The WP Minute+

Matt: Mark Szymanski, welcome
back to the WP minute.

Mark: Uncle Matt, it's
always a pleasure to be here.

Thank you so much.

It's been a while, but it has

Matt: been a while.

Uh, you know, there's, uh, today's
topic we're, we're going to be

talking hopefully a little bit
about, uh, AI and, uh, my exploration

with coding with AI, like getting
a little bit more serious about it.

I saw you put up a video.

Uh, I almost trolled you on that video.

I'll save that.

Uh, for a later conversation.

That's how you put out a video
using AI to code a plugin.

So I'm excited to hear more about that.

And of course I'll link
it up in the show notes.

Um, but, uh, yeah, this time instead of
me leading, I'm just as lost as you are.

So this is going to be an
interesting conversation, right?

Usually I'm here coaching you
through something, helping

you understand something.

I'm just as lost as you are
with AI, uh, maybe even more so.

And, uh, I'm kind of interested to talk
about both of our experiences and maybe

what we're, what we're up to with it.

If that's cool with you.

Mark: That's always great.

Always great to chat with you, brother.

Um, yeah, I mean, um, well, I just want
to say right off the top, I know we

were talking in the green room before
this, uh, that you, I've only known

you for, I think, I can't remember
what our anniversary is, knowing each

other, but it's been a year or very,
just around a year, and you've gone in

that time from, completely dis complete
disdain for AI to now realizing that

it's, it's coming, it's, it's here.

It's gonna continue to get
a little more interesting.

And you've had a chance to
look at it a little deeper.

And now as you're saying off
camera, you're basically a

senior developer, you know, so I

Matt: feel like a senior developer, but
the code output is probably abysmal.

Um,

Mark: we definitely need to dig into that.

Yeah.

Yeah.

We definitely need to dig into that.

Matt: So let, let's talk about, um.

Sort of giving up the reins to
AI and like where my thoughts

are, you know these days.

I still think it's just a
Really good assistant right now.

It is Replicating so I've been using
it Claude I've been using Claude

largely over the last I don't know
six months I generally it's how I

shape a lot of my show notes for the
podcast episodes that I put out Really,

that's what I use it for mostly.

Um, and, uh, it's been
fine for that, right?

It's, it's been okay for that.

Uh, these are things that
are, you know, low value.

Um, you know, I, I just need concise.

I need speed and efficiency.

And that's where, you know, Claude
and his recent iteration of, uh, Uh,

it's, uh, LLM has been really great 3.

5 or whatever it was as of this recording.

Um, but anything else that I was trying
to have it do on the things that I was

looking for, like, I don't know, even
analyzing data, um, throwing a spreadsheet

at it and asking it certain questions.

Um, and I'd say even like six months ago
when I started getting into it, at least

with Claude, asking it to code some stuff
for me, it just wasn't really working.

Um, And fast forward six months, uh,
I've made a lot more headway, probably

because the tools are getting better.

There is one tool that I started
to use, which is like literally

life changing, uh, which I just
discovered over the, over the weekend.

But all of this is to say is like.

It still is just doing the things
I'm asking it to do, writing

words is doing good, writing code
seemingly is doing great because the

shit that I'm building is working.

But it's not really helping me or guiding
me if I'm making the right decisions.

In other words, there still
has to be a layer of human

interaction with this stuff.

Uh, which I think anyone taking it
serious, uh, will know that, uh, and

certainly if you're doing it for your
business, like if you're in marketing and

I don't know, you're trying to like make
blog posts and pages for customers or even

write code for your WordPress website.

I think at the end of the day, like you're
still going to need to know what the hell

you're doing and understand the concepts.

Right now I'm totally
lost with what I'm doing.

I just keep smashing the prompts
and it keeps giving me code.

I try it.

It works, or it doesn't, and then
I ask it to do it again, and then

I just move on to the next step.

It's the worst way of doing it.

It's like, totally, uh, unoptimized.

I'm burning oil.

And, like, I am the world's problem
with greenhouse gases right now,

because I am making NVIDIA CPUs just,
or NVIDIA GPUs just go nuts with

the code that I'm trying to write.

There's no optimization on my side.

Uh, does that make sense?

Like, I feel like it's still not
there to, like, just completely

solve the world's problems.

Mark: That is a really good example of
probably where I would say most of the

world is with these tools right now.

Um, I want to make some quick
disclaimers and maybe just kind of

continue to set the scene a little bit.

The first disclaimer that I'm going
to make, I think I can speak for

both of us, when we say that we don't
know how this stuff works right now.

technically at all.

Um, we can, we can pseudo
describe how it works, right?

It's basically just like they somehow
coded something that went across the

internet, so to speak, all the language
data that we have and basically built kind

of like a little bit of a language model.

We could think of it as almost like a
little working brain, but it doesn't

really expand too much on things.

I feel like we're maybe getting there,
but it doesn't really expand on things.

It's really just like.

Uh, uh, a model of data that it can pull
from and accurately kind of like, like

again, pseudo feel like a person that
you're talking to or something like that.

It's way better than a Google search, uh,
in a lot of ways because it has all that

data and it can express it back to you.

It's kind of like talking to a person
and for people that are completely

removed from technology, it's very new
and very interesting and very cool.

Once you play with it a little bit,
you realize kind of what's going on.

You realize some of the limitations
and all that sort of stuff,

but that's a big disclaimer.

The other big disclaimer that I feel
like that I can kind of say or set the

scene here for, uh, in, as of November
of 2024, Claude, ChatGBT, Gemini, there's

a ton of other ones obviously out there.

None of them are, this is I think the big
part in what you're kind of describing

in the beginning, they're not magic,
they're not mind readers, they're not, um,

independent, like, thinkers in any way,
they are literally just Assistance for

the most part the best way that in the way
that I, I feel like in the last like month

or two, I've really like stepped up my
prompt engineering or just my utilization

of these tools and the mindset change
that I had was like This isn't magic.

This is like my If I was like a, a
senior marketing director or a senior

developer or whatever, this is my junior
sitting next to me and I can be like the

manager or more of the director and they
can do some of the work and they're not

going to be able to do the whole work.

It's, it's actually kind of funny.

I've thought about this many times and I'm
like, I haven't, I don't have a extensive

background in corporate environment,
but it's almost kind of like that.

It can just take the
place a little bit of.

The less senior devel I'm not, I'm
not saying that we should, like, get

rid of junior developers and stuff.

I'm just saying, like, that's the level
of capability that this stuff has,

but like you said, you ha it has to be
directed by somebody, kind of, currently.

It's not really It's not really there yet,
at least the way that most people use it.

And again, the biggest disclaimer
of all is this just started, like

let, like a few years ago, max.

So we are extremely early to this.

And, uh, that's why, that's why
I'm making some of the content.

Cause I know this is where it's headed.

Like, we're just going to keep
getting crazier with this.

Matt: Yeah.

I was having a conversation
with my brother.

Who's been using chat GPT, uh,
for a little while and he uses it

much more for, uh, coding stuff.

And I talked to him a couple of
weeks ago about this and, you know,

he was saying like, Hey man, this
is, you know, changing the world.

This is going to be, you know, totally,
this is really going to make things

weird over the next couple of years of
how people are, whatever, just doing

anything on the web, whether it's,
you know, your search from search to

creating content, to writing code,
to interacting with those sites.

Um, and I agree with him.

But I think at the top of this stuff,
uh, I feel like we're in the, um,

the housing bubble, uh, or like, uh,
a couple of years ago during COVID

when like money was free and everyone
was just like getting mortgages.

It is like 2%, you know, you could do all
this stuff and money was just out there.

I feel like we're in that mode
right now with AI where this stuff

is still too cheap, like 20 bucks
a month to do what I'm doing.

I can't see that sustaining, right?

And I think what we're looking at is in
the near future, a lot of these sites

either just like constantly ramping
up the price like we see with freaking

streaming services with everything, right?

Because they're burning
cash and power literally.

Um, and Inevitably, there'll be some like
some major corporation needs to win as

it always does and we're gonna start to
see like the Tightening of like what's

available at what cost so the way I see
it is all these guys Claude Chat GPT.

It was the only two I really use
I don't use Gemini or anything

else really but those two are just
gonna be like Alright, we're done.

Like, the free ride is over, it's
either now minimally a hundred

bucks a month, two hundred bucks
a month, like, what's the pressure

point where people will pay up to?

Uh, or you're just not gonna get
all this new cool shit that you're

getting now for twenty bucks a month.

And that's really gonna, that's when
things will change, I think dramatically,

with like, who has access to what.

Like, what is your best AI,
and can you afford to use it?

Uh, I think is, is where we're headed.

It's not gonna be, like, this commodity,
like, where, like, where Accessing today.

I mean could be wrong, but
that's the way I see it

Mark: mate could be Um, I wish I knew who
it was and if I find the video again, we

can link it below but there was a video
it was just a short like a one or two

three minute video where this guy was
saying that Around the time that open eye

open ai did they go private or public?

I can't whatever they did there
like They had like a weird

Matt: non profit

Mark: at

Matt: one point sort of
like this mullenweg and

Mark: and automatic stuff
like you're just like what?

Yeah, what yeah what happened there?

Um, but so Around that time, somebody
was, uh, in this video that I saw, like,

kind of like, it was, I wish I could
remember the argument entirely, but

basically they said that they're not
worried about basically what you described

there, because all, because everybody
and their mother, it was so weird how

this happens, like, we heard the word
AI, we saw ChatGPT, and then basically

overnight, everyone started implementing
these things, and it's strange how that

happens, sometimes it's like you need one
person to just be like, oh, I figured it

out, and then somehow, Even though you
couldn't figure it out for many years.

Now everybody is releasing their thing
where they figured it out, I guess

just reverse engineering or whatever.

But, um, but because we're seeing
that and all these things are cheap,

they're thinking that everything is
just going to continue to be cheap.

Now, again, certainly at certain
points, there's going to be more

developments and maybe there'll be
like more expensive, but then they'll

come down in a more commoditized thing.

I just don't think that this version
of AI is ever really going to be

like expensive because it's already
cheap and it doesn't do anything.

Completely groundbreaking.

Like it's definitely a big shift, but
it's not like you can just tell it to

do something like, you know, we, we
think about, um, I don't know if you've

seen, obviously you've heard of Tesla
and the cars and everything like that

and everything you can do, but those
are still like a premium ish product.

I don't know if you saw the
Tesla bots just like that.

They showed like that level of stuff is.

I feel like AI, I don't know, 0,
whatever they call it, you know,

whatever they're going to dub that,
that's going to be the next thing.

Once it starts moving into like hardware,
I feel like, and physical items that are

going to like actually enhance your life,
like more physically, rather than just the

stuff that we can do digitally, um, that's
when the world's really going to change.

Like, because everybody's going to
see that, like when you see robots

walking around and things, it's
like literally iRobot from that Will

Smith movie, that's going to be wild.

But, um,

Matt: Yeah, it doesn't it.

Yeah, I agree.

And I think we're we're
missing this layer right now.

And I want to talk about our experiences
like coding in a moment, but we're

missing this like connected layer.

So like, what would AI be
really great for right now?

Well, I want it to look at my to do list.

I want it to prioritize it based on
like what I've prioritized in the past.

I want it to You know, recommend
what my next steps are, right.

To improve my productivity.

Right.

Because, you know, getting things
done, like, you know, you have this

to do list, it's like write the
blog posts and you're like, okay,

like, where do I start with this?

Like, I want smart intelligence
to like, look at my list, break it

down and then start to prompt me and
say, here's what you should do next.

Like these are, but it's not
gonna, there's no AI to do that.

And, and then you're just siloed.

Let me stop

Mark: you there for a second though.

Do you.

First question, and then
there's a follow up.

Do you think it can't do that right now?

Do you believe that it can't like
from your experience or do you

believe that you just don't have the
tooling or you don't, you don't know?

I just don't think

Matt: that I don't think
that tooling exists.

Like I don't want to give
up my to do list app.

I want it to work with it.

And you know, we've seen some of
these things like, I don't know

if, Folks are just listening.

I'll try to find it and link it up.

But I know Claude launched, um,
some new experimental thing.

I don't think it's generally available
yet, but it'll control your computer

to do things like that's getting to
the next level where like, yeah, okay.

So now I can tell Claude, Hey, you
know, prune my to do list, open up my

email, start responding to that stuff.

But I still think that's so far away.

And then do you really trust back
to that human interaction layer?

Like, do you trust the code that
you're deploying to your server?

Like you're going to spot
check it before you do that.

Um, And are you going to trust it
to manage your, your to do list?

Uh, you know, I, I don't know, but
that, but yeah, I mean, once it gets

incorporated into that layer, yeah,
I mean, I think productivity will,

we're going to see a few years of like,
wow, this is immense productivity.

Um, but then what are the next few
years after that look like where.

You know, it's starting to do things
for you and you're like, holy shit.

Like, do I get paid the same, uh, you
know, because of these efficiencies?

I, you know, I don't know.

It's, that's why, you know,
again, shout out to my brother.

I think he's looking at it going,
things are going to get weird, but, um,

let's just talk about our experiences
with, with coding and, um, I'll start.

So I, I have this idea for this web
app and it's not even like something

I'm making generally available.

Let's say.

It's more of like a tool for myself,
um, but I want it to operate, uh,

as a web app, um, one, because I
want the challenge, but two, like,

there's other ways that I can
extend it and do some other things.

But anyway, the point is, I was
just like, all right, I'm going

to get more serious about it.

Let me see if I can figure
this stuff out because Clawd 3.

5 came out and I heard
it was great for coding.

So my first experience is Clawd.

Here's the idea.

Here's what I want you to build.

Uh, I had some direction.

Uh, I asked it a few things cause
I do have a sort of a, uh, uh,

network operating background.

So I was like, let's compare
this to running on an AWS stack

versus, uh, a CloudFlare stack.

Uh, I've heard some great things
about CloudFlare's efficiencies.

I just don't know if it's as mature.

So I asked it to like compare and
contrast some of this stuff and

I ended up going to CloudFlare.

Path because I just didn't
want to use AWS for this.

I don't even know if that's the
right thing So that's problem.

Number one.

I'm asking AI to like give me advantages
of Both of these technical stacks.

It said, you know cloudflare
is the most efficient.

Okay, you know, so great So I decided
to go with that and then I just started

dumping the ideas and like just hammering
it with Like the questions and what I

need it to do and the first major lesson
that I learned was these interfaces You

know you go to Whatever it is, claude.

ai, chat.

openai, um, com.

These interfaces, you just dump stuff
into it, and it is just like this huge

pool that you're just throwing stuff into.

And what I found with Claude was, I
just started burning all of the tokens.

I'm a pro 20 bucks a month user,
but I would reach my limit.

Uh, In like an hour of asking it
questions and having it code for me.

So I found myself talking to it
and the answers it was giving me.

It was like, ah, it's
like, it's guiding me.

It's like saying, here's
what you should do next.

And then it would prompt me
and say, well, you know, do you

want to extend these features?

Do you want more clarity?

Like do you want, and I'd be talking to it
and like, yeah, man, give me more clarity.

Cause I don't know what
the hell I'm doing.

So it would give me these answers.

And extend upon it, it was just so
much so fast that I was burning the,

the, the tokens and then it would
just say, Hey, you can't use 3.

5 until whatever, 2 PM, you know,
it'd be like a four, three or four

hour window and I'd have to like shut
my laptop and be like, fuck, I can't

talk to this thing anymore, right?

But I could go to three or whatever,
whatever it's called and that, but the

answers were just so vastly different.

Um, the experience was so different.

To like ask a code and it didn't
know like where it left off

with my other conversation.

So that was super frustrating.

But the first lesson I learned was
you still have to be dialed in.

Otherwise you're just at least with Claude
anyway, you're going to burn through those

tokens, you know, just asking a zillion
questions and having it talk to you.

If that makes sense.

Mark: Yeah.

And if you, if you've, if you've tried
Claude, I would recommend you take a look

at it, but what Matt's describing there,
what you're describing, yeah, it's, it's

entirely, um, it's entirely like that.

So I made, uh, a couple of videos maybe
two months ago at this point where I

was just diving into this concept of
literally maybe my first times using

Claude and trying to code when I did that.

I had that exact same
experience that you had.

I was like.

What the fuck?

Uh, I, I have all these questions still.

And like, I, I can't do this
until tomorrow or whatever.

I'm not sure if there's a way around that.

I don't know if you can buy more tokens.

Um, one piece of one thought there
though, is that whatever you're spending

to get that information, you absolutely
would have spent way more money trying

to get somebody to do it for you.

Like, sure.

So, I mean, just to keep that in
perspective, but I, but again, it's,

it's, your point is completely taken.

Like that's kind of annoying
where we're at right now.

I do think it's probably a bit of.

Again, where we're at right now, so
to speak with again, being extremely

early, but the fact that you can do
it, you know, we could stick on that

and that's like the, the main piece.

What I've learned from my second go
around with this video that I just,

uh, posted today, uh, as of recording
this, it's like an hour and a half

long video and, and I did it, and
I did it specifically in two ways.

The first part of it was just kind
of me explaining what I was doing

and the finished product and then
the back half hour or whatever.

As me literally.

like playing around building it.

So there's a, there's a bunch of
little gold nuggets in there, but a

couple of the ones that I found was if
you're going to do this, you have to

realize like you have to continuously
get better at being a prompt engineer.

You have to understand that, like, you
have to give it all that information.

You have to ask it to ask you questions,
like prompted to ask you questions

and particularly one at a time too.

So you can just like kind
of have a conversation.

There's a little bit of stuff
in there about like, you know,

saving tokens and all that.

The first time I did it,
this is a really big one.

The first time I did it, when it'll
give you, let's say you're just

making a simple WordPress plugin, so
it codes a bunch of stuff for you.

It outputs the files and the nice little
art, uh, artifacts feature things.

So you can kind of see it
and everything like that.

You can download the files, you
can save them into your, on your

computer directories and all that.

If you ask it to change something or
it has to rewrite it, what it, I think

what it does on purpose is only gives
you part of the file back because

I think that's just less tokens.

It doesn't want you to rewrite everything.

I was telling it, no, I don't, I want
you to rewrite the whole thing for

me so I can just copy and paste it.

Not really a good idea because you're
going to burn a shit ton more tokens.

Yeah, that's what I was doing

Matt: too.

I was just ripping through it.

This

Mark: is what I mean, like, you
have to be the senior developer,

and it's the junior developer.

Like, you're still working,
you're just guiding it, and

it's doing a lot of this stuff.

So that's one, that's
one thought in my mind.

Um, we could talk about, like,
you know, is the code good or not?

I think that if you're doing something
relatively ubiquitous, it's going

to be pretty good because it's not,
unless it's hallucinating, which

I Hopefully you can kind of tell
maybe sometimes, but I don't know.

You definitely probably want to get
the shit vetted, let me just say

that, but The one other piece of it is
you mentioned where you're just kind

of dumping everything into a pool.

The one cool thing that I've found with
Claude is it has a projects feature.

I don't know what plan it's on or
whatever, but that kind of like

where we're, where we're not at yet.

And kind of, they're kind of not there
is think about like these LLMs of having

like a big language model that they've
scraped the internet and looked through

all the data and everything like that.

Okay.

That's the whole LLM.

What I want though, and I've tried this
and I haven't like put it, I haven't

made a video or anything too much yet.

What I want to do is the LLM doesn't
know what I want in this project or maybe

like about me or my business or whatever.

The secondary piece of this, and people
have been doing this on like, if you've

ever seen somebody do like a personal
AI chat bot where they basically

recreate themselves so people can ask
questions and answer like there, you

need to, what we all should be doing.

This is a really good piece of
content actually that I should write,

but we all were like building our
Facebook pages or our LinkedIn's

or our websites or our blogs.

What we need to be doing now is
like building our own personal LLM

where it's like everything about
us so at any point like we could

call in or somebody else could call
into that and like get the answer.

Cause that is like, I feel like
maybe 75 area of where we need

to go with some of this stuff.

Because then, if you can do that,
and you can create that, your own

little segmented thing there, then now
you've added more information to it,

kinda like how you're doing, right?

Like you threw all that information
in there, and if you could silo that,

you could always come back to that.

I know they do it in the chat formats, but
the projects thing is a separate thing.

You could always come back
to that and you could always

continue that that piece of it.

So a lot of stuff there, but
Um, that's been my experience

a couple little tips that i've
found so far in my short journey.

Matt: Yeah, I started I Purposely
started in a project like when I got

more serious about building this app.

I started in a cloud project um, and
I I I still find it lacking in It,

especially when you hit, uh, uh, I have
to just get better at understanding

like how the, how this stuff works.

But if you start a chat, let's say with 3.

5, uh, sonnet, I believe
is what it's called.

Yeah.

Yeah.

Sonnet.

Um, which is there at this point,
they're most advanced, uh, that

is accessible to average users.

Um, You start that chat, you start,
you know, going down the road of like,

okay, build me this, this script.

And, um, I'm building something
again, runs on CloudFlare.

CloudFlare has workers, which is, I don't
know, the way that I understand it, the

way that I've always perceived it as
is like a serverless code architecture.

Basically I can run tiny, tiny,
tiny little operating systems that

will run these, these scripts.

Uh, these commands.

So I have to build this big script that
does all these API endpoints and routes

and connections and all this other stuff.

And if I'm chatting with it, 3.

5 SONNET, I'm improving it.

I'm adding API endpoints.

I'm testing it.

I'm building it, deploying it.

And then I run out of tokens.

And it's okay.

You can't talk to me until four o'clock.

I can't like back out to the pro like
air quotes back out to the projects

folder and then pick up on it with
like version, uh, you know, LLM 3.

0 and say, okay, let's reference this file
that we've been building over here with 3.

5 and now like, let's
continue building this.

It just, it has no idea.

On how to do that, unless I'm just
maybe not prompting it well enough.

I mean, it creates the artifact, you
see it, this is getting like really

inside baseball here, but if you use
Claude, folks who are listening, use

Claude, you know what I'm talking about.

You have like, the previous files that
it's created in the project that you

can click on at any time and see, even
with a name, because it gives it a name,

or you can actually tell it, give it
this name, uh, and you try to reference

it, it still has no freaking idea.

Which is one of like that most
aggravating things for me unless I

unless their projects don't work the
same as I thought they did Or it just

simply has no idea because the 3.

5 stuff was way more advanced like that
Let me backtrack if you're asking 3.

5 to code something for you It
is literally like having somebody

sit alongside of you and talk to
you Like, it explains everything.

It asks you what you want to do next.

Suggests things that you should
be doing, you know, next.

Uh, for like features.

You're like, oh shit, I never
even thought about that feature.

And you're like, wow, this is amazing.

And then as soon as you go to like
three, It's like, the baseline one,

It's just like, here's the code.

It doesn't say anything else.

And it's just like, shit
man, what the hell happened?

You were so nice to me in this other chat.

And now we got nothing.

Um, it's vastly different.

And um, So what I started to do it
too is like, yes, I started to, as I

understood the limitations I was running
up against, I started, you know, making

the requests a little bit smaller and
then in the, I forget what it's called.

Uh, let's see if I can
pull it up really quick.

There's instructions.

I think you can say, I think
it's called instructions.

Yeah.

Set custom instructions for the project.

So in the upper right hand
corner in the project, it says,

how should Claude respond?

I said that.

I profiled myself, and I said, Before
writing a lot of code, please explain

quickly what you'll be doing before,
before I can confirm, or so I can confirm.

Right, so basically,
don't burn those credits.

Or tokens by writing a shitload of code.

Tell me what you're going to do
first and then I'll approve it.

And even that didn't work all the time.

It did when I asked it bigger
things and it would like bullet

point some features it might do.

Um, you know, uh, it, you know,
it would ask me, it was, it

was like hit or miss with that.

And I have another one in the instructions
that says when writing code, please

use good inline documentation.

I'm a beginner level developer.

It helps me find the areas you need me to
update because what was happening was it

would suggest those changes to the code.

Like you said instead of like writing
out 800 lines of code every single

time which was a complete list of the
file It would say just update these

sections and I'm like, I don't even
know where do I find this section?

Yeah, where the hell is that at?

So I'm trying to say like like
really document this and help me

understand where it needs to go later
on So again hit or miss with that.

I wish it was a little
bit a little bit better

Mark: Yeah, I mean again, I've had similar
things and it probably it sounds like

what you're trying to achieve is probably
more more robust and more advanced than

what I was trying to achieve in that video
specifically that was like building a I

think that maybe had Maybe seven files,
eight files or something like that.

Um, so I'm not sure if there's more,
if there's more to it, but I mean, if

you have more functionality, then the
files are just getting bigger, it's more

tokens, it's more decisions, it's more
features, it's all that sort of stuff.

Um, but I mean, I don't know, I'm,
I'm bullish on it to this point.

Cause I think it's pretty
incredible what we can do so far.

But all of the feedback that you're
giving, I'm sure, like a lot of other

people have like had, and like, I
don't know, the thing, the thing

with this is the part that confuses
me the most is how you actually

make something like this better.

Like, like I, I've, I've, I mean, if
we want to go off topic of coding,

I don't know if you have more on the
coding stuff specifically, but just

like other use cases, I've actually
found a ton of ways that I've used

like Claude specifically and a chat,
GPT and Gemini to certain cases, like.

I have a laundry list of things now just
in the past two months that I've used

this stuff for that has actually made
a difference, I would say, in all of my

workflows and my just thought processes
and the way that I work on things.

Matt: Uh, let me just wrap up
on some of this coding stuff.

So then, so what happened next is
like, here's the things I don't know.

I'm way out of my league on,
uh, on what I'm coding, right?

It's pure JavaScript, you know.

I mean, I'm actually.

I can understand some of it now, right?

Because I've literally been neck
deep in it and it's actually

a great education process.

Sort of like reminds me of
learning WordPress all over again.

Um, which is cool, but way
out of my league in terms of

like what it's developing.

So I found, uh, an app called cursor.

Have you run into this?

I've

Mark: heard of it.

I haven't used it yet

Matt: because when I was coding,
like when I initially started this

project, I'm like, I got to start
like a whole local development.

Environment.

I haven't done that in years.

And when I did it for
WordPress, WordPress is so easy.

Like it's so, uh, you know, I
guess it's the good and bad of

WordPress because it's so archaic.

It's my SQL, PHP and HTML, right?

And CSS and of course now JavaScript.

But the essence of it is like you can get
a local environment up pretty damn easy.

Whereas with like this stuff.

You know, package installers, NPM
install, like you're building all this

stuff, then you have to deploy it.

You have to do all of this stuff
to get this thing going, and

then what are you coding in?

So I've, okay, hey, hey Claude, what's
everyone using for coding, right?

And they're like, oh, VS code.

It's okay.

So I started using VS code, and then
my mind is blowing up because I'm

like, fine, man, I don't even know
how to say, like, then I just feel

way out of my league because I'm
like, this thing isn't set up right.

I don't even know if I'm optimized right.

Am I doing this correctly?

Um, and then I started
hitting that wall of Claude.

Now over the weekend, I started
making a lot of headway.

Things were working.

I built my, what's known as the
Cloudflare Worker, and it was working.

And I could send data to these APIs,
it would talk to other APIs, and send

me the information back that I needed.

I was like, okay, I'm
making, making headway.

And then I was like, okay, now I need
to make this a front end because I can't

just do everything through the terminal.

I need to be able to manage
this in a web browser.

And that's when the freaking wheels
fell off because I was just like,

Hey Claude, build me a react app
that connects up to all these APIs.

I can do this stuff.

And it was just.

It was just a shit storm.

I mean, I just 48 hours of like trying
to just figure out how to deploy this to

cloudflare is the problem with cloudflare.

I don't know is a problem
with my code probably.

Right.

So there's all this stuff happening.

And, um, I found cursor and I
don't even remember how I found it.

I think I was like, I need to find an
alternative to vs code because I just felt

like, I'm, I just using the wrong tool.

Like, I just can't do this stuff anymore.

And then I found Cursor, and
Cursor integrates with Cursor.

Directly with Claude and the
magic is it's not your tokens.

It's their tokens So I can just
freaking pound away at this thing and

just be like, here's the 900 lines.

Look at it Tell me what's wrong.

All right, give me the update.

Here's another 900 lines like give you
know, give me another update And that has

solved it dramatically and basically what
it has stopped and what it has allowed

me to do is in, in cursor, now this is,
you know, anyone who's like a proper

dev is like freaking out flipping tables
right now, but I ask it to do something,

it shows me the code like I was doing
in my Claude web browser and instead of

me copying and pasting it into the file,
you just say, apply it and it writes it.

Hmm.

Right in the file Right.

So I'm just like oh, yeah all day man.

Like I just said yes.

Yes.

Yes.

Yes Like this is all the code I want
And I finally got to a point where

now I'm not using react anymore I'm
using another framework called svelte,

which is like a pure JavaScript play
on it supposed to be more efficient.

Yada Yada, I don't know.

I'm like watching YouTube
videos while I'm coding.

I am you're not coding I'm coding, man.

You're

Mark: prompting.

I'm coding.

Matt: This is the way that I see it is.

I'm the guy with the idea.

We're working together on this.

It's a partnership.

Uh, me, Claude, and, and Cursor.

And, I finally get to a point
where I've, I've launched the app.

It's, the very basic iterations
of it are, are working.

Um, but Cursor is really
the thing that made me.

You know, turn the, turn the corner,
which has allowed me to now I use.

Um, so instead of like, uh, working with
like some of like the real inside, like

if a big issue comes up, right, like,
um, I'm talking about and I'm, I'm asking

it to like help me style the front end.

So instead of doing that even in
cursor, because you've got this tiny

little window, literally, um, I'll,
I'll start asking chat GPT and Claude

what their recommendations are.

So now it's like, I have two
different people in front of me.

You know, big, big question, right?

Like a big question.

Like what should I use tailwind or
should I use like some other CSS

framework, whatever, and I'll, I'll ask
them both, see what the answers are.

And then I'll aggregate that and
then bring that into a cursor

and be like, okay, here's where I
think I should be going with this.

What do you think?

It'll give me the answer
and man, I'm a, I'm a single

person enterprise is what I am.

Okay.

I'm going to be launching
so many apps, dude.

It's not even gonna be funny.

Okay.

All I have to say is
WordPress is in trouble.

Mark: WordPress is in trouble.

Yeah.

Oh my God.

Um, man, people are gonna
be so mad about this.

Um, No,

Matt: it does help me really appreciate.

Like what WordPress has like really
like even like I'm looking at

it now, and I'm like holy shit.

There's Lot of like legacy stuff
here and I can understand why people

who are on the cutting edge look
at WordPress and go That's a joke.

Like look at that dinosaur over there
But at the same time as an end user

like imagine trying to do like imagine.

Oh, yeah AI is gonna change all this
stuff You can just make your own

WordPress, and it ain't gonna happen.

Uh, I mean, not anytime soon, and not
to what you can do with WordPress, but

it does also make me appreciate, like,
what we have as, like, this little

bottled up application that is WordPress.

Mark: I got a lot of thoughts.

The first one I have to say,
though, is that I didn't realize it.

It took less than a year.

It took a week for you to be
completely bullish on AI and be

like literally a senior developer.

That's amazing.

I don't know if anybody's
ever done that in a week.

Um, I'm gonna quickly go

Matt: back to the CEO level in a moment.

Mark: There you go.

Once I smart it up.

Um, yeah, the one random question that
I have that I don't know if we can

really answer, but it would be good to
get other people from the community to

answer this is, We are not developers.

I wouldn't classify myself.

I don't think you'd classify
yourself as developers, right?

Like we know code here and
there, just from being around.

We know what stuff does, but we've never
sat for eight hours, multiple times per

week, per day, per year to like actually
develop software or anything like that.

We know enough to be dangerous.

And now with Claw, we know
enough to be really dangerous.

So I'm wondering.

If we're that camp, so to speak, like
we're that archetype and we're see

this and we're like, Oh shit, we could
like build our own apps and stuff

like that, or like our own plugins.

And it's like, we know the dangers
kind of, if we're being honest with

ourselves, but we're trying it.

Cause it's interesting.

We're trying to learn.

I wonder if software developers, the other
archetype, if they look at these things

and they're leveraging, I wonder if like
that, that, that, what you just described

there with like Claude and cursor, I don't
know this enough, and I'm sure there's.

Videos and people doing this stuff,
but I don't know if our, our software

developers, you would think that if, if
you put that into a software developer's

hands, like the co pilot from GitHub or,
you know, whatever, or whatever, whatever

the, well, there's tons of them now.

If you gave that to somebody like, are
they just like flying at like Mock 10 now.

Yeah.

I think that's

Matt: where the biggest
wins are, for sure.

Yeah.

Mark: You know what I mean?

So it's like, the thing is, it's
not like only we have these tools.

They have the tools too, so they're
not in trouble or anything like that.

I don't think that by any means.

Um, and they still know what they're
doing, so it's like they can go faster

and they know what they're doing.

So it's like if they run into the problems
that you're having and I'm having, then

it's like, oh, I know exactly what to do.

Boom, boom, boom.

I know exactly what to tell it to do.

Da, da da, and, and
they're off to the races.

Yeah.

Um.

Matt: I think that's where
cursor is supposed to really win.

I'm using it from a, I'm using it
from, uh, like it's chat features.

I feel like are supposed to be just
an assist literally, whereas I'm

using it to code the whole thing.

I, they, uh, from all the stuff that I've
been watching and listening to people

talk about how like amazing cursor is,
is because it has like this auto complete

feature, which if you know what you're
doing speeds it up like that 10 X level

where you can start, like if you know
the code, you can start writing the code.

And then it'll start auto completing
the code it thinks is gonna come next.

Mark: Oh wow.

Matt: So, I don't even
know what I'm doing.

So I'm asking it to write stuff for me.

And then, that's what I'm,
that's how I'm using it.

Whereas like, man, if you know
what you're doing, this thing is,

you know, puts you on Mach 10.

And then if you really have a
question, then you can ask AI.

So, yeah, like, it really helps
that kind of person for sure.

Mark: Yeah, I mean, just in general
with the coding stuff, I love

that these tools are available.

It's awesome that they're
available to anybody.

Um, I'm sure some people are
going to be like, well, you

shouldn't be coding your own stuff.

It's like, I mean, I don't
know if you want to play around

and you understand the risks.

Like in that video, I say like,
I'll give you guys this code, but

do not, do not hold me liable.

Like, this is your own, you
know, use it your own risk.

Like, uh, it's, uh, It's there, but
I don't know, you know, it's got an

encryption class for like the API
keys, but like other than that, like I

have no idea what holes you could poke
in this thing or anything like that.

So I would always just say exercise
caution there, but it is pretty

incredible what we're able to do.

And seemingly in such a short time
from when we went from like pretty

much everything being like hand
coded, so to speak, to now, you know,

we have, you know, We can get so
much further so much faster us and

people doing this professionally.

It's, it's pretty, it's pretty insane.

Matt: So how, what other areas are you
using it in outside of content or coding?

Mark: Yeah.

So the other area would be like,
um, so we're not coders, right?

Like we said, but we are content creators.

And I think that like you
and I are probably similar,

we're similar in that realm.

So it's like, we know ways, and we
know the hangups and the limitations.

That creating content, like
of creating content, right?

We don't really know the software side.

So if we switch to this content
side, it's like, How, how, how,

how can we possibly speed this up?

So let me think of a
bunch of different ways.

The first way this is like, again,
some of these are like small

wins because they're not like
groundbreaking things, but they're

really like, actually pretty awesome.

Like if you stack them all up, first thing
is, okay, so we record YouTube videos.

I'm not going to do it in order.

Cause I can't remember off the
top of my head, but I'll just

go from thinking about these.

The first one is like, we
record YouTube videos, right?

And I like to put chapters
in all my YouTube videos.

I was literally doing this.

Like I would watch the whole video
back and put the chapters in.

I don't have to fucking do that anymore.

Like I, I upload the video.

I wait, maybe, sometimes it's right away
depending on the length of the video.

I took that whole hour and a half script.

In ChatGPT, I tried to do
this with ChatGPT but a while

ago it's, it's not as good.

Claude is way better with like
pasting things in and maybe it's

just the new model now, like
more, it can just understand more.

So I, I upload the video.

I go to the bottom, show transcript.

I copy that whole thing
with the timestamps in it.

I paste that into Claude and I say, Hey,
Give me YouTube chapters, one liners for

it, just like formatted perfectly, and
just give me chapters for these videos.

And then I can even say, hey, get a
little more detailed if it doesn't

give me enough, or how I want it to do.

And that saves a shit ton of
time if you were going to sit

there and watch all of that.

And that's just a small win, like I said.

So there's that.

You already mentioned
things like explaining.

You could, you could even say,
I literally just tried this

with the last couple of videos.

Let's stick on that just
from the transcript.

It knows everything your
video is about, right?

And it's an expert at YouTube if you
tell it it's an expert at YouTube.

So you can get the subtitles from that.

I'm sorry.

You can get the chapters from that.

You can tell it, give me 10
really good titles for this video.

Based on just your knowledge of YouTube.

So, and then I'm not saying
like copy and paste that part.

Like I just need the ideas cause I'm not
creative enough to think of the ideas.

So like, give me the titles.

I even experimented with this.

I said, you already mentioned the
description and stuff like that.

So you could obviously get that.

But the, I even experimented
with, Hey, I know you for Claude.

I know you can't, I know
you can't generate images.

Okay.

And we can talk about images.

Cause that's like kind of
another modal of this piece.

That's good and bad.

Very hit or miss more than text.

I know you can't generate images, Claude.

But can you prompt some, another
tool to generate a really good

thumbnail for this, or at least give
me some ideas and stuff like that,

or what a good thumbnail would be.

It does it.

It doesn't do it great.

Cause then I copied that.

I took it over to Grok on X cause it,
cause I think Grok uses flux, which

is probably like one of the best
image generator generative things now.

And I put it over there and it didn't give
me great stuff, but it gave me something

that like, you know, maybe would spark
a little bit, you know what I mean?

But like even with text, you can kind of
get some ideas for imagery and things.

So those are just some of the ways like
in YouTube that I've sped up and at

least gave myself way more ideas for
things because, you know, Those things

are hard to come up with on your own.

I mean, yeah, I mean,

Matt: as an assistant,
it definitely helps.

Um, it, it still hasn't ever really
knocked it out of the park for me

for, for, uh, titles and stuff.

I think it's still very
baseline, but it does bring in.

Uh, like a, a different perspective
and what I'll do is like ask, like,

here's my, here's the type of viewer
I'm hoping to, uh, uh, to, to watch

this video, uh, whatever it might be.

Freelancer, WordPress beginner,
whatever it might be, the keyword is.

And then like, what would appeal to them?

And I, and I use that as like ideation,
you know, and other, like, like other

words, you know, like what are other words
I can use that are, are keywords for this?

Um, yeah, all of that stuff,
uh, you know, works great.

For the chapters, uh,
because I use DS script.

Um, D Descrip just has chapters,
a chapter thing built right in.

So you just click it and
it's still all ai Yeah.

Chat, GPT driven.

But, um, you know, I've always found it
to be a little bit more accurate because

it has the whole video file, uh, and it
just, whatever, just does a better job.

Um, but, uh, yeah, often I'll use, uh,
I'll use that and I use Whisper AI as a.

I think it only runs on Mac, though.

Uh, so Whisper AI, or
Whisper Transcription is

what it's called, uh, on Mac.

It's a Mac app, and, um, you can
just drop in your video or, um, MP3

file, or any, any file, any media
file, and ask it to transcribe it

and then run the same commands on it.

It's just like a local app.

Um, I'll use that from time to time.

Mark: Yeah, the other thing I
try to do, I haven't done it yet,

As much recently as I take the
transcription because it's not great.

I mean, maybe Descript already
does this, but the transcript

isn't great from YouTube.

Um, like it's the words,
but it's not formatted well.

So like the idea would be to take
that text and put it on your website

so you could index that text.

Um, I'm not even sure as we move forward,
if that's even going to be completely

necessary, because it's probably
going to just, you know, somehow get.

You know, SEO juice and ranking and any
sort of index ability from like the actual

words in the YouTube video at this point,
but that's just another random thing.

Um, yeah, there's a, there's a lot
of trying to think of anything else

that I've, that I've tried to do.

I just literally hate making thumbnails.

So I'm waiting for a really
good thumbnail thing.

I've explored, explored some
GPTs, custom GPTs for that.

Um, but, uh, Yeah, I don't know.

This is all kind of like very
interesting and slightly kind

of concerning at the same time.

Cause like at some point, you know, I
mean, I don't know if you've, I don't

know if you've experimented with 11
labs and any sort of, uh, image or

not, not image, but like voice, like
you can very, very soon we're gonna be

able to just like make our own stuff.

Like, let's just generate our own
things just from all of the, the voice

data we have of ourselves, the video
data we have of ourselves, and then

all of the, any LLM stuff that we've.

Put together.

So it's going to, it
is going to get weird.

Um, I'm just hoping it doesn't
get weird super soon or we

could at least capitalize on it.

Um, it's a lot of stuff.

Matt: So are you still going to
build out that, that plugin and

continue like making plugins for
like sites that you're building out?

Like what's, what's the future
for developing plugins for you?

I

Mark: mean, so that one specifically
was, uh, I was trying to create a.

automated via the YouTube API,
just like automatically know

when I'm live on my website.

And I just created like custom
it's an, it checks the YouTube API.

It says, are you live or not?

It gives a couple pieces of
data that it pulls in, like the

YouTube URL and stuff like that.

And then it, it can, it integrates
specifically with bricks to

like use all those things with
dynamic tags and conditions.

Works great.

Perfect.

And in order to do all of that, I
would have either needed to hire a

developer or find a plugin that did it.

So I think for very niche things
like that, you're definitely

probably very well off doing that.

I'm thinking of if you do like an
analysis on like the ROI of doing

something like that, just even personal
ROI, like nothing's going to come

along and like fix that right away.

I would love to have a larger conversation
on like where this is all going.

Cause I don't like, I love WordPress
and I love websites, but I don't think

we're going to be on websites for
the rest of our lives necessarily.

I think there's going to be some
more to that just in like the

different ways that we operate
with things because to go back to.

If you think about like the Google
search era versus now the AI prompting

era, I don't know about you, but
there have been times very recently

where I have not gone to Google.

I've gone to like a chat GPT or a
Claude because I want I don't want

to go search for an answer anymore.

I want the answer to kind of like
come to me and I don't mean like

I'm just going to take the answer.

It's like it has all that information.

So we can explain it to you in a way that
you're actually asking particularly for.

I don't know.

I'm sure you've done this.

Like you go and you have
a very specific problem.

So then you have to search Google and
you have to go through like 10 links

to find like maybe what the answer is.

But with now, it's like it does
that step for you in a lot of ways.

And the other big thing that I do,
and this is one of the things I

don't like about cloud just yet, but
I'm sure they're going to have it.

I love chat GPT and Gemini, which is
Google's for the conversational, because

you know, I hate text and this is a
lot of reading that we're doing so far.

So like, I'll like actually have a
conversation with it and I'll be like,

Hey, what do you think about this topic?

Like I'll do this in preparation
for like, You know, any sort of

podcasting debate, like live topics,
like, hey, you know, like, what do

you think about, like, open source?

Do you think that's a good
thing, a bad thing, or whatever?

You can kind of, like, get a lot of
ideas and a lot of different things that

you would have had to go read, like, a
bunch of different people's opinions.

But now it, it's aggregated all
that and it's just thrown in at you.

I'm not saying just take it as face value.

I'm saying like you can almost do
like a little mock conversation or

mock debate, debate with these things.

And then you can do it
from both sides too.

And then you have all the perspectives and
then you can create your own perspective.

And then you could, you know,
go into the content or the, or

the show or whatever with that.

I mean, I've literally done that before
just because of the ease of that.

Like it's just, I'm just
talking to my phone.

Like it's, it's so easy.

Matt: Yeah.

Yeah.

Uh, it's, it's certainly.

Um, you know, some, it's many different
ways, you know, to, to use it.

Um, I was also thinking too, like while
I was using cursor, I'd be like, man, if

I could just talk to this thing, which
it probably even has, I don't, I don't

know if I could just like talk to this
and, and explain what I, what I want.

Uh, but I still have a lot of like
uncertainty on, you know, what,

what the answers are going to be.

Um, even from a, uh,
you know, I don't know.

I like to see a ton of options.

I don't mind sifting through it as
long as that the options I have are,

you know, Are are close to you know,
giving me the answer, you know, argue

if google even still does that anymore.

I don't know Um, but I don't mind like
sifting through that through that stuff,

but I totally get it like just give
me the answer I signed up for Uh only

because Kevin rose, uh from dignation
He has he's uh, I guess an advisor or

his the vc firm that he works for is
invested in perplexity And you could

get a free year Of Perplexity Pro,
uh, if you were on his email list.

Um, I've heard a lot of people talk
about it, like, oh, it's like, you

know, search results and all this stuff,
and like, that's supposed to think, be

the thing that really crushes Google.

You know, I did it for, for this
project that I'm working on.

I mean, it didn't really give
me anything that Claude and

ChatGPT wasn't already giving me.

And I mean, it's a cool interface,
and I'm sure it's like, if I

was doing more, it would give
me some, you know, better stuff.

But once again, I'm just like, I,
I, I'm, I'm not, I haven't tapped in

to the full potential of this stuff.

Uh, cause you know, effectively,
I feel like I'm just getting

started, but that's where that is.

Uh,

Mark: We're all just
getting started, Matt.

You know, it's kind of, it's early.

We're all just getting started

Matt: and we're all just on the
way out all at the same time.

Mark: You know, I, that
is a larger, larger topic.

Um, and I don't, I'm sure this has
already happened though, kind of in

history and waves, obviously not AI
related, but again, like with so many

other things, I think we're just going
to have to continue to, I think that.

If we're talking about this now, we are
perhaps potentially the best position.

So I'm trying to get more people, you
know, like just from this type of content,

if we can get more people think about this
and understanding that it's real, the hope

would be that we can continue to leverage
it for as much as we need to leverage it.

And then we could continue to evolve
and adapt to obviously fill in the

gaps that like this thing, you know,
these types of things are going to do,

like we could, there's always going
to be a reason to, um, continue to

try to innovate and do other things.

So, um, I don't know.

It's going to be interesting
though, regardless.

Yeah.

We'll see what, uh, we'll see what the
future holds with, with all of this.

Matt: It's the WP minute, the WP minute.

com slash subscribe.

Still not powered by, uh, AI yet.

Yet.

Yes, maybe someday, uh, hit
that, uh, hit that newsletter.

I've, um, I think I mentioned
this to you a couple of weeks ago.

I switched over to convert kit.

It's going to be kit.

Is it, has it already, excuse me?

Sorry.

It's just, it's just kit now.

That's right.

It's just kit, kit.

com.

Um, one, because this is a semi tangent
to what we're talking about today, but,

um, you know, everything that happened
with WordPress and, um, just my, I've

always had this feeling anyway, but
like everything, especially with like

WordPress and understanding, like who,
Where you're doing your business like

who owns that business that you're
doing business with Um, and yeah, I

was running on mailchimp for a while.

It was fine.

I never really loved like making the
newsletters there uh, but I switched

to uh kit because I Had nathan on
the show many many years ago and

he'll dm me every once in a while to
ask a couple questions And uh, I was

like, yeah, you know i'll use it.

We got a new add on for it for gravity
forms You Uh, and so far, so good.

Uh, consolidated two MailChimp accounts.

It was a little bit of a pain in the neck.

Uh, because it flagged me for like,
where did you get these users from?

And it's like, well, I merged
two, I had two different lists.

My mat report list and my WP minute list.

And I was always mailing both.

So I just consolidated into one
and it flagged a bunch of stuff.

Um, but crafting the templates there.

Much easier, uh, happier about that.

So if you do receive the email
and you have any feedback, hit

reply on the next email that I
send, let me know what you think.

Mark, what about you?

What are you, what are you up to?

Where can folks find you?

Mark: Um, just trying to double, triple,
quadruple down on content on YouTube.

You can just go to mjs.

bio.

You can sign up for my newsletter.

All my social links and everything like
that are there just trying to provide

as much value as possible to you guys.

So appreciate it.

Matt: Awesome stuff.

Thanks for watching everybody.

See you.

Or you'll hear us in the next episode.