turns

out that reality strike. We should call this

phase of the Internet reality strikes back. The

whole idea of the cyberspace is its own

independent, separate thing is not held true. Right. Because

at the end of the day, everything virtual has to exist

somewhere on planet Earth, right? Yeah. What will be

interesting is that if they. If the talk of building,

you know, data centers in orbit becomes true,

that will be interesting. It'll be interesting, right? Will they.

Will they flag them? Will it work like Merchant? I don't know anything about space

law, but, you know, I would imagine

it would probably be kind of like maritime law in a sense. Right? Like, would

it. Would the data centers be. Would they be

flagged? Would they be like, you know, this is operating under US law,

this is operating under, you know, EU

law. That'll be interesting to see how. Interesting

point. Yeah. See how it all works out. It's almost like we should have a

podcast about future facing tech and AI.

Foreshadow much. Foreshadow much.

But once I get Podzi finished and I have a

nice. I have an even smoother workflow,

the barriers that I currently have will no longer

apply. So if you think we're doing a lot barriers,

we don't need no stinking barriers.

So with that in mind, I have multiple

text windows, like people needing to reach me.

Definitely. Stay tuned. We'll talk more about Podzi in the future episode.

And Andy's. Andy's still with us. And

we'll look forward to wrapping up season nine and then kicking

off season 10. It's going to be

awesome, Frank, 100%. And thank you, everyone. Once again,

we'll let Bailey finish the show. That wraps up

episode 400 of Data Driven. Thanks to everyone who's

listened, shared, and supported the show over the years. It truly

means a lot. We've got more coming as we close out Season

9 and head into Season 10. Until then, thanks for

listening and we'll catch you on the next one.