Redemptive Marketing Podcast

Snap market.

Brandon Giella: Hey, Parker.

Parker Smith: Well, hey Brandon.

How are you?

Brandon Giella: I'm doing okay.

I, uh, grew some tomatoes and
squash this weekend with my

daughter, which was really cool.

Parker Smith: Wow, that's amazing.

Brandon Giella: Yeah, it was fun.

Uh, so anyway, we were just talking
about a, uh, AI paper that was

released from Anthropic recently.

This was about a month ago in late
March, where the researchers were.

Essentially trying to understand how an
LLMs brain, quote unquote, brain works.

And, um, they found that it works
somewhat like a human brain, that

it's like, you know, making certain
connections and, and um, basically

breaking down thoughts in order to, you
know, assess the next word or phrase.

And.

It's interesting because you have been
talking about the ways that we could

be publishing or using content either
generated from AI through our podcast

process, things like that, but that are
published on chain in order to create

attribution to solve the copyright
problem, to basically create a, uh.

If you will, like a record of things
that have been published that, that then

are used in training data for purposes
like royalties in the future, and.

It's very interesting.

I think that's where my brain stops.

So you have some Web3
blockchain background.

You, you work for a, a Web3 startup
back in the day and so your brain,

you know, kind of naturally fits
in this space 'cause you have

a lot of experience with it.

So tell me your ideas on that because
ever since you mentioned it, I've been

so fascinated and we haven't really
sat and talked about it too much.

And so our podcast is the way
we're actually like fleshing

this out in real time.

Um, yeah.

So what, what have you been
thinking about with like publishing,

uh, on the blockchain and Web3?

Using AI and using the kind of tools and
processes that we have at Snap Market.

Parker Smith: Yeah, so you
led with, with all the, the

Brandon Giella: I,

Parker Smith: but really there's a lot of
buzz to put it very modestly around ai,

Brandon Giella: mm-hmm.

Parker Smith: but most people
do perceive it as a black box.

There's concerns about
copyright infringement and,

plagiarism and things like that.

so, uh, when this article really,
it's a research paper by Anthropic

came up, uh, late March of 2025.

It's called Tracing the Thoughts
of a Large Language Model.

They have a good two and a half minute
overview video, and then they proceed

to give examples of poetry, math facts.

How does.

The brain of a large language
model, the intelligence of the

artificial kind process and answer.

And it's, it's fascinating because
the first thing they do is, um,

they, they define the black box as a
brain and less of a machine, which,

Brandon Giella: Hmm.

Parker Smith: um, the
trial and error, the.

The weighted logic.

Um, and by means of their research and
development, they're able to peek into

that black box and map the, the steps
of, uh, of a completion or a prompt

to the point where if they have a poem
and um, it's something to the effect

of there's a carrot and there's a.

Something that grabs it, that they could
see where it would wanna rhyme with

rabbit and they could actually fork
that logic and rhyme it with habit.

Um, and being able to manipulate even
that thought process along those lines.

Um, so this is timely with a theory
that I've had coming from the Web3

space, um, of, um, our friend Michael
Mercer is actually defined, so Web3.

Might be unfamiliar, but it,
it encompasses like blockchain.

Um, if you've heard of
like Bitcoin, Ethereum,

Brandon Giella: Crypto.

Mm-hmm.

Parker Smith: Those are all onchain
So, um, Web2 or kind of the, if you're

to think in that paradigm, would
be read and write on the internet.

all So all that we've been doing,
you can read a blog, you can write

a blog that, that sort of thing.

Web3 is read, write, and own.

The ability to own something where you
have a crypto wallet and you have a wallet

ID that's unique to you, you own the keys.

Um, you can own things like NFTs, uh,
like JPEGs and monkeys, sorts of things.

You can pay however much
money you want for those.

Um, but.

It starts becoming useful when
those things can be monetized.

Brandon Giella: Mm-hmm.

Parker Smith: you've, you see,
I haven't looked at it in a

while, but like mirror xyz,

Brandon Giella: Mm-hmm.

Parker Smith: a blog post on chain.

It doesn't have to be a jpeg.

You can publish knowledge, content,
authorship onchain to the point where

you own it and your subscribers pay into
it, and then they own shares in that.

It's this owner economy type type thing.

Um, it actually, uh, it's, it's not
totally one-to-one, but the, the movie

sinners that just came out, um, was
very much like a crowdsource, uh,

like actor ownership, type model.

Um, and, and I think that's, that's going
to be a changing shift in general, just.

ownership across, across the board.

So if we're considering publishing,
um, then connecting these dots here.

I know we've, we've circled around a
couple things, but if you're able to

trace the thoughts of an LLM to the point
where you can attribute the sources of

those thoughts via a Web3 on chain token,
id, then you can start proceduralizing.

Royalties, citations, ownership, um,
the, the vision that I see is like a

Spotify for, for content, if you're
an author, all of a sudden, if you're,

if you're content is surfacing in
these thought patterns of an LLM, you

start getting, you know, fractional.

Percentage royalties and the
same way that musicians do.

You know, you stream one
time, it's like 0 cent.

You know, it might look something
like that it might be millions of

attributions across a single completion.

But over time and, and given the quality
of content, you can imagine a scenario

where you subscribe to chat GPT or
whatever, anthropic, and you do your work.

In the background, there's a, a
royalty layer to the internet to one

Brandon Giella: All right.

Parker Smith: So that's, that's
the, the full circle thought there

on, on where this might be going.

Um, and the fact that, that we're
able to trace that line of thought,

is kind of that first step in my mind.

Brandon Giella: I think this
is so fascinating because.

Well, there's like, there's so many
different problems that are created

out of the solutions of AI tools,
and one of those is in publishing.

So this is, this has been very common,
but, so you have a, you know, New York

Times, for example, I think they're
suing, um, Google at the moment

because, uh, their AI generated.

Search results are bypassing
people actually clicking on

the New York Times article.

So where Google search results would
actually pull up that New York Times

article that people could click on and
give a little, like, you know, um, give

like the meta description of that page
and people would click there and go.

New York Times, obviously run by
advertising, gets paid for those

clicks, those people to view their page.

But if Google is just summarizing
those pages and then citing it,

but it's, it's giving enough
information where people don't feel

the need to click into the article.

The, the New York Times is losing
tremendous amounts of money.

This is happening across like.

Mommy blogs and food blogs and recipes
and, you know, all this, these other

kind of like publishing, um, uh,
spaces, publishers out there, you

know, because web two made it to where
anybody could publish material anywhere.

And so a lot of people have built
their lives off of this one.

One solution for that was
something like a substack.

Where you could have a direct
relationship with your subscribers and

you could, they could pay, you know, $5
a month or something to see your work.

But again, AI can disrupt that.

Where your work could be published
online, be scraped by an AI training

model, and then it gets surfaced in
their results such that people don't

need to actually click into the article.

So, uh, what's fascinating about this
is that same material you could publish,

but you publish it on chain and that
goes into the model and every time

it's cited or summarized or pulled into
some kind of search results from one of

these AI tools, you get paid for that.

Immediately, directly.

That's the idea.

That's the idea.

The second thing that I find even
more interesting than that, um,

though as a publisher, as a writer,
you know, I, I co-founded a magazine

once and have tried my hand at the
monetized newsletter writing and

all that kinda stuff, which I love.

I think that that direct relationship
with your readers is fantastic.

But what I love even more
is, uh, the idea that.

Um, where you have like, AI can generate
its own thoughts and, um, can summarize

and plagiarize and use others' works.

Um, like a deep fake, for example.

So you, you have a video of somebody
that is not actually real, um, because,

uh, it's just a made up video that
pulled together a bunch of images

and speech and it created something
new that doesn't actually exist.

But it looks like that person,
it sounds like that person.

That's a massive.

Problem in the, in the future?

Probably.

Now, I mean, I, I've heard some
scams related to stuff like

that that are actually really
sophisticated and impressive.

But what is cool is if you can publish
all of your material onchain you can.

Uh, you can create a digital permanent
record to say, I was the first person

to say this, or write this, or have
that idea, or put these connections

together, and I can prove that because
here's my record onchain that said

that I published this and this was the
first, you know, de demonstrable case

of this phrase or this idea being used.

Such that when you see something
in the future that was generated or

summarized or a deep fake or something
like that, you can actually point to

the, the provenance or like the, the
original authorship of that material.

And I find that really interesting.

Um, and it just solves so
many different problems.

Parker Smith: In the fact that
you're describing it, modally.

been talking about it,
plagiarism, you think in terms of.

Text, you think

Brandon Giella: Yeah.

Parker Smith: essay or, or whatever, but
multimodal where you can put anything

in, get anything out, that's where
it's gonna get really fascinating.

Um, where and, and really
complicated ethically.

So

Brandon Giella: Yeah.

Parker Smith: visually, if somebody's
doing a deep fake of, of us, like where's

the on chain record of our likeness?

Like.

Do we upload our, you know, 360 portfolio,

Brandon Giella: Yeah.

Yeah.

Parker Smith: like on chain like likeness

distributor.

Like that's, that's where it gets
really weird and kind of like

Blade Runner But, there, there is a
route that we see in terms of, um,

that, that attribution and royalty.

So,

Brandon Giella: Yeah.

Parker Smith: so really.

Where we're thinking is
like future proofing this

Brandon Giella: Mm-hmm.

Parker Smith: do we post on chain now

Brandon Giella: Mm-hmm.

Parker Smith: because you know
that once it becomes popular and

once people understand and can get
paid, people are gonna be scraping

everything and claiming it as their own

Brandon Giella: Mm-hmm.

Parker Smith: and publishing it on chain.

So if you start publishing on chain now

Brandon Giella: Mm-hmm.

Mm-hmm.

Parker Smith: um.

You have that claim of ownership.

So it's, it's gonna be a wild ride if,
if it, if things do progress to that.

but

Brandon Giella: It's, it's to stake
your claim in the internet, you know,

as all of this sense of truth and
reality gets so warped and distorted.

And then there's the economic and business
models that get distorted as well.

Like, like Parker kind of led with, I.

But I like the, I like the, the
truth and and reality angle where

like I can stamp and say, I said
this, this is when I said that.

This is how I said that.

I said it on this platform, on this date.

And anybody that pulls from that.

Phrasing or that thinking or
whatever, like they will know

that, that that was my claim.

You know, I don't know if that's
like the most modest way of like

thinking about these things, but
I just think like as a writer, as

a thinker, that stuff is your ip.

That is your livelihood is to
say, I had this thought first.

And I phrased it this way.

And so as, as like books are gonna be
coming out that are like mirroring,

you know, mark Twain sound like
him, you know, write like him, uh,

has the same kind of characters and
setting and all of that kind of stuff.

Like I.

The Mark Twain estate now
will start to lose, you know,

money or something like that.

So like, it's important for Mark
Twain and his, and, and that estate

to like say that this is the work
that we, that Mark Twain has created.

This is his ip.

And similarly for, for today's creators,
that's, that's the case, you know.

Parker Smith: like a couple
weeks ago, studio Ghibli

would've made a bajillion dollars

Brandon Giella: Right.

A per great example.

Yes, exactly right.

Exactly right.

Parker Smith: OpenAI released their
new SOA image generation model powered

by OpenAI Image one API, um, probably
just by soa, but they, they released

that in tandem, which was super cool.

Um, and everybody was just studio Gib.

Everything, you know, like do Lord
of the Rings in studio, Gib leave

Brandon Giella: Mm-hmm.

Parker Smith: Um, that of course
if there was that, that logic and

attribution, now it's like free
marketing probably for Studio Gibble.

Brandon Giella: Yeah.

Right.

That's fair.

That's fair.

Parker Smith: um, you know, being
able to attribute that would be great.

Brandon Giella: But I,
yeah, well, I do see it.

Yeah, I do think that you see
some creators catching onto this.

So if you go to, um, abc.com,

which is a very, very famous like, uh,
Silicon Valley Tech, um, blog, um, I

don't even know the guy's name, actually.

I just know him as a c.

Um, but he moved all of his writing
over to Paragraph, which is an on chain

newsletter publishing, uh, platform.

And so it's like Substack,
but it's on chain.

Um, and then he's written about
Mirror, which you mentioned earlier.

Mirror XYZ is an on chain blog format.

So there's the on chain blog format,
on chain newsletter, uh, platform.

And then there's gonna be others.

And I, we haven't like researched a ton,
but what we're trying to think through

is like, how can we have these different
channels and media that we create?

And put them all on chain.

So we have like a, a web two process
and we have a Web3 process for folks

that wanna start putting all their
content and, and media and brand on,

on chain, which I think would be cool.

Parker Smith: So the last I, I
think that's, that's exactly right.

The last consideration, because
I know some people when they

think of this conversation will.

Go to

Brandon Giella: Uh,

Parker Smith: uh, environmental impact.

Brandon Giella: yeah.

Mm-hmm.

Parker Smith: AI and Web3 are
infamous for their energy consumption,

Brandon Giella: Mm-hmm.

Parker Smith: just a modest completion
could power a, you know, small or

Brandon Giella: I read, I read last week.

I read last week that you saying
please and thank you to ChatGPT costs

OpenAI Millions of dollars a year.

Isn't that crazy?

Parker Smith: to process.

Yeah.

Brandon Giella: I,

remember posting

Parker Smith: that on LinkedIn where I
was like, but nobody's trying to piss

off the AI we wanna be on their good

side, We're

Brandon Giella: so polite.

We're just costing
millions of dollars to ai.

So yeah, if you want to see
OpenAI completely collapse,

just be really polite.

Parker Smith: Yeah.

Well, so the thought in in this too
is it's going to have to also pair a,

along with aligning these two giant
tech paradigms of AI and Web3, you

also have to, you also have to align
the third leg of the stool, which is.

Energy efficiency.

you see your, your metas and your Amazons
of the world investing heavily in nuclear

energy and trying to get ahead, in, in
those costs because Lord knows how much

the, the server cost is for OpenAI and I'm

Brandon Giella: Seriously.

Parker Smith: probably
benefiting from that, or

Brandon Giella: Mm-hmm.

Parker Smith: you know,
owns, owns those servers.

But that's the.

Those are, that's the three
legged stool I think is AI

Web3 and, and energy efficiency

Brandon Giella: Hmm,

Parker Smith: us to this next level.

Brandon Giella: hmm.

Parker Smith: in the meantime,
uh, exploring, publishing on

chain, I think is maybe the next
step from a marketing standpoint.

Brandon Giella: And I think that's a
challenge for us too, if we continue to

explore the, the on chain route is how,
how do we take care of God's creation

by, you know, like, 'cause yeah, that
would be really resource intensive.

So it would have to do some.

We would have to do some kind of
legwork or research on finding some

kind of, there's like maybe there's
like the Web3 publishing process,

but then trying to find vendors or
platforms that are the most energy

efficient or renewable or something.

I mean, it, it would be an
interesting concept that,

again, I think these are pretty.

If, if I dare say innovative ideas
that I don't think a lot of people have

answers for, um, but I'm with you and
ever since you've said it, I, it's,

it's caught my attention where I think
that is a very realistic future where

you have to publish on chain for you to
survive as a business, as a publisher,

as a creator, um, because of, of the.

Of AI eating everything.

So, yeah, it's kind of fascinating.

Parker Smith: we know it, are
likely going to evolve and change

where at least search input boxes

Brandon Giella: Yeah,

Parker Smith: in the same
way that we expect So.

Brandon Giella: yeah, yeah.

It's, it is weird.

The next 10 years are gonna be really
radical and I think, I think Web3 is,

uh, an answer to how to do that well.

So we'll see.

We'll see.

Well, anyways, thanks for, uh.

Thanks for having that idea because
I've been fascinated by it ever since

and I'm glad we get to talk about it
today 'cause um, I think there's more

here and I want to keep exploring it.

So pay attention to Snap Market
V four in the coming years

where we're gonna be on chain.

We'll see you soon.

Parker Smith: Awesome.

Brandon Giella: All right.

See ya.

Parker Smith: So yeah.