7/7/2025 - Chain of Craft: Embracing AI in the Creative Process

Description

Chapters

Transcript

0:02 la.
0:08 [Music]
0:14 [Music]
0:36 [Applause]
0:38 Every time I see you now,
0:40 get that look in my eyes.
0:44 Every time I see your mouth,
0:48 hear that smile.
0:51 Early misty morning right after the
0:55 engine turning
0:57 the old f outside.
1:02 [Music]
1:05 You were leaving me
1:08 again today.
1:11 You will convince me
1:14 again today.
1:17 Leaving this heart town looking for
1:20 someone else's golden ring.
1:25 [Music]
1:27 So long Suzan
1:33 now don't cry.
1:36 [Music]
1:38 So long
1:40 I
1:44 don't you cry for me.
1:47 [Music]
1:53 Sharing jeans and cigarettes and keeping
1:56 warm out on the road.
1:59 Chasing down a lifestyle out on Highway
2:02 24.
2:04 New York State was a rolling breeze and
2:07 a sunshine with a blue sky falling. The
2:10 chill of old September cre
2:14 [Music]
2:31 looking for someone else's golden ring
2:37 should say so
2:40 Susanna
2:42 [Music]
2:45 now don't you cry.
2:47 [Music]
2:56 [Music]
2:59 Oh, Champy, you're such a good boy. Are
3:02 you a good boy or you big bad dog? You
3:05 fatty. Are you fatty? You fat. You You
3:09 had too many burgers. You You're a fat
3:11 burger boy, huh? Yeah.
3:16 Yeah.
3:19 [Music]
3:27 Well, I
3:29 heard there was a secret
3:33 David played and it pleased the Lord.
3:37 You don't really care for music, do you?
3:42 Well, it goes like this. The fourth, the
3:44 fifth, the manifall and major lift
3:49 king composing. Hallelujah.
3:53 [Music]
3:54 Hallelujah.
3:55 [Music]
3:57 Hallelujah.
4:00 Hallelu.
4:03 Hallelujah.
4:05 [Music]
4:17 Faith was strong.
4:22 You saw her bathing on the roof. Her
4:24 beauty and the moonlight overthrew you.
4:30 Tied you to the kitchen chair. Smash
4:34 your throat and cut your hair from your
4:37 lips. Shed your hallelujah.
4:40 [Music]
4:42 Hallelujah.
4:45 Hallelujah.
4:49 Hallelujah.
4:51 Hallelu.
4:54 [Music]
5:04 Does Satan wear a suit and tie? And does
5:06 it work at the Dairy Queen?
5:09 Does he listen and rock and roll? Does
5:12 he feed the mean? Singing hallelujah,
5:16 [Music]
5:17 hallelujah,
5:20 hallelujah,
5:22 hallelujah.
5:24 [Music]
5:26 Source camp, I understand that you're on
5:28 time. Congratulations. One of the rare
5:30 moments.
5:32 One of the rare moments. Kelly Camp
5:36 showed up at the uh
5:40 at the uh what do they call it?
5:43 A religious festival. The
5:48 what's it called? No, not Burning Man.
5:52 No, like a a revival. At the revival.
5:55 We're We're here at the AI revival
5:57 preaching the gospel. Good lord people.
6:00 I have been getting my ass. Oh, Source
6:04 Camp is hiring. Okay. Prediction,
6:07 prediction,
6:08 prediction. If Source Camp is hiring, if
6:11 Kelly Camp is hiring because two and a
6:13 half years ago, she started an AI agency
6:15 because she was like, "Uh, I think this
6:17 is going to be a [ __ ] thing." And she
6:20 [ __ ] ground it out for what, a year
6:21 and a half where people were like,
6:22 "That's cute."
6:24 So, if she's hiring,
6:27 that probably means she got a job or
6:29 two, more than she can handle. So,
6:32 awesome prediction.
6:37 You think she's going to be reading
6:39 resumes?
6:44 You think she's going to be
6:48 going on LinkedIn and saying, "I'm I'm
6:50 hiring."
6:53 Nope.
6:56 What she's likely going to do, my
6:57 prediction is she's going to hire
6:59 someone from the AI salon community or
7:02 this community, the irregulars here, or
7:07 someone in one of the other circles she
7:09 travels in, one of the communities she
7:10 travels in.
7:12 And you know she who she's going to hire
7:16 someone that she knows, someone that she
7:18 trusts, someone that's got integrity,
7:21 someone that's got a [ __ ] clue about
7:23 how to do all this AI stuff.
7:28 So, one of the things you should do if
7:30 you're new here, there's not that many
7:32 people here, so it's probably mostly
7:33 it's mo it's it's probably mostly
7:35 saloners in here right now, but if
7:37 you're new here and someone who knows
7:39 manufacturing, there you go.
7:42 Um, so there might be a vertical
7:45 expertise to it, but she's going to need
7:46 someone that that knows what they're
7:47 doing with this AI stuff. Um, you got to
7:51 get your ass in community. You got to
7:53 get your ass as to be part of the AI
7:56 salon to show up for things like this
7:58 and events like after hours and things
8:00 that we do.
8:04 Um,
8:06 so one of the things going on today,
8:09 yeah, speaking of LinkedIn,
8:11 um, so,
8:14 so it's meltdown Monday. Appropriate.
8:19 So I think it was Friday. Friday or
8:22 Saturday? No, it was it was on the
8:23 weekend. on the weekend because I know
8:25 how to make good life decisions, I
8:27 decided to write a LinkedIn post,
8:30 an article, and because because I've had
8:33 this this this burning idea in my head
8:37 called Chain of Craft is what I came up
8:40 with.
8:42 And what Chain of Craft is is
8:48 there's kind of
8:51 there's two ways to use AI, right?
8:53 There's the lazy way. You push a button
8:55 and out comes slop.
8:57 And then there's
8:59 the considered way where you're actually
9:01 crafting,
9:03 you know, multiple steps and you're
9:05 stitching different tools together. No
9:08 screen share. Yeah, I don't have a
9:09 screen share right now. I I'll I'll
9:11 share. Thanks.
9:13 Um
9:15 Oh, yeah. Maybe I can go share my I'll
9:17 go share my article.
9:21 [Laughter]
9:24 So,
9:26 so there's my little article. There's my
9:28 little There's my little masterpiece,
9:32 right?
9:34 And and what the article basically says
9:36 is
9:37 if someone asks you
9:40 what AI tool did you use for that?
9:42 They're kind of diminishing de they're
9:44 diminishing whatever you might have done
9:47 whatever the chain of craft is. They're
9:48 diminishing it into oh you just pushed
9:52 the simple button because you used AI.
9:56 Um how many comments do I have here? I
9:59 have 181 comments, five reposts, and 47
10:02 reactions. So, my LinkedIn posts
10:05 generally get I don't know, I a couple
10:07 of likes from some irregulars, an
10:10 occasional repost, and never any
10:12 comments. Occasionally, it's pretty
10:14 cool. Uh 181 comments. Most of the
10:19 comments, well, there's a bunch in there
10:21 that are my reactions, but most of the
10:24 comments are
10:26 um
10:28 just abject.
10:32 No room for argument. If you use AI, you
10:35 are a thief. You are an evil person. You
10:37 are an untalented, uncreative
10:40 person. Even if I only used AI for one
10:44 little tool in a 27
10:47 se 27st step sequence, I you if if I
10:51 used AI in one of them, it's oh, it's
10:53 just garbage. You're just a thief.
10:55 You're condoning the use of this thing.
10:58 Um,
11:00 and so it's just it's just fascinating
11:02 and it's, you know, and there's there's
11:03 no
11:05 like I try to make logical arguments,
11:07 but there's no arguing. Like it's just
11:09 it's a very binary thing. Paint pin. I
11:12 don't think someone asking about
11:14 tools automatically makes it negative.
11:16 No, I I don't either. Tool. Uh, Pate, I
11:18 I don't either. I
11:22 But but but if if the answer if the
11:25 answer is just so if you if you put
11:29 together a video, for example, Pate, and
11:34 and they and they say, "Hey, what tool
11:36 did you use?" In the end, you you might
11:38 have used whatever uh Lumal Labs, right?
11:42 That was the one that worked best for
11:43 you. But leading up to that was you had
11:48 an idea for for a story, you wrote a
11:50 script, you came up with visual
11:52 concepts, you maybe used an image
11:55 program or two, you used 11 Labs for
11:58 sound effects, things like that. So it
12:01 this was more like like take ownership
12:04 of your chain of craft that when someone
12:07 asks you a question that kind of implies
12:10 a simple answer, don't just feel like
12:12 you have to give the simple answer like
12:14 talk about your chain of craft. So, the
12:15 whole point of the article was if if
12:18 you've got this if if you've got a a
12:20 good complicated um creative process to
12:24 how you got to something, whether it's
12:26 an application that you made or an image
12:27 or a movie or whatever it is, share
12:30 that. Um, and that was the point of the
12:32 article, but boy, good lord,
12:35 Vicki is offering to come to my defense.
12:41 I can't shrink my video down enough to
12:43 post it. Oh, did you do a video, Vicki,
12:46 of it? That's awesome.
12:53 Um, no, you you know what hit me, Pate,
12:56 is like part of part of why I wrote the
12:59 article was
13:06 like I fell into the trap of when
13:09 someone would say, "Oh, what tool did
13:11 you use?" I would answer them. And I
13:14 would just sort of say what was the most
13:15 prevalent tool, right? But I wouldn't
13:18 talk about
13:20 that this wasn't just push a button and
13:22 squirt [ __ ] out because I think there's
13:24 there's an educational element in that
13:26 that if you if someone says what tool
13:28 did you use and you say hey there there
13:30 were 17 steps along the way like which
13:33 tool would you like to talk about or
13:35 which process would you like to talk
13:36 about? Because one of the most important
13:38 tools I used was the whiteboard. Huh?
13:41 Right. like like that's not an answer
13:43 they're expecting, right? And I think
13:45 it's something what I found is that I
13:47 was diminishing. It wasn't I was
13:49 diminishing. I was not
13:53 consciously
13:56 thinking about all of what I did to
13:57 produce something. I had kind of uh
14:02 dismissed
14:04 my human contribution in that chain of
14:07 craft. So that was why I wrote the
14:09 article. I thought it was just going to
14:11 be a helpful little thing that no one
14:13 would give a [ __ ] about. Oh, do they
14:15 give a [ __ ]
14:18 Oh my god.
14:22 So, if you want if you want some
14:23 entertainment, go jump in the comments.
14:25 And listen, I tried to defend myself a
14:27 couple of times and then they went off
14:28 and did research and everything I did,
14:30 they refuted and they said, "Uh, it just
14:32 comes back to if if you're talking about
14:34 using AI, you're a thief and you're just
14:36 stealing from us." And like some of the
14:38 people are like they're like talking to
14:40 me like I [ __ ] invented chat GPT or
14:43 midjourney. I'm like, "Hey kids, I I
14:46 just I just used this [ __ ]
14:51 If you use it, you're a thief." Okay.
14:54 All right. If uh if the medical
14:57 establishment cures cancer, we we
15:00 shouldn't take that cure for cancer,
15:02 right? Because they just they just
15:04 trained all those medical models on all
15:07 that PhD research that they stole.
15:09 Right. Right. Cuz it's the same thing,
15:13 right? Oh, no. They can do medical
15:15 stuff. That's fine. Ah, I see. I see.
15:18 So, all the all the data analysis, they
15:21 shouldn't do that because they just
15:22 stole all the all the work that data
15:24 analysts did over No, no, that's fine.
15:26 Okay. But so, it's just when it hits
15:28 your backyard that it's a problem. I
15:31 see.
15:33 So anyway, absolutely crazy. Um, how's
15:36 everyone doing? I hope you all had a
15:37 good FTH. Had a good fourth fall. Um, I
15:43 don't know what we'll do tonight. I
15:44 think. Um,
15:48 [Music]
16:06 you know what's funny?
16:08 given gi given that I just had my ass
16:12 handed to me. Like the thing that the
16:14 the tool that I continue to be enamored
16:16 with right now is Midjourney.
16:20 Um,
16:23 and you know, it's kind of fascinating
16:24 because
16:27 if you put together mood boards of your
16:29 work,
16:31 you can get MidJourney to to behave in a
16:34 way that it very much mimics your work.
16:36 Right? So, I've got I've got this mood
16:39 board of my sky photography that I've
16:42 done over 20, you know, two decades.
16:48 And I've got another one for street
16:49 photography
16:51 and another one for architectural stuff
16:54 and another one for close-ups, macro
16:56 stuff.
16:57 And
17:02 if I really wanted to,
17:05 I could use this as a tool to
17:08 explore all different kinds of
17:14 imaginations of my work, right? Where
17:18 I'm not really leveraging other people's
17:21 work. But that's not true, right? the
17:23 core model
17:25 has been trained on all this stuff. Like
17:27 one thing I don't disagree with, I agree
17:31 that all of these models were trained on
17:36 [ __ ] that they didn't pay for. I
17:38 absolutely agree with that. I absolutely
17:41 agree that it is shitty
17:47 and it's not going away and now we have
17:50 these tools. And then if it's not going
17:53 away and we have these tools,
17:56 then you only have two options.
18:00 You can deal with it or not.
18:03 And so, you know, what I'm in the throws
18:05 of right now is is interacting with a
18:07 lot of people who are not only choosing
18:10 not to deal with it, they are
18:12 vehemently,
18:14 you know, they're going to go down
18:15 swinging that this will never be okay.
18:18 It will never be acceptable.
18:23 I don't know that that's a sustainable
18:25 position
18:27 because
18:28 and this was on LinkedIn. Like if this
18:30 were on Facebook, I'd get it. Facebook
18:32 is full of people that just don't give a
18:35 [ __ ] about AI or care about it. But
18:37 LinkedIn, like a lot of the people that
18:40 I was looking at are creative
18:41 professionals, right? That have like
18:43 creative professional jobs.
18:46 And if they're refusing to use it,
18:48 they're going to end up get hosing
18:50 themselves and and getting blasted out
18:52 of those creative jobs by a bunch of
18:55 little snot-nosed, you know, 20somes
18:57 that learn to use AI or some, you know,
19:00 pissed-off Gen Xers that are like, "This
19:02 is the coolest [ __ ] thing ever. Move
19:04 over, Skippy."
19:07 You know? So, but I couldn't like I
19:10 couldn't even make that argument. It's
19:11 just like you can hate it all you want,
19:14 just learn it. learn what these things
19:16 are. No, no d no room for dialogue. It
19:19 is evil.
19:23 Which is kind of fascinating. I just
19:25 it's like I just don't I think because
19:28 I've been swimming in these waters for
19:30 so long where it's like it's clear to
19:33 it's you know it's clear to us what its
19:35 value is. Um, and that it is absolutely
19:39 a craft to understand how these tools
19:42 work, which ones to use when, how to
19:44 manipulate them to give you what you
19:46 want, how to mix your own voice into the
19:49 prompts so that what you're getting
19:51 isn't derivative [ __ ] but it is
19:53 actually interesting and unique and
19:55 original.
19:57 But there is like there is no
19:59 possibility in in their minds that
20:02 anything original can come from this
20:04 thievery machine.
20:08 [Laughter]
20:11 Oh, good lordy. Good lordy. Good lordy.
20:14 Good lordy. All right.
20:22 [Music]
20:31 What's up, Mr. It?
20:35 9,000 layoffs at Microsoft. Oh, the the
20:39 hits are going to keep on coming.
20:46 What are we doing? Hey, Henry from Henry
20:48 from Brazil. Welcome, Henry. Um, I don't
20:52 know what we're going to do tonight. I
20:53 wish it would steal ideas from me. So So
20:56 Kelly, what's going on with uh what
20:58 what's hiring? Who who you hiring? What
21:00 are you looking for? Let's uh we'll get
21:02 the word out for you for the folks that
21:04 are here. Um if anyone has any uh any
21:10 questions about AI, thoughts about it,
21:12 if anyone's here who read my article and
21:15 stumbled over here to say, "Let me see
21:16 what this asshole's all about." Jump in
21:19 like like you know, let me know your
21:20 thoughts. I'm happy to happy to try to
21:22 engage.
21:24 Now hiring elite sales partner
21:27 manufacturing focus. Okay, cool. All
21:30 right, so Kelly's looking for
21:33 someone who can walk and talk and shoot
21:34 gum at the same times, understands the
21:36 manufacturing sector,
21:39 and doesn't have their head up their ass
21:40 about AI. There you go. Oh, and isn't an
21:43 [ __ ]
21:45 right? Has integrity, shows up, does
21:49 professional [ __ ]
21:52 commissionbased sales partner. Lots of
21:54 opportunity. Very very cool.
22:06 Mr. Te's looking for work. He's got to
22:08 stay home.
22:22 [Music]
22:31 Hello.
22:32 [Music]
22:38 Hallelujah.
22:41 [Music]
22:55 Yeah.
22:56 [Music]
23:06 Heat.
23:09 [Music]
23:27 Oh. Uh, Henry from Detroit was asking,
23:29 "What are we going to do tonight?" I
23:31 don't know. Um, I'm happy to do
23:33 anything.
23:36 my my gravitational pull keeps going
23:39 back to midjourney.
23:42 And this is very weird for me. If you
23:44 know this channel,
23:47 when chat GPT first came out, it was all
23:49 chat GPT for probably six months. We
23:52 didn't do almost anything else other
23:54 than that. So, there is precedent in
23:56 this channel for me just sticking with a
23:57 single tool. Um,
24:02 I find it fascinating that that Oh,
24:05 yeah. Oh, by the way, um, if you're here
24:09 because I told you to be here because
24:10 tonight at 9:00 Mountain time, so in in
24:14 half an hour,
24:16 Grock 4 is dropping.
24:18 I didn't pay attention to the day.
24:22 Elon Musk said it was dropping. I
24:24 assumed it was tonight because normally
24:26 when he drops tech, he sort of announces
24:29 it that day and drops it that night.
24:31 It's Wednesday. So, if you're just here
24:34 because I told you to be here,
24:39 welcome. Welcome. Uh, this is the
24:42 madness. The madness here, the nickname
24:44 of the channel is Chad. I will
24:46 occasionally finish a sentence. I have
24:49 no qualifications. If you go read the
24:51 comments in that LinkedIn article, it
24:53 will be clear to you I have I I am
24:55 beyond unqualified. I am a scourge of
24:58 the earth
25:00 for daring to say that you could
25:03 possibly declare anything you make with
25:05 AI a creation, not a theft. Anyway,
25:10 Michael Laut, how do you get midjourney
25:12 to replicate your product image? Always
25:14 changes it, especially bottles. Uh, you
25:17 you kind of don't, Michael.
25:20 um you you can do it and and it's got
25:22 this thing called omni reference now um
25:25 which we could go play with it. I
25:27 haven't played with a ton of Omni
25:29 reference. Um,
25:32 but what I would go play with, what I
25:35 would go look up, Michael, is um, flux
25:39 context.
25:41 Flux ko n
25:44 xt.
25:46 Um, go look up flux context. The other
25:50 thing you can use is runway ML has a new
25:54 thing called elements. I think it's
25:57 called elements or ingredients. I think
25:58 it's called elements. And you can put
26:00 three different images. So you could put
26:02 like a location, you could put two
26:04 different actors, and then you could put
26:06 a product shot. And then you could say
26:07 the woman is holding the bottle of
26:09 perfume or whatever. Um, and it'll make
26:12 that image. And then you can turn that
26:13 image into a video. So I would say try
26:16 RunwayML's
26:18 elements feature. and I can show how to
26:20 use that. Uh we can we can try
26:22 midjourney um omni reference but like
26:25 it's it ain't great but flux context is
26:29 remarkable and it's in a bunch of tools
26:33 like it's in a bunch of image generation
26:35 tools. Um so I don't know who's been
26:37 playing with flux context. Midjourney
26:40 isn't good for product images or text.
26:42 Yeah. Yeah. Midjourney. Midjourney is
26:44 clearly going for we want a model that
26:48 understands
26:50 like the immersive world, a world that
26:53 you will enter visually. They don't
26:55 necessarily care about text and product
26:57 demos and [ __ ] like that. They want it
26:59 to be like art. Um,
27:04 and they're and it's it's really good at
27:06 that. And I think I'm the more I'm
27:08 learning about what where MidJourney is
27:11 headed, I think the reason that their
27:13 their image generations are just more
27:16 artistic than others, I don't think they
27:18 were just trained on images. I think
27:19 they were trained on 3D worlds and games
27:22 and movies and all sorts of [ __ ] So
27:25 anyway,
27:27 um
27:30 All right.
27:35 [Music]
27:41 Um,
27:42 oh, Source Cam, do me a favor. Can you
27:45 connect with Tobias? Um, he's got
27:48 something for um, Serena's husband. Um,
27:52 I don't have his contact info. And if
27:55 you do, um, or if you know someone who
27:57 does, um, can you connect with Tobias to
28:00 arrange all that? Thank you.
28:06 I've had Flex Context open in my browser
28:08 for many days but haven't played with
28:10 it. Maybe we'll go play with that
28:11 tonight.
28:14 So, what would you use to make product
28:16 videos? See, it's such a spec Oh, you
28:19 sent the flower. Vicki sent the flowers.
28:21 Okay, great. Terrific. So, Vicki, could
28:23 you connect with Tobias and just give
28:24 contact information to him because he's
28:26 got the the money that was raised from
28:27 the stickers he wants to send over. Um,
28:32 okay. Um, let's see. Where are we going
28:34 to go?
28:38 Uh,
28:40 let me go ask chat GBT for a second
28:42 here.
28:48 [Music]
28:52 Does he always sing like that? That's
28:54 rather annoying.
28:57 This is what you know what is this sort
28:59 of nondescript singing world.
29:10 All right, dear Chad GPT, why don't you
29:14 research
29:17 um
29:19 the image
29:23 and
29:26 video generation tools that have
29:31 incorporated
29:35 I see Tik Tok question there Brandon
29:37 have incorporated
29:38 Flux F L Ux context K O N T X T into
29:45 their
29:48 um
29:49 tools slash
29:52 application
29:58 suites. All right, let's see if we can
30:00 get a list there. All right. Uh Iman,
30:03 what's the best vibe coding tool? Okay,
30:05 Iman, great question.
30:12 It depends on your level of geekitude.
30:16 Um,
30:17 if if you're an actual coder and you
30:21 just want to experience it, something
30:23 like cursor or something like replet
30:26 agent um where you still have to kind of
30:29 set up environments like replet agent is
30:31 a little more friendly. Um,
30:35 those are biggies.
30:37 Um,
30:39 you can also just use Claude AI's uh
30:43 artifact, right? So, you just ask Claude
30:46 to make you an application and it will
30:48 make you an application in the artifact
30:50 window. Now, those tend to be web apps.
30:53 the one the the Vibe coding app. If
30:55 you're not a coder, if you're if you
30:58 don't understand setting up databases
30:59 and web servers and and all that sort of
31:02 happy horseshit, authentication schemas,
31:08 if if you're not excited about the
31:10 plumbing,
31:12 then the the absolute number one vibe
31:14 coding tool right now is lovable.dev.
31:18 Lovable.dev.
31:22 deev.
31:23 Um, it does really decent design. It
31:27 does very good coding. It just will
31:32 make you [ __ ] Um, Brandon has a lovable
31:36 story. Look what I made. Yeah, happy to
31:38 happy to have you share that if you want
31:39 to if you want to hop up. Um, yeah.
31:43 Yeah. So, the this weekend we went out
31:45 of town and we were at the beach and um
31:48 my kids were digging in the sand and
31:50 they invented a game where they were
31:51 trying to keep zombies from digging up
31:54 through the sand and they were throwing
31:55 sand at them to beat them back down. And
31:58 so, uh later on that evening, I thought,
32:01 well, that's an idea for a game right
32:03 there.
32:04 Sure.
32:05 I gave uh lovable one prompt that
32:09 basically described the game and said
32:12 make this a game and it is posted in the
32:15 salon under look what I made. Um
32:16 Oh, beautiful. All right, hang on. Let
32:18 me go let me go grab it. Um okay, so so
32:21 and this is this is actually a really
32:23 good example
32:25 of what I was talking about
32:29 with my chain of craft prompt. AI
32:32 salon
32:34 dot and then Okay, that's that.
32:40 I've got the URL on the screen if you
32:41 actually need it.
32:43 Okay.
32:44 Um,
32:46 look what I made. All right. And where I
32:48 am right now is I'm Oh, yeah. At the
32:50 salon.ai. If if you go there and then
32:52 click on join our community, that'll
32:54 take you to where I am. So, so here's
32:56 the sand zombie smashdown
33:00 game. Start game. Okay. So, what do we
33:02 do here? Z zombies are digging up from
33:04 the sand. Click anywhere to make the
33:07 nearest kid throw the sand at the
33:09 zombies. Beautiful. You have a 100 sand
33:11 throws. Um, hit send buckets that fall
33:16 every 26 zombie to refill. Okay, so
33:19 we're going. Okay, so
33:22 all right, there's sand being thrown at
33:24 that zombie. Now, do I click on this guy
33:25 and go here?
33:28 Yeah, there we go. I see. So, this is
33:30 like this is like a beachy a beachy
33:33 space invaders. I like it.
33:35 Yeah. It's uh it's DigDug meets Missile
33:38 Command for those of you back in the
33:40 era. I lost a dude. I lost a child.
33:42 That's
33:43 You're not doing very good.
33:44 Oh, no. Beach defense. I did bad. Okay.
33:47 So, now here, let me tell you. And And
33:50 so, wait. So, the kids invented this
33:52 game on the beach. You got home and you
33:55 said, "Oh, this would be a cool game."
33:56 You made this. I assume you showed it to
33:58 them and they've played it.
34:00 Yes, they now enjoy playing that more
34:02 than they do at the beach, which
34:03 presents other challenges.
34:06 But let me let me take this as a let me
34:08 take this as a as a as a context
34:12 providing moment where
34:15 um
34:18 if we go back to chain of craft part of
34:21 the point I was making in the chain of
34:23 craft article was
34:27 Brandon
34:29 his his kids inventing that game on the
34:32 beach
34:34 was the idea.
34:36 Brandon
34:38 because he knows
34:40 lovable.dev dev exists
34:44 and he knows that he can make a video
34:46 game without having to know coding.
34:49 Said, "Huh, that would be an interesting
34:51 video game." He sort of parked that in
34:53 his mind. Probably cleaned up some
34:55 sticky [ __ ] on the way home after the
34:57 beach, right? Clean clean some sticky
35:00 [ __ ] out of the car after the beach run.
35:04 and then sat down and said, "Now, let me
35:07 make this." And he put in a prompt and
35:09 made the thing. So, this wasn't just him
35:12 making AI slop.
35:14 This was a single prompt that created
35:17 this game. And I'm sure he did some
35:18 iteration of it.
35:22 But
35:24 human expression, the ideas for what you
35:27 want to create in the world are all over
35:29 the place. And that now that the game is
35:31 made, his kids now play the game that
35:34 they invented on the computer
35:38 the same day it happened.
35:45 You know,
35:55 I'm just trying to I like I I I am
35:59 I'm kind of in this in this
36:03 flumxed state
36:07 where there is such vitrial
36:10 over
36:12 one one one
36:21 element
36:23 of
36:25 how all of this AI stuff's possible. And
36:28 it's an important one, right? These
36:30 models were trained on data that the
36:32 companies didn't pay for. I absolutely
36:35 acknowledge that.
36:38 Um,
36:43 but the the amount of good that that
36:46 this technology is going to be able to
36:48 do in society is is [ __ ] massive. And
36:52 the amount of opportunity right now is
36:54 [ __ ] massive.
36:56 And
36:58 like it or hate it, AI ain't going
37:01 anywhere. It's here.
37:05 So if you're worried about your
37:06 livelihood,
37:08 you know, fighting a fight that was lost
37:11 four years ago,
37:14 I like I just I don't get it. I don't
37:16 get I don't get why learning new skills
37:21 like you can still be pissed off at this
37:23 stuff. The other thing that you can do
37:24 with AI, and I I talk about this a lot,
37:32 you've got creative product and you've
37:34 got creative process,
37:37 and then you've got business product and
37:39 you've got business process.
37:41 So, I think a lot of the things that the
37:43 artists are most pissed off about is the
37:46 creative product.
37:48 They're saying if you make images,
37:50 you're a thief.
37:52 Okay. Okay. Let's let's give you that. I
37:55 will never make images again. Or you as
37:57 an artist can choose. I'm never going to
37:59 use AI tools to make images because I'm
38:02 an artist and [ __ ] all that other stuff.
38:04 Okay, great.
38:06 Would you be willing to use AI for
38:08 ideiation?
38:10 Like maybe rather than you having to go
38:13 to a whiteboard and come up with 20
38:14 ideas, you could go to chat GPT and and
38:17 ideulate some things over there or go
38:19 over and do some research
38:21 of different kinds of artists that you
38:23 want to go look at their images on
38:25 Google images to inspire your work,
38:28 right? So, so one of the things you can
38:29 do with AI is you can cordon off, you
38:32 know, no go zones. You can say,
38:34 "Ethically, I am not okay with using AI
38:37 to make art." Okay, don't ever do that.
38:41 There there is this massive other world
38:44 of things you can do with AI that are
38:47 not that,
38:49 right?
38:52 You could on business process,
38:55 let's say you're trying to sell yourself
38:57 as an artist, you could use chat GPT or
39:00 claude or one of the things Gemini to
39:03 help you write personalized emails to
39:06 people that might hire you to do
39:08 illustrations,
39:09 right?
39:11 You might use um code interpreter and
39:15 chat GPT to analyze your business
39:18 documents at the end of the year to help
39:19 you do taxes easier. There's a lot of
39:22 things that you can do with AI. Like you
39:25 might hate AI for images but love AI
39:28 because it cured some disease,
39:32 right? We're on the verge of curing type
39:34 1 diabetes right now. I don't know if
39:36 you've seen that.
39:40 They figured out a way to genetically
39:42 modify
39:44 insulin generating cells
39:48 that they inject into the forearm of
39:50 this patient.
39:52 And the way they genetically modified
39:54 these cells, the body's immune system
39:57 can't find them.
39:59 What happens in type 1 diabetes is the
40:02 immune system attacks all the cells that
40:05 generate insulin which is which is sucks
40:10 right?
40:12 So do you forego? So if you say listen
40:16 it is absolutely anathema to use AI at
40:20 all
40:22 if it was trained at all on any art that
40:26 some artist made that's copywritten.
40:29 Okay. Put a big giant X through that.
40:34 Should we also put an X through medical
40:37 breakthroughs? Oh no that's fine. Oh is
40:39 it? Should we also put an X through all
40:42 of the books that were trained into chat
40:44 GPT and Claude?
40:50 Over the next, I don't know, 5, 10, 20
40:53 years, IP attorneys are going to work
40:55 this [ __ ] out. This is not at all
40:59 dissimilar from the early days of hiphop
41:02 and sampling, right? Where where the
41:05 music industry said this is going to be
41:07 the death of us. musicians were like,
41:09 "They're not real musicians. All they're
41:11 doing is copying and pasting,
41:14 you know, intellectual property from
41:16 other artists. They're not real
41:18 musicians,
41:21 right?
41:23 Very similar. Very similar. It took 20
41:27 years for for all of that to shake out,
41:31 but they they ultimately shook it out.
41:33 It's going to take 20 years, 30 years,
41:35 50 years for all this AI
41:38 stuff to to shake out. Here's what I can
41:41 promise you. The big copyright holders
41:43 like Disney and and Sony and Lionsgate
41:46 and the big copyright holders,
41:49 they're going to get paid off at some
41:50 point. The small guys will not. Why?
41:54 Because you don't have [ __ ] $20
41:55 billion to spend on lawyers,
41:58 right? So, so at some point someone will
42:02 get paid. um individual artists
42:07 will not.
42:09 And then what happens? The individual
42:11 artists that embrace AI and figure out
42:13 how to use these tools are are going to
42:16 have superpowers
42:19 in this whatever this new world becomes.
42:21 And the ones that ignore it are going to
42:22 get blindsided by it. It's going to
42:24 suck. And I hate it. I have like listen,
42:26 I have [ __ ] deep deep deep empathy
42:30 for this [ __ ] for for people that are
42:32 pissed off about it. My mission for the
42:36 past nearly three [ __ ] years
42:39 is just to scream as loud as I can, this
42:42 [ __ ] isn't going away. Love it or hate
42:46 it, it's not going away.
42:49 Learn it. Learn what it makes possible.
42:52 Learn if you've got sacred areas you
42:55 never want to touch AI with. Fine, but
42:58 don't throw the baby out with the bath
43:00 bath water with all the other things you
43:02 can do with it. Anyway, Suzanne Welker
43:05 Jurgens, isn't that why some designers
43:08 might choose Adobe Firefly over
43:09 MidJourney? Yeah, I tried to make that
43:11 argument, Suzanne, and I got blasted.
43:13 They they posted Bloomberg articles
43:15 about how Adobe didn't really do it
43:17 ethically. they really did steal from
43:19 people and Adobe are still [ __ ] and
43:22 they're just now they're just
43:23 manipulating the art community to try to
43:25 make it okay to do AI so that they can
43:28 steal from more artists. So, but yes,
43:31 like one of the things that you can do
43:33 is you could
43:36 if you want to there are open-source
43:39 image generation models. I don't know
43:41 the names of them right now but you can
43:43 go to Reddit and find them. You can go
43:44 to hugging face and find them. There are
43:47 open-source image generation models that
43:49 have been ethically trained.
43:52 They've been trained on open-source
43:54 public domain copyright free images and
43:58 videos and [ __ ] like that.
44:00 Now, are they going to be as good as
44:02 ones that were trained on all the [ __ ]
44:04 No.
44:06 But but are they probably good enough?
44:08 Like, are they probably better than
44:10 anything that we had two years ago?
44:13 Yeah, they are. absolutely usable.
44:18 So that there there are ways that you
44:20 can take a stand
44:24 and not shoot yourself in the foot,
44:26 right? And that's the thing about a
44:28 forum like LinkedIn or anywhere you've
44:30 got comments and you've got anonymity to
44:32 some degree, you know, people just stand
44:34 on their soap box and scream. You know,
44:36 a lot of the comments, it was clear they
44:38 didn't read the article
44:40 or look at who I was, right?
44:44 You'll never ever be a creative. All
44:46 right. Well,
44:48 past 40 years, you know,
44:52 says different, but you know, maybe
44:53 you're right. Maybe I'm just a
44:55 talentless hack. It is. It's very
44:57 possible.
44:59 Quen has one that's open source. There
45:01 you go. Shuttertock's another one. There
45:02 you go.
45:06 Not going away like Photoshop. It's
45:08 staying. Yeah. When you know when
45:10 Photoshop first came out, well, that's
45:12 not real design. You're just pushing
45:13 pixels around.
45:18 Yeah. With your digital camera, you
45:19 know, it's not real photography. It's
45:20 not if it's not on film stock, right?
45:23 The the whole point of photography is
45:26 silver halli crystals in in an emulsion
45:30 on top of a carrying unit and then uh
45:35 use light projected through the negative
45:37 onto photographic photosensitive paper.
45:41 Uh, the art is in the dark room. Okay,
45:44 that's photography. Okay. Okay. None of
45:48 this. Push the button and out come
45:50 pixels.
45:51 Have you seen the pixels? You can
45:53 actually see the pixels. You can't see
45:55 the pixels with film.
45:57 Film is real photography. Come back to
46:00 me when they have enough pixels in in
46:02 one of those cameras that you can't
46:03 actually see the pixels.
46:06 Three years later. Oh,
46:09 right. And 10 years later. Okay. Yeah,
46:11 that's just another form of photography.
46:13 We accept it now. Every [ __ ] time a
46:16 new technology comes out. That's not
46:18 real stuff.
46:20 AI art's just theft. Well, it's just
46:24 theft. And it's as if
46:30 someone
46:31 figured out how to take every output
46:35 that they could that has been digitized
46:40 of the humanity of the history of
46:42 humanity and they figured out how to
46:45 smash it into this little ball that we
46:48 get to plug our little brains into.
46:53 And so, yeah, in this one little corner
46:55 over here are these 10 illustrators of
46:59 comic books work. And that's in there,
47:03 but so are the entire history of
47:06 humanities works. So, like all of the
47:09 museums, all of the galleries, all of
47:11 the all of the uh the uh the uh just
47:17 anywhere where there's cultural
47:20 um items,
47:23 the Smithsonian, right? All the
47:26 archives, it's all in there and we get
47:29 to tap into it,
47:34 you know? It it's just it's just
47:37 staggering to me that like the the the
47:40 this the self harm that that people are
47:44 going to do like you know take a stand
47:46 but don't be uneducated. Anyway, all
47:48 right thoughts.
47:51 I was cleaning out my storage last week.
47:53 Found a list of serial numbers to for
47:55 Photoshop for in PageMaker. I remember,
47:58 you know, I remember having those you
48:00 you'd have like the the notepad with
48:02 like your friend your friend would buy
48:04 page maker and someone else would buy
48:06 Photoshop and you'd trade serial numbers
48:09 and and so we would all have the list of
48:11 the serial numbers. Yeah, exactly.
48:16 Oh man.
48:22 All right, cool. Um, side hustle. Mimi
48:25 has a movie and a regulars. All right.
48:42 Side hustle mimi. My first fulllength
48:44 video. Lia
48:46 may leo eldo
48:49 dollar.
48:51 I I I'm not good with Spanish.
48:54 [Laughter]
48:57 Yeah, I'll I'll flip over to YouTube
48:58 here.
49:00 Um,
49:04 okay.
49:06 So, let's go watch this. Now again,
49:11 if you were to
49:15 take in what all of the haters in my
49:20 LinkedIn article said today,
49:26 in their mind,
49:28 there is not nor can there ever be any
49:33 any
49:36 true human creativity in what I'm about
49:38 to show you. There can be none.
49:42 Um, and this is AI slop.
49:47 So, let's see it now. It might be it's
49:49 possible side hustle Mimi just pushed a
49:51 button and out came a movie, but it
49:53 looks like she at least went into um
49:56 iMovie or or Final Cut Pro and learned
49:58 how to do titles.
50:02 Okay. Although, it does say title text
50:04 here. You should probably go fix that.
50:06 Um, okay.
50:08 Let's watch this Tik Tok pin from Side
50:11 Hustle Mimi. It's my first one. Be
50:13 gentle. No. No. I'm really excited about
50:15 this. I just it
50:19 the whole purpose of that article is
50:22 that it struck me
50:24 just with my personal work, right? I've
50:27 been doing this stuff now for I don't
50:29 know three years.
50:32 And I
50:35 dismissed my own contribution in most of
50:39 this stuff and gave credit to the AI
50:41 tool.
50:43 And I just the whole point of the
50:44 article was no, don't it's not just the
50:48 AI tool doing the work. Like give
50:50 yourself credit for the the the chain of
50:53 craft that led you to get here.
50:56 Um, and that's, you know,
51:00 I think it's kind of obvious. Someone,
51:02 someone in my comments said, "Oh, that's
51:04 really obvious." In fact, it's so
51:06 obvious that we've built a whole
51:07 business around this. It's called chain
51:09 of creation. I think they said, and
51:12 they've got a whole thing about um
51:14 they've got a whole business around
51:16 people be being able to document their
51:18 chain of creation for copyright uh
51:21 arguments. I'm like, there you go. So
51:24 anyway, whatever. Okay, here we go. I'm
51:27 excited about this.
51:51 I'll apply
51:53 your
51:56 [Music]
52:06 Foreign
52:08 [Music]
52:15 [Music]
52:18 speech. Foreign speech. Foreign speech.
52:19 [Music]
52:40 [Music]
52:43 Wonderful.
52:48 [Music]
52:59 I apply
53:03 for
53:04 [Music]
53:15 [Music]
53:36 [Music]
53:46 Okay, now I want to point something out.
53:49 So couple of things. So creative
53:53 director Naomi Reyes images midjourney
53:56 music lyrics chat GBT.
54:00 In no other
54:02 in no other medium do we give credit to
54:05 the tools right
54:08 unless it's contractually required.
54:10 Right? Technicolor if you want to use
54:12 technicolor technology in your movie you
54:16 have to say this was made with
54:17 technicolor. That's that's a that's a
54:20 licensing agreement, right? But in
54:22 general, we're not saying filmed on red
54:25 cameras with these lenses with these
54:27 lighting packages.
54:30 The reason it makes sense to to call out
54:33 the tools now is because everyone's
54:35 trying to figure out what the how the
54:38 [ __ ] do we do this? But I actually want
54:40 to go back. I side hustle, I hope you
54:43 don't mind. I'm gonna um I want to go
54:45 back and I I wanna I w to talk about
54:48 some creative choices you made
54:52 because I think it's really really
54:54 really
54:56 important
54:58 that this is not an AI movie,
55:03 right? This is a music video
55:07 that it was crafted with AI tools is
55:10 interesting,
55:13 but it's not an AI movie. It's not an AI
55:16 video. It's just a video. It's a music
55:18 video. And I want to I want to talk
55:22 about the chain of craft. Some of the
55:23 things that that I see in this. I think
55:26 just the the title, the black, you know,
55:28 white text on the black thing.
55:30 Interesting creative choice. Um,
55:34 she starts out. So, first of all, she
55:37 picked an aesthetic, right? She found a
55:40 look and feel that resonated with her. I
55:43 don't know how she got to this.
55:46 Were there were there 50 different
55:48 styles that she tried? Did she start
55:50 with the song first or the images first?
55:52 Did she create this image or one of
55:54 those images of the women dancing and
55:55 go, "Oh, I should make a song around
55:57 that. Let me make some other images and
55:59 then let me go make the song." or was
56:01 she over in Sunno and made the song? And
56:02 I know you're here. You could answer
56:03 this, but I'm just I just want to go
56:05 through this kind of this hypothetical
56:08 chain of of craft
56:10 that you did here because so much of
56:13 what's here
56:15 is your human point of view,
56:20 right? Why did you make a song in that
56:22 style? Why did you have it sung in
56:24 Spanish? Why did you pick this kind of
56:26 50s Hawaii looking um vibe, right? It it
56:32 looks like a postcard, right? Was that
56:33 from something from your childhood or
56:35 did you just happen to groove on this
56:37 aesthetic? We don't know. It doesn't it
56:39 it ultimately doesn't matter. But what's
56:42 clear is you've got a point of view.
56:44 There's a creative point of view. Notice
56:45 we're also starting, you know, the beach
56:48 the beach is off in the distance, the
56:50 clouds are off in the distance, and
56:51 there's these butterflies flying around.
56:53 So, so we start out there. So, that was
56:56 one of the first things I noticed. And
56:57 then as the music gets kicking, we zoom
57:00 in
57:02 and we go from, you know, farther away.
57:05 We're zooming in. Then we go kind of
57:08 framing
57:09 where, okay, now we've gone into the
57:12 jungle, right? Now we've gone into the
57:13 woods. We've come off the beach. Maybe
57:16 we can see it in the distance there, but
57:17 now we're in the jungle, right?
57:21 Really interesting choice. So, she's
57:22 inviting us into this world. We start
57:24 out broad. She's inviting us into the
57:26 world. Oh, cool. Is this going to be
57:28 flowers and bees and birds and things
57:30 like that
57:31 and then very quickly, no,
57:34 there's people dancing and having fun?
57:36 It's a whole different world.
57:39 Oh, Puerto Rico. Perfect. Perfect.
57:43 Um,
57:44 there's another cool thing that you did
57:46 which I think is wild, which is this.
57:53 A lot of times in videos, directors will
57:55 do a crossfade or a fade to black. And
57:59 she's doing fade to white. Why? Doesn't
58:01 matter. But she chose to do this. AI
58:04 didn't do do this. She chose to do this.
58:07 Whatever this signifies with her, this
58:09 white might be now we're entering the
58:11 world of dancing, right? Or now we're
58:13 going to change time frame or we're
58:15 going to change a different topic in the
58:16 song. Whatever it might be.
58:22 And there's just these beautiful women
58:24 just dancing, just freely dancing in
58:26 their white dresses.
58:36 [Music]
58:37 So there's a cross fade, right? She goes
58:40 from
58:42 scene to that scene. Cross fade.
58:47 [Music]
58:58 [Music]
59:06 fade to white. Why? Oh, we get to meet
59:10 someone at a different level of
59:12 intimacy. We're going deeper into this
59:14 world, right?
59:16 So,
59:17 so yes, she used midjourney. Yes, she
59:21 used Sununo and she made all these other
59:24 choices.
59:27 That's the craft.
59:29 The craft is knowing which tool to use
59:33 and ultimately having a vision. Side
59:35 hustle Mimi's been coming to these these
59:38 lives for what, a year or more? You've
59:40 been here for a while. This is her first
59:43 video. She's been taking this stuff in,
59:46 learning how to do it, playing, learning
59:48 how to make images, learning how to do
59:50 this, learning how to make videos,
59:51 learning how to make songs, learning how
59:53 to do [ __ ] in chat GPT, learning trying
59:55 something over with code interpreter and
59:57 realizing, "Oh [ __ ] I hate data
59:59 analysis and I don't care about that."
1:00:00 Coming back to this, right? At some
1:00:03 point
1:00:04 she learned the tools well enough.
1:00:08 She has enough of a technical vocabulary
1:00:10 with the tools to be able to go, you
1:00:13 know, I want to make something.
1:00:17 And she put together this chain of
1:00:19 craft, these many, many steps, which
1:00:22 part of it included all of the learning
1:00:25 that got her comfortable enough with
1:00:27 these tools to be able to [ __ ] do
1:00:29 this, right?
1:00:31 the art of AI. Well, I you know, I would
1:00:34 say no. This is the this is the art of
1:00:37 side hustle Mimi
1:00:39 using some AI tools.
1:00:44 But this is her vision, right? This is
1:00:49 this is not just AI slop. This is very
1:00:51 intentional storytelling,
1:00:54 right? as the song goes on. And then
1:00:56 look, we bookend it as the song starts
1:01:00 to resolve.
1:01:02 [Music]
1:01:03 We're going back out, right?
1:01:07 We're leaving the intimate world of the
1:01:09 dancers. We're going back out into the
1:01:10 woods.
1:01:15 Now we're back on the beach.
1:01:19 That's good [ __ ] storytelling.
1:01:22 It's good [ __ ] storytelling. Look.
1:01:26 We're out on the beach.
1:01:28 We go We go into the forest. We meet
1:01:31 some ladies. They dance. They dance.
1:01:35 We get to meet her. Who the hell is she?
1:01:37 I want to know who she is. Is she the
1:01:38 one singing? Why is she not moving? Why
1:01:41 are the butterflies singing and she's
1:01:42 not moving?
1:01:44 Well, that was probably just an AI
1:01:46 hallucination.
1:01:47 Or it was an artistic choice.
1:01:51 It's [ __ ] fascinating. Why doesn't
1:01:54 she move?
1:01:58 [Music]
1:02:00 The butterflies are moving and she's
1:02:01 not. That's a creative choice. Why? I
1:02:04 don't know.
1:02:07 But it's certainly interesting. It
1:02:08 certainly makes you [ __ ] your head and
1:02:09 go, "Huh?"
1:02:12 Right.
1:02:15 We go on. Ladies are dancing. The
1:02:18 butterflies are butterflying. There she
1:02:20 is again. Does she move this time?
1:02:28 What's that mean? She didn't move
1:02:29 before. She does now. I don't know.
1:02:33 Then we get head back out.
1:02:36 It's called bookending, right?
1:02:39 You're going to bookend.
1:02:41 Here's how we started. Here's how we
1:02:43 end.
1:02:48 So much of this So much of this has
1:02:52 nothing to do with AI.
1:02:57 It's got to do with the fact that Mimi
1:03:00 side hustle Mimi Naomi
1:03:04 took the time to learn the tools well
1:03:07 enough to integrate them into her
1:03:10 creative vocabulary.
1:03:14 So that at the point she wanted to make
1:03:16 this,
1:03:18 she could make this. Now she might have
1:03:20 done this over a weekend. This might
1:03:22 have taken 3 months. It doesn't matter.
1:03:24 Like
1:03:26 this was not just push a button and outs
1:03:29 squirted some crap,
1:03:32 right?
1:03:36 That's the point of the article.
1:03:39 Anyway,
1:03:41 I think you're right. Right before I
1:03:43 played this, you know, you did some some
1:03:45 version of an apology like it's my first
1:03:47 thing. Don't be too hard on me. I think
1:03:49 this is [ __ ] beautiful. I think it's
1:03:51 beautiful. Like, objectively,
1:03:54 independent of how you made it and what
1:03:56 tools you use to make it. It's
1:03:58 beautiful. The music's got this weird
1:04:02 stilted
1:04:03 like it it it's like
1:04:07 like it almost feels like it's just too
1:04:09 slow
1:04:11 like like you almost feel like you you
1:04:13 you need like a a motor scooter scooter
1:04:17 to push it forward like like and there's
1:04:21 something fascinating about that
1:04:25 right it's got this haltingness to it.
1:04:28 That's part of it. Like the add part of
1:04:32 me was like, "Get on with it." Like, "Up
1:04:34 the tempo, up the tempo." But no, she's
1:04:36 sort of forcing us to sit in the pocket
1:04:38 of this music
1:04:44 and just take in this beauty of this
1:04:46 place of Puerto Rico. [ __ ] gorgeous.
1:04:50 Gorgeous. So, congratulations.
1:04:55 According to the haters on my LinkedIn
1:04:58 article, um nothing you did here is
1:05:00 valid and you're not creative and you're
1:05:03 a talentless slop pusher.
1:05:08 But according to me, because I am an
1:05:11 evil AI lover,
1:05:13 I'm gonna I'm gonna go ahead and give
1:05:15 you an A+ on this one.
1:05:19 I think it's quite beautiful. Wait, it
1:05:21 was supposed to be It was supposed to
1:05:24 convey a little slowness but happy
1:05:25 vibes. Yeah, exactly.
1:05:28 Right. It's got this this laziness to
1:05:31 it, but there's something about it that
1:05:33 makes you want to dance and and get up
1:05:35 and move. Sort of pulling you into that
1:05:38 vibe.
1:05:40 Very successful. And I don't even know
1:05:42 what the words mean,
1:05:45 right? I'm pasty white guy number seven
1:05:48 from the courtroom drama.
1:05:51 But it's [ __ ] beautiful. Congrats.
1:05:58 I'll translate the lyrics. You don't
1:06:00 have to translate the lyrics. [ __ ] that.
1:06:02 [ __ ] those people.
1:06:07 All right, let me change out my Oh, look
1:06:09 at my hair. That's unfortunate. I You
1:06:11 know what I need? I need some I need
1:06:13 some AI assistance on my [ __ ] hair.
1:06:15 That's what I mean.
1:06:19 [Music]
1:06:26 Irregular is now defending you on
1:06:28 LinkedIn. Good. Thank you.
1:06:32 It's not black and white.
1:06:36 The thing that's frustrating about
1:06:37 almost every one of those arguments is
1:06:40 that it's it's it's black and white.
1:06:43 AI evil, not AI good.
1:06:47 No, sorry. There's shades of gray. And
1:06:51 there's absolutely creative expression
1:06:53 that can be had. This video is absolute
1:06:57 proof of the argument I make in that in
1:07:00 that article.
1:07:03 Can we get shades of black bar?
1:07:08 Oh my god.
1:07:11 Really nice work. Really nice work.
1:07:12 Okay, wait. I'm supposed to go to a
1:07:13 regulars now.
1:07:29 Right. There's almost like a It's It's
1:07:32 almost like a Latin pula. One, two,
1:07:34 three. One, two, three.
1:07:38 It is. It's It's a It's a It's a waltz,
1:07:41 right? Isn't it a waltz?
1:07:51 One, two, three. One, two, three,
1:07:55 two, three. It's it it's something it's
1:07:58 something Walt's like. It's not in four.
1:08:06 Not for nothing but last Thursday I
1:08:08 thought Kyle was saying chain of crap.
1:08:10 Well, the the the you you will fit in
1:08:14 perfect with the with the commenters on
1:08:16 on LinkedIn.
1:08:18 Chain of craft. Craft, not crap. That's
1:08:22 really funny. Yeah. But that their whole
1:08:25 thing is if you're using AI, it is chain
1:08:27 of crap and you're a thief and go [ __ ]
1:08:29 yourself.
1:08:34 That's very cool. That song's very cool.
1:08:36 Waltz, a waltz of a sort.
1:08:41 YouTube comments,
1:08:44 the negativity is like talking to
1:08:46 flatearthers. It is. Just wait till they
1:08:47 find out the creativity is round.
1:08:54 Oh my god, that's funny. What am I
1:08:56 looking at in your regulars? Oh, long
1:08:58 live embrace the Jang.
1:09:01 That's nice. That's cool. You got the AI
1:09:04 Salon logo in there. Love that. Who did
1:09:06 that? Steo, of course. Beautiful. Daniel
1:09:10 Togos. Very nice.
1:09:13 Singularity
1:09:15 USA reserved for the AI salon.
1:09:23 Oh my god, my AI character. Oh, it's
1:09:27 hilarious. Um,
1:09:31 all right. Uh, what are we gonna do?
1:09:35 What are we gonna do? What are we going
1:09:36 to do? What are we gonna do? What are we
1:09:38 gonna do? What are we gonna do?
1:09:41 What are we gonna do, people?
1:09:47 Uh uh uh uh uh uh uh uh. We haven't used
1:09:52 ideogram in ages. That's funny.
1:09:56 Oh,
1:09:58 here's an indepth look at flux context.
1:10:00 Flux one context is a multimodal flow
1:10:03 matching model from Black Forest Labs.
1:10:08 It's available in three flavors. De Oh,
1:10:10 so this is I don't know if the person
1:10:12 that asked for how do we do product
1:10:14 shots is still here. YouTube comment
1:10:17 about Tik Tok.
1:10:20 Wait, what's with the What's with Oracle
1:10:22 buying Tik Tok? Oh, okay. So, I didn't
1:10:26 know it was Oracle, but doesn't surprise
1:10:28 me. So, that's is that Larry Ellison or
1:10:31 is Yeah, that's Larry Ellison, right?
1:10:33 Doesn't surprise me. He's he's a he's in
1:10:36 the billionaire class.
1:10:38 Um, so what it looks like is going to
1:10:42 happen is that I think starting in the
1:10:44 fall.
1:10:46 So apparently right now Tik Tok is being
1:10:49 rewritten
1:10:53 so that it's a USonly app
1:10:58 that the current Tik Tok app is going to
1:11:00 keep existing
1:11:04 and as of March 2026
1:11:09 people in the US will not be able to use
1:11:11 it.
1:11:13 We
1:11:15 lucky
1:11:17 freedomloving individuals
1:11:20 will be on an app called M2 that we have
1:11:24 to install and I assume start over all
1:11:28 of our I assume we can't keep our
1:11:30 followers and we certainly can't keep
1:11:32 our followers followers if they're
1:11:34 outside of the US. So what we're going
1:11:36 to have is effectively a
1:11:39 governmentsponsored
1:11:42 US-only social media platform
1:11:46 that cannot reach anyone else in the
1:11:48 world and they can't reach us.
1:11:52 And
1:11:54 I would be
1:12:00 I cannot imagine a world in which the
1:12:04 government that is saving
1:12:07 this app is not going to build full-on
1:12:13 citizen surveillance into it.
1:12:17 They just cut a deal with Palunteer
1:12:21 to do a database of every citizen in the
1:12:23 United States.
1:12:25 I would be shocked
1:12:27 if this app weren't tied to Palunteer.
1:12:32 So if
1:12:35 if one of your questions is we'll still
1:12:38 be here, right, Kyle? Well, anyone who's
1:12:41 here like Steve O that's not from the US
1:12:43 won't be here. But here's what I can
1:12:45 tell you. This channel will not exist on
1:12:48 M2. One of the reasons we simoc cast on
1:12:52 YouTube
1:12:55 is for exactly this the fear of what
1:12:58 they're doing. So as of March 2026,
1:13:04 the Tik Tok version of of this channel
1:13:06 is going to go. I'm not I'm not
1:13:09 continuing this on some surveillance
1:13:12 state piece of [ __ ] technology. Um
1:13:15 there there's no other social media
1:13:17 platform in the world that that has
1:13:20 national borders.
1:13:23 I think that's horshit.
1:13:25 So that's the Oracle thing. I didn't
1:13:27 know it was Oracle. Geez, time to go to
1:13:30 England. Exactly. Like honest to [ __ ]
1:13:33 god,
1:13:35 that's like China coming in and taking
1:13:37 LinkedIn and saying Americans can't use
1:13:38 it. Yeah, exactly. It's it's just it,
1:13:41 you know, it it's just political
1:13:44 [ __ ]
1:13:46 and and you know, and it is it is it is
1:13:48 going to absolutely turn into a
1:13:50 surveillance tool. So, I know I sound
1:13:53 like a conspiracy theorist, but
1:13:54 whatever.
1:13:58 Now that I've got every everything
1:14:00 warmed up, uh, Ann is backstage. I don't
1:14:02 see Ann backstage. Is Ann backstage? Oh,
1:14:04 yes, she is.
1:14:05 Ah, Ann Murphy.
1:14:08 How's it going?
1:14:09 It's going good. I'm just ranting about
1:14:11 the what Tik Tok's about to do.
1:14:14 Oh, okay. So, I'm a smidgen behind on
1:14:17 that, but guess what? What
1:14:20 you I can we just have like a little
1:14:23 smidg in touch base with the with the
1:14:25 nice people about our both getting
1:14:28 dragged on LinkedIn in the past 48
1:14:31 hours.
1:14:32 Oh my god.
1:14:34 I will be respectful right now in my
1:14:36 comments. You know how I really feel,
1:14:38 but I'll be nice.
1:14:41 I do. Yeah. No, listen. I I I mean, I've
1:14:46 been doing this long enough that people
1:14:48 know who I am, right? They know what I
1:14:49 stand for. And it's just like whatever
1:14:51 it was about that post, you know, it it
1:14:54 just it hit a different a clearly a
1:14:57 different community. Um and it's a
1:14:59 community. They don't know who the [ __ ]
1:15:01 You were you were in a different
1:15:02 neighborhood.
1:15:03 I was you were in a different
1:15:05 neighborhood.
1:15:06 I took a left at Albuquerque and I
1:15:08 shouldn't have.
1:15:09 Yeah. Yeah. That was in the wrong
1:15:12 neighborhood.
1:15:15 The algorithm. But guess what?
1:15:18 It's so good for your It's so good for
1:15:21 your algorithm to have that kind of
1:15:26 Yeah.
1:15:26 Go so much wider now.
1:15:28 I think it's I think it is also
1:15:33 like like one of the reasons I'm
1:15:34 engaging and and you know trying to make
1:15:37 counterarguments and trying to apply
1:15:38 logic to the situation. I actually do
1:15:41 think it's important. I, you know, I I
1:15:43 think that um I deeply empathize with
1:15:48 shit's changing. Um there are absolutely
1:15:52 ethical concerns about how this stuff
1:15:54 was created, but but but
1:15:58 was created is the operative term. These
1:16:00 models exist, right? The the damage was
1:16:03 done. the lawyers are going to work this
1:16:05 out moving forward. But but like to to
1:16:08 just use that as a reason to not, you
1:16:11 know,
1:16:13 let yourself understand what's happening
1:16:15 just just feels really shortsighted to
1:16:17 me. So I feel like my
1:16:21 my ability to engage
1:16:24 in a highly vitriolic
1:16:27 conversation and and and just maintain
1:16:30 my point of view. it actually feels like
1:16:32 a really important thing to learn. So,
1:16:35 um, and to practice.
1:16:36 Okay. So, yeah. So, that was what I
1:16:39 wanted to bring up
1:16:40 is and and the reason why I think it's
1:16:43 relevant for the community and for the
1:16:47 folks, you know, here on the show. Um,
1:16:51 these are like pretty extreme public
1:16:54 examples of having to decide how you
1:16:58 want to lead in AI.
1:17:00 Yeah.
1:17:02 And you took you took the approach
1:17:06 of this is an opportunity
1:17:09 and you in the mo in the most respectful
1:17:12 of ways articulated facts.
1:17:16 Well well tried.
1:17:20 Yeah. I mean
1:17:22 you can
1:17:24 obviously you cannot make them drink and
1:17:26 those people are going to get it you
1:17:28 know in a while and and they're in a
1:17:32 phase they're in a phase that so many of
1:17:35 us and particularly a lot of our
1:17:37 creatives have gone through and some of
1:17:39 them are still in the middle of of like
1:17:41 [ __ ] this to your point things are
1:17:45 changing
1:17:47 and they feel justifiably threatened Ed
1:17:50 and it's a very scary thing.
1:17:52 Yeah. Yeah. And it's and it's um as
1:18:00 as as our communities an get more
1:18:03 educated like like I don't know if you
1:18:05 were here. Did you see Side Hustle
1:18:06 Mimi's video that we we showed
1:18:10 Side Hustle Mimi? No, I did not. I'll
1:18:13 look back on the YouTube.
1:18:15 Oh, yeah. Look back on YouTube. It's
1:18:17 it's really really good. But um
1:18:21 as our communities get more and more
1:18:23 sophisticated with with you know this is
1:18:26 not just an AI tool. It's not just like
1:18:28 push a button make a video. It's like in
1:18:31 order to do interesting sophisticated
1:18:33 work you need to know lots of tools and
1:18:35 lots of like what works and what doesn't
1:18:37 and when you would use this tool over
1:18:38 that tool and ideas and things like
1:18:40 that. As our community gets more and
1:18:42 more sophisticated with those
1:18:44 complicated chains of creation or chains
1:18:46 of craft as I call them, um the divide
1:18:51 between people who haven't even like
1:18:53 like who are actively avoiding AI, that
1:18:56 gap is going to get increasingly large
1:18:58 and and I I
1:19:00 exponentially large,
1:19:02 right? And I feel I sense that those
1:19:04 people are going to get increasingly
1:19:05 pissed off and increasingly
1:19:08 disenfranchised and
1:19:09 Oh, yeah. Right. My whole thing is just
1:19:13 stay in the lane. Just just keep, you
1:19:15 know, just keep saying what you're
1:19:17 saying is AI is not going away. And
1:19:20 there's only two choices. If it's not
1:19:21 going away, you either deal with it or
1:19:23 it happens to you. And and I am not I am
1:19:27 not okay with the former. And like even
1:19:29 if people come in and get pissed off at
1:19:31 me, I'm still going to hold my ground to
1:19:33 say
1:19:34 you should still learn this. I get that
1:19:35 you're pissed off. I I feel [ __ ]
1:19:38 horrible that what you're going through
1:19:40 and how scary it must be. And you should
1:19:42 [ __ ] learn this stuff, right?
1:19:46 Well, yeah. Okay. And the thing the good
1:19:50 thing about not like, you know, burning
1:19:53 the house down with to to to go back and
1:19:56 forth with people is that they're in
1:19:59 grief phase right now.
1:20:01 Yeah. and they're super vulnerable and
1:20:05 superw.
1:20:13 Um, and for the PE for the people got an
1:20:17 I thought an interesting note. Um,
1:20:21 I got some important feedback uh from
1:20:24 the social medias. It wasn't none of it.
1:20:27 It wasn't like I had a post and people
1:20:29 were going back and forth like that, but
1:20:32 it was on other people's posts something
1:20:34 topical
1:20:36 and never you nobody named my name or
1:20:39 whatever. But
1:20:40 it was a roller coaster of like emotions
1:20:44 around it because my feelings were hurt.
1:20:46 Yeah.
1:20:46 Right. Like I'm a I'm a human too
1:20:48 everybody.
1:20:50 Like
1:20:51 if when you drag me I feel dragged. Like
1:20:54 that feels heart heavy on my heart. I I
1:20:57 cried. You saw me cry.
1:20:59 I did. I saw you
1:21:01 over time. I was like,
1:21:02 "Yeah."
1:21:05 Yeah. I was like, I can't talk.
1:21:10 But, you know, now I understand like how
1:21:13 important it is to listen to people and
1:21:16 I can make changes as a result.
1:21:18 Yeah.
1:21:19 And we have to. But I do wish for both
1:21:23 of us. Sorry, I hate to say it, but I
1:21:25 wish that people could be a little bit
1:21:27 nicer sometimes.
1:21:30 Yeah.
1:21:33 Yeah.
1:21:34 I I don't know that that's going to
1:21:35 happen. I mean, I just think that I I
1:21:37 think especially when when people are in
1:21:40 existential threat, which which
1:21:42 Yeah.
1:21:43 the creative community. I mean, listen,
1:21:45 if you've been an illustrator for 20
1:21:47 years and all of a sudden along comes
1:21:50 Exactly.
1:21:51 that can literally take your style and
1:21:53 replicate it instantly and and
1:21:56 efficiently
1:21:56 instantly.
1:21:57 Um, it is that is existential, right?
1:22:00 That is existential crisis. So, I get I
1:22:03 get where the anger is coming from. Um,
1:22:05 but but yeah, I think it is a stages of
1:22:07 grief thing. And and you know, here's
1:22:09 the thing that that post now, I mean, I
1:22:12 took it I took it very personally, too.
1:22:14 It was funny. I was like halfway through
1:22:16 Sunday and I'm like, I'm really bummed
1:22:18 out. Why am I bummed out? And I realized
1:22:19 like I was taking that [ __ ] personally,
1:22:22 right? Because, you know, they're not
1:22:24 just saying they're not just saying your
1:22:26 argument's flawed. They're saying, you
1:22:27 know, you're an evil piece of [ __ ]
1:22:29 right? For for for even dealing with
1:22:32 this at all. You you know, you you are
1:22:34 the problem. Um, and I and I took it
1:22:37 personally, but um, that's out there and
1:22:40 I think that's something we're going to
1:22:41 have to manage through. And if, you
1:22:44 know, in some of the 2,000 people that
1:22:46 have seen and commented and, you know,
1:22:49 done things with that post, if in there
1:22:53 five different people are like, "Huh,
1:22:55 maybe I should just check this stuff
1:22:56 out." Then I think it's then I think
1:22:58 it's worth it, right? Because those will
1:23:00 be five less people that are blindsided
1:23:02 by this stuff. So anyway, yeah. No, it's
1:23:05 been really interesting couple of days.
1:23:06 So, yeah, I'm sorry. Sorry you've gone
1:23:08 through it, but you know, now you have a
1:23:12 in uh in
1:23:14 the bludgeon the budgeoning. What is it?
1:23:16 The beatings will continue until morale
1:23:17 improves.
1:23:22 Yeah, exactly. That's the thing of like,
1:23:24 well, where do we go from here,
1:23:25 everybody?
1:23:28 All right. Well, I just want to say hi
1:23:30 and and good job and keep up the good
1:23:34 work. Um,
1:23:37 and I'll see you in a while because
1:23:41 tomorrow morning to drive across
1:23:42 country. So, I'll see you in a little
1:23:44 while, everybody.
1:23:46 Beautiful. All right. Thanks, Ann.
1:23:51 Beautiful. Beautiful. Uh, it sounded
1:23:53 like we had some audio issues there. I
1:23:55 don't know what's going on. Sometimes
1:23:56 Streamyard's just weird. Um, one thing
1:23:58 that I did, we did discover, we were
1:24:02 recording, um, Ann and I were recording
1:24:04 a podcast over the weekend
1:24:07 with Vicki, and if Vicki had a
1:24:10 background on her her video, it would
1:24:13 crash Streamyard, it would kick her out
1:24:15 the minute we brought her live. So,
1:24:17 there's some weird instability issues
1:24:19 with the current um the current version
1:24:22 of StreamYard that we use here. Um, so
1:24:25 maybe that audio thing is just part of
1:24:26 maybe they've just got a crappy version
1:24:28 of the software. You know what it
1:24:29 probably is? They probably vibecoded
1:24:31 this version of Streamyard.
1:24:35 Glad I grew up Gen X when bullying was
1:24:38 still encouraged. Yeah, exactly. I'm
1:24:39 totally being bullied over there on
1:24:41 LinkedIn. What's up? Anything new with
1:24:43 Grockport? Ah, John Harris. Fantastic.
1:24:45 Interesting ha thing happened from when
1:24:47 I made the video this morning or this
1:24:50 afternoon that said come to my live
1:24:52 tonight. will be live with with Gro 4.
1:24:56 Um,
1:24:59 what I what I got correct is that Gro 4
1:25:01 will be dropping at 8:00 PM Pacific
1:25:04 time. The part that I slightly got wrong
1:25:07 was it was not tonight. It's Wednesday.
1:25:10 So,
1:25:13 so, so you just get me tonight. You
1:25:15 don't get Grock for tonight. But, uh, it
1:25:18 looks like it's coming and then who
1:25:20 knows if it's any good. Um, Elon did put
1:25:24 out something about video generations
1:25:26 getting really good. So, now there's
1:25:28 speculation that Grock 4 will have video
1:25:30 generation in it, which who knows, maybe
1:25:32 it does, maybe it doesn't. Um,
1:25:35 [Music]
1:25:41 oh, LTX Studio has flux flux core flux
1:25:45 context in it as does foul.ai
1:25:50 iterative local. I don't think I've Have
1:25:52 I used FA? I think you've got to
1:25:54 subscribe to FAL and I think I'm not a
1:25:56 subscriber.
1:25:57 foul.ai generative media platform for
1:26:00 developers. Yeah, I don't think I've
1:26:02 done this.
1:26:04 Continue with Google.
1:26:18 Oh, this is the thing where you have to
1:26:19 make the workflows. Yeah. No, I'm not.
1:26:22 I've got I've got no credits here. Um, I
1:26:25 think you can also go to Flux
1:26:29 Flux.pro fluxai.pro
1:26:34 flux image AI
1:26:38 number of images.
1:26:43 Flux text new. Okay.
1:26:48 reference image
1:26:50 more image.
1:26:55 Okay. So, what we're gonna do here, so
1:26:58 if if anyone was here before,
1:27:02 what was it? Um, an hour and a half ago,
1:27:05 someone asked a question.
1:27:09 They said, "Hey, how can I do a product
1:27:12 shot um and not have it suck in
1:27:15 midJourney?" and I said, "Well, you
1:27:17 shouldn't use MidJourney. You should use
1:27:19 Flux Context or you should use um Runway
1:27:23 ML's elements thing." Um, let's see. Can
1:27:26 you tell me why lately chat GPT is
1:27:28 giving wrong answers consistently and
1:27:30 Grock is getting technical questions
1:27:32 right the first time, but then when
1:27:35 verifying with Chat GPT, it says that
1:27:38 it's right. Um,
1:27:42 something is up with chat GPT.
1:27:45 um
1:27:48 very often before they do a major
1:27:50 upgrade.
1:27:52 So here's here's my theory. This is pure
1:27:57 chat TMZ speculation. Pate will probably
1:28:00 laugh at me.
1:28:02 My speculation is that when OpenAI is
1:28:04 getting ready to launch something new,
1:28:07 so let's say that they've got
1:28:10 a million a million servers, right, that
1:28:13 that they're doing inferences inference
1:28:14 on that they're doing chat GPT serving
1:28:17 on.
1:28:19 when they're getting ready to do
1:28:20 something new, my instinct is they
1:28:23 probably take down, you know, a h
1:28:25 100,000 a 100,000 and they're they're
1:28:29 sort of uninst they're they're like
1:28:31 upgrading they're taking servers offline
1:28:34 to upgrade them which degrades
1:28:39 the the
1:28:40 [ __ ] over here. or they're doing
1:28:43 something like they're they're upgrading
1:28:45 components in real time across what's
1:28:47 live on ChatGBT and that's [ __ ]
1:28:49 something up. I don't know what they're
1:28:51 doing, but ChatGpt tends to get unstable
1:28:56 and weird and shitty
1:28:59 about two weeks before they launch
1:29:00 something. Apparently, 10 days from now,
1:29:03 July 17th, GPT5 is coming out. That's
1:29:08 that's the rumor. That's the chat TMZ
1:29:11 rumor. I I I I heard it on the Twitter
1:29:14 and therefore it's fact. Um but it looks
1:29:17 like July 17th is going to be new chat
1:29:20 GPT. So my my gut is telling me that
1:29:23 they're doing something on the server
1:29:25 side. They're up updating something.
1:29:27 Maybe they're updating architectures,
1:29:29 things like that. And in doing that,
1:29:30 they're maybe having to to degrade
1:29:32 service or or you know, sort of limit
1:29:35 the number of cycles it does thinking um
1:29:38 in order to not have the thing come
1:29:40 crashing to the ground. So that's just
1:29:43 my guess. Um or it could just suck.
1:29:45 Maybe it's just going through a phase
1:29:46 where it sucks. I feel like I've been
1:29:49 waiting for GPT5 longer than GTA 6.
1:29:53 That's really funny.
1:29:56 Okay.
1:29:58 So, let's let's do flex. Let's go over
1:30:00 to let's go to let's go to the Google.
1:30:03 We're gonna actually go to the Google
1:30:04 and we're going to go um
1:30:08 uh let's see. Let's see. Uh uh CPG
1:30:11 product
1:30:13 shot.
1:30:15 And then we'll go images.
1:30:21 [Laughter]
1:30:24 Producer Brandon's hangry. I think
1:30:26 producer Brandon needs a snack.
1:30:32 Okay, let's see. These scope bottles are
1:30:35 pretty cool.
1:30:56 Um, all right. Let's save this image.
1:30:59 Aroma Spa. Oh, this is a good one. Okay.
1:31:03 Aroma Spa. So, we'll go Aroma Spa.
1:31:08 Save that to the desktop. And then let's
1:31:11 go. Um
1:31:13 we'll go um
1:31:31 [Music]
1:31:35 side hustle Mimi. I can't get that song
1:31:37 out of my [ __ ] head.
1:31:39 Um, let's do um middleaged
1:31:44 woman
1:31:47 in
1:31:50 say middle-aged woman studio
1:31:54 portrait.
1:31:58 Okay.
1:31:59 So, let's do She looks pleasant enough.
1:32:02 We'll go grab her. We'll do save image
1:32:04 as.
1:32:06 uh we'll say uh woman. So we've got
1:32:10 aroma aroma spa and woman. We've got two
1:32:13 images. And so now we're going to go
1:32:15 back to Flux.
1:32:17 So click, drag, or paste an image. Okay.
1:32:20 So
1:32:22 we're going to go downloads, no desktop.
1:32:26 We're going to grab Aroma Spa is going
1:32:29 to go there
1:32:31 and woman is going to go there. Okay.
1:32:35 Context Prompt
1:32:38 strength, we'll go
1:32:40 7.5. Private mode, we don't care.
1:32:43 Advanced settings. Number of inference
1:32:46 steps, let's bump that up to 35. Safety
1:32:50 tolerance, let's drop that low. Oh, you
1:32:52 got to pay for that. Okay, fine. Okay,
1:32:55 so we've got we've got our product and
1:32:58 we've got our lady. So then we say um
1:33:04 woman
1:33:08 uses
1:33:12 spray bottle.
1:33:17 Um
1:33:19 in
1:33:21 bathroom
1:33:24 that looks like a high
1:33:29 and
1:33:32 luury
1:33:33 spa.
1:33:41 I don't I think I think Flux can do text
1:33:44 as well. So let's say um title text
1:33:50 over the
1:33:54 image
1:34:07 says
1:34:10 what is it? What's the product called?
1:34:17 Aroma Spa. Okay.
1:34:20 Um
1:34:36 Spray. Ah, here we go. Spray your way to
1:34:41 bliss.
1:34:47 All right, let's see. Let's see what we
1:34:49 get.
1:35:00 If I have to redo this prompt, I'm going
1:35:03 to be pissed off.
1:35:09 You [ __ ] kidding me?
1:35:28 woman.
1:35:30 Uh, woman sprays
1:35:38 from bottle
1:35:41 in a bathroom that looks like a
1:35:47 No, I haven't seen that yet. in a in a
1:35:51 that looks like a luxury
1:35:55 vacation spa. Vacation
1:35:58 spa title text
1:36:04 on top of image
1:36:08 says, what was my tagline? Spray
1:36:12 away.
1:36:15 What did I say? Anybody?
1:36:18 Anybody remember the brilliant
1:36:20 copywriting I did?
1:36:24 Uh, spray
1:36:26 Oh, wait. Spray yourself to bliss or
1:36:28 something like that. Spray
1:36:33 yourself to bliss. That wasn't quite it.
1:36:37 Anybody?
1:36:39 Anybody? Anybody? Is anybody listening?
1:36:42 Spray your way to bliss. Yeah, spray.
1:36:45 That's it. your way to bliss.
1:36:50 Um, by the way
1:36:54 to
1:36:56 to bliss.
1:36:58 Remember when I joke all the time about
1:37:00 having a pissed-off Gen Xer in the in
1:37:03 the corner office?
1:37:06 The fact that this site lets me come
1:37:08 here, upload images, type in a prompt,
1:37:11 and then when I hit generate, then it
1:37:14 forces me to log in, and then it erases
1:37:17 the [ __ ] that I just typed in there.
1:37:20 That's the kind of [ __ ] the Gen Xer in
1:37:22 the corner should walk in and ring the
1:37:24 [ __ ] neck of the engineer that put
1:37:25 this site together.
1:37:31 It didn't remember my settings.
1:37:34 Damn it. All right, let's see what we
1:37:36 got here. We'll see if this works.
1:37:39 Sorry, I was getting a snack. That's
1:37:40 okay. Spray your way to bliss. All
1:37:42 right. Wait, John Harris, by the way, I
1:37:45 scored a new position at my company as a
1:37:50 lead AI architect.
1:37:52 Embrace the jank. Yes.
1:37:56 All right, we got we got an image.
1:38:06 it. First of all, it doesn't look like
1:38:08 her at all. Um,
1:38:15 [Laughter]
1:38:18 look at the look at the [ __ ] the
1:38:20 double spray bottle. And she's she's
1:38:22 spraying she's spray spraying it on
1:38:25 herself.
1:38:31 Um, but anyway, look at the product with
1:38:34 with the dramatic exception that the the
1:38:38 uh product is [ __ ] wrong, right? It
1:38:41 it's got two spray handles and it's
1:38:43 spraying in her face. Um, that is the
1:38:46 shape of the bottle. That is the the
1:38:48 logo, right? So, one of the nice things
1:38:51 about Flux context is that you can um
1:38:56 you can change things in the image. Like
1:38:58 I should be able now
1:39:01 like how would I do I put that there?
1:39:04 No. Let's see.
1:39:08 How do I get rid of this?
1:39:20 Okay.
1:39:22 So now I can take that as the image and
1:39:24 say um
1:39:27 the top of the bottle
1:39:31 should only have one handle
1:39:36 and should
1:39:40 be spraying away from her,
1:39:45 not in her face.
1:39:50 Yeah, exactly.
1:39:52 Plot twist is chloroform. That's pretty
1:39:54 good.
1:40:02 Okay. Are we generating not safe for
1:40:04 work? What do you mean it's not safe for
1:40:07 work? What's not safe for work? What the
1:40:08 [ __ ] are you talking about?
1:40:15 generate.
1:40:21 Oh, all right. Whatever. Okay. Embrace
1:40:23 the jank. Sometimes [ __ ] just doesn't
1:40:25 work.
1:40:28 Oh my god. Um, okay. Let's uh So,
1:40:35 yeah, play around with flux context. It
1:40:38 It's It's pretty good. There's There's
1:40:41 other sites that have it. Um, I just
1:40:44 forget what they are. I gotta I gotta
1:40:45 get my [ __ ] together and go learn it
1:40:47 because it's it's one that's actually I
1:40:49 feel like this is a really important
1:40:51 tool. Um, because
1:40:54 chat GPT is okay at this. It's not great
1:40:57 at this. Um, so anyway, whatever. All
1:40:59 right,
1:41:01 I'm done now.
1:41:04 Um, it sounds like some some folks from
1:41:07 in here ran over to LinkedIn and and put
1:41:09 up some supportive posts. That's fine. I
1:41:12 appreciate that. Um my request would be
1:41:15 if you're going to do that um
1:41:19 you know do it with empathy and do it
1:41:22 with um
1:41:24 just like the sense of optimism and
1:41:26 embrace the jank that you guys already
1:41:28 have. Um just share your experience. Um
1:41:32 you don't have to you don't have to
1:41:33 fight the fight. Um but if you have if
1:41:35 you have something to say just you know
1:41:37 I would say please please don't get
1:41:39 personal in there. um to the extent that
1:41:41 you can. There's plenty of that going in
1:41:42 on already in there. Um just um just
1:41:47 have empathy. Like we we are in scary
1:41:50 times and and people who feel like their
1:41:52 livelihood is at fundamental risk um are
1:41:58 [ __ ] terrified, right? And I get
1:42:00 that. So um so anyway, that's that.
1:42:05 Um tomorrow is just a normal Tuesday. We
1:42:08 don't have an AI salon tomorrow. So, if
1:42:10 you're around, you should come here.
1:42:13 It's It's
1:42:15 24 hours before
1:42:17 Gro 4 launches, which is on Wednesday,
1:42:21 which if you look at my TikTok video, I
1:42:23 said it was Wednesday in there. Some
1:42:25 people are claiming I said it was
1:42:26 tonight. I didn't. Prove it. Prove it.
1:42:29 Go play the video back and see what I
1:42:32 say.
1:42:37 Tik Tok question. Have you tried
1:42:40 Flora.ai? I think I have. Thi This feels
1:42:43 like one of those ones that comes up
1:42:45 once a month. Flora.ai.
1:42:53 No. Oh, maybe it's flora.com.
1:42:58 Flora Aai.
1:43:02 Flora AI logo redesign.
1:43:07 up. Um, oh, this is the one with the Oh,
1:43:10 the intelligent um canvas. These things
1:43:13 are really cool. Um,
1:43:17 they make my ADD hurt.
1:43:22 They require a lot of follow through and
1:43:24 a lot of
1:43:25 a lot of thinking. Oh, look. I've done
1:43:27 stuff in here.
1:43:32 Look at look at me. So, not only have I
1:43:34 been to Flora AI, I've created some
1:43:37 workflows. I started with a Mustang.
1:43:40 Oh, this has got Flux in it. This has
1:43:43 got Flux.dev in it. All right. Yes, I
1:43:46 Yes, I've absolutely used flora. It is
1:43:50 we are definitely not at a stage of AI
1:43:52 where there are so many [ __ ] tools
1:43:55 that it is possible that you have
1:43:57 created an account, generated workflows,
1:44:01 and can't even remember that you did
1:44:03 that. Yes, that's that's the stage of AI
1:44:06 we're in. Oh, look, a muscle car in a
1:44:09 warehouse. I know, right?
1:44:15 Um, this site, the interface on this
1:44:19 looks really, really good.
1:44:22 My challenge with these sort of
1:44:24 interfaces is that they've got all sorts
1:44:27 of micro
1:44:29 um micro features and micro like if you
1:44:32 if you click this thing and then drag
1:44:34 the line then it gives you six options
1:44:36 which you can choose the one. Like it's
1:44:38 just it's what these things end up
1:44:40 becoming is
1:44:43 you're not just doing AI, you're
1:44:45 designing the dashboard of a 747 at the
1:44:48 same time and it just makes my head
1:44:51 hurt. Like I either want to be in
1:44:52 creative mode or I want to be in
1:44:54 programming mode. And these things are
1:44:56 this weird hybrid of creative and
1:44:58 programming. Um they're not bad. I just
1:45:02 find myself
1:45:04 I want to spend more time in these like
1:45:06 N8N is one of these that I want to be
1:45:08 able to spend more time in and and make
1:45:11 some visual workflows, but I just don't
1:45:13 end up getting there. So, uh I will do
1:45:16 that at some point. That's it's a really
1:45:17 good Thanks for the reminder. Uh because
1:45:20 I I just this actually does look like a
1:45:22 really good site.
1:45:24 Okay. Um I've had so much fun tracking
1:45:27 down subscriptions I signed up for and
1:45:29 canceling them. can't even remember all
1:45:31 the things that I would play with. I
1:45:33 know. Me, too. I heard Wednesday. Yes,
1:45:35 John Harris. It's Wednesday. Uh the the
1:45:37 Gro four thing is Wednesday at 8:00 PM
1:45:39 Pacific. So, we'll we'll do that live
1:45:41 here. Um all right, I'm going to get out
1:45:43 of here. Hope you had fun tonight. Uh
1:45:46 and uh and uh if you stumble into an AI
1:45:49 hater out there, they're just scared.
1:45:51 Give them give them some grace. All
1:45:53 right, peace out and I will see you
1:45:55 tomorrow. Bye.
1:45:56 [Music]