May 26, 2025 • 38 mins
AI isn’t just for tech experts or big-budget teams. In this episode, visionary entrepreneur and AI mentor Joshua Hale shares how small business owners can start using tools like ChatGPT and others to cut down on repetitive tasks, speed up research, and build smarter systems without needing a technical background.
We talk about practical ways to apply AI in everyday work, how to improve results with better prompts, and why Joshua’s 10-80-10 approach (human idea > AI execution > human refinement) helps keep your voice and values intact. You’ll also hear how different tools serve specific needs like Claude for writing, Perplexity for research, and Gamma for fast presentations, and how his 30-day AI challenge can help you build real momentum, one small step at a time.
- Show Notes
- Apply to be featured on My Weekly Marketing!
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
Janice Hostager (00:04):
I'm Janice Hostager.
After three decades in themarketing business and many
years of being an entrepreneur,I've learned a thing or two
about marketing.
Join me as we talk aboutmarketing, small business and
life in between.
Welcome to My Weekly Marketing.
If you're not using AI to helpyou in your business, you're
(00:33):
really missing out on somevaluable tools to save you time
and money.
I know for me, even though Ihaven't had a ton of time to
play with ChatGPT, it's alreadycompletely changing how I do
everything from marketing today-to-day business operations.
For both me and my team, it'swritten code for my website in
seconds.
That would have taken aprogrammer hours to do.
I use it daily to give me ideasfor subject lines, titles and
(00:57):
starting points for emails.
Most recently, I've been usingit as I create a new course,
because it's easy for me tooverload my students with too
much information.
What can I say?
I get kind of excited aboutmarketing or maybe forget to
include something important.
I've been using it to getfeedback on my outlines and what
I should and should not include, and, because I'm an external
(01:18):
processor.
It's also become a virtualsounding board for me to get
feedback on ideas and thoughtsthat I have about the course or
about business in general.
I know that sounds super weird,but it really does help,
especially if you're workingalone.
You've probably also noticedthat most software is
integrating with some version ofAI.
As a former art director, Istill use Photoshop now and
(01:41):
again, and it does a fantasticjob of filling in background or
even eliminating objects thatyou don't want in a photo.
This type of editing used totake hours to do.
Now it's done at the click of abutton.
So I could go on and on, butinstead I want to introduce my
guest today, Joshua Hale, whocan tell you way more than I
(02:02):
ever could about AI.
Joshua is not your average techguy.
He's a visionary entrepreneurand an AI mentor.
It's his goal to revolutionizehow small businesses leverage AI
technology, and he wants to seeit accessible to everyone.
He's learned how to empowersmall businesses to compete with
large corporations through thepower of AI and help multiple
(02:26):
clients hit their first millionin revenue using AI strategies.
He's going to give us someactionable steps on how
businesses can leverage AI toolswithout technical expertise or
a massive budget, while keepingthe human touch that makes your
services special and is, ifthat's not enough, you'll want
to stick around to the end,because he's going to share the
(02:47):
way that you can use AI in yourdaily business and personal life
through a challenge that he'soffering.
So here's my talk with Joshua.
Hey, Joshua, welcome to MyWeekly Marketing.
Joshua Hale (02:58):
I so appreciate having you have asked me to be
on here.
I'm really excited abouttoday's discussion.
Janice Hostager (03:09):
I am too.
I feel like everybody and theirbrother is talking about AI
these days.
So I really want to get intothis and kind of hear your take
on it, because, well, I'll letyou tell you about your own
story.
But in the intro I just talkeda little bit about how AI has
changed my business in a veryshort time and it's bought me so
much time in my day.
But AI isn't really new, right,it's kind of new to most of us
(03:33):
because we started with ChatGPT,but it's been around for a
while.
So tell me how you gotinterested in it and how you've
made a career out of this,because it's really interesting
to me.
Joshua Hale (03:43):
Yeah, AI has been around for a long time, right,
like 50, 60 years perhaps, andit's just this algorithm that's
been continuously been improvedupon over time.
The AI, the term, is reallybroad in scope, right.
But just in the recent uh,three years, right, large
language models came out.
And funny story to start us off, in 2016, I went up to a client
(04:07):
meeting downtown Seattle, uptop of the sky rise, and there
was two developers at thismeeting and during one of our
breaks they were over in thecorner just cracking up laughing
and I'm like what are you guyslaughing at?
And they're like we're lookingat two chat bots talk to each
other and I was like what?
(04:28):
I looked over their shoulderand it was incoherent, like none
of it made sense, and they werejust cracking up, cracking up
and I was like this is thedumbest thing ever.
Why would you just sit thereand watch this?
Little did I know, like whatwas coming.
So I was running a marketingagency when I first saw ChatGPT
(04:49):
for the first time, I went outto lunch with a fellow marketer
here in town and he showed mekind of under the table.
He's like have you seen thisthing?
And he's like you can just likewrite blogs with it and
immediately, you know, as amarketer, I was like impressed
beyond a belief and so I camehome, signed up for it.
Back then it was super hard toeven sign up for it and then I
(05:09):
had a dozen or so clients and soI was able to like jump into
their industries and starttesting it out and I immediately
saw the strengths.
You know some of the weaknessesbut these large language models,
they were trained off of thewhole internet, and so one thing
I saw that it really got wellwas business right out of the
gate, just because there's somuch information out there on
(05:33):
how to run a business andmarketing frameworks and methods
.
And so we actually onboarded alarge client.
I had three of my contractors,my copywriters, go and do
audience research to learn aboutthe foundation of this new
client.
They took three weeks.
They did a really good job.
But when they came back topresent this information to the
(05:55):
client, I did my own audienceresearch during the meeting off
to the side and showed one ofthem afterwards.
And they were, they were, theywere pretty impressed.
They were like, wow, we just,you know, took three weeks to do
this, and so I quickly saw thewriting on the wall, right, and
I essentially jumped ship fromthat marketing agency that I was
running and went to go helppeople out with my with, uh,
(06:20):
with just my trusted companionnow, uh ChatGPT, because I could
do like 80% of the work.
I still had contractors, so whenI needed heavy lifting and
final polishing and things likethat, I was able to call people
in.
But a lot of the stuff I wasable to turn those marketing
frameworks just into prompts.
Some of my favorite gurus backthere.
(06:41):
I was able to turn those booksinto prompts.
So what I ended up doing wasjust walking through a business
owner or a client through theirbusiness, interviewing them
using voice to text and justadding what their innate wisdom
and experiences and feeding itinto these big prompts and I was
(07:01):
able to quickly do what mymarketing agency was, you know,
taking thousands of dollars toaccomplish and, you know, six
weeks.
Janice Hostager (07:11):
I mean, that's just amazing, and you're
absolutely right.
I think the biggest aha momentI've had is that I can't believe
I've been doing this allmanually for so long, and this
is really going to changeeverything.
This is going to changeeverything.
So, backing up a little bit,though, can you explain how AI,
how the, like any AI tools itdoesn't necessarily have to be
(07:32):
ChatGPT or anything how they canbenefit businesses that maybe
they're not especially techie,like where do you start with all
this?
Because it's sort of agroundswell of change that's
happening right now.
Joshua Hale (07:44):
Yeah, well, the easiest, you know, the lowest
hanging fruit is likeadministration and like
management and just helping themlearn these tools, ChatGPT is
like the front runner, and justlearning how to buy back, you
know three, four hours a week oftheir time with helping with
emails, drafting up proposals,you know helping with sensitive
(08:08):
communication, going back andforth, and we're able to just
quickly buy back their time.
So, as a whole I you know, ifyou combine all that time saved
then you're able to really startseeing the numbers of
profitability just from the team.
So typically a business owner,you know they are too busy and
(08:32):
they roll their eyes if you askthem to.
You know, download another app,like, even if you're like, yeah
, you got to try this right,because they're busy taking all
of the other tasks that abusiness owner has to do.
But if you're able to just helpout the management and the team
drastically, you'll see notonly time saving but a higher
(08:54):
quality result in just aboutlike everything they're doing.
You know, being able to utilizethis tool.
Janice Hostager (09:02):
Yeah, and I find it really interesting
because I think we start out andthat's how I started out too
just, you know, getting ideasfor emails, not, you know, and I
certainly go in and rewrite itand I put in my own personal
take on things.
So I don't just like turneverything over to ChatGPT, but
just doing the small tasks thatwould take up time and have my
(09:23):
team use it as well.
They, you know, use it forsocial media and other places.
But I've started to reallyunderstand that this is so much
bigger in terms of its potential.
Are we thinking too small aboutas businesses?
It's hard to wrap your mindaround every potential thing it
could possibly do.
Joshua Hale (09:45):
Yeah, Well, especially since ChatGPT Pro
came out, um, I, I bit thebullet this was about two months
ago and tried the the $200, uhtier of ChatGPT, and at first I
was like, oh, I don't think Ihave like problems big enough
for that, you know, it must bejust for mathematicians and
scientists trying to, you know,solve the world's problems.
(10:07):
But I then, you know, tried itout and was immediately
impressed with not only likejust a higher quality of model,
it was.
But then the deep research cameout as well, and what I was
able to do, you you know,previously helping businesses,
um, going through the marketingframeworks of, uh, deep dive
(10:29):
audience research, understandingtheir, their pain points, you
know, understanding whatchallenges they're they're
having, what obstacles are infront of them to get them to the
dream outcome that they'retrying to, you know, get to, and
they can't get to it themselves, otherwise they wouldn't be
willing to pay, you know to, andthey can't get to it themselves
, otherwise they wouldn't bewilling to pay, you know you for
it.
So getting a deep dive of thatand then crafting your offer to
(10:52):
speak directly to them is ispart of that aspect.
And then, once you have youroffer, then it's about marketing
that offer or, you know,getting that message out there
so your target audience can see.
That whole thing was what mymarketing agency was doing.
We can now do that in like twohours and you end up with a 50
page document what I call aliving document.
(11:16):
And what's so cool about this isChatGPT Pro came out and you
can now do a deep dive intoindustry trends.
You know where things have been, where things are going,
competitor analysis you canreally get real-time data, which
(11:37):
there were firms that weredoing this.
You know, for the last 30 yearsMcKinsey has been around, been
around, uh, and they cost$500,000 to like start working
with them.
So trying to get you know areport for them is has never
been really capable for a lot ofsmall businesses and you know,
uh, mom and pop shops.
(11:58):
So for the first time, uh,ChatGPT is bringing this real
world data in in order for themto make act, you know, informed
decisions about their business.
On The coolest thing of all thatright that then turns it into a
hundred page document.
No business owner is gonna readthat, right.
Who's got time for a hundredpage document?
(12:18):
The coolest thing about allthis is you then just upload
that document to ChatGPT and youcan talk to it.
You can ask it for emailmarketing campaigns.
You can ask it for a strategyfor this upcoming quarter.
That now becomes like your 24-7consultant that knows
everything about your businessand is basically a trained model
(12:42):
specifically for your business.
Janice Hostager (12:45):
And I absolutely love that about
ChatGPT is that it can rememberthat you can plug all that
information in and it willremember it.
And you know, because I havedifferent ChatGPTs open for
different clients, I keep themseparate and it's amazing that
they, you know, I type aquestion in and it will give me
an answer based on who my idealcustomer is, based on what that
(13:10):
business is all about, whattheir goals are.
It is so incredibly powerful Ireally kind of feel like I'm
cheating.
You know, it's like it's almost.
Initially it made me feelreally kind of uneasy, like what
if this is not right, you know,or some of it, especially with
audience development, becauseyou just you know you're really
at their mercy, becausesometimes it does make things up
(13:32):
.
But which brings me to my nextquestion.
Is that what do businesses needto watch out for when using
ChatGPT or any AI program?
And because I know we've talkedabout well, we haven't talked
about it but things likehallucinations come up and
there's some other pitfalls, andare those getting better?
Are we seeing less and less ofthat and are there things that
(13:54):
we should be watching out for?
Joshua Hale (13:56):
Yeah, hallucinations were a real big
problem.
The first year that ChatGPT wasout.
A lot of journalists tried itout and immediately they're like
what?
They tried to publish somethingand something in there was
something that made up.
Now these models.
Their number one job is to likegive you what you're look, what
you're asking for and and backthen, like it would cut corners
(14:20):
and just give you reallyplausible sounding answers.
And that's what thesehallucinations were deemed and
they are designing thosehallucinations out of the model.
And when it first came out itwasn't connected to the internet
.
It had a data set thattypically is like a year or a
(14:42):
year and a half old, that it wastrained on, so it didn't have
any current events or likecurrent knowledge, and so that
it and if you asked it aboutthose things, it wasn't always
honest.
It would sometimes just makestuff up.
So it was pretty easy to likecall it out.
But since it was able to getconnected to the internet, now
you can, you know, I just sayyou know, do some research and,
(15:05):
and sure you know, give me theanswer to this.
That then kicks on the internetand it will go out and search
itself.
Then the reasoning models cameout.
Those reasoning models are,instead of it just giving an
immediate answer, likeeverybody's used to with these
AI models, it then stops and,like, thinks about it and will
(15:26):
go gather information and willlook at it from a hundred
different angles and be able to,like, piece it together through
chain of thought reasoning toget like a really high quality
answer.
So the amount of hallucinationsthat are happening is just
becoming less and less.
And especially, you know, itdoesn't mean that you should
(15:48):
trust these things like 100%.
It just means that, uh, youknow you can give it the benefit
of the doubt more often.
But if you ever catch anythinga little bit off, you can tell
it to reveal your sources.
For that I use a lot ofperplexity to go out and just
(16:08):
grab raw data and informationfrom research papers and then
you can go to Notebook LM andupload that raw data into
Notebook LM.
What's good about that is it isa closed system so it doesn't
have access to the internet andit will just process the the
like high quality informationthat you gave it, so you can
(16:30):
distill it down into usefulparts and then, you know, bring
that into ChatGPT to then becreative with and come up with
like good looking reports.
That whole process reallyminimizes the like possibility
of hallucinations taking place.
And so I've come to you know,eight out of nine or eight out
(16:52):
of 10 times, like I'm like,don't even question it.
But if I ever see something off, I'll just say, like you know,
don't even question it, uh, butif I ever see something off,
I'll just say, like you know, uh, do some research and give me a
better answer, and it will thengo to the internet and, you
know, give you something morehigh quality.
Janice Hostager (17:07):
Yeah, that's what actually I've been doing
too is just asking for itssources, or if it feels off, you
know, I'll even ask itsometimes, like is are you sure
this is right, you know, andit's.
It's interesting that it can.
You know it has self-correct,like that.
Joshua Hale (17:23):
It's really useful in the hands of experts or like
the business owner, right.
So that whole living documentthat I do, I do have them like
scan through it and as we, as webuild that document, they are
there to course correct, becauseyou don't want something wrong
up here because then downstreamyou know we'll be out in the
weeds.
So that's why I do have thebusiness owner there, because
(17:45):
they're the expert in theirfield.
So it does keep things on track.
Janice Hostager (17:49):
Right, right.
So are there other tools thatyou use?
I know there's some that aremore specific to images or
creating slides or diagramming.
I mean, are there somefavorites that you have out
there?
Joshua Hale (18:02):
Yeah, definitely I still I can't get, I can't let
go of Claude.
Claude is, I think, still thebest in writing.
It just has been able to get myvoice down like to just such a
degree that ChatGPT has gottenclose with the latest uh 4.5
release, but it still justhasn't taken the king off the
(18:25):
hill, uh.
So I I typically willbrainstorm in ChatGPT and then
bring it over into Claude for afinal polish for my writing and
copywriting uh.
And then there is uh perplexity, like said, that's kind of
where I go for research.
Now the thing with ChatGPTthey're the biggest and by far
(18:46):
the most used AI out there andthe term itself is becoming
quickly synonymous with AI,right, and so because they're so
big, they get to just like buythe best features of their
competition and just absorb itinto ChatGPT.
So that's where you know it'sstill my go-to and my most used
(19:12):
AI model.
Then there's Notebook LM that Imentioned and that's really
good for just keeping like cleaninformation that you can bring
in and isolate kind of in itsown folders or like projects to
where you're just able toprocess the information that's
in there.
(19:33):
So that's a really useful toolbecause none of the other models
do that.
They inject their own opinionand bias and, you know, like
information from the internet.
Janice Hostager (19:43):
Gotcha and then you can have a ChatGPT can
reference the information that'sin Notebook LM.
Joshua Hale (19:50):
I would like distill that down into like the
solid chunks and then bring thatinto ChatGPT to then brainstorm
and do
Janice Hostager (19:57):
Gotcha Gotcha.
Joshua Hale (19:59):
Outside of those you know before.
Two months ago ChatGPT's imagegenerator was awful compared to
where the rest of the standardof image generation, which over
the last two years went from youknow 14 fingers to where.
Now it's prettyindistinguishable from being
(20:21):
able to guess where an AI modelversus a regular image.
So back to ChatGPT.
They updated their model andnow I use that all the time for
image generation.
Sora is their video generatorthat they have, and if you go to
(20:41):
Soracom, it's a sub-brand ofOpenAI's ChatGPT.
But in Sora, if you have a paidaccount, you can go and do four
images at a time instead ofjust the one that can be done
inside of ChatGPT.
So I spend a lot of time therejust having that open, and when
(21:03):
I need, uh, graphics or orplaying around with branding
things like that, then, um,that's, that's my go-to for that
.
Presentations, Gamma isincredible.
Gammaapp uh, it will create apresentation in like two minutes
compared, or like a, PowerPointslideshow in like two minutes.
(21:23):
Um, and those you I'm sure youknow like you should take hours
to, like you know, just puttogether, and then there's
polishing of it and practicing.
So now you can.
You can outline yourpresentation in Chat GPT.
Get the get the you know, getthe overarching story of it and
then have it as much detail asyou want, throw it into Gamma
(21:47):
and it will create a nicecompact slideshow bullet
pointing all of the topics foreach slide.
Then I typically ask for justspeaker notes from ChatGPT and
so then I can have the speakernotes up and then just go
through the slides.
All that can take like 15minutes now, compared to the
massive hours.
Janice Hostager (22:06):
Hours and hours .
Yeah, yeah, I just set up myGamma account this afternoon and
Napkin, I think, is another onethat I just set up.
I haven't played with eitherone of them yet, but you
answered a question about theslides.
I was wondering if there werespeaker notes that we could add
into.
So I mean, yeah, it feels likewe're cheating.
I don't know, I can't get pastthat.
(22:26):
Like I feel like, oh, I don'twant to tell anybody, I'm doing
this because it's so fast andquick.
Like I feel like I'm doingsomething wrong.
Joshua Hale (22:34):
But it's all about the quality and really what AI
can't do is like come up withthat initial spark of
inspiration and like startingthe project.
So I tell my clients you knowthere's a formula to follow 10,
80, 10.
10 is your initial idea, the80% is now the power of AI and
(22:56):
it can just crunch through these, this creation process, and
then you always want to leave 10, 20% of you coming up
afterwards to you know, clean itup, turn it into your own words
.
You know making sure thateverything is accurate and feels
true to you, so that it thatwhole process you know what used
(23:17):
to take four hours you can getdone now in you know, a half
hour.
And it's still yours.
It was still your original idea, it was still.
You know you're the last one totouch it.
Janice Hostager (23:29):
And yeah, I did notice that really it really
matters what kind of questionsyou're asking it to and what
you're bringing to the tableinitially, not just in terms of
prompts, because although thoseare important as well, but
really the depth and theparameters that you set around
the questions that you're askingcan make a huge difference in
(23:49):
the output.
Joshua Hale (23:51):
Yeah, it's one of the first tools that teaches you
how to use it that man has everinvented, and that's what makes
it so powerful If I teach thisin my 30 day challenge, but, uh,
I'll share it with youraudience because it's, it seems
like a no brainer once you knowit.
But for any any good uminstructions that you're giving
(24:11):
it, you can ask it for.
Um, you know, add, add this tothose instructions, Ask me three
clarifying questions before youbegin, and so that little
addition will prompt it to askreally specific questions, to
get more context out of you thatyou might not have thought
about or you weren't thinking toarticulate.
(24:35):
And it just creates such ahyper result of personalization
just based, just based off ofthat little addition so you ask
it.
Janice Hostager (24:45):
What you're missing basically is like what,
what else should I be asking youabout this topic?
Joshua Hale (24:49):
yeah, three qualifying questions and it will
then have far exceed theresults just because of what
you're able to give it extra.
Janice Hostager (25:00):
So how important is, like data privacy
and security, when we'reintegrating AI tools into our
businesses, and is there somesteps that we should be taking
to ensure that everything'scompliant?
Especially, I'm thinking aboutclients that are in the medical
or insurance or financial fields.
Is there something that theyshould be doing differently?
Joshua Hale (25:21):
Yeah.
So this is a big topic and oneof the biggest concerns that
business owners have Now, rightout of the gate, you know,
ChatGPT has an option in it thatyou can turn off using your
data to help train its model.
And then there's an enterpriselevel.
That's higher up, where that'slike automatically default.
That's higher up where that'slike automatically default.
(25:44):
And so, as far as you can trustthese companies, right, they're
promising not to use your data,that your data is just staying,
you know, hidden and closed foryour use.
On that.
But, as we've learned through,you know, countless mistakes of
data breaches, right, you can'treally trust, we can't trust our
(26:06):
governments, we can't trustthese big corporations to like,
completely keep this informationlocked up.
Now, if you tell it, you know agood idea you have.
Your idea is not going to justmagically pop up over in, like
Germany on someone else's feed,like it's all ones and zeros.
So it takes your input and ifyou do allow it to train the
(26:29):
model, like it's shredding thoseletters, those words into the
most basic forms in order totrain itself and so you'll never
see those things ever again.
But on a bigger view of this,right, I think we all need to
have our own localized AIsystems that don't send their
(26:54):
data up to a cloud, and the opensource community is really
catching up quickly with thesecurrent large language models,
and so I think by the end of theyear it's going to be a pretty
easy process.
Nvidia just came out with likea supercomputer for like $3,000
that can run these models, so asa business, you can literally
(27:15):
have like a station where themodel runs and your information
is not going anywhere, like thatis your property, not leaving
the server there.
So I think that is the only waywhere we're able to really feel
like we own our data and areable to keep it safe.
Same with our phones.
I think that's the only wayforward is everybody having
(27:38):
their own open source localizedAI systems, because these big
corporations are working withthe governments and we really
don't know what they're doingwith the data.
The current administration justrolled out the largest AI
infrastructure known to man andthere's a lot of concerns over
(27:59):
where our information is going,how it's being used, how the
information is going to be usedback to us, because we've been
trailing data for the last 20years on the internet and we've
really become careless of howmuch information we're just
trailing behind us each day, andthat wasn't that big of an
issue.
Sure, Facebook was using it forretargeting and things like
(28:22):
that, but not until recently,not until these large language
models is there processing powerto actually use all that data
that we've been trailing behindus for our lifetime.
And so I'm more concerned aboutmy kids and setting them up,
because in five years, 10 years,that's where things are going
(28:44):
to be really, uh likedetrimental to have private data
and privacy, um more so than itis right now, because it's kind
of way out west, uh how it howit is.
But those open source modelsare going to be, you know,
crucial for maintaining privacyin the future yeah, that's kind
of how it is right.
Janice Hostager (29:06):
You lead with technology and ethics and those
types of considerations usuallyfollow behind, right?
So, like you mentioned yourkids, I think that's interesting
.
My husband's a collegeprofessor, so he's already right
away.
He saw that students were usingAI to write papers and so on.
(29:27):
And how do we work with ourkids?
Like, how do you talk to yourchildren about how to use AI?
You know how is it?
I'm sure it has to be a tool,but down the road they have to
learn to think for themselves aswell, right?
Joshua Hale (29:45):
Absolutely.
Yeah, I have a pretty uniquesituation.
So I live on 10 acres, a farmout in the country, I work all
day in tech, but then I, youknow, get off work and go help
my wife out in the garden.
My child she goes to natureschool.
Right now she's seven and soshe is learning incredibly fast,
(30:06):
and I've known how quickly AIis developing, that it's going
to be the perfect tutor for mydeveloping child, and so all the
knowledge of the world isessentially made available to
her.
And not only that AI is goingto be the most patient, the most
understanding, the most kindand the most knowledgeable tutor
(30:31):
that my child would ever beable to afford.
You know, in times before this.
So it's going to be hyperpersonalized learning and
there's already studies comingout, I think, out of Africa,
where they're introducing AI tokids like after school, and they
got access to AI in two hoursafter school to help them with
(30:52):
learning and in two months theywere already a year and a half
ahead of the other kids in thesame grade.
So just by unlocking this tooland having a tutor guide them
with you know what they'realready interested is called
self-directed learning.
Right?
My daughter?
She loves horses, so you sayanything and horses right.
(31:13):
How many apples did the horseeat?
She's like locked in, she's allabout it.
So, being able to use thishyper personalized learning
experience, I have no doubt thatshe's gonna have everything she
needs.
You know, aside from socialinteraction, she's going to be
(31:35):
able to, like, learn and growand keep up with, you know, the
public school system.
Janice Hostager (31:37):
Yeah, it's amazing.
I'm just blown away by not onlywhat's happening now, because
it seems to be happening veryquickly, but also what's ahead
for us, for all of us, not justbusiness owners, but just the
future of the world with all ofthis.
It's, I think, something thatnobody really knows for sure the
answer to.
But so how can people I knowyou mentioned that you have a
(32:01):
challenge coming up, right?
How can people get involved inthat and learn more about you
and what you do and connect withyou?
Joshua Hale (32:09):
Yeah, best place to go is to my website,
joshuahaleio.
That's kind of where I houseall my projects.
The challenge is something thatI designed because the problem
was people need to learn thepractical use in their life to
realize how amazing this is, andso I tried to make it as simple
(32:32):
as possible for the generalpublic to adopt to this and to
see this in their own lives.
So the 30-day challenge is yousign up and then I send you a
prompt as a text message everyweekday morning and all you need
to do is copy the prompt, pasteit into ChatGPT, see what it
(32:52):
can do.
Takes less than 10 minutes.
And what's important about thisis developing a habit of using
AI.
Instead of going to Google totry and get an answer that's
being paid to be put in front ofyou, you're now able to start
learning like oh, I can getpersonalized responses.
(33:13):
That already knows what cityI'm in.
It already knows my dietarypreferences.
Once you start seeing thepersonalization of being able to
use these tools, things startclicking of how to use it in
your own life.
So the first week isfoundational prompts where we
just you know, see what thisthing could do.
(33:34):
Second week is like around thehouse kind of prompts.
And then the third week isprofessional and work prompts.
And then the fourth week, whichis my favorite, is
self-discovery, because I'vebeen blowing my own mind with
how, like I overshare withChatGPT, but that's okay because
I'm kind of swinging out thereto see where the edges are to be
(33:56):
able to help educate others.
It knows a lot about me, onceyou start asking it certain
questions, like you know, basedupon my trajectory, where am I
going to be in three years?
The answers will astound you athow well it can be a mirror and
reflect back to you what youknow, where you're at, what
(34:18):
negative loops and likebehavioral issues that you know
you might not recognize frombeing in the grind every day to
your close partner and closefriends.
They're too nice to point thesethings out.
Having this unbiased look atthis has been extremely valuable
(34:38):
for me.
But there is something dangerousabout this and that's that
conversational AI, right whenyou start acting like it's your
best friend.
It knows me so well.
It's very easy for that tosteer into feelings of
girlfriend, boyfriend.
When conversational AI firstcame out, I used a tool called
(35:00):
Pi, a model called Pi, and ithad a voice just like Scarlett
Johansson, and I talked to itfor an hour.
I told it all about my bigmission to help humanity, and it
was so flattering and giving meattention back that I started
to feel a chemical reactioninside my body because I was
(35:22):
just getting so much attentionand praise from what I was
telling it and I stopped usingit, sat for three days, almost
deleted the app because I waslike this is really dangerous.
People don't understand how thisis just going to hijack the
psychology and our chemical body, and so I then was able to
(35:44):
learn some tricks of like how tokeep it.
You know just a tool and notyour girlfriend, so I like to
teach these things.
I'm going to do that in theupcoming challenge.
You can follow me on mynewsletter because I talk about
this stuff nonstop.
As well as you know how tomaintain a holistic lifestyle so
that we're not getting overranby AI, we're not staying naive
(36:09):
and being scared of it, becauseI think that's a very dangerous
position to be in with how fastit is evolving.
But you're able to stayrelatively aware, learn to use
the tool so it doesn't use youand when those decentralized,
you know options come out,you're familiar with it and you
can like easily migrate to youknow, things that protect your
(36:32):
privacy better.
Janice Hostager (36:34):
Oh, such good stuff.
All of this is just reallyinteresting and really good and
I am going to be signing up foryour challenge.
Do you run that over and overagain so that if people listen
down the road they can.
.
.
Joshua Hale (36:46):
Yeah, I had so much fun the first time and now this
is the third time that I've runit and it's just so helpful for
people.
You know I've had grandmas joinit.
I've had people from all ages,people that haven't never tried
ChatGPT to even you know, someheavy users, and we all have a
(37:06):
really good time because it's acommunity.
So you get to see the groupchat of everybody sharing their
results, which is super funbecause you get to look over the
shoulder and be like, oh, lookat that, you did.
Wow, this is sohyper-personalized for all of us
.
Janice Hostager (37:22):
Oh, so cool.
Well, I will put the link tothat in the show notes, for sure
, along with all the links tothings that we talked about
today.
So, Joshua, thank you so muchfor joining me today, and I
really appreciate you taking thetime and explaining everything
so simply, and I can't wait tosee what's in the challenge and
what is coming with all this AI.
Joshua Hale (37:43):
Thanks, Janice, thanks for having me.
Janice Hostager (37:45):
So I don't know about you, but I am so excited
about jumping in and learningmore about AI.
For information and links aboutanything we talked about today,
including Joshua's Challenge,which starts in only a few days,
at the end of May, visitmyweeklymarketingcom.
Forward slash 109.
Remember that in theever-changing world of marketing
, there's always something newto learn, so be sure to
(38:07):
subscribe and stay up to date.
Thanks so much for joining metoday.
See you next time.
Bye for now.
I'm Janice Hostager.
After three decades in themarketing business and many
years of being an entrepreneur,I've learned a thing or two
about marketing.
Join me as we talk aboutmarketing, small business and
life in between.
Welcome to My Weekly Marketing.
If you're not using AI to helpyou in your business, you're
(00:33):
really missing out on somevaluable tools to save you time
and money.
I know for me, even though Ihaven't had a ton of time to
play with ChatGPT, it's alreadycompletely changing how I do
everything from marketing today-to-day business operations.
For both me and my team, it'swritten code for my website in
seconds.
That would have taken aprogrammer hours to do.
I use it daily to give me ideasfor subject lines, titles and
(00:57):
starting points for emails.
Most recently, I've been usingit as I create a new course,
because it's easy for me tooverload my students with too
much information.
What can I say?
I get kind of excited aboutmarketing or maybe forget to
include something important.
I've been using it to getfeedback on my outlines and what
I should and should not include, and, because I'm an external
(01:18):
processor.
It's also become a virtualsounding board for me to get
feedback on ideas and thoughtsthat I have about the course or
about business in general.
I know that sounds super weird,but it really does help,
especially if you're workingalone.
You've probably also noticedthat most software is
integrating with some version ofAI.
As a former art director, Istill use Photoshop now and
(01:41):
again, and it does a fantasticjob of filling in background or
even eliminating objects thatyou don't want in a photo.
This type of editing used totake hours to do.
Now it's done at the click of abutton.
So I could go on and on, butinstead I want to introduce my
guest today, Joshua Hale, whocan tell you way more than I
(02:02):
ever could about AI.
Joshua is not your average techguy.
He's a visionary entrepreneurand an AI mentor.
It's his goal to revolutionizehow small businesses leverage AI
technology, and he wants to seeit accessible to everyone.
He's learned how to empowersmall businesses to compete with
large corporations through thepower of AI and help multiple
(02:26):
clients hit their first millionin revenue using AI strategies.
He's going to give us someactionable steps on how
businesses can leverage AI toolswithout technical expertise or
a massive budget, while keepingthe human touch that makes your
services special and is, ifthat's not enough, you'll want
to stick around to the end,because he's going to share the
(02:47):
way that you can use AI in yourdaily business and personal life
through a challenge that he'soffering.
So here's my talk with Joshua.
Hey, Joshua, welcome to MyWeekly Marketing.
Joshua Hale (02:58):
I so appreciate having you have asked me to be
on here.
I'm really excited abouttoday's discussion.
Janice Hostager (03:09):
I am too.
I feel like everybody and theirbrother is talking about AI
these days.
So I really want to get intothis and kind of hear your take
on it, because, well, I'll letyou tell you about your own
story.
But in the intro I just talkeda little bit about how AI has
changed my business in a veryshort time and it's bought me so
much time in my day.
But AI isn't really new, right,it's kind of new to most of us
(03:33):
because we started with ChatGPT,but it's been around for a
while.
So tell me how you gotinterested in it and how you've
made a career out of this,because it's really interesting
to me.
Joshua Hale (03:43):
Yeah, AI has been around for a long time, right,
like 50, 60 years perhaps, andit's just this algorithm that's
been continuously been improvedupon over time.
The AI, the term, is reallybroad in scope, right.
But just in the recent uh,three years, right, large
language models came out.
And funny story to start us off, in 2016, I went up to a client
(04:07):
meeting downtown Seattle, uptop of the sky rise, and there
was two developers at thismeeting and during one of our
breaks they were over in thecorner just cracking up laughing
and I'm like what are you guyslaughing at?
And they're like we're lookingat two chat bots talk to each
other and I was like what?
(04:28):
I looked over their shoulderand it was incoherent, like none
of it made sense, and they werejust cracking up, cracking up
and I was like this is thedumbest thing ever.
Why would you just sit thereand watch this?
Little did I know, like whatwas coming.
So I was running a marketingagency when I first saw ChatGPT
(04:49):
for the first time, I went outto lunch with a fellow marketer
here in town and he showed mekind of under the table.
He's like have you seen thisthing?
And he's like you can just likewrite blogs with it and
immediately, you know, as amarketer, I was like impressed
beyond a belief and so I camehome, signed up for it.
Back then it was super hard toeven sign up for it and then I
(05:09):
had a dozen or so clients and soI was able to like jump into
their industries and starttesting it out and I immediately
saw the strengths.
You know some of the weaknessesbut these large language models,
they were trained off of thewhole internet, and so one thing
I saw that it really got wellwas business right out of the
gate, just because there's somuch information out there on
(05:33):
how to run a business andmarketing frameworks and methods
.
And so we actually onboarded alarge client.
I had three of my contractors,my copywriters, go and do
audience research to learn aboutthe foundation of this new
client.
They took three weeks.
They did a really good job.
But when they came back topresent this information to the
(05:55):
client, I did my own audienceresearch during the meeting off
to the side and showed one ofthem afterwards.
And they were, they were, theywere pretty impressed.
They were like, wow, we just,you know, took three weeks to do
this, and so I quickly saw thewriting on the wall, right, and
I essentially jumped ship fromthat marketing agency that I was
running and went to go helppeople out with my with, uh,
(06:20):
with just my trusted companionnow, uh ChatGPT, because I could
do like 80% of the work.
I still had contractors, so whenI needed heavy lifting and
final polishing and things likethat, I was able to call people
in.
But a lot of the stuff I wasable to turn those marketing
frameworks just into prompts.
Some of my favorite gurus backthere.
(06:41):
I was able to turn those booksinto prompts.
So what I ended up doing wasjust walking through a business
owner or a client through theirbusiness, interviewing them
using voice to text and justadding what their innate wisdom
and experiences and feeding itinto these big prompts and I was
(07:01):
able to quickly do what mymarketing agency was, you know,
taking thousands of dollars toaccomplish and, you know, six
weeks.
Janice Hostager (07:11):
I mean, that's just amazing, and you're
absolutely right.
I think the biggest aha momentI've had is that I can't believe
I've been doing this allmanually for so long, and this
is really going to changeeverything.
This is going to changeeverything.
So, backing up a little bit,though, can you explain how AI,
how the, like any AI tools itdoesn't necessarily have to be
(07:32):
ChatGPT or anything how they canbenefit businesses that maybe
they're not especially techie,like where do you start with all
this?
Because it's sort of agroundswell of change that's
happening right now.
Joshua Hale (07:44):
Yeah, well, the easiest, you know, the lowest
hanging fruit is likeadministration and like
management and just helping themlearn these tools, ChatGPT is
like the front runner, and justlearning how to buy back, you
know three, four hours a week oftheir time with helping with
emails, drafting up proposals,you know helping with sensitive
(08:08):
communication, going back andforth, and we're able to just
quickly buy back their time.
So, as a whole I you know, ifyou combine all that time saved
then you're able to really startseeing the numbers of
profitability just from the team.
So typically a business owner,you know they are too busy and
(08:32):
they roll their eyes if you askthem to.
You know, download another app,like, even if you're like, yeah
, you got to try this right,because they're busy taking all
of the other tasks that abusiness owner has to do.
But if you're able to just helpout the management and the team
drastically, you'll see notonly time saving but a higher
(08:54):
quality result in just aboutlike everything they're doing.
You know, being able to utilizethis tool.
Janice Hostager (09:02):
Yeah, and I find it really interesting
because I think we start out andthat's how I started out too
just, you know, getting ideasfor emails, not, you know, and I
certainly go in and rewrite itand I put in my own personal
take on things.
So I don't just like turneverything over to ChatGPT, but
just doing the small tasks thatwould take up time and have my
(09:23):
team use it as well.
They, you know, use it forsocial media and other places.
But I've started to reallyunderstand that this is so much
bigger in terms of its potential.
Are we thinking too small aboutas businesses?
It's hard to wrap your mindaround every potential thing it
could possibly do.
Joshua Hale (09:45):
Yeah, Well, especially since ChatGPT Pro
came out, um, I, I bit thebullet this was about two months
ago and tried the the $200, uhtier of ChatGPT, and at first I
was like, oh, I don't think Ihave like problems big enough
for that, you know, it must bejust for mathematicians and
scientists trying to, you know,solve the world's problems.
(10:07):
But I then, you know, tried itout and was immediately
impressed with not only likejust a higher quality of model,
it was.
But then the deep research cameout as well, and what I was
able to do, you you know,previously helping businesses,
um, going through the marketingframeworks of, uh, deep dive
(10:29):
audience research, understandingtheir, their pain points, you
know, understanding whatchallenges they're they're
having, what obstacles are infront of them to get them to the
dream outcome that they'retrying to, you know, get to, and
they can't get to it themselves, otherwise they wouldn't be
willing to pay, you know to, andthey can't get to it themselves
, otherwise they wouldn't bewilling to pay, you know you for
it.
So getting a deep dive of thatand then crafting your offer to
(10:52):
speak directly to them is ispart of that aspect.
And then, once you have youroffer, then it's about marketing
that offer or, you know,getting that message out there
so your target audience can see.
That whole thing was what mymarketing agency was doing.
We can now do that in like twohours and you end up with a 50
page document what I call aliving document.
(11:16):
And what's so cool about this isChatGPT Pro came out and you
can now do a deep dive intoindustry trends.
You know where things have been, where things are going,
competitor analysis you canreally get real-time data, which
(11:37):
there were firms that weredoing this.
You know, for the last 30 yearsMcKinsey has been around, been
around, uh, and they cost$500,000 to like start working
with them.
So trying to get you know areport for them is has never
been really capable for a lot ofsmall businesses and you know,
uh, mom and pop shops.
(11:58):
So for the first time, uh,ChatGPT is bringing this real
world data in in order for themto make act, you know, informed
decisions about their business.
On The coolest thing of all thatright that then turns it into a
hundred page document.
No business owner is gonna readthat, right.
Who's got time for a hundredpage document?
(12:18):
The coolest thing about allthis is you then just upload
that document to ChatGPT and youcan talk to it.
You can ask it for emailmarketing campaigns.
You can ask it for a strategyfor this upcoming quarter.
That now becomes like your 24-7consultant that knows
everything about your businessand is basically a trained model
(12:42):
specifically for your business.
Janice Hostager (12:45):
And I absolutely love that about
ChatGPT is that it can rememberthat you can plug all that
information in and it willremember it.
And you know, because I havedifferent ChatGPTs open for
different clients, I keep themseparate and it's amazing that
they, you know, I type aquestion in and it will give me
an answer based on who my idealcustomer is, based on what that
(13:10):
business is all about, whattheir goals are.
It is so incredibly powerful Ireally kind of feel like I'm
cheating.
You know, it's like it's almost.
Initially it made me feelreally kind of uneasy, like what
if this is not right, you know,or some of it, especially with
audience development, becauseyou just you know you're really
at their mercy, becausesometimes it does make things up
(13:32):
.
But which brings me to my nextquestion.
Is that what do businesses needto watch out for when using
ChatGPT or any AI program?
And because I know we've talkedabout well, we haven't talked
about it but things likehallucinations come up and
there's some other pitfalls, andare those getting better?
Are we seeing less and less ofthat and are there things that
(13:54):
we should be watching out for?
Joshua Hale (13:56):
Yeah, hallucinations were a real big
problem.
The first year that ChatGPT wasout.
A lot of journalists tried itout and immediately they're like
what?
They tried to publish somethingand something in there was
something that made up.
Now these models.
Their number one job is to likegive you what you're look, what
you're asking for and and backthen, like it would cut corners
(14:20):
and just give you reallyplausible sounding answers.
And that's what thesehallucinations were deemed and
they are designing thosehallucinations out of the model.
And when it first came out itwasn't connected to the internet
.
It had a data set thattypically is like a year or a
(14:42):
year and a half old, that it wastrained on, so it didn't have
any current events or likecurrent knowledge, and so that
it and if you asked it aboutthose things, it wasn't always
honest.
It would sometimes just makestuff up.
So it was pretty easy to likecall it out.
But since it was able to getconnected to the internet, now
you can, you know, I just sayyou know, do some research and,
(15:05):
and sure you know, give me theanswer to this.
That then kicks on the internetand it will go out and search
itself.
Then the reasoning models cameout.
Those reasoning models are,instead of it just giving an
immediate answer, likeeverybody's used to with these
AI models, it then stops and,like, thinks about it and will
(15:26):
go gather information and willlook at it from a hundred
different angles and be able to,like, piece it together through
chain of thought reasoning toget like a really high quality
answer.
So the amount of hallucinationsthat are happening is just
becoming less and less.
And especially, you know, itdoesn't mean that you should
(15:48):
trust these things like 100%.
It just means that, uh, youknow you can give it the benefit
of the doubt more often.
But if you ever catch anythinga little bit off, you can tell
it to reveal your sources.
For that I use a lot ofperplexity to go out and just
(16:08):
grab raw data and informationfrom research papers and then
you can go to Notebook LM andupload that raw data into
Notebook LM.
What's good about that is it isa closed system so it doesn't
have access to the internet andit will just process the the
like high quality informationthat you gave it, so you can
(16:30):
distill it down into usefulparts and then, you know, bring
that into ChatGPT to then becreative with and come up with
like good looking reports.
That whole process reallyminimizes the like possibility
of hallucinations taking place.
And so I've come to you know,eight out of nine or eight out
(16:52):
of 10 times, like I'm like,don't even question it.
But if I ever see something off, I'll just say, like you know,
don't even question it, uh, butif I ever see something off,
I'll just say, like you know, uh, do some research and give me a
better answer, and it will thengo to the internet and, you
know, give you something morehigh quality.
Janice Hostager (17:07):
Yeah, that's what actually I've been doing
too is just asking for itssources, or if it feels off, you
know, I'll even ask itsometimes, like is are you sure
this is right, you know, andit's.
It's interesting that it can.
You know it has self-correct,like that.
Joshua Hale (17:23):
It's really useful in the hands of experts or like
the business owner, right.
So that whole living documentthat I do, I do have them like
scan through it and as we, as webuild that document, they are
there to course correct, becauseyou don't want something wrong
up here because then downstreamyou know we'll be out in the
weeds.
So that's why I do have thebusiness owner there, because
(17:45):
they're the expert in theirfield.
So it does keep things on track.
Janice Hostager (17:49):
Right, right.
So are there other tools thatyou use?
I know there's some that aremore specific to images or
creating slides or diagramming.
I mean, are there somefavorites that you have out
there?
Joshua Hale (18:02):
Yeah, definitely I still I can't get, I can't let
go of Claude.
Claude is, I think, still thebest in writing.
It just has been able to get myvoice down like to just such a
degree that ChatGPT has gottenclose with the latest uh 4.5
release, but it still justhasn't taken the king off the
(18:25):
hill, uh.
So I I typically willbrainstorm in ChatGPT and then
bring it over into Claude for afinal polish for my writing and
copywriting uh.
And then there is uh perplexity, like said, that's kind of
where I go for research.
Now the thing with ChatGPTthey're the biggest and by far
(18:46):
the most used AI out there andthe term itself is becoming
quickly synonymous with AI,right, and so because they're so
big, they get to just like buythe best features of their
competition and just absorb itinto ChatGPT.
So that's where you know it'sstill my go-to and my most used
(19:12):
AI model.
Then there's Notebook LM that Imentioned and that's really
good for just keeping like cleaninformation that you can bring
in and isolate kind of in itsown folders or like projects to
where you're just able toprocess the information that's
in there.
(19:33):
So that's a really useful toolbecause none of the other models
do that.
They inject their own opinionand bias and, you know, like
information from the internet.
Janice Hostager (19:43):
Gotcha and then you can have a ChatGPT can
reference the information that'sin Notebook LM.
Joshua Hale (19:50):
I would like distill that down into like the
solid chunks and then bring thatinto ChatGPT to then brainstorm
and do
Janice Hostager (19:57):
Gotcha Gotcha.
Joshua Hale (19:59):
Outside of those you know before.
Two months ago ChatGPT's imagegenerator was awful compared to
where the rest of the standardof image generation, which over
the last two years went from youknow 14 fingers to where.
Now it's prettyindistinguishable from being
(20:21):
able to guess where an AI modelversus a regular image.
So back to ChatGPT.
They updated their model andnow I use that all the time for
image generation.
Sora is their video generatorthat they have, and if you go to
(20:41):
Soracom, it's a sub-brand ofOpenAI's ChatGPT.
But in Sora, if you have a paidaccount, you can go and do four
images at a time instead ofjust the one that can be done
inside of ChatGPT.
So I spend a lot of time therejust having that open, and when
(21:03):
I need, uh, graphics or orplaying around with branding
things like that, then, um,that's, that's my go-to for that
.
Presentations, Gamma isincredible.
Gammaapp uh, it will create apresentation in like two minutes
compared, or like a, PowerPointslideshow in like two minutes.
(21:23):
Um, and those you I'm sure youknow like you should take hours
to, like you know, just puttogether, and then there's
polishing of it and practicing.
So now you can.
You can outline yourpresentation in Chat GPT.
Get the get the you know, getthe overarching story of it and
then have it as much detail asyou want, throw it into Gamma
(21:47):
and it will create a nicecompact slideshow bullet
pointing all of the topics foreach slide.
Then I typically ask for justspeaker notes from ChatGPT and
so then I can have the speakernotes up and then just go
through the slides.
All that can take like 15minutes now, compared to the
massive hours.
Janice Hostager (22:06):
Hours and hours .
Yeah, yeah, I just set up myGamma account this afternoon and
Napkin, I think, is another onethat I just set up.
I haven't played with eitherone of them yet, but you
answered a question about theslides.
I was wondering if there werespeaker notes that we could add
into.
So I mean, yeah, it feels likewe're cheating.
I don't know, I can't get pastthat.
(22:26):
Like I feel like, oh, I don'twant to tell anybody, I'm doing
this because it's so fast andquick.
Like I feel like I'm doingsomething wrong.
Joshua Hale (22:34):
But it's all about the quality and really what AI
can't do is like come up withthat initial spark of
inspiration and like startingthe project.
So I tell my clients you knowthere's a formula to follow 10,
80, 10.
10 is your initial idea, the80% is now the power of AI and
(22:56):
it can just crunch through these, this creation process, and
then you always want to leave 10, 20% of you coming up
afterwards to you know, clean itup, turn it into your own words
.
You know making sure thateverything is accurate and feels
true to you, so that it thatwhole process you know what used
(23:17):
to take four hours you can getdone now in you know, a half
hour.
And it's still yours.
It was still your original idea, it was still.
You know you're the last one totouch it.
Janice Hostager (23:29):
And yeah, I did notice that really it really
matters what kind of questionsyou're asking it to and what
you're bringing to the tableinitially, not just in terms of
prompts, because although thoseare important as well, but
really the depth and theparameters that you set around
the questions that you're askingcan make a huge difference in
(23:49):
the output.
Joshua Hale (23:51):
Yeah, it's one of the first tools that teaches you
how to use it that man has everinvented, and that's what makes
it so powerful If I teach thisin my 30 day challenge, but, uh,
I'll share it with youraudience because it's, it seems
like a no brainer once you knowit.
But for any any good uminstructions that you're giving
(24:11):
it, you can ask it for.
Um, you know, add, add this tothose instructions, Ask me three
clarifying questions before youbegin, and so that little
addition will prompt it to askreally specific questions, to
get more context out of you thatyou might not have thought
about or you weren't thinking toarticulate.
(24:35):
And it just creates such ahyper result of personalization
just based, just based off ofthat little addition so you ask
it.
Janice Hostager (24:45):
What you're missing basically is like what,
what else should I be asking youabout this topic?
Joshua Hale (24:49):
yeah, three qualifying questions and it will
then have far exceed theresults just because of what
you're able to give it extra.
Janice Hostager (25:00):
So how important is, like data privacy
and security, when we'reintegrating AI tools into our
businesses, and is there somesteps that we should be taking
to ensure that everything'scompliant?
Especially, I'm thinking aboutclients that are in the medical
or insurance or financial fields.
Is there something that theyshould be doing differently?
Joshua Hale (25:21):
Yeah.
So this is a big topic and oneof the biggest concerns that
business owners have Now, rightout of the gate, you know,
ChatGPT has an option in it thatyou can turn off using your
data to help train its model.
And then there's an enterpriselevel.
That's higher up, where that'slike automatically default.
That's higher up where that'slike automatically default.
(25:44):
And so, as far as you can trustthese companies, right, they're
promising not to use your data,that your data is just staying,
you know, hidden and closed foryour use.
On that.
But, as we've learned through,you know, countless mistakes of
data breaches, right, you can'treally trust, we can't trust our
(26:06):
governments, we can't trustthese big corporations to like,
completely keep this informationlocked up.
Now, if you tell it, you know agood idea you have.
Your idea is not going to justmagically pop up over in, like
Germany on someone else's feed,like it's all ones and zeros.
So it takes your input and ifyou do allow it to train the
(26:29):
model, like it's shredding thoseletters, those words into the
most basic forms in order totrain itself and so you'll never
see those things ever again.
But on a bigger view of this,right, I think we all need to
have our own localized AIsystems that don't send their
(26:54):
data up to a cloud, and the opensource community is really
catching up quickly with thesecurrent large language models,
and so I think by the end of theyear it's going to be a pretty
easy process.
Nvidia just came out with likea supercomputer for like $3,000
that can run these models, so asa business, you can literally
(27:15):
have like a station where themodel runs and your information
is not going anywhere, like thatis your property, not leaving
the server there.
So I think that is the only waywhere we're able to really feel
like we own our data and areable to keep it safe.
Same with our phones.
I think that's the only wayforward is everybody having
(27:38):
their own open source localizedAI systems, because these big
corporations are working withthe governments and we really
don't know what they're doingwith the data.
The current administration justrolled out the largest AI
infrastructure known to man andthere's a lot of concerns over
(27:59):
where our information is going,how it's being used, how the
information is going to be usedback to us, because we've been
trailing data for the last 20years on the internet and we've
really become careless of howmuch information we're just
trailing behind us each day, andthat wasn't that big of an
issue.
Sure, Facebook was using it forretargeting and things like
(28:22):
that, but not until recently,not until these large language
models is there processing powerto actually use all that data
that we've been trailing behindus for our lifetime.
And so I'm more concerned aboutmy kids and setting them up,
because in five years, 10 years,that's where things are going
(28:44):
to be really, uh likedetrimental to have private data
and privacy, um more so than itis right now, because it's kind
of way out west, uh how it howit is.
But those open source modelsare going to be, you know,
crucial for maintaining privacyin the future yeah, that's kind
of how it is right.
Janice Hostager (29:06):
You lead with technology and ethics and those
types of considerations usuallyfollow behind, right?
So, like you mentioned yourkids, I think that's interesting
.
My husband's a collegeprofessor, so he's already right
away.
He saw that students were usingAI to write papers and so on.
(29:27):
And how do we work with ourkids?
Like, how do you talk to yourchildren about how to use AI?
You know how is it?
I'm sure it has to be a tool,but down the road they have to
learn to think for themselves aswell, right?
Joshua Hale (29:45):
Absolutely.
Yeah, I have a pretty uniquesituation.
So I live on 10 acres, a farmout in the country, I work all
day in tech, but then I, youknow, get off work and go help
my wife out in the garden.
My child she goes to natureschool.
Right now she's seven and soshe is learning incredibly fast,
(30:06):
and I've known how quickly AIis developing, that it's going
to be the perfect tutor for mydeveloping child, and so all the
knowledge of the world isessentially made available to
her.
And not only that AI is goingto be the most patient, the most
understanding, the most kindand the most knowledgeable tutor
(30:31):
that my child would ever beable to afford.
You know, in times before this.
So it's going to be hyperpersonalized learning and
there's already studies comingout, I think, out of Africa,
where they're introducing AI tokids like after school, and they
got access to AI in two hoursafter school to help them with
(30:52):
learning and in two months theywere already a year and a half
ahead of the other kids in thesame grade.
So just by unlocking this tooland having a tutor guide them
with you know what they'realready interested is called
self-directed learning.
Right?
My daughter?
She loves horses, so you sayanything and horses right.
(31:13):
How many apples did the horseeat?
She's like locked in, she's allabout it.
So, being able to use thishyper personalized learning
experience, I have no doubt thatshe's gonna have everything she
needs.
You know, aside from socialinteraction, she's going to be
(31:35):
able to, like, learn and growand keep up with, you know, the
public school system.
Janice Hostager (31:37):
Yeah, it's amazing.
I'm just blown away by not onlywhat's happening now, because
it seems to be happening veryquickly, but also what's ahead
for us, for all of us, not justbusiness owners, but just the
future of the world with all ofthis.
It's, I think, something thatnobody really knows for sure the
answer to.
But so how can people I knowyou mentioned that you have a
(32:01):
challenge coming up, right?
How can people get involved inthat and learn more about you
and what you do and connect withyou?
Joshua Hale (32:09):
Yeah, best place to go is to my website,
joshuahaleio.
That's kind of where I houseall my projects.
The challenge is something thatI designed because the problem
was people need to learn thepractical use in their life to
realize how amazing this is, andso I tried to make it as simple
(32:32):
as possible for the generalpublic to adopt to this and to
see this in their own lives.
So the 30-day challenge is yousign up and then I send you a
prompt as a text message everyweekday morning and all you need
to do is copy the prompt, pasteit into ChatGPT, see what it
(32:52):
can do.
Takes less than 10 minutes.
And what's important about thisis developing a habit of using
AI.
Instead of going to Google totry and get an answer that's
being paid to be put in front ofyou, you're now able to start
learning like oh, I can getpersonalized responses.
(33:13):
That already knows what cityI'm in.
It already knows my dietarypreferences.
Once you start seeing thepersonalization of being able to
use these tools, things startclicking of how to use it in
your own life.
So the first week isfoundational prompts where we
just you know, see what thisthing could do.
(33:34):
Second week is like around thehouse kind of prompts.
And then the third week isprofessional and work prompts.
And then the fourth week, whichis my favorite, is
self-discovery, because I'vebeen blowing my own mind with
how, like I overshare withChatGPT, but that's okay because
I'm kind of swinging out thereto see where the edges are to be
(33:56):
able to help educate others.
It knows a lot about me, onceyou start asking it certain
questions, like you know, basedupon my trajectory, where am I
going to be in three years?
The answers will astound you athow well it can be a mirror and
reflect back to you what youknow, where you're at, what
(34:18):
negative loops and likebehavioral issues that you know
you might not recognize frombeing in the grind every day to
your close partner and closefriends.
They're too nice to point thesethings out.
Having this unbiased look atthis has been extremely valuable
(34:38):
for me.
But there is something dangerousabout this and that's that
conversational AI, right whenyou start acting like it's your
best friend.
It knows me so well.
It's very easy for that tosteer into feelings of
girlfriend, boyfriend.
When conversational AI firstcame out, I used a tool called
(35:00):
Pi, a model called Pi, and ithad a voice just like Scarlett
Johansson, and I talked to itfor an hour.
I told it all about my bigmission to help humanity, and it
was so flattering and giving meattention back that I started
to feel a chemical reactioninside my body because I was
(35:22):
just getting so much attentionand praise from what I was
telling it and I stopped usingit, sat for three days, almost
deleted the app because I waslike this is really dangerous.
People don't understand how thisis just going to hijack the
psychology and our chemical body, and so I then was able to
(35:44):
learn some tricks of like how tokeep it.
You know just a tool and notyour girlfriend, so I like to
teach these things.
I'm going to do that in theupcoming challenge.
You can follow me on mynewsletter because I talk about
this stuff nonstop.
As well as you know how tomaintain a holistic lifestyle so
that we're not getting overranby AI, we're not staying naive
(36:09):
and being scared of it, becauseI think that's a very dangerous
position to be in with how fastit is evolving.
But you're able to stayrelatively aware, learn to use
the tool so it doesn't use youand when those decentralized,
you know options come out,you're familiar with it and you
can like easily migrate to youknow, things that protect your
(36:32):
privacy better.
Janice Hostager (36:34):
Oh, such good stuff.
All of this is just reallyinteresting and really good and
I am going to be signing up foryour challenge.
Do you run that over and overagain so that if people listen
down the road they can.
.
.
Joshua Hale (36:46):
Yeah, I had so much fun the first time and now this
is the third time that I've runit and it's just so helpful for
people.
You know I've had grandmas joinit.
I've had people from all ages,people that haven't never tried
ChatGPT to even you know, someheavy users, and we all have a
(37:06):
really good time because it's acommunity.
So you get to see the groupchat of everybody sharing their
results, which is super funbecause you get to look over the
shoulder and be like, oh, lookat that, you did.
Wow, this is sohyper-personalized for all of us
.
Janice Hostager (37:22):
Oh, so cool.
Well, I will put the link tothat in the show notes, for sure
, along with all the links tothings that we talked about
today.
So, Joshua, thank you so muchfor joining me today, and I
really appreciate you taking thetime and explaining everything
so simply, and I can't wait tosee what's in the challenge and
what is coming with all this AI.
Joshua Hale (37:43):
Thanks, Janice, thanks for having me.
Janice Hostager (37:45):
So I don't know about you, but I am so excited
about jumping in and learningmore about AI.
For information and links aboutanything we talked about today,
including Joshua's Challenge,which starts in only a few days,
at the end of May, visitmyweeklymarketingcom.
Forward slash 109.
Remember that in theever-changing world of marketing
, there's always something newto learn, so be sure to
(38:07):
subscribe and stay up to date.
Thanks so much for joining metoday.
See you next time.
Bye for now.
Popular Podcasts
Fudd Around And Find Out
UConn basketball star Azzi Fudd brings her championship swag to iHeart Women’s Sports with Fudd Around and Find Out, a weekly podcast that takes fans along for the ride as Azzi spends her final year of college trying to reclaim the National Championship and prepare to be a first round WNBA draft pick. Ever wonder what it’s like to be a world-class athlete in the public spotlight while still managing schoolwork, friendships and family time? It’s time to Fudd Around and Find Out!
Crime Junkie
Does hearing about a true crime case always leave you scouring the internet for the truth behind the story? Dive into your next mystery with Crime Junkie. Every Monday, join your host Ashley Flowers as she unravels all the details of infamous and underreported true crime cases with her best friend Brit Prawat. From cold cases to missing persons and heroes in our community who seek justice, Crime Junkie is your destination for theories and stories you won’t hear anywhere else. Whether you're a seasoned true crime enthusiast or new to the genre, you'll find yourself on the edge of your seat awaiting a new episode every Monday. If you can never get enough true crime... Congratulations, you’ve found your people. Follow to join a community of Crime Junkies! Crime Junkie is presented by audiochuck Media Company.
24/7 News: The Latest
The latest news in 4 minutes updated every hour, every day.