Latest / Found in Interpretation Podcast / Ep. 58 - Joachim Lepine - Stop Asking ChatGPT. Ask Yourself First.

Transcript

Ep. 58 - Joachim Lepine - Stop Asking ChatGPT. Ask Yourself First. · Found in Interpretation Podcast

0:00
Hello, hello and welcome to another episode of Found in
0:03
Interpretation podcast, the podcast that talks about
0:05
translation, interpretation, a I guess technology and I'm here
0:11
with Brian. Brian, how you doing?
0:13
I'm doing great, thanks. And this is a first for us.
0:16
This is our third time having a guest, the same guest speaker on
0:20
and we've got Joe coming back for 1/3 time and it's but it's
0:23
going to be totally new content. So I'm really excited.
0:26
How you doing, Joe? I'm doing great.
0:28
How are you? Thank you so much for having me
0:29
on. All right, excellent.
0:32
Uh, don't forget everyone like share, subscribe, comment.
0:38
It really helps us and encourages us to continue.
0:41
We got great feedback from our episodes, so now let's kick it.
0:57
We're here with Joshi. So Joshi is here to present or
1:04
introduce us to his latest book, isn't it?
1:08
Right on yes, you GPT before a ChatGPT is gonna release in
1:12
March and I'm so thankful to have this opportunity to tell
1:15
you all about it and just to catch up with you guys.
1:18
I love watching your podcast because it's always fun and
1:20
fresh and I love your vibes. So it's just awesome having any
1:24
opportunity to to hang out with you cool dudes.
1:26
Yeah, likewise, it's always a pleasure to see you evolve
1:30
throughout your career too. Like the last couple of years
1:33
it's been because you're one of our first guests and we were
1:35
talking about Lying Academy and all that.
1:38
And then you launched the first book.
1:40
Are you published a first book about AI and now you're doing a
1:45
second book, which is pretty monumental feat considering how
1:49
busy we are in our everyday lives to have an another
1:51
publication so quickly after the next.
1:54
So we have a a few questions for you.
1:58
Firstly, um, do you have to read the first book about AI in order
2:03
to understand the new book? Great question.
2:07
So no, absolutely not. So AI Resilient was all about
2:11
giving professionals and experts the tools to continue to
2:14
showcase their value and get hired and booked solid by
2:17
clients in the AI era, where clients essentially can go to AI
2:21
and get a deliverable. And so sometimes they're
2:23
wondering if they still need to hire a professional and
2:25
sometimes they're they're foregoing professional services
2:28
and trying out AI. And so that book was like tools
2:31
for the experts to help talk to our clients a little bit more
2:34
strategically. U GPT before ChatGPT is totally
2:38
different. It's a book that talks about the
2:41
difference between the human mind and the AI mind or you
2:46
know, semi mind or whatever you want to call it, the difference
2:49
between the two. And it's also tailored toward
2:52
experts, but also decision makers and businesses and
2:55
organizations. So all of all different sizes to
2:59
help them understand humans versus machines.
3:02
And also by using our own brains more effectively, we can use AI
3:07
much more strategically and much more sparingly.
3:10
So I'm sort of aiming to hit a couple of different birds with
3:14
one stone. OK.
3:15
So then the the first time when you published AI Resilient, we
3:20
read the book and we had a kind of a book review session if you
3:24
want to call it that way. And this time is when is this
3:29
new book going to be available? Yeah, well, thank you for having
3:32
me on before the books even out. It's super nice of you.
3:34
It's gonna come out in late March.
3:36
I'm aiming for the third week of March.
3:38
I don't have the exact launch date, but when it comes out
3:40
it'll be up for $0.99 for a couple days.
3:44
And yeah, right now the manuscript is undergoing
3:47
editing. It's a little bit longer than
3:48
the first book, a little bit more complex in different ways.
3:50
So it's been a little bit nerve wracking going through the
3:53
paces. But if all goes well, one month
3:55
from today it should be out on Amazon.
3:58
Cool. Now Full disclosure, did you use
4:01
ChatGPT to help you write this book?
4:04
No, I did not use chatpad to write the book, but I think it's
4:06
such an important and good question.
4:09
I've noticed that in different places people are starting to
4:11
use labels like human only. I was just watching like
4:14
relaxation YouTube videos this morning.
4:17
Like I take a little time before I start my day and like people
4:20
are starting to say like human made music or no AI video right
4:25
And and right in the 1st 2 pages of the book, I'm going to talk
4:28
about how I created the book and I did use AI, but I did not use
4:31
it by going to chat PT, for example.
4:34
I think that's the the opposite of what the book preaches.
4:37
So I did use AI, but I used it very, very strategically.
4:40
What what when you're saying you you didn't use Chaji PT but you
4:44
used I? Most people associate AI with
4:48
ChatGPT, So what do you mean by that?
4:51
Yeah, it's a great question. So actually what I decided to do
4:53
for the second book, since it's all about turning to our own
4:57
hearts and minds before we turn to machines.
5:00
What I did is I did like you know, I felt there was a point I
5:04
literally woke up 1 morning, I was at my in laws place and I
5:07
was like Oh my God, I need paper.
5:08
Like now I need paper and a pen. And I just had I had like all
5:12
these ideas for a book on why we need to put people before
5:15
machines. And I just you know, was
5:17
scrambling around my inlaws place looking for something
5:19
right on. And just like I think I found
5:21
some like old sketching paper or something, I just did like a
5:24
brain dump. Like I wrote down all these
5:25
ideas and stories and anecdotes I'd like to share.
5:27
Your mother-in-law say. You know, they were, they were
5:30
very patient with me. And it's like, oh, Joe's, Joe's
5:33
off in his corner. Yeah, yeah.
5:36
So they're used to this. Yeah, they're used to it.
5:39
Yeah, everyone's used to it. So but yeah, essentially like I
5:41
took the ideas that I knew I wanted to share and things that
5:44
I've also been like cogitating and talking to people about.
5:47
I'm including with you guys in in our, the podcast that you've
5:50
done with me. I took all these different ideas
5:52
and I put them down and I used a wonderful app that let me
5:55
dictate the ideas and then put them into categories.
5:58
So that's one of the ways I used AI, not to come up with the
6:01
ideas, but to do sort of the grunt work of putting things
6:04
into buckets to help to organize the content.
6:06
So that's what I mean by I used AI.
6:08
And that's really what the whole entire book is about is about.
6:12
Don't go to the machine to do stuff for you.
6:15
Come up with your own ideas, own your stuff, own your passion,
6:19
own your concepts, own your intellectual property, and then
6:22
go to the machine to help when and if useful with the grunt
6:27
work and not the. Opposite.
6:28
Do you give specific instructions to the machine to
6:31
say don't give me any suggestion, don't go outside of
6:34
what I just told you, because naturally the artificial
6:37
intelligence will give you a lot of propositions to go into a
6:42
certain way. So that might influence you.
6:44
Did it try to do that or did you give specific instructions?
6:48
Yeah, it's an awesome question, actually.
6:49
Use a specialized app for this. I did not use ChatGPT
6:52
specifically because ChatGPT is prone to hallucinating.
6:56
And you can give it instructions, but it's
6:57
stochastic, right? So it'll give you one response
7:00
one time, and then you put in the same prompt another time and
7:02
you get a different answer. And so it's not very.
7:05
Language app. I'm curious about the app.
7:08
Yeah, so the app is called the Letterly app.
7:11
Literally the nice thing about it is that you can record up to
7:15
75 minutes. So this lets me go off in my
7:18
crazy corner at my inlaws place and talk to myself and and it
7:22
will take all of that content and it will very, very
7:24
effective, effectively structure it.
7:26
And yes, you can enter your own custom prompts to make sure it
7:29
works well. Well, well, kudos to your in
7:32
laws for giving you this creativity and for allowing you
7:36
to have a special space for you to talk to yourself.
7:40
You're used to it. Like sometimes I'll disappear
7:42
and go take a nap and it's like, OK, well, you know, no one does
7:45
that in our family but Joseph taking a nap.
7:47
And it's like, you know what? Yeah, I'm off taking my naps.
7:49
Like it's cool, guys. I'm gonna do me.
7:53
So with literally, it's literally if I remember
7:56
correctly because I've heard of it, it's a text to speech app,
8:00
right? It's actually the opposite
8:03
speech to text, but the great thing is.
8:05
That sorry. Sorry.
8:05
Exactly. Yeah, yeah, yeah, sorry.
8:07
It's a speech to text app. Right.
8:10
But it actually goes further because what it does that's
8:12
really special is it? It will do the speech to text.
8:15
So it will get down what you said and produce a minimum of
8:18
errors. But also once it has all that
8:21
content, you can like custom prompt it and it has built in
8:24
prompts and you can create your own.
8:26
So this is just one of like countless ways in the book that
8:30
I talk about how I'm using AI. So I don't want to go too much
8:33
in the weeds, but it really is a great app because it has those
8:36
really high character limits. Like you can talk to it for a
8:38
long time and I'll sometimes walk around in my neighborhood
8:42
and I'll use different AI programs and speak to them.
8:45
And then it's taking my ideas and it's it's able to basically
8:48
apply the three years that I've been teaching.
8:50
So research, rounding out and refining for my content and my
8:54
ideas. I love that because it's like I
8:57
hear so much junk about like use AI to write your blog in one
9:00
click, replace your translator, right?
9:03
I've seen ads on Facebook replace your interpreter, use
9:05
our AI, go to AI to replace your graphic designer with ChatGPT
9:10
connected to Canva, right? I see that every day in my
9:13
Facebook feed and I'm like, no, guys, this is not the future.
9:16
This is not, this is not right. This is not how we do it.
9:19
This is not how we make AI a win for everybody.
9:22
It's not about replacing people. That's, you know, that's going
9:24
to lead to slop and mediocrity and bad feedback loops and 100
9:28
other problems. And meanwhile there is a smarter
9:31
way you could use it. Absolutely.
9:33
And so you have this, the AI resilient, that was the message
9:39
behind that basically is like use your human, the human touch
9:43
to be able to connect with your clients and.
9:46
Totally. And basically grow your business
9:50
because you are human and you cannot offer extra that a
9:54
machine can't. Is there like a key message to
9:56
retain from your new book? Right on.
10:00
So the message with the new book is humans have incredible power
10:07
that we can tap into. And if we clearly understand our
10:10
strengths and weaknesses and the machine strengths and weaknesses
10:13
and the basics of machine learning, then we can a, be more
10:17
strategic in the decisions we make.
10:20
Um, and often that will lead to not even needing AI if we, if
10:24
we've prompted ourselves first, right, We can prompt ourselves.
10:26
Sometimes it's helpful to like apply the way we think about
10:28
machines to us just to sort of like get it, like we can tap
10:31
into ourselves and get better results.
10:33
And then when we do use AI, because we're coming to the AI
10:37
with something quality already, then it can do what it's really
10:41
good at, which is pattern recognition and repurposing and
10:45
a lot of other stuff that can be very complimentary.
10:48
But I think that often in society, people are getting the
10:51
order backwards. It's like, let's just ask
10:53
ChatGPT. And I'm like, let's not just ask
10:56
ChatGPT because I'm studying AI right now.
10:58
And I can tell you it's not the place to go for critical
11:01
judgment. It's not the place to go for
11:04
innovation and new ideas. That's not what it.
11:06
Does it's a recombining me all the time that's wrong?
11:09
I know it says that had the greatest podcast ever.
11:13
It praises me too. Yeah, the flattery.
11:16
And it's. Programmed to do that.
11:19
There's something about that. I mean, like it's it's trained
11:24
to tell you what you want to hear because that maximizes
11:27
engagement. And it's trained to sound
11:28
authoritative, of course, as we all know by now, because because
11:33
people will use it more and go to it more if it sounds like it
11:35
knows what it's talking about. Yeah, that was doing something
11:38
with York University the other day and, and someone asked a
11:42
question and it gave a response and then she asked the second
11:47
follow up questions like did you make this up or is this that
11:50
like actual research? And actually, it did confess
11:52
that it was made-up. Yeah, yeah, it will do that to
11:56
me too. And it'll reply and it's like,
11:57
oh, Joe, you have such an eagle eye.
11:59
Congrats. Like bravo for spotting that.
12:01
You're so good. Let me let me try again and I'll
12:03
do better. But each time.
12:05
Yeah, this program had our gesture.
12:09
Here, yes. I love it.
12:13
For those who who are only listening to us and not watching
12:16
us, you've missed out on an epic.
12:17
Yeah, here. I was playing the violin and
12:21
Brian was tapping his back. Putting myself in the shoulder,
12:26
yeah. Right.
12:28
Yeah. So that's great.
12:29
And now quick question, are you, have you run this by ChatGPT?
12:34
Like is there any issue about putting a trademark name on a
12:37
title of a book? Yeah, so originally I had
12:41
thought about the title Me GPT before ChatGPT because I started
12:46
telling people like, hey, guys, look at how I'm using AI these
12:49
days. I've tested it a lot, and I'm
12:50
getting better results by asking Joe GPT first.
12:53
But then, you know, that won't resonate with other human
12:55
beings. So I started talking about Me
12:56
GPT, and then I put that in the title of the book.
12:59
But, you know, before you go publish a book, you have to
13:01
check copyright infringement and stuff and patents.
13:04
And I found out that Me GPT is actually a cloning service that
13:08
lets people create an AI clone of themselves.
13:11
So that would be going exactly against the message of the book.
13:15
So I changed it to 2 GPT and there are a couple of like
13:19
little. Sounds like a the premise of a
13:21
dystopian movie here. It does, doesn't it?
13:23
Yeah. Yeah, like I don't wanna, like,
13:26
I don't know about you guys, but once I was having a beer with a
13:30
friend and I spilled my beer and he was filming or something.
13:35
And then he he put it into an AI and then he created.
13:41
Now he took a picture and then he did a little AI video with me
13:46
spilling my beer and. I would pay good money for that
13:49
by the way. If it's if it's for sale, I'd I
13:52
would. Like let us know if it's
13:53
available from the. Public well, I should ask him,
13:55
but when I saw myself in the situation that it didn't happen,
14:00
did the guy just typed it up, typed the prompt and then I and
14:03
the video I spilled my beer and I felt really weird.
14:07
I it, I knew it wasn't me. You know when, when if I film
14:11
you, you know it's you. But if the if you see yourself
14:15
doing something you've never done before, it's pretty creepy,
14:19
I thought. I don't know about you guys.
14:22
It is creepy. I see a lot of that AI generated
14:24
like, um, you know, people are using it for personal branding.
14:26
So they'll take a photo of themselves.
14:28
Then I will put them in all different places.
14:30
I don't know about you guys, but what I see, like I see some
14:32
influencers that do that with like every post, it's like them
14:35
in a different location on a beach or something and you can
14:37
just barely tell that it's I maybe in a few months we won't
14:39
be able to tell anymore. But it does not resonate with
14:42
me. Like, I see that stuff and I'm
14:43
like, no, like, if you're going to go on social media, share
14:46
something about yourself and don't don't be putting AI stuff
14:49
up there. Like, hey, there's maybe after
14:52
they get a call from the CRA and the CR get on their revenue
14:55
agency says, hey, how how come you have all this money to go
14:59
around the world, you know, and then maybe they'll stop.
15:02
Yeah, and maybe your insurance won't, won't, you know, grant
15:06
your claim because you've been drinking alcohol and that's it.
15:08
That's when the fine print it says you shouldn't be drinking.
15:11
So yeah, there are many ways like insurance uses those kinds
15:14
of images for their purposes as well.
15:15
And there are all kinds of nebulous and fascinating ethical
15:19
issues with creating image like likenesses of people, right, and
15:22
then putting them out there for the world to to see.
15:25
But see, the issue is too is that when those things happen,
15:29
they don't put a disclaimer that this is AI generated.
15:31
Whereas like now we're putting a lot more disclaimers at this.
15:35
We're not using AI. So there's kind of a double
15:39
standard there when it comes to to that aspect.
15:42
So that is changing. If you generate an image of
15:45
yourself, for example with ChatGPT, now there is a label.
15:48
So I did this recently. I created an image on ChatGPT
15:52
and put it straight into LinkedIn and a little icon
15:55
appeared and was like was in the upper left hand corner saying
15:58
like this was ChatGPT generated so and you YouTube as well.
16:03
I don't know if you've noticed this, but in the videos now, if
16:06
AI was used very significantly, then like there has to be a
16:10
label now that says like AI was significantly used in the
16:13
creation of this video. It's still a little bit vague,
16:15
but it's starting, right? Like AI labels are starting.
16:19
Yeah, well, that's good because it's it's getting everyone's
16:22
really confused. Like, I mean, you look at like
16:26
world leaders and doing like these crazy things and I'm like,
16:29
did this really happen or not? Like, are you see someone like,
16:31
you know, jump in front of a subway or something?
16:34
Like it's like, Oh my God. Like I think I know this person
16:37
and you know, it's very it's, it's scary.
16:40
Like the way I understand, like a lot of the stuff like Facebook
16:43
and everything, when it first appeared, it was like the maybe
16:46
create a network and create like a something kind of positive.
16:51
And it's become so negatively used that like it's just, um, I
16:56
don't know how to explain it. It's become very perverse, I
16:58
guess. Yeah, there there are many risks
17:01
too, you know, for misinformation and
17:02
disinformation for election interference.
17:05
Like, you know, we can add this as a new layer on top of the
17:08
bots that already were previously interfering in, in
17:11
major elections, for example, in North America.
17:14
So yeah, I mean, there are tons of issues.
17:17
It's quite a can of worms. And there's precious little
17:19
legislation to actually prevent people from putting that stuff
17:22
out in the world for now. And then there's the
17:24
environmental cost, right, Which a lot of people don't talk
17:27
about. But when you generate like, I
17:29
don't know if you saw those cute little monkey videos or
17:32
whatever, like or Bunny videos where people create like a 10
17:35
seconds real of like some bunnies jumping and it's super
17:38
cute and whatever, you can tell right away it's AI generated and
17:42
it's like for a week or two everyone was doing that.
17:44
Well, I did a little bit of research and one, I think it was
17:47
like 10 seconds video of those jumping bunnies or whatever.
17:52
It takes as much power to create as running your microwave oven
17:57
for one hour. And that's before you get into
18:00
the carbon dioxide emissions and the habitat destruction and you
18:05
know, so many other things. So while the tech is getting
18:09
more efficient, it's not keeping up with the rate of usage when
18:12
there's basically dumb stuff like, you know, AI videos, you
18:16
know, cute AI videos going viral, like that's that's very
18:19
destructive to our planet. And I look at those videos and I
18:21
think, is it worth the trade off?
18:23
Especially if like they're portraying animals.
18:24
It's like, you think this is cute, but are you really helping
18:29
animals with this video or are you actually really doing
18:32
something very destructive? Like are people are thinking
18:34
about this, you know? Hmm, absolutely.
18:37
Like I was gonna, I've been looking into getting solar
18:40
panels for my house and I was talking to one of the, the
18:45
companies that sells them and, and he was telling me like that
18:48
the the grid is just, was not expecting this boom in power
18:53
surge just because of all the new AI that's coming out and how
18:57
much electricity, Australia electricity it takes to be able
19:00
to generate all of these things that are like, like you said,
19:02
like it's not necessarily useful information.
19:05
It's just, but it does take a lot of energy.
19:08
Right, exactly. And you know, I'm, I'm a
19:10
pragmatist at heart, so I'm not on a crusade against AI.
19:13
I don't think that that's going to work.
19:14
You know, just like when cars came out, they were wildly
19:18
inefficient, right? And you could go on a crusade
19:20
against cars, but it's like cars are coming, guys.
19:23
And it's very much the same thing today.
19:24
Like AI. I'm not sure it's going to be
19:28
effective to like go all out on a crusade against it or try to
19:32
prevent it or try to put the genie back in the bottle.
19:34
As a pragmatist, I'm more like, you know, it's out there, people
19:37
are using it. Let's have a conversation about
19:40
not using it for dumb stuff. And conversely, how can we use
19:45
it strategically in a way that actually is intelligent and
19:48
ethical? And I find it's kind of a
19:49
fascinating intersection there. I was reading a or I was
19:53
listening to a podcast yesterday and there's there's a special
19:56
word, I forget what it is for all of these large corporations
19:59
that have like billions of dollars to invest in marketing
20:04
AI, for example. There's a special word for it,
20:08
for what it is. I wish I knew it now because it
20:10
would be very relevant to today's podcast.
20:12
Would you like me to ask JPT? So maybe the AI bubble, is that
20:18
what you're thinking? Of.
20:19
No, no, no, it's like another term instead of marketing, it's
20:23
like just like it's like, yeah. Anyways, I'll maybe I'll put it
20:27
in the description if I can find out which podcast I was
20:31
listening to yesterday. So my my other question would
20:35
be, so I'm a business owner, Elanza, A freelancer.
20:40
Could this book be beneficial to both of us and are different
20:45
professional activities? Absolutely.
20:48
I mean, I think I've, I've tried to provide a lot of value for
20:52
really all professionals who can use the book directly because
20:56
there are workflows in there for how to get better results from
21:00
AI when you do use it. So there are tips for doing
21:02
better research. There are tips for finding blind
21:05
spots and things that you can improve, or testing your
21:07
assumptions or stress testing ideas or doing market research.
21:11
But all of this stuff starts with tapping into your own mind
21:14
first. So that stuff should be useful.
21:17
There's an entire chapter with really, really great down to
21:20
earth things you can do with AI strategically and sparingly.
21:25
And then the other side of it is I'm going to be encouraging
21:30
people as well to not hesitate to recommend the book to their
21:33
clients because I think that this, you know, I've primarily
21:37
written this for business owners, for organizations, so
21:40
that they can understand the difference between the human
21:42
brain and the AI, you know, output production, I guess you
21:47
could call it, or generation. And so I'm hoping that it will
21:50
help with advocacy, which is not really what the first book did.
21:53
The first book was exclusively for professionals.
21:56
The second one is really intended to to build broader
21:59
societal awareness of where humans excel, where machines
22:02
excel. And let's give it a little a
22:04
little bit of thought, you know, before we jump on to AI, let's
22:06
not ask ChatGPT first. So I think that by building
22:10
awareness of all the value that humans provide, I'm hoping that
22:14
I'm going to be able to help organizations to make better
22:17
choices and to also simply continue hiring humans.
22:20
Because, for example, a recent MIT article showed that because
22:25
everyone's using AI, AI in itself is not a competitive
22:28
advantage. So there's nothing to be gained
22:31
from starting to use AI to replace services of highly
22:34
qualified professionals. Quite the opposite.
22:36
What the research is starting to show, and it's very encouraging,
22:39
is that the market leaders of the future are the ones that are
22:41
going to power their organizations with passion and
22:45
innovation and the human elephant and creativity.
22:48
And so that's a big part of it is like, read it for you to get
22:51
awesome cool tips of how to use AI better.
22:54
But then think about recommending this to your
22:56
clients as well because this is some advocacy for them.
22:59
And if all goes well, we can hopefully make them a little bit
23:03
more aware in a tactful and empowering way.
23:06
We can make them aware of the importance of continuing to hire
23:08
human as well. So where, where does a feature
23:13
end in an AI feature comes out for you?
23:16
Because a lot of the things we see that are AI, it's basically
23:20
a marketing term basically for a feature that can organize your
23:25
thoughts, your ideas, your text or things like that.
23:29
Like if I use AI to correct, let's say say proofread this,
23:32
well, basically the I, is it really an AI or proofreader now?
23:36
So I'm I I had a hard time sometimes seeing where the
23:41
future ends and the real AI begins.
23:44
Totally. Proofreading is a fascinating
23:46
sort of case study because AI is not especially good at
23:50
proofreading. And the reason is, again,
23:53
there's always an element of randomness in its answers.
23:55
It's stochastic. So if you have it revised the
23:57
same text twice, it will give you different things and it will
24:01
always recommend changes that are inappropriate and that would
24:04
make the text worse. You know, I've noticed that
24:06
consistently, but when I use it effectively, I do have
24:09
specialized prompts where I basically will not tell it to
24:12
correct the text. I never asked AI to correct the
24:14
text. That's, that's a very poor use
24:16
in fact, of AI, right? The correcting requires human
24:19
judgment. But what I will ask for is
24:20
strategic suggestions and stress testing and I'm checking blind
24:24
spots and things like that. And it's actually quite good at
24:29
giving ideas again, sometimes useful ideas, right?
24:33
So it should, again, This is why I think it's so important to
24:36
start with human first. Now there the other side of this
24:39
is that there are apps that use AI proofreading that our
24:43
specialized and these are more things that are typically in the
24:46
hands of humans. So if we think of Antidot, for
24:48
example, in their online version, now they have an AI tab
24:52
and it will suggest reformulations.
24:54
But essentially, this is, you know, used by professionals
24:57
typically who will be able to look through the AI suggestions
25:00
and then make a determination on what you know, what it's worth
25:03
correcting and what is not. And I think that's really
25:05
interesting. That's a more specialized way.
25:07
We're going to see this more in translation.
25:09
For example, we have QA checkers that are really, really trained
25:14
to proofread text and that do a better job.
25:16
And I'm all in favor of like, keeping these tools in the hands
25:19
of the professionals, right? When you go to the dentist, you
25:21
don't give them the tool, you don't suggest their tools.
25:24
You respect them and you say, hey, you know, I need a root
25:27
canal. You guys do your thing.
25:30
I trust you guys, right? And it's kind of the same thing
25:33
like come to us, the language professionals and we'll choose
25:36
the tool. I think that's that's really the
25:37
message that I would like to to share with all translation and
25:41
interpreting and editing all language, you know, kind of task
25:45
clients. Yes, don't tell a dentist how to
25:48
do his job. That's a good message.
25:50
Yeah, exactly. I'm just joking.
25:52
No, but I like, I like the tagline like tap into your own
25:55
brain first before you tap into Chat GPT's brain.
25:59
That that's, that's really sound advice.
26:01
Because when I go into ChatGPT, because I want like in my, the
26:04
back of my mind, I'm like, I don't want to think too much
26:06
about this. I'm going to ask ChatGPT.
26:09
But yeah, you're right. Like you do have to tap into
26:11
your own brain, especially in like in a professional setting.
26:15
Totally. And I mean there are there have
26:16
now been studies that find that if we over rely on it, we go to
26:19
it first and we experience cognitive atrophy.
26:21
So for example, students who will write an essay and 1st go
26:25
to AI compared to students who have to actually think and then
26:28
are allowed to use AI. Like in testing they found that
26:31
you're basically you experience brain rot and produce worse
26:36
stuff if you over rely on it. So I think that's going to be
26:39
true in a professional setting, just like you said.
26:41
So we have to be very like watchful and decide to be
26:44
deliberate with our workflows. Yeah, it happens with like,
26:48
driving too, like when you automatically put Google Maps
26:51
on. I mean, you know how to get from
26:53
point A to point B, but you still do it.
26:55
Like, it just has an instinct for a lot of people, like, just
26:58
people that intuitively do it. And like, you just forget
27:01
sometimes. Like, Oh yeah, I should have
27:03
turned right here. I should turn left here.
27:04
Like it's just weird how your brain works out when you over
27:08
rely on electronics. Totally.
27:10
And also think about when you take the taxi, they all use a
27:13
GPS now and we don't pay them less.
27:16
No, am I right? Right.
27:17
We pay them the same thing. They're using the AI.
27:20
It's we're still getting our result that we asked for.
27:23
In fact, if anything, we're getting, we're paying little bit
27:25
less or we're getting service, we're getting to where we need
27:27
to go a little bit faster, right?
27:29
Both are usually true at the same time with taking a taxi.
27:32
And so it's sort of like, it's very interesting, like what you
27:36
just did is very useful because when we go and take other
27:39
occupations, other professions, other areas of endeavor, then we
27:43
suddenly can like shed a new light on, on the language
27:46
professions and say, Hey, hold up, hold up.
27:48
You know, if, if you're, if we're not paying our taxi driver
27:52
less or if like we're paying the same, we would have paid before,
27:56
but we're getting better service in some way, Why would it be any
27:58
different for us? You know, why should we start
28:00
telling clients that we're going to lower our rates, for example,
28:03
like a lot of people are doing that.
28:05
And I'm like, no, guys, stop, don't do this.
28:08
It's not going, it's not in your interest and it's not in your
28:10
client's interest. Well, it reminds me, you know,
28:13
when you're using Uber and then you know your area and you see
28:17
the Uber driver doing this and you're like, Oh my God, he's
28:20
using Google Maps because Google Maps for you go this way and
28:23
they're not using their brains. Yeah, right.
28:26
You're like, good point. I'm not even sure we're really
28:28
getting better service in the end because you're like, hey, I
28:31
know my neighborhood. You gotta go this way.
28:33
And then they go all the way and you're like, you're watching
28:36
them going all the way there. So, you know, Google Maps or
28:40
Apple Maps or whatever, not always to be trusted.
28:45
I just want my tacos, guys. I just want my tacos.
28:48
I see you're going the wrong way.
28:49
And I just, I'm hungry. Such a great point, Yeah.
28:53
It's like there's this assumption that we're going to
28:57
save time or that we're going to get a better result, but that's
28:59
not always the case. Yeah, and, and I think that
29:03
there's we're starting to see this, like people are starting
29:05
to get it, but we're still, it's early days with AI, right?
29:08
A lot of people are still experiencing what I call the Wow
29:10
effect, right? They put something into chat GPC
29:13
and they're like wow, right? And that Wow makes them think
29:15
that I I think they are, you know, plus, you know, chat GT5,
29:19
we did see some improvements. They were incremental, but you
29:22
know, I definitely noticed them. So I think we're still very much
29:26
in the wow phase. Feel that in a way we've I don't
29:28
know, maybe there's going to be a big jump with whatever future
29:31
version comes out, but I feel that every feature now is less
29:36
impressive than the previous one.
29:38
I think like a few months ago, the biggest thing was ohe AI
29:42
agents look, you won't able you won't have to do anything on the
29:45
Internet. You have a prompt to go on its
29:47
way for half an hour and I'll come back, but I'm like, at
29:50
first I was like, oh, wow. And I'm like now, Oh, I
29:53
basically haven't found anything that I really wanted in the end.
29:57
So, so I, I think, and I think that when Chachi first came or
30:02
whatever AI agent, we were impressed by the interaction and
30:07
having to, instead of having to use machine language or like
30:12
programming or code to interact with the machine.
30:14
Now you could basically do it with text.
30:17
And I think that was the impressive part because most of
30:19
us are not computer programmers and we use text based
30:23
communication. So I think that's a that was a
30:26
very natural way to interact with the machine.
30:29
But then after that leap, every other version that comes up,
30:33
unless you're researcher doing some, I don't know, amazing
30:36
things in science. As a regular user and
30:41
interpreter, I basically plateaued in my interaction with
30:46
AI. I don't know how you guys feel
30:47
about that. Yeah, I think, I mean, it comes
30:50
back to Joe's point last podcast, the 9010 rule, right?
30:54
Like they, they gave us something that was like 90%
30:57
complete. And then but it's the last 10%
30:59
that really counts to be able to get to like that 100%, to be
31:03
able to get that accuracy that you wanted to be able to get the
31:06
the exact thing that you're looking for.
31:09
And that's the part that's missing right now.
31:12
Yeah, 100%. I, I could not agree more.
31:14
And you're right, it was like a, an exponential boom, right when
31:19
chat T came out and like the, the, the couple of years after
31:21
that now, I mean, it's been out for a few years, right?
31:24
And we do see incremental improvements.
31:26
But a lot of people, for example, were disappointed with
31:30
the last version of PT because it's like it didn't redefine the
31:34
world. And it's like, yeah, like it's,
31:35
it's exponential. It's not going to do that every
31:37
year. But I think that also, like Alan
31:40
pointed out, something that is so important, which is I think
31:43
that there are a lot of newfangled apps and features
31:46
that are a solution in search of a problem, right?
31:50
So like there's this massive boom in AI.
31:53
So everyone and their cousin is vibe coding their own AI.
31:57
And so you have a gazillion AI's and like, but is it really doing
32:02
the right thing? You know, like I was talking to
32:04
a transcriber yesterday, Linda Grace Fisher.
32:06
I don't know. She's brilliant person and, and
32:09
trans creator and marketer. And she was talking about this
32:12
too, about how like the she, she was using AI to get ideas for
32:18
some transformation she was doing.
32:19
And then at some point she was like, wait a second, I've been
32:22
spending more time on this job because I'm using AI and I look
32:26
at the end result and I'm not as happy with it.
32:28
And she was talking about blog writing, A blog blog she was
32:31
writing. And so she went back to writing
32:33
it herself. And she said, yeah, it was like
32:34
hard to come back to writing. She hadn't written A blog, I
32:36
think in three years. So she found it hard.
32:39
But it's like sometimes hard is good, sometimes struggle is
32:43
good, right? Like I like.
32:45
Say worth in life takes time and effort.
32:48
Exactly. It's not worth it.
32:50
Yeah, that's totally, totally. It's like I, I love that saying,
32:53
you know, your mess is your message.
32:54
Also, like you could write about a blog, about how hard it is to
32:57
write a blog and how you're stuck.
32:58
And that would resonate with like so many people, right?
33:02
So I think there's always a way to bring things back to our
33:05
humanity, and that's going to resonate more because I think
33:08
that we're all sick and tired of people putting up AI content
33:11
anyway, right? There are a lot of people that
33:13
are creating their content with AI and you could just tell,
33:16
right? Even in 2026, you can tell
33:19
someone wrote something with AI and it just does not land the
33:22
same as someone that wrote with her heart and soul.
33:24
And I think most of us would rather read something that was
33:27
written or produced with someone's heart and soul, even
33:30
if it's imperfect. Maybe it's even better if it's
33:32
imperfect because it's human. We connect with that more and
33:35
where we don't want the AI, you know, kind of slob and glossy
33:39
stuff. Like I don't know about you
33:40
guys. That's how I feel anyway.
33:41
Maybe. I'm old fashioned.
33:42
I mean, we're all in languages. I studies as a translator,
33:46
Brian, as well. We're familiar with text.
33:49
And when you read something online that was generated by an
33:54
AI, you feel it, right? Like I, I can't point my finger
33:58
on what exactly, but I read it and I'm like, oh, you know what?
34:02
I'm just gonna scroll and skip it because I know it was written
34:05
by an AI. And as soon as I feel that, I
34:09
lose interest and I move on. But I don't know if the general
34:12
public really sees that as well. That's a great question, yeah.
34:18
Sorry, I'm gonna transition a bit to something a little bit
34:22
different, but still on topic, AI Resilient and your and your
34:26
new book, are those going to be part of like a your curriculum
34:30
and line Academy? They're going to form a book
34:35
series, so up on Amazon, they're going to be like, um, structured
34:38
that way. They're not part of the
34:40
curriculum per se. I haven't really decided what I
34:42
want to do with all of these. Like my goal has always been to
34:46
help people. I, I write because I stumble on
34:49
something where I discover something or I have a great
34:51
conversation with someone and I'm like, we've got to get this
34:53
idea into more people's hands because this can really help.
34:56
Like that's always my goal when I write something.
34:59
I know that in conjunction with UGT for ChatGPT, I've offered a
35:05
free workshop so I can show people like an over shoulder
35:09
view. They're going to see me doing
35:11
the UGT before ChatGPT process and seeing the difference
35:15
compared to just going to AI. So that should be really fun.
35:18
I'm just offering that for free next month to to to show people
35:22
what this is like in action and to make it a little bit less
35:24
abstract in the long term. I haven't really thought about
35:27
what what offer this might lead to like for people who might
35:31
want to go further. I may offer a more complete like
35:34
training suite around human first processes.
35:37
I think that could that could be a really inspiring project, but
35:40
I haven't really decided yet at this stage.
35:41
I'm just like, you know, guys, I found a really cool thing.
35:44
I'm getting amazing results with AI that I don't see anyone else
35:47
getting. Here's how I'm doing it and I
35:49
want to help as many people as possible.
35:52
You call it you GPT so why OU? Yeah, exactly right.
35:58
Not the plant. No, I mean, I first heard me
36:02
thought it was HUQ. No, no, it wasn't huge.
36:05
And then the letter U Oh no, he didn't tell you.
36:07
And then I figured why OU? I know it's early in the
36:10
morning. I haven't.
36:10
Had coffee? That's OK, that's OK.
36:12
But slow. It's still a new concept and.
36:17
Yeah, I like it better than me. GPT for men like me, no T.
36:22
Right, but that's not to mention the the GPT that's a really
36:26
important part of it is that when I say U GPT before ChatGPT,
36:29
I also define what GPT means in that context.
36:32
And I say like the strength of the AI is that it has a
36:35
generative pre trained transformer, right?
36:37
So it's run like hundreds and actually thousands of training
36:41
iterations. Look at chat GTA 3, for example.
36:43
It trained on 500 billion tokens.
36:46
And that's where where AI is very, very powerful is that pre
36:49
trained transformer that's able to even take multimodal input
36:53
and create stuff from it and recognize patterns.
36:56
And that's amazing. But what we've got is something
36:58
that AI can never take from us really, which is, I'm calling it
37:02
UGT for give it your personal touch.
37:05
Yes. That's what the.
37:08
Machine should remember then generative pre trained
37:11
transformer. Exactly.
37:12
Exactly. So we've got something that I
37:14
think is a lot more powerful. But sometimes we forget, right?
37:17
And we, it's so convenient to go to AI.
37:19
Or if you're tired, you can go to AI.
37:20
You're gonna have Gemini write your e-mail for you.
37:22
And we forget how powerful we are.
37:24
We forget. Hey, my ideas are pretty damn
37:26
good too. Why?
37:29
Why should we defer to machines? It doesn't make a lot of sense,
37:32
so that's what I'm pushing back against.
37:34
Absolutely yeah, Sometimes when I need help like crafty message,
37:37
I'll ask ChatGPT and I'll put all the stuff in and then
37:40
they'll make it pretty. But I'll I'll go back and and
37:44
fix it, you know, after. So they, yeah, if you don't do
37:47
that, if you don't have that human touch, you're going to.
37:49
Well, I don't know about you guys, but in French, like in
37:53
English works, you know much that I have a higher accuracy in
37:58
what I want in English. But when it comes to, let's say,
38:01
rephrasing, reviewing or correcting a French message,
38:05
yeah, that's more of a hit and miss.
38:08
Yes, totally. And it will continue to be more
38:09
of a hit or miss in French. And the reason is simply that
38:12
there's less training data with the way that these AI's are
38:15
trained. You know, I'm currently taking a
38:17
course on this. It's the MIT course AI
38:19
Implications for Business Strategy.
38:21
And they really, really go through how exactly the models
38:24
work. And one of the things that I've
38:26
understood a lot better since taking that course is that your,
38:31
the, the raw amount of data you have matters a lot.
38:35
It really, really matters a lot for the quality.
38:37
And of course, the the quality of the input is important too,
38:42
but the raw quantity of even semi good about data, as long as
38:47
it's like relatively good data data, it's going to improve the
38:51
model a lot more. So even if you have a little bit
38:54
of quality data, you're still going to end up with a better
38:57
trained model if you have more data, even if the quality is not
39:00
all five star. So the French will continue to
39:02
lag in terms of writing. I've noticed that as well.
39:06
There's a lot of like all like comma present participle
39:09
structures that are directly borrowed from English.
39:12
There are a lot of things that are not very natural or very
39:14
interesting. The sentence structure, very
39:16
poor variety. It's just because there's less
39:18
data. That's that's the very, very
39:20
simple reality behind it. Can we assume then then like
39:24
basically Internet is the main language or English is the main
39:27
language of the Internet? So basically there's more data,
39:30
there's more material to be able to get information from from and
39:35
if you're using an AI tool. Yes, exactly.
39:38
There's a lot more data out there in English and there's a
39:41
lot less text out there in French.
39:43
It's really a numbers game in terms of the training, because
39:46
the way that the the tests are run, it's doing things like
39:51
checking to see, OK, if we take out this token and we put
39:53
something else like a random number and then we run it
39:55
backwards, like can we predict what was originally there?
40:00
And then based on that, you train the model.
40:02
And so this only works if you have huge amounts of data and
40:06
that's, that's where you start to get convincing pattern
40:08
recognition. So.
40:11
Yeah, if you just understand you're taking the end result and
40:15
then they're walking their way backwards in order to basically
40:19
break it down. Yes, that's one of the ways that
40:22
they train the AIS. There are other ways too, but at
40:24
its core, it's all about recognizing patterns and about
40:28
running predictions. It's all about prediction.
40:31
And so the more stuff you have to predict on in the future to
40:34
produce future answers, the better the outcome will
40:37
typically be. It's it's, it's literally a
40:40
numbers game. And in fact, when we talk to
40:43
ChatGPT, for example, in text, it doesn't understand text.
40:47
That text is being converted into tokens and it's, it's
40:50
thinking, if you can call it that, it's actually just
40:52
recognizing patterns in tokens. So there is literally 0
40:56
understanding in the human sense.
40:57
The way the AI works, it is a really impressive technology.
41:00
It's a modern marvel. You know, if you think about how
41:04
great this is for medicine, for example, where a radiologist can
41:07
get like a pre diagnosis from an AI that can help them make a
41:10
faster determination based on preset criteria.
41:13
It's really great. But it does not have the, the
41:17
the professional judgment to know when is something is the
41:19
right answer and when it's inappropriate.
41:21
That's the the key thing. And that's what I think that we
41:24
also help have to help our clients understand like it the
41:27
the machine doesn't understand whether it, what it's saying is
41:29
good and it will not because it's pattern recognition.
41:33
You know, that's that's what it's all about.
41:34
It's not what we thought I would be what we think about.
41:37
Like I grew up on Isaac Asimov. You know, there were like those
41:40
androids and they had rules for them.
41:42
It's not that, you know, the AI habits is essentially a very,
41:47
very sophisticated system of pattern recognition and training
41:51
based on patterns to be able to create multimodal content with
41:54
foundation models has nothing to do with the kind of, you know,
41:58
potentially self aware AI based on like rules.
42:03
You know, it's not rule based rule rules didn't work.
42:05
You know, programmers tried that and what we have is
42:08
sophisticated algorithms instead.
42:11
But is that? Sorry sorry.
42:13
So is that explained in in one of the chapters of your book?
42:16
Like what Chappie T ChatGPT does and how it finds information for
42:19
you? Yeah, I explained some of the
42:22
basics of of how it works. We know from some rush.
42:25
For example, they did a study last summer and found that for
42:28
some of the prompts they did 40% of the so-called facts that they
42:31
were given by Chegg PT came from Reddit.
42:34
And this is because Reddit is a partner of open AI.
42:37
So they were able to scrape their data for, you know, you
42:40
know, massive amounts of data. So even though they have
42:42
guardrail prompts so you won't directly see like our rants like
42:45
ChatGPT will not often go on a rant right like someone would on
42:49
Reddit. But they may still use all kinds
42:51
of data that's coming from very dubious sources.
42:57
And yeah, a, a very basic, even rudimentary, rudimentary
43:01
understanding of this is so helpful, I think for us and for
43:05
also all organizations today that are contending with AI in
43:07
some form. Absolutely, because I tell you
43:10
that I'm, I'm learning a lot today about all this AI stuff.
43:13
And I feel like for someone like myself, I like to listen to it.
43:17
I like to read about it and then maybe put it into my own
43:20
thoughts and then tell it to someone.
43:22
And that's how I, I retain things personally.
43:25
So I'm, I'm excited to read your, your new book because I,
43:28
I, I'm curious to see how I'm gonna remember this because the
43:33
way that you say it is so succinct and so well put that
43:36
I'm like, I need to use this for like to explain to my clients
43:38
why they shouldn't say I for like interpretation or
43:41
translation or other language functions.
43:44
I'm I'm super happy to hear that.
43:45
That's really my, my only goal is to, to take all this
43:50
information and all my own learning and good conversations
43:53
I have also with people like you.
43:54
And to be able to put that into a hopefully palatable form where
43:58
this can be useful for us. And also, this time around,
44:02
useful for clients who may get a lot of value out of some of the
44:05
workflows, but who may also come around and go, oh, wait a
44:08
second, really. 40 percent, 40% from Reddit.
44:12
OK, maybe let's give this some thought.
44:13
We're another example of this is confidentiality.
44:17
Like I've talked to people, I'm not talking about translators.
44:19
I'm talking about people who hire translators, some of my own
44:21
clients. And they're like, Oh yeah, we
44:23
use like Gemini like all the time for stuff.
44:26
And I'm like, did you know that your data is going through the
44:29
US government, you know, because of the Patriot Act?
44:31
They have access to your, your content and they can, they can
44:35
vary it. They can basically prompt to
44:37
find what they want in your data and find everything you know,
44:40
like everything is just an open book.
44:42
So art, is that what? You want the Patriot Act?
44:45
Isn't that the bill that was passed after September 11?
44:48
Yeah, yeah, basically like it's still out there.
44:50
And you have to be really careful because my understanding
44:53
is that you're, when you're putting data into these AI,
44:57
because the US government is using it for intelligence
44:59
purposes, they can access your data.
45:02
And a lot of people don't know that.
45:04
Like they're sending out their data.
45:05
Yes, it's going to Google, yes, it's going to Microsoft or
45:08
whoever. But in many cases, it's
45:10
transiting through the US also, the US government has access.
45:12
And so that's also just something to keep in mind.
45:15
Sometimes businesses don't even realize they're breaking their
45:17
own rules, right? Like third, they're not
45:20
following their own rules because they're not aware, you
45:22
know, they're, they're using AI and they're, they just have not
45:25
been really properly guided on what's what this technology is
45:29
and, and where your info is going.
45:31
Well, well, the, the, the, the, the thing sometimes that I am
45:36
surprised about is that in our case, you were saying earlier
45:39
that the like Judge BT or any other AI doesn't understand
45:47
something, right? But it gives the illusion of
45:49
understanding, correct? And it's, it's really sometimes
45:54
hard to go against that illusion, like to really tell
45:58
people, listen, you may feel that it understands what you're
46:02
saying, but it's not because for most people, what matters is
46:06
basically the result. And some people say, well, what
46:08
do I care if it doesn't understand, right?
46:11
I, it did what I wanted in the end.
46:13
So how do you fight against that?
46:16
That's a great point. And in some cases, I don't fight
46:19
against that. I would say there are purposes
46:21
where AI can be really useful, genuinely useful.
46:23
And it's OK that it doesn't understand per se.
46:26
But I guess part of the way that we counter this on a technical
46:30
level is to say, like, what's the outcome that you want?
46:33
And often the outcome requires a human.
46:35
So I can be part of the equation, but you wouldn't go to
46:39
ChatGPT for your medical diagnosis, right?
46:41
You might, you might go to it for like to ask for opinion, but
46:45
you're probably not going to take it at face value and start,
46:49
you know, performing surgery on yourself because chat TV like
46:52
told you how to do it, right. It's kind of the same thing.
46:55
Like, yeah, the the technology really is impressive.
46:57
Frankly, it's amazing for a lot of things.
46:59
But it does not mean you can take humans out of the equation.
47:02
I think that's the wrong direction.
47:03
And businesses will shoot themselves in the foot if that's
47:06
the road that they start going down.
47:08
Yeah, Yeah. No, definitely like plan your if
47:11
you're gonna use it, like use it for your meal plan for the week,
47:14
but don't use it to maybe see how much profit you could make
47:17
in a week by by applying this into your own business.
47:21
Yeah, I find in particular with business advice, I've used it a
47:24
lot for that to to test different things, different
47:26
AI's. I've also tried some like pre
47:29
trained chat bots to for, for, for business advice.
47:33
It's pretty bad, you know, even if you do a few shot prompting,
47:36
even if you give it lots of information about your business,
47:39
it's pretty bad. So there are some.
47:41
Things you S is watching. Yeah, I should watch what I say.
47:47
When Eagle eye. Bald eagle eye.
47:50
Eagle Eye. But there there are ways that
47:53
you can use it like one of the things I teach in the book is
47:55
what I call interview prompting. So you stop going to AI as a
47:59
slot machine and instead you configure it to ask you
48:03
questions to draw out the best of yourself.
48:05
So if you do this right, this can actually be very effective
48:09
because you are going to give things your some thought and
48:12
think about it and then the AI will start to actually sometimes
48:16
produce useful output. But business strategy is really
48:20
interesting because even if you have lots and lots of data, it's
48:24
hard to even really say what's like, what's your criteria for
48:28
the right choice? Because as you know, it's not
48:30
just about making more money. Your business has to be
48:33
sustainable as well. You have values to abide by.
48:37
There are timelines. There are limits on your human
48:39
energy, on what you can do in a business.
48:41
You know, this is true for big businesses, just like small
48:44
ones. And AI is especially bad at
48:47
synthesizing that into something useful because essentially there
48:50
are so many. It's, it's so it's not like a
48:53
hard and fast thing like like we were talking about radiology,
48:57
like an X-ray diagnosis where certain markers have been found
49:01
over billions of data points. Certain markers can be
49:03
identified to be precancerous, right?
49:06
With business, there are so many variables.
49:08
And it's also a very personal thing, right?
49:10
The way that you you want to live and run your business might
49:14
be very different from Starbucks.
49:15
It's like while they pull the pulls apart.
49:18
And so I found that business is very revealing of the
49:21
limitations. But if you use interview
49:23
prompting, you can have it ask you smart questions and those
49:27
can help you to find better answers.
49:29
Not because the AI is giving you the right answer necessarily.
49:33
Often it will still come up with a bad solution, but it's useful
49:36
because it's helping you think better and to ask questions you
49:40
didn't ask before and maybe pointing some things out to you.
49:42
You can have it do market research and it's extremely good
49:45
at market research. So I find, again, I would never
49:48
go to AI to come up with the answers for me.
49:51
But if you use it strategically with interview prompting and
49:54
some other things that I I detail in the book, then it can
49:58
be an extremely useful sort of sparring partner for your your
50:01
thinking. Cool, looking forward to read
50:04
it. Likewise, yeah, now, yeah,
50:06
before we we end we didn't do this last time when we were
50:09
talking about a resilient, but acknowledgements, thank yous.
50:13
Do you have anyone in particular who you would like to to thank
50:16
for other than your in laws for for the creation of of your,
50:20
your two, your two bucks? Well, thank you guys.
50:24
Thank you for for helping me and, and helping me to get this
50:27
message out there, just like you did for my first book, like that
50:29
means the world to me. Also, I mean, I could take the
50:32
opportunity to thank my coach Michelle, who's Michelle is
50:36
helping me with the promotional part of it.
50:39
It's very important to get it into more people's hands.
50:41
I have Veronica Goldspiel who's going to be helping with some of
50:44
the editing processes, my colleague Ann Marie, as well as
50:46
helping with with some content to pieces and editing.
50:51
I have Roberta. Roberta where is currently doing
50:53
the manuscript editing, which is I'm so thankful to her.
50:56
It's a big job. I dictated the text, so it does
50:59
require like we're cleaning up in my first book did, and she's
51:02
doing an incredible job. And I'm currently working with
51:05
Jessica Anderson as well, a book coach who is helping with
51:07
different sides of things, different side of things, like
51:09
in terms of what to do once the book is out there.
51:13
Because it really does take a village like you need, you need
51:15
to have a team. I think like anyone can launch a
51:17
book. You can upload a book in like 2
51:19
minutes on Amazon. But if you want to reach
51:22
people's hearts and minds and you want to have a message and
51:25
you want to make a difference in the world, it really does take a
51:28
team to help you with different aspects.
51:30
I also have Zelja. I can't pronounce her name.
51:32
I'm not going to try. But Zeljka is my wonderful book
51:34
cover designer. I hire humans for everything for
51:36
this book. It's very important to me, of
51:38
course, in terms of my values. So yeah.
51:41
And I'm thankful to the missed team over at Mercy because
51:44
they've signed a website that goes with it as well.
51:47
So yeah, lots and lots of people who have helped me.
51:50
And I'm incredibly grateful because, you know, I think
51:54
publishing, like I see a lot of people on Facebook saying
51:56
publish your book in one click. And I'm just cringing so much
51:58
and I'm going, I don't, I would never want to read an AI written
52:01
book. I still believe that we should
52:03
do business with humans, all of us.
52:05
That's how we can succeed. That's how we can help each
52:07
other. It's how we can make a better
52:09
world. It's like going to the best of
52:11
what every person has to offer. And and to me that will never
52:14
change. Alright, that is a great way to
52:20
end the episode. It's a great message and not
52:23
only I look forward to reading it, but once we've read it,
52:28
we're going to bring you back for 4th time here on Fountain
52:32
Interpretations. So we can ask, we can do a deep.
52:35
Dive I would be so thrilled that anytime it's you know what
52:40
you're doing in the profession is very valuable.
52:42
You guys have a good reach and you've worked hard for it and
52:45
you're, you're helping our community to develop in a time
52:48
that's particularly difficult, right, with AI.
52:51
So like, kudos, really sincerely kudos for everything you guys
52:54
are doing. Thank you and for those who are
52:56
subscribed to our podcast, I forgot what the work we're
53:01
doing, but no for those who are who are subscribed to our
53:04
podcast will put all of it in the description.
53:07
Joe, keep us in informed as to when the if there's a soft
53:11
launch, a hard launch and any sessions and we'll just add it
53:14
to your description so people are aware of what's happening
53:17
and in the world of of our GPT. Wonderful.
53:23
Thank you so much, Brian. Thank you everyone OHS of
53:27
course, of course, it's always fun to to have you.
53:31
So everyone don't forget again, like share, subscribe and read.
53:38
People forget we have to read. We don't scroll.
53:40
Read a book. Buy a book.
53:43
Encourage an author. Don't read Judging beauty too
53:46
much, alright? Thank you.
53:48
I appreciate that. Alright, take care.
53:50
Alright guys, take care. Bye now.