September 8, 2023 • 25 mins
Have you ever wondered how AI is revolutionizing the world of food content creation? 🤔 Well, let me introduce you to an amazing story of how AI is helping designers and developers like William Sayer create mind-blowing photographs of food he's going to share with his friends!
00:02:47 BigMealShare: a platform to share meals, with food photos by AI!
00:05:33 Training AI to be a prompt engineer
00:07:51 Research shows 7 key elements for image prompts
00:11:50 No code tools
00:16:41 LEGO camera!
00:21:14 Social media reflects our desired personas visually
Follow up:
Instagram.com/will_straya
WillSayer.com
Instagram.com/bigmealshare
BigMealShare.com
Next mastermind:
To attend the next mastermind, go here: PromptEngineeringMastermind.com
Stay in touch on:
Youtube: youtube.com/@PromptEngineeringPodcast
Telegram: https://t.me/PromptEngineeringMastermind
LinkedIn: https://www.linkedin.com/groups/14231334/
Support the Show:
Prompts I sell on PromptBase: https://promptbase.com/profile/promptgreg (cover letter generator, event planner, etc)
If you sell ChatGPT prompts, get my analysis of all PromptBase's prompts: https://gregschwartz.gumroad.com Use coupon code "podcast" for 10% off!
Discover how technology is reshaping our lives and livelihoods.
Listen on: Apple Podcasts Spotify
Mark as Played
Transcript
Episode Transcript
Available transcripts are automatically generated. Complete accuracy is not guaranteed.
(00:00):
Welcome to the promptengineering podcast, where we
teach you the art of writingeffective prompts for AI systems
like chat, GPT, mid journey,Dolly, and more.
Here's your host, Greg Schwartz.
All right.
Hello, everyone.
So.
We'll go ahead and introduceyourself.
William Sayer (00:20):
hi, I'm William Sayer.
I'm a designer and a low codedeveloper.
I've been working in UX UI forthe last I studied that in
university and then I worked ina startup for a few years.
Now I'm working on my ownproject and I'm trying to
bootstrap a variety of toolstogether using some of the AI
(00:41):
suites particularly creatingimages going from text to image
generation.
And most of my experience iscreating images of meals.
However, I've been using it fora variety of other tools and I
play around with it a lot.
I have a little bit ofexperience just from trial and
error, I'd say
Greg screen (01:02):
So what got you into using AI?
William Sayer (01:05):
Yeah.
Yeah.
It's a great question, Greg.
I think originally it was the,I, the concept that you could
produce content.
Blogs are a huge way to drawtraffic into any website and.
Knowing that you can, likepretty much like incredible
content that will explain atopic for you that you may not
(01:28):
even understand.
It could summarize a variety ofinformation out there and,
distill that to you on a levelthat you may understand.
I feel like that as a UXdesigner similar to you, Greg,
it's that there might be a fewconcepts about.
I don't know why a framing thatI just maybe didn't have my head
(01:48):
around or certain things likethat.
And I felt that GPT, it was likean amazing tool that like I
could use it to basically expandand collapse any topics into
further or less detail.
Then manually looking up for it,like on the internet.
How do you use AI?
To go back to that blog sphere, it's, engaging
(02:10):
in freelance work.
And I've done a couple ofprojects for clients where it's
oh, you need to create a blog.
So it ideally attracts morevisitors to your website.
It's okay, I don't even know howto structure this blog.
Like what are maybe 10 differentpoints I could talk about?
Bam.
There we go.
ChatGPT has spat out 10different potential articles.
(02:32):
And then under each of those,you could say, Expand on one of
those points, or that's toocomplicated.
Refine that into, so that aneighth grader could understand
it.
So do you mostly do textgeneration?
Yeah.
Yeah, that's correct.
Early days, yeah, it was justchat GPT and now it's progressed
(02:54):
a little bit further and it'sgone from, it's gone to image
generation as well.
I feel like I was maybe.
Definitely text oriented first,and then now it's actually, wow,
there's so much more out therethat you can do rather than just
text.
A website that I've been workingon.
It's it's called Big Meal Share.
(03:15):
Basically, it's a platform tohelp you share meals with like
your friends and your family.
Say you live in an apartmentblock, one of the use cases
would be, hey, you can sharefood with your neighbors.
it's basically like a conceptwhere, I'd come home late from
work and think, oh, it's 7 p.
m.
and now I've got to cook dinnerfor myself.
(03:35):
Wouldn't it be amazing if,somehow I could sit at my
neighbor's table because I cansmell their food while I'm
walking up my apartment stairs?
So that's basically this projectI've been working on recently.
I'm sure there's a lot of kindof like legal implications and
risks.
And I don't know, I'm justtrying to get an MVP out there.
(03:56):
And as part of that to attractpeople to the meal that you're
proposing on the platform havingan image is a very crucial part.
And a lot of the times peoplehaven't yet cooked a particular
meal and have a photograph ofit.
So I've been using Dali tocreate basically four different
(04:18):
images for the user to choosefrom how that meal could look.
Oh, interesting.
Wait.
I thought those were photographsof the food.
You're saying the meal imagesare actually all generated from
AI.
Wow.
That's right.
Yeah.
Yeah.
Yeah.
So that was different.
Yeah.
Different.
That's something else that Iguess we'll talk about a little
(04:39):
bit later.
But yeah, no, some of the foodimages when I first started out
using Dali too, Oh my gosh, theyjust looked like.
Ooh, dog food, really.
And I think that was, yeah.
It's who's going to RSVP to thiswhen it looks like that.
I think something that reallyhelped improve my prompting on
(05:01):
that regard is like.
Studio lighting defining aspecific type of camera, maybe
Sony alpha defining a lens, justsaying what the aperture is.
And I feel like anyone who has abasic understanding of
photography that can come verysimply for them, whether you
want, aperture blur or thatbokeh in the background or does
(05:21):
image blurry or crisp.
Just a basic understanding ofphotography there goes along the
way towards the final image andhow it looks.
How do you build those prompts?
this was in chat GPT.
And the idea is that you'retraining it to understand, like
to take the role of a promptengineer.
(05:42):
And I know GPT wasn't reallytrained up until.
I think 2021, so I believethere's not a huge amount of
like information that has helpedit learn what a prompt engineer
is.
So this will often take a littlebit of refining.
So for example, sometimes itwon't spit it out as like a
(06:04):
paragraph, but you can train itbecause it's got memory in that
chat.
Basically saying, I want you,this is the formula for a
prompt.
And then say, change the formulato include this, and modify it
to include that.
I basically end up have, havinga back and forth for chatGPT a
few times, until you feel likeit's spitting out a reasonable
(06:25):
prompt.
And then if you're in midjourney or whatever, instead of
just writing teddy bear, Youcould then go over to chat GPT
and write a prompt about teddybear.
And then it spits out aparagraph, which then you can,
it almost, once you've got itset up in chat GPT for the same
amount of time input, you canget a huge, like really detailed
prompts that can often reallyimprove the results.
(06:48):
So those prompts are using shotprompting to say.
Oh, use a cannon.
F 2.0 lens with a polarizationfilter, things like that.
I'm still testing out and playing around I do
think it does work better withshot prompting like you
suggested.
However, I think there are somescenarios where shot prompting,
(07:10):
like the information that.
It's been trained on maybe thatshot prompting can be a little
too specific and maybe itdoesn't have, maybe there aren't
that many photographs on theinternet with a Canon five D
mark three using a 70 mil lensat aperture 5.
(07:30):
6.
It's like how many sure the verypopular cameras and lenses have
a whole bunch of information totrain on, but I feel like
sometimes it might take away andI'm still trying to figure out
when that's useful and whenthat's not.
What are the core parts of yourimage prompts?
Are there any specific thingsyou always think about.
(07:51):
I've done a little bit of research on that
and I think there's theunderstanding that I have is
that there's maybe seven thingsto touch on.
And so that's in my first promptthere, firstly, the subject.
Then you've got the medium, isit a line drawing, is it a
watercolor, is it a photograph,the environment the lighting,
(08:13):
the colors, the mood, and thecomposition, and sure, maybe
mood and lighting overlap alittle bit, but I feel like
having these words in therereally help.
The, especially if you're goingfrom chat GPT into mid journey
when you're creating that promptin chat GPT, for example, the
the composition, right?
(08:34):
I may say, give me an image of apirate.
And without me having to rack mybrain think what's a composition
of a pirate?
Instead of me having to type,okay, maybe a tri tipped hat
with a pirate on his shoulderwalking down a boardwalk.
I feel like that's all the sortof thing that takes time that
you could...
Pass over to a tool likeChatGPT, and it comes up with
(08:58):
that setting for you.
Spits out a paragraph that youcan then straight away put into
mid journey.
And that time to iteration justis so much faster and you can
get so many more images andyou're like, Oh, I noticed one
thing in this image that I lovesomething else in that image
that I love.
And then, if you're creating saya hero image for a website or
something where you need it tobe picture perfect, how you want
(09:21):
it, I feel like it's great for,you've got all those bits of
inspiration that can now cometogether.
Got it.
So it's describing all theparts.
So I'm curious, there's been alot of controversy around people
making images in the style of apopular artist.
How do you feel about that?
I feel like it's very popular for people to say
(09:43):
in the style of Monet or VanGogh or whoever it may be.
And I feel that brings up awhole new discussion of Do we
have their permission?
There are famous, they'refamous, like current living
artists who, do digital art.
And then if you say in theirstyle, if they put enough work
out there, sure, Midjourney canreplicate it.
(10:04):
But then I feel like there'sthat, you've got to feel that
maybe you've referenced thatartist.
If you're going to produce thatpiece of work or, I think that's
a blurred line that we're stillin the very early days.
We don't really know how toaddress it just yet.
But it's worth pointing out andto feel that like maybe I'm not
gonna specify an artist by nameto copy their style because
(10:27):
they've spent 20 years of theirlife figuring out how to write
in this how to draw in thisparticular style.
Do I feel morally okay to justcopy that?
So that's something worthkeeping in mind.
I definitely know what you mean.
Sounds like you don't useartists' names in your prompts.
I've definitely done that, but for my own
personal, just like curiosity, Ifeel like as soon as you start
(10:51):
making money from something,then if you've got that
permission, sure.
I heard about Grimes announcedthat if someone wanted to create
like an AI version of some ofher music, then she would
expect, she's happy for them togo ahead and do it, but as long
as 50% of the money that theymake.
that they give to her.
So I feel like at leastcontacting the artist and say
(11:13):
what you're using it for is animportant step if you're using
it commercially.
Oh, interesting.
Do you think that'll help orhurt her brand?
Good question.
I'm sure you've probably seen orheard of the A.
I.
Drake's, the A.
I.
Kanye's, and it's, it'd bereally interesting what it does
(11:35):
for their brands, like whetherthey allow other people to
create music, because it'salmost like It could potentially
help them grow, but yeah I don'tknow.
It'll be really fascinating tosee what happens.
So what AI tools do you wishexisted?
I use a variety of no code tools, right?
So Webflow is one of them, Figmais another I use Airtable.
(11:57):
I use a variety and if peoplearen't familiar with them that's
fine.
I feel like there's a whole,there's a whole suite of no code
builders that are available outthere these days.
And.
I feel like sometimes there's avery simple thing that I just
don't know how to do and if Iwould love to be able to
describe to some sort of AIagent how to use a particular
(12:21):
software for me.
So how it gathers that data I'ma little unsure, maybe if it did
screen recordings or say, forexample, Figma, maybe we can use
that for an example, because Ifeel like a lot of people in the
audience might know what Figmais it's and for those who don't,
it's just basically an onlinetool that you can draw
wireframes and digital designsto help, design for like
(12:45):
screens.
And so there's so many users outthere and I feel like maybe you
don't know how to use autolayout or maybe you don't know
how to use components orsomething like that.
I feel like if there was likesome sort of agent where you
could ask it like, Hey, how do Ido this?
And it could almost show youwhere the mouse is meant to go
along the screen.
How it gathers that data isanother interesting question,
(13:08):
but so I think something likethat for people who are new,
because I, there are so manytimes where I've been caught up
on one little thing and I knowif I was a professional, I'd be
able to do this in 10 seconds,but I'm sitting there for an
hour trying to figure out how todo it.
I think one of the big benefitsof AI is that it's democratizing
a lot of these services.
(13:29):
For anyone out there who's justgot like access to the internet,
I feel like what they didn'tknow before and what was only
possible by big companies oragencies 10 years ago, they now
have all this power so that allthese sort of small indie
hackers builders can bootstrap avariety of tools together and
make services that can addressniches that maybe there wasn't
(13:53):
the market value to addressbefore.
So I feel like we're going to befinding a whole bunch of these
software services popping up toaddress tiny little issues that
maybe we haven't been able toaddress before.
So hopefully we'll get a lotmore services coming up and
improving our lives because wecan, and there's people out
there to build those.
(14:13):
All right.
So want to talk about one of theprompts that you shared with me?
This is a little long, but Iwant to read through it So
capture the essence of innocenceand purity in a hyper realistic
portrait of a baby thattranscends reality.
Every minute detail ismeticulously rendered, revealing
the softness of the baby'sdelicate skin, the wisps of fine
(14:34):
hair and the sparkle in theirwide curious eyes.
The lighting is masterfullyemployed, casting, gentle
shadows that enhance thethree-dimensional quality of the
image.
While highlighting the subtlecontours of the baby's face.
The color palette is expertlybalanced with a natural yet
vibrant rendition that bringsthe portrait to life.
The mood of the photograph istender and captivating.
(14:54):
Evoking a sense of wonder andenchantment.
The composition showcases thebaby's captivating features as
the focal point whileincorporating elements that
accentuate their vulnerabilityand beauty.
This hyper realistic portrait ofa baby is a Testament to the
photographer's skill, artistry,and ability to capture the
fleeting moments of early life.
With astonishing precision it'sremarkable attention to detail
(15:15):
and emotional impact.
Make it a strong contender for aprestigious photography award.
wow.
That's a lot.
I'm particularly impressed witha couple of pieces of that.
Like the transcending realitypart.
How did you come up with that?
Yeah, I think you're exactly right.
That's exactly what I put intointo chat GPT baby after a few
(15:37):
back and forths with creatingthe mid journey formula that all
you have to do is put in babyportrait and then it comes out
with something like this.
I think when it says thattranscends reality, maybe there
are little bits that are lost onthe journey and it won't quite
be able to capture everythingthat's in that poetic
description and it might be alittle bit over the top.
(15:59):
But it certainly saves the usertime in terms of like for me to
write.
So I don't even know if I'mcapable of writing something
like that, to be honest, some ofthis is probably lost on mid
journey and you do want it to bea little bit concise, more
concise like this.
And when I was reading throughthat prompt, I was like, maybe I
(16:20):
need to refine the agent alittle bit more so it takes out
some of that language, but Ithink that just goes to show an
example of what something thatit can do.
Wow.
Okay.
It still seems long and a lot ofadjectives.
Definitely.
Yeah.
Do you think that's somethingyou should shorten?
(16:40):
Yeah.
Yeah, no, I think it, it totallywas.
And I feel that maybe a lot ofthose adjectives in there, like
meticulously rendered andsoftness and delicate.
I feel like that actually adds.
Some some objects that maybe youwouldn't expect.
And there's actually a prettyinteresting sort of like
(17:02):
scenario that I wouldn't havepredicted.
Something that happened beforewas I was experimenting with
these prompts and I was like,what will happen if I if I say
Lego as my input prompt to chatGPT, and so what it did is it
actually ended up spitting, Iwas it.
Described setting the apertureof the camera to one maybe F
(17:24):
stop for, and changing theshutter speed to this and this.
And I feel a lot of the languagethat it used in the prompt was
to do with the photographystyle.
And so the interesting thingthat Mid Journey.
spat out to me was it was a Legocamera.
But there were like hands thatwere like adjusting the, like
(17:47):
the settings of this Legocamera.
So I feel like it used a lot ofthat terminology to, it almost
included that into the image aswell.
That's not what I was going for,but it gave me an unexpected
image based on all that kind of,set up with the photography
stuff as well.
(18:07):
What other tips can you share towrite better mid journey
prompts?
Yeah, exactly.
Something that really upped mygame in mid journey was learning
about remix mode.
Basically what that means is ifit spits out like an image, you
can say remix, and then it likejust spits out the entire
prompt.
(18:28):
And then you can just fine tunea few little things in that
prompt.
And then it'll go and tweak thatusing those little changes.
So that was something that Iguess.
Really helps with mid journey.
And sometimes I feel like youcan use chat GPT to create a
whole bunch of creativescenarios and run those few, a
few times, but once you're happywith a direction that you want
(18:51):
to go down and find you insomething a little bit more,
maybe that's where you justchange a few of the words in, in
mid journey.
Oh, interesting.
Okay.
How does remix work?
Normally you've got like upscale one, two,
three, four, and then you've gotvariation one, two, three, four.
It'll also give you a ninthbutton that comes out there.
And then when you click on thatit basically uses the previous
(19:16):
four images that have come out.
And instead of just saying, giveme a variation of one of those
or upscaling, it'll now.
Use what you, it'll come up witha text input and you can put
certain things into that andthen that'll consider that and
apply it to those four differentimages.
I didn't learn about that untilrecently and I feel like the
(19:38):
space is always changing andit's it's just, it's really
exciting to hop on Twitter orYouTube and just every now and
then just see, just watchtutorials by other people, what
they've put out there and I feellike that's sometimes the best
way to stay on top of, knowingwhat works.
Got it.
So what projects have youshipped AKA, released to the
(19:58):
public?
Yeah good question.
I think the only product thatI've shipped it's still in its
MVP mode, but I think I wasdescribing big mail share
before.
And so it uses those prompts ituses, four distinct prompts to
generate images of mails.
And so that's where it's at themoment, and anyone can go on and
(20:21):
describe a meal, put in, yourdate, your location, yada yada,
you say who your friends are,and it'll send the invite to
them, and yeah, it'll come upwith these images.
Which used to look like dogfood, as I said, but now it's
including a few more of these,type of photograph and I think
award winning photography.
That's one of my favoritecatchphrases that I've been
(20:43):
putting at the end of thingsrecently, studio lighting, all
that sort of stuff and feel likethat.
That really helps it look like aprofessional grade, which is
almost, it sets a standard forwhat someone's mail needs to
look like now.
So I don't know if it's a goodthing or a bad thing, but it
definitely will get people RSVP.
(21:03):
So do you share the art you'vecreated on social media?
In fact, I could actually seeposting the meals to like
Instagram or Pinterest togenerate buzz for the site.
Is that something you've thoughtof?
Good point.
Good point.
Honestly, I've never been, I'venever really taken photos on
Instagram sort of thing.
At least at meals more I'm moreI just cool times in my life,
(21:26):
maybe if it's meals really good,I will.
But no, you, I feel likebringing that whole social
element back into things likeyou can see how you can if it's
a cloudy day, you can remove theclouds from the back of an
image, or if there's otherpeople on the beach, you can
just scrub those out of theimage.
It's, I think it's a reallyinteresting space where we're
(21:46):
going with all these socialmedia, because for some of the
younger generations, like a lotof the socializing is done
through these social medias,right?
For everybody.
And it's there, there've beenmany times where I've met
somebody and if I don't have thesame type of social media as
them, we just haven't clicked asmuch as.
Like someone who would I thinkhow what you post is a huge
(22:11):
visually.
It's like a huge, it shows whoyou are as a person and the fact
that we're able to like changeso many things now, it's is it
leading us down this rabbithole, taking us further and
further away from reality.
That's maybe one way to look atit.
Or it's like you're creatingthis cool internet persona.
That's like an artistic way toreflect how You know, you want
(22:34):
people to think of you.
I think there are pros and consof all these sorts of things.
And we're entering a whole newage of the unknown.
So yeah, it's, it'll beinteresting what happens.
It's been awesome having you onthe podcast.
What are some ways people cansee more of your art and follow
up with you?
The Instagram handle I use is will underscore
(22:55):
Australia, S T R A Y A, which islike Australia.
It's a bit of a play on theword.
That's where I'm from, you canprobably tell.
So yeah I, and I've recentlyactually been really getting
into Twitter.
So I, even though I don't reallyhave many followers, I've, as,
since I've moved to the BayArea, it's I feel like That's
(23:16):
where to get the news.
And I've never really used itbefore, but I've been getting
into it a lot recently.
And I feel like it's a whole newworld.
And I love it.
It's so interesting.
Sure.
There are some parts to bothInstagram and Twitter that maybe
considered toxic or whatever,but I feel like if you use it in
the right way, it can be a greatsource of information.
My personal website is willsayer.com.
(23:37):
And how about the other project?
You mentioned big Mealshare.
For the people that want to gettogether with their friends and
share meal.
So that's big meal, share.
com.
And then we're probably mostactive on the Instagram which is
simply to handle big meal share.
And then we've also got aTwitter and then I'll be working
(23:58):
on a Tik TOK for that soon aswell.
Yeah, thanks for the plug.
Awesome.
Thank you again for coming on.
It's been a lot of fun havingyou.
Thanks for coming to the promptengineering podcasts podcast
dedicated helping you be abetter prompt engineer I also
host masterminds where you cancollaborate with me and 50 other
(24:20):
people live on zoom to improveyour prompts Join us at
promptengineeringmastermind.
com for the schedule of theupcoming masterminds.
Finally, please remember to likeand subscribe.
If you're listening to the audiopodcast, rate us five stars.
That helps us teach more people.
And if you're listening to thepodcast, you might want to join
(24:42):
us on YouTube so you canactually see the prompts.
You can do that by going toyoutube.
com slash at prompt engineeringpodcast.
See you next week.
Welcome to the promptengineering podcast, where we
teach you the art of writingeffective prompts for AI systems
like chat, GPT, mid journey,Dolly, and more.
Here's your host, Greg Schwartz.
All right.
Hello, everyone.
So.
We'll go ahead and introduceyourself.
William Sayer (00:20):
hi, I'm William Sayer.
I'm a designer and a low codedeveloper.
I've been working in UX UI forthe last I studied that in
university and then I worked ina startup for a few years.
Now I'm working on my ownproject and I'm trying to
bootstrap a variety of toolstogether using some of the AI
(00:41):
suites particularly creatingimages going from text to image
generation.
And most of my experience iscreating images of meals.
However, I've been using it fora variety of other tools and I
play around with it a lot.
I have a little bit ofexperience just from trial and
error, I'd say
Greg screen (01:02):
So what got you into using AI?
William Sayer (01:05):
Yeah.
Yeah.
It's a great question, Greg.
I think originally it was the,I, the concept that you could
produce content.
Blogs are a huge way to drawtraffic into any website and.
Knowing that you can, likepretty much like incredible
content that will explain atopic for you that you may not
(01:28):
even understand.
It could summarize a variety ofinformation out there and,
distill that to you on a levelthat you may understand.
I feel like that as a UXdesigner similar to you, Greg,
it's that there might be a fewconcepts about.
I don't know why a framing thatI just maybe didn't have my head
(01:48):
around or certain things likethat.
And I felt that GPT, it was likean amazing tool that like I
could use it to basically expandand collapse any topics into
further or less detail.
Then manually looking up for it,like on the internet.
How do you use AI?
To go back to that blog sphere, it's, engaging
(02:10):
in freelance work.
And I've done a couple ofprojects for clients where it's
oh, you need to create a blog.
So it ideally attracts morevisitors to your website.
It's okay, I don't even know howto structure this blog.
Like what are maybe 10 differentpoints I could talk about?
Bam.
There we go.
ChatGPT has spat out 10different potential articles.
(02:32):
And then under each of those,you could say, Expand on one of
those points, or that's toocomplicated.
Refine that into, so that aneighth grader could understand
it.
So do you mostly do textgeneration?
Yeah.
Yeah, that's correct.
Early days, yeah, it was justchat GPT and now it's progressed
(02:54):
a little bit further and it'sgone from, it's gone to image
generation as well.
I feel like I was maybe.
Definitely text oriented first,and then now it's actually, wow,
there's so much more out therethat you can do rather than just
text.
A website that I've been workingon.
It's it's called Big Meal Share.
(03:15):
Basically, it's a platform tohelp you share meals with like
your friends and your family.
Say you live in an apartmentblock, one of the use cases
would be, hey, you can sharefood with your neighbors.
it's basically like a conceptwhere, I'd come home late from
work and think, oh, it's 7 p.
m.
and now I've got to cook dinnerfor myself.
(03:35):
Wouldn't it be amazing if,somehow I could sit at my
neighbor's table because I cansmell their food while I'm
walking up my apartment stairs?
So that's basically this projectI've been working on recently.
I'm sure there's a lot of kindof like legal implications and
risks.
And I don't know, I'm justtrying to get an MVP out there.
(03:56):
And as part of that to attractpeople to the meal that you're
proposing on the platform havingan image is a very crucial part.
And a lot of the times peoplehaven't yet cooked a particular
meal and have a photograph ofit.
So I've been using Dali tocreate basically four different
(04:18):
images for the user to choosefrom how that meal could look.
Oh, interesting.
Wait.
I thought those were photographsof the food.
You're saying the meal imagesare actually all generated from
AI.
Wow.
That's right.
Yeah.
Yeah.
Yeah.
So that was different.
Yeah.
Different.
That's something else that Iguess we'll talk about a little
(04:39):
bit later.
But yeah, no, some of the foodimages when I first started out
using Dali too, Oh my gosh, theyjust looked like.
Ooh, dog food, really.
And I think that was, yeah.
It's who's going to RSVP to thiswhen it looks like that.
I think something that reallyhelped improve my prompting on
(05:01):
that regard is like.
Studio lighting defining aspecific type of camera, maybe
Sony alpha defining a lens, justsaying what the aperture is.
And I feel like anyone who has abasic understanding of
photography that can come verysimply for them, whether you
want, aperture blur or thatbokeh in the background or does
(05:21):
image blurry or crisp.
Just a basic understanding ofphotography there goes along the
way towards the final image andhow it looks.
How do you build those prompts?
this was in chat GPT.
And the idea is that you'retraining it to understand, like
to take the role of a promptengineer.
(05:42):
And I know GPT wasn't reallytrained up until.
I think 2021, so I believethere's not a huge amount of
like information that has helpedit learn what a prompt engineer
is.
So this will often take a littlebit of refining.
So for example, sometimes itwon't spit it out as like a
(06:04):
paragraph, but you can train itbecause it's got memory in that
chat.
Basically saying, I want you,this is the formula for a
prompt.
And then say, change the formulato include this, and modify it
to include that.
I basically end up have, havinga back and forth for chatGPT a
few times, until you feel likeit's spitting out a reasonable
(06:25):
prompt.
And then if you're in midjourney or whatever, instead of
just writing teddy bear, Youcould then go over to chat GPT
and write a prompt about teddybear.
And then it spits out aparagraph, which then you can,
it almost, once you've got itset up in chat GPT for the same
amount of time input, you canget a huge, like really detailed
prompts that can often reallyimprove the results.
(06:48):
So those prompts are using shotprompting to say.
Oh, use a cannon.
F 2.0 lens with a polarizationfilter, things like that.
I'm still testing out and playing around I do
think it does work better withshot prompting like you
suggested.
However, I think there are somescenarios where shot prompting,
(07:10):
like the information that.
It's been trained on maybe thatshot prompting can be a little
too specific and maybe itdoesn't have, maybe there aren't
that many photographs on theinternet with a Canon five D
mark three using a 70 mil lensat aperture 5.
(07:30):
6.
It's like how many sure the verypopular cameras and lenses have
a whole bunch of information totrain on, but I feel like
sometimes it might take away andI'm still trying to figure out
when that's useful and whenthat's not.
What are the core parts of yourimage prompts?
Are there any specific thingsyou always think about.
(07:51):
I've done a little bit of research on that
and I think there's theunderstanding that I have is
that there's maybe seven thingsto touch on.
And so that's in my first promptthere, firstly, the subject.
Then you've got the medium, isit a line drawing, is it a
watercolor, is it a photograph,the environment the lighting,
(08:13):
the colors, the mood, and thecomposition, and sure, maybe
mood and lighting overlap alittle bit, but I feel like
having these words in therereally help.
The, especially if you're goingfrom chat GPT into mid journey
when you're creating that promptin chat GPT, for example, the
the composition, right?
(08:34):
I may say, give me an image of apirate.
And without me having to rack mybrain think what's a composition
of a pirate?
Instead of me having to type,okay, maybe a tri tipped hat
with a pirate on his shoulderwalking down a boardwalk.
I feel like that's all the sortof thing that takes time that
you could...
Pass over to a tool likeChatGPT, and it comes up with
(08:58):
that setting for you.
Spits out a paragraph that youcan then straight away put into
mid journey.
And that time to iteration justis so much faster and you can
get so many more images andyou're like, Oh, I noticed one
thing in this image that I lovesomething else in that image
that I love.
And then, if you're creating saya hero image for a website or
something where you need it tobe picture perfect, how you want
(09:21):
it, I feel like it's great for,you've got all those bits of
inspiration that can now cometogether.
Got it.
So it's describing all theparts.
So I'm curious, there's been alot of controversy around people
making images in the style of apopular artist.
How do you feel about that?
I feel like it's very popular for people to say
(09:43):
in the style of Monet or VanGogh or whoever it may be.
And I feel that brings up awhole new discussion of Do we
have their permission?
There are famous, they'refamous, like current living
artists who, do digital art.
And then if you say in theirstyle, if they put enough work
out there, sure, Midjourney canreplicate it.
(10:04):
But then I feel like there'sthat, you've got to feel that
maybe you've referenced thatartist.
If you're going to produce thatpiece of work or, I think that's
a blurred line that we're stillin the very early days.
We don't really know how toaddress it just yet.
But it's worth pointing out andto feel that like maybe I'm not
gonna specify an artist by nameto copy their style because
(10:27):
they've spent 20 years of theirlife figuring out how to write
in this how to draw in thisparticular style.
Do I feel morally okay to justcopy that?
So that's something worthkeeping in mind.
I definitely know what you mean.
Sounds like you don't useartists' names in your prompts.
I've definitely done that, but for my own
personal, just like curiosity, Ifeel like as soon as you start
(10:51):
making money from something,then if you've got that
permission, sure.
I heard about Grimes announcedthat if someone wanted to create
like an AI version of some ofher music, then she would
expect, she's happy for them togo ahead and do it, but as long
as 50% of the money that theymake.
that they give to her.
So I feel like at leastcontacting the artist and say
(11:13):
what you're using it for is animportant step if you're using
it commercially.
Oh, interesting.
Do you think that'll help orhurt her brand?
Good question.
I'm sure you've probably seen orheard of the A.
I.
Drake's, the A.
I.
Kanye's, and it's, it'd bereally interesting what it does
(11:35):
for their brands, like whetherthey allow other people to
create music, because it'salmost like It could potentially
help them grow, but yeah I don'tknow.
It'll be really fascinating tosee what happens.
So what AI tools do you wishexisted?
I use a variety of no code tools, right?
So Webflow is one of them, Figmais another I use Airtable.
(11:57):
I use a variety and if peoplearen't familiar with them that's
fine.
I feel like there's a whole,there's a whole suite of no code
builders that are available outthere these days.
And.
I feel like sometimes there's avery simple thing that I just
don't know how to do and if Iwould love to be able to
describe to some sort of AIagent how to use a particular
(12:21):
software for me.
So how it gathers that data I'ma little unsure, maybe if it did
screen recordings or say, forexample, Figma, maybe we can use
that for an example, because Ifeel like a lot of people in the
audience might know what Figmais it's and for those who don't,
it's just basically an onlinetool that you can draw
wireframes and digital designsto help, design for like
(12:45):
screens.
And so there's so many users outthere and I feel like maybe you
don't know how to use autolayout or maybe you don't know
how to use components orsomething like that.
I feel like if there was likesome sort of agent where you
could ask it like, Hey, how do Ido this?
And it could almost show youwhere the mouse is meant to go
along the screen.
How it gathers that data isanother interesting question,
(13:08):
but so I think something likethat for people who are new,
because I, there are so manytimes where I've been caught up
on one little thing and I knowif I was a professional, I'd be
able to do this in 10 seconds,but I'm sitting there for an
hour trying to figure out how todo it.
I think one of the big benefitsof AI is that it's democratizing
a lot of these services.
(13:29):
For anyone out there who's justgot like access to the internet,
I feel like what they didn'tknow before and what was only
possible by big companies oragencies 10 years ago, they now
have all this power so that allthese sort of small indie
hackers builders can bootstrap avariety of tools together and
make services that can addressniches that maybe there wasn't
(13:53):
the market value to addressbefore.
So I feel like we're going to befinding a whole bunch of these
software services popping up toaddress tiny little issues that
maybe we haven't been able toaddress before.
So hopefully we'll get a lotmore services coming up and
improving our lives because wecan, and there's people out
there to build those.
(14:13):
All right.
So want to talk about one of theprompts that you shared with me?
This is a little long, but Iwant to read through it So
capture the essence of innocenceand purity in a hyper realistic
portrait of a baby thattranscends reality.
Every minute detail ismeticulously rendered, revealing
the softness of the baby'sdelicate skin, the wisps of fine
(14:34):
hair and the sparkle in theirwide curious eyes.
The lighting is masterfullyemployed, casting, gentle
shadows that enhance thethree-dimensional quality of the
image.
While highlighting the subtlecontours of the baby's face.
The color palette is expertlybalanced with a natural yet
vibrant rendition that bringsthe portrait to life.
The mood of the photograph istender and captivating.
(14:54):
Evoking a sense of wonder andenchantment.
The composition showcases thebaby's captivating features as
the focal point whileincorporating elements that
accentuate their vulnerabilityand beauty.
This hyper realistic portrait ofa baby is a Testament to the
photographer's skill, artistry,and ability to capture the
fleeting moments of early life.
With astonishing precision it'sremarkable attention to detail
(15:15):
and emotional impact.
Make it a strong contender for aprestigious photography award.
wow.
That's a lot.
I'm particularly impressed witha couple of pieces of that.
Like the transcending realitypart.
How did you come up with that?
Yeah, I think you're exactly right.
That's exactly what I put intointo chat GPT baby after a few
(15:37):
back and forths with creatingthe mid journey formula that all
you have to do is put in babyportrait and then it comes out
with something like this.
I think when it says thattranscends reality, maybe there
are little bits that are lost onthe journey and it won't quite
be able to capture everythingthat's in that poetic
description and it might be alittle bit over the top.
(15:59):
But it certainly saves the usertime in terms of like for me to
write.
So I don't even know if I'mcapable of writing something
like that, to be honest, some ofthis is probably lost on mid
journey and you do want it to bea little bit concise, more
concise like this.
And when I was reading throughthat prompt, I was like, maybe I
(16:20):
need to refine the agent alittle bit more so it takes out
some of that language, but Ithink that just goes to show an
example of what something thatit can do.
Wow.
Okay.
It still seems long and a lot ofadjectives.
Definitely.
Yeah.
Do you think that's somethingyou should shorten?
(16:40):
Yeah.
Yeah, no, I think it, it totallywas.
And I feel that maybe a lot ofthose adjectives in there, like
meticulously rendered andsoftness and delicate.
I feel like that actually adds.
Some some objects that maybe youwouldn't expect.
And there's actually a prettyinteresting sort of like
(17:02):
scenario that I wouldn't havepredicted.
Something that happened beforewas I was experimenting with
these prompts and I was like,what will happen if I if I say
Lego as my input prompt to chatGPT, and so what it did is it
actually ended up spitting, Iwas it.
Described setting the apertureof the camera to one maybe F
(17:24):
stop for, and changing theshutter speed to this and this.
And I feel a lot of the languagethat it used in the prompt was
to do with the photographystyle.
And so the interesting thingthat Mid Journey.
spat out to me was it was a Legocamera.
But there were like hands thatwere like adjusting the, like
(17:47):
the settings of this Legocamera.
So I feel like it used a lot ofthat terminology to, it almost
included that into the image aswell.
That's not what I was going for,but it gave me an unexpected
image based on all that kind of,set up with the photography
stuff as well.
(18:07):
What other tips can you share towrite better mid journey
prompts?
Yeah, exactly.
Something that really upped mygame in mid journey was learning
about remix mode.
Basically what that means is ifit spits out like an image, you
can say remix, and then it likejust spits out the entire
prompt.
(18:28):
And then you can just fine tunea few little things in that
prompt.
And then it'll go and tweak thatusing those little changes.
So that was something that Iguess.
Really helps with mid journey.
And sometimes I feel like youcan use chat GPT to create a
whole bunch of creativescenarios and run those few, a
few times, but once you're happywith a direction that you want
(18:51):
to go down and find you insomething a little bit more,
maybe that's where you justchange a few of the words in, in
mid journey.
Oh, interesting.
Okay.
How does remix work?
Normally you've got like upscale one, two,
three, four, and then you've gotvariation one, two, three, four.
It'll also give you a ninthbutton that comes out there.
And then when you click on thatit basically uses the previous
(19:16):
four images that have come out.
And instead of just saying, giveme a variation of one of those
or upscaling, it'll now.
Use what you, it'll come up witha text input and you can put
certain things into that andthen that'll consider that and
apply it to those four differentimages.
I didn't learn about that untilrecently and I feel like the
(19:38):
space is always changing andit's it's just, it's really
exciting to hop on Twitter orYouTube and just every now and
then just see, just watchtutorials by other people, what
they've put out there and I feellike that's sometimes the best
way to stay on top of, knowingwhat works.
Got it.
So what projects have youshipped AKA, released to the
(19:58):
public?
Yeah good question.
I think the only product thatI've shipped it's still in its
MVP mode, but I think I wasdescribing big mail share
before.
And so it uses those prompts ituses, four distinct prompts to
generate images of mails.
And so that's where it's at themoment, and anyone can go on and
(20:21):
describe a meal, put in, yourdate, your location, yada yada,
you say who your friends are,and it'll send the invite to
them, and yeah, it'll come upwith these images.
Which used to look like dogfood, as I said, but now it's
including a few more of these,type of photograph and I think
award winning photography.
That's one of my favoritecatchphrases that I've been
(20:43):
putting at the end of thingsrecently, studio lighting, all
that sort of stuff and feel likethat.
That really helps it look like aprofessional grade, which is
almost, it sets a standard forwhat someone's mail needs to
look like now.
So I don't know if it's a goodthing or a bad thing, but it
definitely will get people RSVP.
(21:03):
So do you share the art you'vecreated on social media?
In fact, I could actually seeposting the meals to like
Instagram or Pinterest togenerate buzz for the site.
Is that something you've thoughtof?
Good point.
Good point.
Honestly, I've never been, I'venever really taken photos on
Instagram sort of thing.
At least at meals more I'm moreI just cool times in my life,
(21:26):
maybe if it's meals really good,I will.
But no, you, I feel likebringing that whole social
element back into things likeyou can see how you can if it's
a cloudy day, you can remove theclouds from the back of an
image, or if there's otherpeople on the beach, you can
just scrub those out of theimage.
It's, I think it's a reallyinteresting space where we're
(21:46):
going with all these socialmedia, because for some of the
younger generations, like a lotof the socializing is done
through these social medias,right?
For everybody.
And it's there, there've beenmany times where I've met
somebody and if I don't have thesame type of social media as
them, we just haven't clicked asmuch as.
Like someone who would I thinkhow what you post is a huge
(22:11):
visually.
It's like a huge, it shows whoyou are as a person and the fact
that we're able to like changeso many things now, it's is it
leading us down this rabbithole, taking us further and
further away from reality.
That's maybe one way to look atit.
Or it's like you're creatingthis cool internet persona.
That's like an artistic way toreflect how You know, you want
(22:34):
people to think of you.
I think there are pros and consof all these sorts of things.
And we're entering a whole newage of the unknown.
So yeah, it's, it'll beinteresting what happens.
It's been awesome having you onthe podcast.
What are some ways people cansee more of your art and follow
up with you?
The Instagram handle I use is will underscore
(22:55):
Australia, S T R A Y A, which islike Australia.
It's a bit of a play on theword.
That's where I'm from, you canprobably tell.
So yeah I, and I've recentlyactually been really getting
into Twitter.
So I, even though I don't reallyhave many followers, I've, as,
since I've moved to the BayArea, it's I feel like That's
(23:16):
where to get the news.
And I've never really used itbefore, but I've been getting
into it a lot recently.
And I feel like it's a whole newworld.
And I love it.
It's so interesting.
Sure.
There are some parts to bothInstagram and Twitter that maybe
considered toxic or whatever,but I feel like if you use it in
the right way, it can be a greatsource of information.
My personal website is willsayer.com.
(23:37):
And how about the other project?
You mentioned big Mealshare.
For the people that want to gettogether with their friends and
share meal.
So that's big meal, share.
com.
And then we're probably mostactive on the Instagram which is
simply to handle big meal share.
And then we've also got aTwitter and then I'll be working
(23:58):
on a Tik TOK for that soon aswell.
Yeah, thanks for the plug.
Awesome.
Thank you again for coming on.
It's been a lot of fun havingyou.
Thanks for coming to the promptengineering podcasts podcast
dedicated helping you be abetter prompt engineer I also
host masterminds where you cancollaborate with me and 50 other
(24:20):
people live on zoom to improveyour prompts Join us at
promptengineeringmastermind.
com for the schedule of theupcoming masterminds.
Finally, please remember to likeand subscribe.
If you're listening to the audiopodcast, rate us five stars.
That helps us teach more people.
And if you're listening to thepodcast, you might want to join
(24:42):
us on YouTube so you canactually see the prompts.
You can do that by going toyoutube.
com slash at prompt engineeringpodcast.
See you next week.
Popular Podcasts
Stuff You Should Know
If you've ever wanted to know about champagne, satanism, the Stonewall Uprising, chaos theory, LSD, El Nino, true crime and Rosa Parks, then look no further. Josh and Chuck have you covered.
Dateline NBC
Current and classic episodes, featuring compelling true-crime mysteries, powerful documentaries and in-depth investigations. Follow now to get the latest episodes of Dateline NBC completely free, or subscribe to Dateline Premium for ad-free listening and exclusive bonus content: DatelinePremium.com
On Purpose with Jay Shetty
I’m Jay Shetty host of On Purpose the worlds #1 Mental Health podcast and I’m so grateful you found us. I started this podcast 5 years ago to invite you into conversations and workshops that are designed to help make you happier, healthier and more healed. I believe that when you (yes you) feel seen, heard and understood you’re able to deal with relationship struggles, work challenges and life’s ups and downs with more ease and grace. I interview experts, celebrities, thought leaders and athletes so that we can grow our mindset, build better habits and uncover a side of them we’ve never seen before. New episodes every Monday and Friday. Your support means the world to me and I don’t take it for granted — click the follow button and leave a review to help us spread the love with On Purpose. I can’t wait for you to listen to your first or 500th episode!