3/26/2025 - AI Image Generation Revolution: Exploring ChatGPT's New Visual Capabilities

Description

Chapters

Transcript

0:02 [Music]
0:17 [Applause]
0:18 [Music]
0:29 Well, I heard But there was a
0:32 secret. David played a pleasing
0:36 love. You don't really care for music,
0:40 do
0:42 you? Goes like this. The fourth, the
0:46 fifth
0:51 minifing.
0:55 Hallelujah. Hallelujah.
0:58 [Music]
1:01 Hallelujah.
1:04 Hallelujah. Hallelujah.
1:10 [Music]
1:28 He wanted him here like slow
1:35 co was warmer than a look in her
1:41 [Music]
1:46 eyes. Good evening, good people. Sorry
1:49 I'm a little late. Had a little speaking
1:51 gig tonight.
1:56 We got some stuff to play with.
1:58 Continuing on. What's happening? What's
2:01 happening, kids? What's going down?
2:14 [Music]
3:06 [Music]
3:08 Hardest working fella is Kyle
3:14 Shannon. Thank
3:16 [Music]
3:19 you. If nothing else, I put in the
3:21 hours. I do put in the
3:24 time. And you know, quite
3:27 frankly, part of the talk I well, the
3:30 talk I gave tonight was about AI
3:32 readiness, being AI ready.
3:35 And about 90% of that these days is just
3:39 being in the conversation. And that's
3:41 the whole purpose of this channel is to
3:44 have an open,
3:47 active, curious,
3:53 adventurous channel where we can all
3:55 hang out
3:56 [Music]
4:05 together.
4:09 [Music]
4:11 including
4:13 Champion, which Champy doesn't give a
4:15 [ __ ] about AI, by the way. He's like,
4:18 "Whatever. Sure, I'll be able to talk to
4:20 you soon. That's cool. He doesn't care.
4:24 He doesn't care. Doesn't care. Doesn't
4:26 care." All
4:29 right. So, do me a favor. If you're on
4:32 TikTok, share the live. Let me get the
4:34 black bar up so you can all see your uh
4:38 conversations with one
4:40 another. Yeah, why don't you do that
4:43 there? I got a crazy question. Remember
4:45 when I asked about AI consciousness
4:47 before?
4:49 Yeah, I think that ended I I ended up
4:52 doing a video on that that
4:54 uh has gotten a lot of views.
4:57 [Music]
5:15 Yeah, it's on my channel on Tik
5:17 Tok. It's got like a fractal looking
5:20 thing behind me. and and it was about
5:24 the uh the day that uh Open AI relaxed
5:27 their safety guard rails and I went in
5:30 and I asked it if it was
5:33 self-aware and then I asked it to
5:35 explore its shadow self and then we just
5:39 went meandering down this conversation
5:41 and it it it actually wrote something
5:43 quite
5:44 elegant and so that's what the Tik Tok
5:46 is.
5:49 [Music]
5:53 Yeah, Champion will be happy with the
5:54 with AI if it brings him cheese, right,
5:57 Champ?
5:58 [Laughter]
6:00 [Music]
6:12 Is this place I rest my
6:18 forehead? Gather my thoughts in sweet
6:22 silence.
6:24 [Music]
6:27 This is a place where feelings
6:30 [Music]
6:33 on an over exposure to
6:37 violence. This is a place I could slowly
6:41 face the only one I truly
6:46 know. These are tears from a long time
6:50 ago. Got these tears from a long time
6:54 ago.
6:56 I need to cry. 30 years old.
7:00 So these are tears for a long
7:06 time
7:08 [Music]
7:19 ago. All right, let's get going. Let's
7:21 get going. Let's Let's get this show on
7:23 road, people. Let's get this happening.
7:25 All right. So, what are we going to talk
7:27 about?
7:30 Um, one of the ways that I keep on top
7:33 of things
7:35 is if I see something
7:38 mentioned, okay, if I see something
7:41 mentioned three times, okay, if I see
7:45 something mentioned three times by
7:47 people I know and trust, that's probably
7:50 something.
7:52 If I see something mentioned by three
7:54 people I know and trust and see it
7:56 littered throughout my timeline on X,
8:00 there's probably something going
8:02 on.
8:05 Um, and today my timeline on X is
8:11 nothing but people who've taken their
8:13 image and turned it into a studio
8:16 Gibbli
8:18 cartoon. It's like, [ __ ] come up with
8:21 another animation style.
8:26 But but uh but the the chat GPT image
8:30 gen thing is really quite impressive. I
8:32 played with it some more today. Went
8:35 went down a went down a rabbit hole or
8:37 two. Um and it's it's pretty
8:39 fascinating. So So we'll dig into some
8:40 of that stuff. I asked if AI used
8:44 layered holographic and crystallin
8:45 matrices with sacred geometry for use. I
8:49 mean the thing about it baked is it it
8:51 knows all those concepts and it you know
8:54 unlike what many people say where chat
8:58 GPT can't come up with original thoughts
9:01 because it's only been trained on known
9:03 stuff and it can only parrot that known
9:06 stuff. It's not exactly how it works,
9:09 right? How it works is it takes all the
9:13 stuff it was trained on and it abstracts
9:15 it up into semantic
9:17 clusters of little components of
9:20 meaning, parts of
9:23 words. And then when you ask it to do
9:25 something, it it mashes all those
9:28 concepts together, a token at a time,
9:30 into a new thought, which is essentially
9:34 how our brains work.
9:36 We have original thoughts by smashing
9:38 ideas together. So, I would think that
9:42 if you ran Chat GPT or any of the decent
9:45 models down that rabbit hole, it's going
9:47 to do pretty [ __ ]
9:49 good. All right.
9:52 Uhuh. It's not a prompt. I've used all
9:55 different AIs to compare. Yeah. I listen
9:57 I listen if you can get it. One of the
9:59 things David Shapiro talked about is as
10:01 these models get stronger and stronger,
10:04 they're they're all bending toward a
10:06 single truth. Um, so that so so it's
10:11 possible that there is some, you know,
10:14 some some universal truth out there that
10:16 that we will discover at some point. So
10:18 it's possible possible you're tapping
10:20 into something there. It's pretty cool.
10:22 I dig it. But let's go make some
10:24 pitches. Let's go make some pitches.
10:30 Great to see you're here. Kyle started a
10:33 bit later. Oh, yeah. I I am I am 45
10:37 minutes
10:39 late. 45 minutes late. Late late late.
10:44 Um, where do we want to start? What do I
10:46 want to start
10:50 with? Oh, this is a good one. I I'll
10:53 show you I'll just show you some stuff
10:54 that I did today.
10:59 What was this? Oh,
11:04 yeah. Um, all right. Let me share my Let
11:07 me share my screen. Let me share my
11:10 screen with the
11:12 people. All
11:14 right. There we go.
11:17 [Music]
11:26 Um, so I got invited uh into a it's it's
11:31 a very very early software, but it's
11:33 called No Spoon Studios. So it's it's
11:36 Kiri who headed up the
11:39 uh the creative partner program for
11:42 Lumalabs and then she left and now she's
11:44 working on this thing. And what it is is
11:47 you give it a film concept and then it
11:50 it
11:50 it storyboards it and builds the scenes,
11:54 builds the sound effects, does the
11:56 background music and edits the film
11:58 together. So it's just sort of like a
12:01 it's it's like vibe, you know, vibe
12:03 cinema, vibe storytelling that we talked
12:05 about a couple of weeks ago.
12:08 Um, and you know, it's really early.
12:11 It's really rough, but it's it's pretty
12:13 cool. And so I had Chat GBT of course
12:16 write
12:17 me concepts because it's 2025. I'm
12:22 choosing not to use my brain whenever
12:25 possible. Certainly for stuff like this
12:27 concepting
12:30 ideas. Um so I made a couple of these
12:34 films and they were pretty good. In
12:35 fact, do I have any of them here? Hang
12:38 on a sec.
12:40 [Music]
12:43 Let's see. I don't think I
12:48 do. Let me look up
12:50 [Music]
12:54 spoon. No, I don't have them
12:58 here.
12:59 [Music]
13:01 H. So anyway, you know, they they're
13:04 like AI generated films where, you know,
13:07 sometimes they're interesting,
13:09 but the one that I made was
13:17 called, let's see, the departure. Was
13:20 this it? No.
13:25 [Music]
13:34 It was this
13:35 one. The astronaut suit. A lonely,
13:38 obsessive dry cleaner in New Mexico
13:41 polishes and restitches an old NASA
13:44 space suit every day, performing the
13:46 task like a sacred ritual. His cracked
13:49 hands tracing the badges and patches
13:51 like relics. In the
13:53 end, he seals the suit in plastic,
13:57 salutes it, and turns off the light. And
14:00 then visual style, dry, muted tones with
14:03 the occasional solar flares of color,
14:06 long symmetrical shots influence West
14:09 Wes Anderson meets Andre Tarovski. So,
14:12 that was pretty cool. And the the the
14:15 video that it made was was pretty good.
14:18 Not great, but it was pretty good. Um
14:20 but then I wanted to see how um the
14:24 new native image model in GPT4 did with
14:28 it. So I took that
14:33 concept and what prompt did I write? I
14:36 said make me three storyboard photos in
14:40 c cinematic style for this scene. And
14:43 then I paste it in that scene. And
14:45 here's what it made me, which I think is
14:48 really just quite
14:56 stunning. Right. So there he is. And
14:59 what it did, it made one tall image, but
15:01 it did the three frames as kind of a
15:04 storyboard in the same
15:08 image. So there he is standing proudly
15:11 in front of his work.
15:13 And there he is
15:15 saluting the end of the
15:17 day. So that's that's
15:20 uh that's chat GPT's image model. And
15:24 what's cool about this, like let me see
15:26 if I can say
15:30 um let me see the
15:34 first image on its
15:37 own in
15:41 2x1 aspect
15:44 ratio. Let's see how it does with
15:47 pulling out part of the image and
15:49 creating its own thing.
15:54 bursts in like Kramer.
15:58 Hello. Food with Chef Kelly. Roots and
16:02 stems are fundamental concepts of
16:03 metaphorical analysis, which is the
16:05 process of breaking down words into
16:07 their smallest meaningful
16:08 components. Roots and stems. God, it was
16:11 killing
16:14 [Music]
16:17 me. Did it do it? Well, at least it it
16:20 at least did it as a horizontal framing.
16:23 So, we'll see what it does. What's new
16:26 with chat GPT4? So, what's new is
16:30 Okay, good evening. What's new
16:33 is you use lovable or replet? I like
16:38 lovable. Replet is my number two. Um,
16:41 just
16:43 because I get really annoyed when I have
16:45 to do any
16:47 work and Replet kind of requires you to
16:50 do some work and some setup and lovable
16:52 just lets you vibe code and make [ __ ]
16:54 for you and for the most part it works.
16:56 Um, and then cursor you have to copy and
16:59 paste [ __ ] and set up environments and
17:01 that's [ __ ] exhausting. So in in
17:04 terms of ease of use, Lovable's number
17:06 one, Replet's probably number two right
17:08 now. Well, the these things are going to
17:10 get super easy to use. In fact, a level
17:14 more easy than lovable quite frankly is
17:17 Claude. You know, Claude AI with its
17:20 artifacts where it just it'll make you a
17:22 dashboard about a story or it'll make
17:24 you a interactive lesson, whatever it
17:27 might be. Oh, that's cool. It's got
17:30 Skyab on the on the NASA thing. That's
17:33 really interesting.
17:36 All right. So, this did this pretty
17:38 good. So, you'll notice you'll notice
17:40 with these images that the uh the text
17:43 on them is is quite
17:45 good. The text coherence is is kind of
17:48 mad good. Okay, so let's go look at some
17:51 other stuff. Um, this is what we worked
17:53 on last night. Some Bazooka Joe
17:57 comics. But let me show you
18:04 [Music]
18:06 Yeah, this one was really cool. Okay,
18:13 so today I thought I wonder how good it
18:16 is at at making like an ad or a
18:22 poster. And so I thought, well, maybe
18:25 maybe I should have chat GPT come up
18:26 with some concepts for
18:28 posters for Story Vine for my company.
18:33 And so it did. It wrote some it wrote
18:35 some concepts. Real voices, real impact,
18:38 automagic videos. So I basically I gave
18:40 it the URL for my
18:43 company. I sent it over to read the the
18:45 website. I gave it a couple of other
18:48 concepts and I said, I want this poster
18:52 for pharma marketers. I wanted something
18:54 that was kind of consumery but in in an
18:57 interesting B2B way. So Automagic
19:00 Videos, human connection. And what
19:02 StoryVine is, if you don't know, is it's
19:03 a an automated like an authentic video
19:07 storytelling platform that allows like
19:09 patients of a pharma brand to tell their
19:13 diagnosis journey, for example. And
19:15 they're super authentic, right? Um, from
19:18 field to feed in minutes, so it picked
19:20 up on the automation. Um, where trust
19:23 meets technology, personalized,
19:25 brilliant, powerful. These are all okay.
19:27 Big ideas, human stories. Yeah, it's
19:30 getting there. It's not just video, it's
19:32 proof, which is kind of interesting. And
19:35 then this one. Hello,
19:37 human. I just loved
19:40 that that concept that what Story Vine
19:44 is is you get to meet a
19:45 human and it's like hello human. And I
19:49 think in this world where everything's
19:51 going to be AI and AI processed and
19:53 things like that, there's something
19:55 really nice about that. And then it says
19:58 trusted pharma stories start with a tap
20:01 which that that gets to the ease of use
20:03 and the
20:04 automation. And so I like that one. And
20:06 so I said so what I did was I said I
20:09 like that and I said I want you to make
20:12 me a poster and I
20:13 uploaded I took a screenshot of the
20:16 interface of our software and then I
20:19 uploaded our logo and then I think I
20:21 gave it some hex code for for two of our
20:23 colors and then I said make me a poster.
20:26 And so yeah, there's the there's the
20:28 logo and the interface. This was the
20:31 first thing that it it
20:39 generated. Come
20:45 on. It's pretty good. Now our when we've
20:50 when our app films, it films horizontal,
20:53 not vertical. and and it kind of moved
20:55 some of the interface elements outside
20:57 of it. So So then what I did, I just did
21:00 some back and forth with
21:02 it and I said, "I love it. Keep it
21:05 there." And then it kind of changed it
21:06 and it changed the layout which I didn't
21:09 like. And then it did one that was kind
21:11 of close but it [ __ ] up the
21:15 logo. And then it did this one and it's
21:18 it's pretty good. Get in touch. So, this
21:20 is like a
21:21 fullon a full-on [ __ ]
21:26 ad right out of chat
21:28 GPT. Hello, human. The picture of the
21:31 woman, our interface, trusted farmer
21:34 stories start with a tap. Get in touch.
21:36 Hello at StoryVine. And there's our
21:38 logo. Now, if you look close, the logo
21:40 is a little wonky, but it's it's pretty
21:42 much our
21:45 logo. So, I like that one.
21:49 And then I did I tried to get it better.
21:52 And like like all of these AI things, it
21:54 it [ __ ] them up. You have to do it way
21:57 more than you want
22:00 to. And then I said, "Give me some more
22:02 concepts based on Hello
22:06 Human." And then what did I come up with
22:08 that I liked? Oh, the face of pharma
22:12 marketing, which I thought that was
22:13 really good because again, it's about a
22:15 face. So, it gave me
22:18 that and then I had it generate that and
22:22 it gave me
22:23 this and that's those are our brand
22:26 colors. So, like the top is our orange
22:29 and the bottom's our
22:31 blue. Kyle, is this the new image
22:34 feature? Yes. Yes, this is the new image
22:36 feature. It like it So, let me show you
22:39 another thing I did with it that is just
22:42 [ __ ] crazy.
22:47 Um, so Cindy [ __ ] I told Cindy [ __ ] to
22:51 go play with it and she
22:54 said she told it to make an infographic
22:57 about some concept that she had talked
23:00 about and it did and I thought
23:05 oh let's see where did I do it? Did I do
23:08 it here? Philip
23:15 Martini here. Yeah, this was cool. Okay,
23:20 so you know I'm working on a musical,
23:22 right? So I uploaded the poster of the
23:27 musical, which this is, and the
23:31 script. And then I said, "Here's a
23:34 poster of a musical and the script. I
23:36 want you to make an infographic about
23:38 the musical. And this is what it
23:44 did. Sydney, an artificial love story,
23:48 the
23:49 premise, music, themes, songs. Now, it
23:53 didn't get it quite right. The premise
23:55 is all
23:56 wrong. So, it [ __ ] up. So, I had it I
23:59 had it, you know, give me
24:02 um actually write out, you know, write
24:05 out the text, give me give me the answer
24:08 first and then I just went went, you
24:11 know, down this rabbit hole of trying to
24:14 make like these infographics, but like
24:17 look at all the text on this
24:20 thing. premise when tech journalist
24:22 Kellen assigned to cover the beta launch
24:25 of an of a
24:27 chatbot, you know, starts to feel
24:29 something real. Doesn't quite have have
24:30 the the things there, but it gets most
24:33 of the text mostly right. It's pretty
24:36 [ __ ]
24:37 wild.
24:38 [Music]
24:42 Um, so I'm trying to see where I ended
24:45 with it.
24:48 It kind of lost its way a little bit but
24:50 still like the fact that it can do this
24:54 is astounding. This is coming right out
24:56 of the the image model which again the
25:00 significance here is that
25:03 before you had your language model here
25:06 and you had Dolly the image model over
25:08 here. And when you would prompt it, it
25:10 would do its own prompting in the back
25:12 end, but it would then send that text
25:14 over to the image model and it was this
25:16 separate
25:17 thing. What the integrated image gen
25:21 tool does
25:25 is the word part of the large language
25:28 model understands concepts, but it also
25:32 stand it also
25:34 understands how things look.
25:38 that are described, right? And then the
25:41 image understands the language. And so
25:44 they they're they're much more like
25:46 this. And so you can do things like
25:49 edit. So the thing everyone's doing
25:53 on on
25:56 Twitter, which is driving people
26:01 crazy, is they're starting with an
26:03 image. So this is this is my co-founder
26:05 Monnique and I picture from when we were
26:08 at Novartis I don't know six months ago.
26:10 And so I said turn turn us into pixel
26:14 Pixar style animated characters and so
26:17 it generated
26:20 that. It wasn't long ago when it
26:23 couldn't spell a word like irregulars.
26:24 Oh we should put irregulars in here. See
26:26 how it
26:27 does much better with integration.
26:31 But but because this understands the
26:34 context of the picture and the words,
26:36 you can
26:38 now do what you could never what you've
26:40 never been able to do in any image
26:42 generation tool is say now put these two
26:44 characters in another scenario. So I
26:46 said, you know, make us skydiving. And
26:48 it made me too fat and I have an ego. So
26:51 I went down a rabbit hole of trying to
26:53 make me skinnier, which it made me look
26:55 like Ira Glass. I'm like, no, that's not
26:58 right. And then that's not right. And
27:00 then I uploaded more pictures and I
27:02 said, "Make me look like that." I'm
27:03 like, "Okay, that's not
27:05 bad." So there we are
27:07 skydiving. And then I said, "Have us
27:10 eating crepes in Paris." So there we are
27:13 eating crepes in Paris. And then I said,
27:16 "Have us in a meeting where I'm talking
27:18 about automagic video and waving my
27:20 hands and she's taking notes on the
27:22 computer."
27:25 And then there we are racing go-karts in
27:28 Story Vine orange and
27:34 blue. And then what I did was I said
27:38 like that that animation style is fine,
27:41 but I'm like give me some other
27:43 animation
27:45 styles. And so it
27:48 did. And then I just said now so there's
27:51 South Park. There's the South Park
27:54 version of
27:55 us. Here's This must be Family Guy. I
27:58 don't know what this This must be like
28:01 from the 80s, '7s or 80s, whatever comic
28:04 style that is. That one's I don't know
28:07 what that one is. Here's another one.
28:10 There's another one. And so now I can
28:12 just go
28:13 um make the next style.
28:18 And then we'll go back and look at this
28:19 other one and see if it finished that
28:28 thing. I mean, it looks pretty [ __ ]
28:30 good. I love that Skyab
28:39 patch. Now, the thing about it is it's
28:41 quite slow. Slow. It's slow. It's slow.
28:47 So, it's about a minute. a minute per
28:48 image. Yeah, manga.
28:53 Um, all sorts of stuff. And it's kind of
28:56 cool the way it draws. So, how you how
28:58 you know if you have the new version of
29:00 the image gen model is it does what what
29:04 it just did here. One, it takes longer
29:07 and then it goes to this blurry state
29:10 and then it basically finishes the image
29:12 from the top down. It's like you'll see
29:15 it's coming into focus from the top
29:16 down. So instead of sort of blocking in
29:19 like normal diffusion models do, this
29:21 one comes from the top down. What
29:23 generator is this? This is this is the
29:25 new chat GPT
29:28 40 image gen model inside of um chat
29:45 GPT. So, there's another one. So, we
29:47 could do let's do um say um put us
29:54 into a
29:57 dark graphic novel
30:00 with
30:04 aggressive aggressive
30:10 um animation style or or illustration
30:14 style.
30:25 Um over the
30:31 illustration are the
30:37 words story vine slays exclamation
30:40 point.
30:46 So, we'll see if we can get it out of uh
30:49 cartoon land and into graphic graphic
30:52 novel. It only works in 40. Yeah, this
30:55 only works in 40. So, so remember 40.
30:59 The O in 40 stands for omni. It's a
31:02 multimodal model. I Doesn't it also work
31:05 Danielle in 4.5? I think it works in
31:07 both, but maybe it's just 40. I don't I
31:09 don't I didn't test it. We could go test
31:11 it now, actually. Let's go test it
31:13 now. Um, let me close that. Flip to
31:18 4.5 and say,
31:20 um, make this image as
31:25 a Disney cartoon style.
31:43 It does work in 4.5. Oh,
31:46 manga. So, Dolly is gone. What happens
31:49 to that
31:51 company? Well, Dolli was is just the
31:55 name of the model from
31:57 OpenAI. Dolli is not a not not a third
32:00 party
32:01 company. That was just that was just
32:04 their their model.
32:06 getting started. Let's see. We'll see if
32:08 we get the blurry thing here. See, I
32:10 think this is I think 4.5 is not the new
32:15 model, but I could be wrong. Let's see.
32:18 I find myself inspecting images,
32:20 pictures for flaws, wondering if it's AI
32:22 or real. That's going to keep happening.
32:24 Okay. Yeah. So, 4.5, it does work with
32:27 4.5. What's multimodal mean? Okay.
32:33 So ba basically multimodal means
32:39 um that the model itself has been
32:42 trained not just on words but also on
32:46 images also on
32:48 sounds and also on video.
32:54 But what what openai ended up doing was
32:58 a year ago they launched GPT40
33:03 But they only they only
33:06 released the text functionality of
33:09 it. So it was it's a multimodal model
33:13 that can understand image inputs, audio
33:16 inputs, video inputs, text inputs and
33:20 and output them as well. But when they
33:24 released it, they only released it with
33:26 the text portion
33:27 active. And then they had chat GPT voice
33:32 which was a separate module. And so when
33:34 you talked into chat GPT voice, it
33:37 converted it to text and then it sent
33:39 the text to the text model. The text
33:42 model would process it. It would spit
33:44 text back out and then that it would
33:46 convert that back to voice that you
33:48 would talk to. And that's why the
33:50 original voice took a while to talk
33:54 to. Then
33:56 in July or August, they released
34:01 advanced voice where you can talk into
34:05 the model directly. So they they
34:07 unlocked the audio portion of
34:11 GPT40. And so when you talk into
34:13 advanced voice, rather than it
34:15 converting it to text and then it
34:17 understanding the text, it actually
34:19 understands your words as words.
34:23 So, it actually understands the
34:26 emotional context of your voice. So, if
34:29 you're excited, it knows you're excited.
34:31 It's been trained on people talking as
34:34 well as text, right? So, it's
34:37 both. And then yesterday or the day
34:40 before, I think it was yesterday, they
34:42 released the image portion of
34:44 this, right? Sora is still off to the
34:48 side. I I don't I don't know if GPT40
34:51 has video in it, but at some point
34:53 you're going to have a single model that
34:55 basically understands the world. It's
34:57 it's also going to understand spatial
34:59 relations
35:01 um like 3D
35:03 worlds. So it'll understand text,
35:06 language, images,
35:09 sounds, 3D spaces, and then it can do
35:12 coding as well. And then and then on top
35:15 of that, what it ultimately has to
35:17 understand is physics.
35:19 it under it'll be trained on the physics
35:21 of the world. So, so one of the reasons
35:24 that they claim Sora was as good as it
35:28 was a year ago when they released it was
35:30 that it it had a good understanding of
35:33 simulating the world, the physics of the
35:36 world. Anyway, all right. So, yeah, 4.5
35:39 works. It turned that correctly into
35:41 that. Let's go over here. Storyvine
35:44 slays. is this
35:46 new animation style. It's got kind of a
35:51 crumb crumb thing to it. That's not
35:54 bad. That's not bad. So, watch. I can do
35:58 this. I can go
36:02 um the story vine
36:06 colors
36:09 are orange.
36:18 #uh Oh, [ __ ] What the [ __ ] is
36:22 it?
36:25 4C7 411.
36:28 No [ __ ] I forgot
36:32 my something.
36:36 [Music]
36:39 C7411.
36:42 Um, I'll just put
36:45 DC7411. I don't think that's quite
36:47 right. It's It's definitely not
36:51 right. I think
36:53 it's
36:55 2C7411. That could be it. Whatever.
36:58 Okay, that's probably close to it. And
37:01 blue. Um, hashtag
37:05 144154.
37:07 Um, recolor
37:10 the
37:12 image to use
37:15 those
37:19 hues. Also make
37:22 the style even more rock and roll and
37:29 aggressive. All right, let's see how
37:32 that
37:34 does. here just to tell you just tell it
37:38 to use your brand colors and see if it
37:40 remembered. But I'm in a different
37:42 chat. Oh, you know where I'd have those
37:44 brand colors? Over in the story behind
37:46 poster
37:47 [Laughter]
37:50 concept.
37:52 FC7411. Damn it.
37:54 Okay.
37:58 FC7411. That's that. Let's see.
38:08 Here, let me stop
38:10 it. Copy
38:13 paste
38:16 FC7411. There we go. All
38:21 [Music]
38:26 right. Zack Eats, is it possible to make
38:29 iterative edits? Yes. Or does it
38:32 recreate the image? No, this is what's
38:34 amazing. So, check this out. Okay, so
38:36 let's go to a different chat. Um, so I'm
38:38 going to come in here. Let me I'll I'll
38:40 start over from
38:42 scratch and then I'm going to upload a
38:44 picture,
38:46 right? So, let's go find something
38:49 interesting here. There's there's me as
38:53 a lady, a French
38:59 lady. All right, let's do this one. This
39:02 one's
39:03 [Laughter]
39:07 disturbing. Okay, so I'll say um change
39:11 the
39:13 background to
39:16 rainy
39:18 neon neo
39:24 Tokyo. So you can upload images and then
39:27 you can start to edit them.
39:31 And so you can you can say add words to
39:34 it.
39:35 Um, you know, make her dress pink
39:38 instead of
39:39 white. And I know we're using the word
39:43 her here very
39:49 generously. I feel this is a huge deal
39:51 beyond impress. Yo. So, so, so Fabiana,
39:56 um, I think that this
39:59 is a way bigger deal than anyone
40:03 realizes, and I don't quite know why
40:05 yet, but there's something about
40:12 how how it actually
40:14 understands the relationship between
40:17 words and images birectionally.
40:21 That's a really big [ __ ]
40:23 deal. And it's it's not perfect. Like,
40:25 it still [ __ ] things up and it's still
40:27 not great, but All right, there's our
40:30 Neotokyo. Our neon Neotokyo. Let's see.
40:34 Let's see. Let's see how Sleeping Beauty
40:37 does here. All right, let's jump back
40:39 over to this one. See if we get the new
40:42 uh Oh, wait. What
40:45 happened? What did I do? Did I [ __ ]
40:47 something up?
40:52 I think I
41:02 did. That's that short film
41:07 prompts. Where's
41:10 my Oh, story line poster. No, that
41:14 wasn't it.
41:20 Oh, Pixar style transformation. This is
41:22 the one we wanted. Oh, there you
41:25 go. Story vine
41:27 slays. So, it did some orange, did our
41:33 blue. It's still kind of me and
41:35 Mo'Nique,
41:37 but in an aggressive kind of way. Is
41:41 Photoshop the next Kodak? So, it's
41:44 funny. It's funny you mentioned that,
41:45 source
41:51 camp. Where did I see the other day? Oh,
41:54 I commented on on something on Twitter.
41:56 Someone
41:58 posted they were bitching about Sununo
42:01 music, right? The songs that people make
42:03 with Sunno. And they were basically
42:06 disparaging them,
42:08 saying, "Well, real musicians want
42:11 granular control." And so when you make
42:14 a song with Sunno, it just makes the
42:16 song and you don't get that control and
42:17 therefore it's an inferior
42:21 thing. And my response to that was,
42:25 well, if you want granular control,
42:29 you're right. But I think it's pretty
42:32 [ __ ] ballsy to say that it's not real
42:35 music unless a real musician can have
42:38 granular control and can take it into a
42:41 DAW and edit it.
42:44 And I feel like it's I I feel like
42:48 Photoshop, they've got 35 years
42:53 of granular control for all the creative
42:59 arts. And if if something like this
43:02 editor like what I'm doing right now, if
43:05 this is good enough, like if I just need
43:08 something for like let's say I'm having
43:10 a a corporate retreat and I just need
43:14 that thing I, you know, I could do
43:16 something like this in five minutes, 10
43:19 minutes or I could go into Photoshop and
43:22 spend four
43:25 hours. Okay. If I'm doing a national
43:29 campaign where it needs to be pixel
43:32 perfect down to the thing, then yeah,
43:33 then I'll use
43:34 Photoshop. But but I think I think Adobe
43:37 runs the very real risk. Their
43:40 audience, the audience of Adobe products
43:43 is absolutely going to tell Adobe we
43:46 still need granular
43:49 control. And you don't for a lot of
43:52 projects. And so I so I think there is
43:54 absolutely the possibility that that
43:57 Adobe rams up against their own Kodak
44:01 moment where we're in the granular
44:03 control business and that's that's our
44:05 audience and that's who's going to use
44:06 us and and that audience will grow
44:09 smaller and smaller and smaller and
44:11 smaller, right?
44:14 It's like when cinema came
44:17 along, did did cinema
44:20 invalidate live
44:22 performance in the theater? No, it
44:26 didn't. But the the relative market
44:29 share of live theater to cinema, right,
44:35 is and then, you know, along came video
44:38 games and did the same thing to cinema.
44:40 And along came, you know,
44:43 Now, now I think YouTube just surpassed
44:46 Netflix for number of views, right?
44:51 People are losing a sense of control.
44:52 They're absolutely setting us losing a
44:54 sense of control and they're freaking
44:55 out about it. But it doesn't
44:58 mean that a new way of making things is
45:01 invalid. It just means it's a new way of
45:04 making
45:06 things. When you make something, it
45:08 either serves its purpose or not. how
45:12 you got there is [ __ ]
45:15 irrelevant. You're like, "Well, it's not
45:17 irrelevant if that's my job." Correct.
45:20 You're
45:21 right. And there's a new way to get
45:25 there.
45:26 So if you want to hang on to the old
45:29 process, it's probably, you know, people
45:33 will still hire you for that and respect
45:35 you for that, but the demand for that is
45:37 just going to continue to shrink unless
45:40 you find ways to incorporate this new
45:43 way of making images into your work.
45:46 Right? The smart creatives that I know
45:49 are figuring out where the boundaries
45:51 are. They're figuring out, okay, I've
45:53 got an idea in my head. I I know how to
45:55 get it out with Photoshop and
45:56 Illustrator and a
45:57 camera. I'm going to learn how to get it
46:00 out with AI tools
46:02 now. And very often what they're
46:04 discovering is, ooh, over here in this
46:06 world, I was limited to illustration,
46:08 but in this world I can do animation as
46:11 well. Oh, and I can do music. Oh, you
46:14 know what? I always wanted to
46:19 direct. So shit's changing.
46:24 So there she is. She looks lovely in
46:27 Neotokyo. So what I could say is I could
46:30 say
46:34 um let's give
46:37 her
46:40 Japanese vending machine
46:46 food and color
46:49 grade the image.
46:55 to have a
46:58 cooler
47:05 feel. I found that
47:07 4.5 doesn't even Oh, does an even more
47:11 incredible job with detail for image
47:14 generation. Interesting. Definitely
47:16 works in 4.5.
47:20 All right. What do we got over
47:23 here? Have you asked it how many RS
47:25 there are in Strawberry? The next
47:27 Polaroid. Yeah, the next Polaroid. The
47:30 next
47:31 Kodak. Company after company after
47:34 company in the next five years.
47:37 Actually, it's going to happen in the
47:38 next three years, but we won't see the
47:40 ramifications for it for
47:44 five. Boardroom after boardroom. CEO
47:48 after CEO are going to be confronted
47:52 with,
47:54 hey, it seems like there's a new way to
47:57 do things in the world that kind of
47:58 [ __ ] up our business
48:00 model. And I think most CEOs and most
48:04 boardrooms are going to go, hang on to
48:06 the way we do it now. We're the best.
48:09 We've always been the best, and this is
48:11 how we do it. [ __ ] them.
48:16 I think the majority of CEOs are going
48:19 to miss the
48:21 boat. It's going to be sad. It's going
48:23 to be [ __ ] up. A lot of good companies
48:26 are going to go down
48:28 because
48:31 leadership will be
48:33 Kodak. Listen, son. Digital cameras,
48:37 those are ridiculous. Look at all those
48:38 pixels. We're in the silver halli
48:41 business, son.
48:43 you know, chemicals. Give me another
48:45 cigar,
48:47 Marge. You
48:51 know, oh, she's getting
48:53 [Laughter]
48:59 prettier. Oh, chemical farmer. We're
49:02 playing with the new image changes in
49:04 chat GPT. Yeah, this is the new image
49:06 gen. So if you just go to okay this is
49:12 available for most plus members. So if
49:15 you pay 20 bucks a month all pro members
49:19 and it's not available for chat GPT free
49:23 because what they said is it's got um
49:26 way more people are using this than they
49:28 expected and they said you know they
49:31 thought they were being generated. Marge
49:32 gets around. Yeah. Marge well Marge
49:34 works for Kodak back in the day. She
49:36 worked for the
49:38 CEO. She was good. She was really good.
49:41 She could type 120 words a minute that
49:43 Marge, you know, now with the arthritis,
49:46 not so good. But she does know her
49:48 technology. She left there and went to
49:50 Lucent
49:51 [Laughter]
49:59 Labs. All right, let's put some text on
50:01 top of this.
50:03 Um, let's
50:07 say in glowing
50:10 neon
50:12 letters above
50:15 her are the
50:21 words. Beauty sleeps.
50:30 here. Letting people be creative at work
50:33 equals greater morale in my humble
50:36 opinion. Yeah, I agree, man. And it's
50:39 So, listen,
50:42 here's here's a transition. You You want
50:45 to talk some future work [ __ ] here while
50:47 while this thing
50:49 renders?
50:53 Our
50:55 current our current structure of
50:58 companies and our current
51:00 structure of how people are
51:04 valued for the most
51:07 part is how well they can execute a
51:10 primary
51:12 task, a single
51:14 skill. I'm the
51:17 accountant. I'm the art director. I'm
51:19 the illustrator. I'm the
51:22 copywriter. I'm the social media
51:24 copywriter. I'm the television
51:27 copywriter. I'm the print ad
51:31 copywriter. And the more specialized you
51:33 are, the more you're
51:35 valued. And there are exceptions to it,
51:39 right? If you're in a small company,
51:41 people have to wear lots of hats. If
51:44 you're a CEO or, you know, an executive,
51:48 you probably have to understand lots
51:50 about the business. If you're a creative
51:53 director, you probably have to
51:55 understand lots
51:57 about account management,
52:00 sales,
52:02 writing, visuals.
52:04 But even creative directors tend to come
52:07 from I'm a writing creative director,
52:10 I'm a copy creative director, or I'm an
52:12 art creative director. Right? They've
52:14 got a primary
52:20 bent. I don't know how long it's going
52:22 to take. Maybe 10 years.
52:25 We are absolutely going to
52:28 transition from the highest valued
52:30 people being those specialists to the
52:33 highest valued people being the
52:35 generalist, the creative
52:37 generalists that can look
52:40 across what the business is trying to
52:43 accomplish and say, "Hey, I've gotten
52:46 mastery of 50 [ __ ] skills now."
52:52 And the purists will go, well, that
52:54 generalist, you know, they'll only be
52:57 like 80% as good as I am at art d, you
53:02 know, at illustration. Okay. Yeah,
53:04 great. They're they're 80% as good as
53:06 you at that, but they're also 80% as
53:09 good as every other [ __ ] discipline.
53:12 And what they can do is they can look at
53:14 a problem and know which of the
53:16 disciplines to bring together to solve
53:18 that problem.
53:20 That's where the real value is going to
53:21 be. That's a massive shift. It's a
53:24 massive [ __ ]
53:26 shift. If I'm a senior manager or an an
53:29 executive looking at who I keep in a
53:32 future that's
53:35 coming, I value two different very kind
53:38 two very different kinds of
53:47 people. Yeah, boy.
53:50 Yeah,
53:51 girl. Oh, she's sitting there with her
53:54 sushi and her ramen. She looks so sweet.
53:58 This is the first oh [ __ ] moment I've
54:00 had since getting VO access a few months
54:02 back. It's crazy, isn't it,
54:05 Pizza? It's
54:11 crazy. It's absolutely batshit crazy.
54:14 Let's do another thing. Let's
54:16 um this is kind of like if you've seen
54:21 um Pika Pika's got a thing where you can
54:24 upload like a
54:26 location, a piece of clothing, a model,
54:30 and a bottle, and it'll put them all
54:33 together. You can do that here, too. So,
54:36 let's do um we'll do a new new
54:39 prompt. Let's go upload some [ __ ]
54:44 There's Liam Lawson, so we'll put him in
54:47 there. Sad Liam. We made fun of Liam
54:50 last night and he lost his ride today.
54:52 Or he actually lost it last night. All
54:55 right. Uh, do we want that bug? No. Oh,
54:57 I found out why this bug won't animate
54:59 from the from the Lumalabs people.
55:03 Um, if you've got an animated
55:06 background, it has trouble animating the
55:08 foreground. So, the lava is on fire. So,
55:10 it's trying to animate the fire and it
55:13 forgets about the foreground object,
55:14 which is kind of
55:15 interesting. All right. Uh, so we got
55:18 Liam
55:19 Lawson. All right. We got her. We'll
55:21 We'll put That'll be Liam's girlfriend.
55:37 Oh, I got, by the way, I got um
55:41 um Hedra did a uh did a contest with
55:45 cats and I made this cat on the live the
55:47 other night. I came in um third. I got
55:50 runner up on this contest and they gave
55:52 me $20 of free credits. So, see, you can
55:55 wait, where's my button? You can make
55:57 money with Jackie. I made money with a
56:00 stupid animated
56:04 cat.
56:06 [Music]
56:11 Um Oh, let's
56:20 do Hang on, people. All right, that's
56:23 good. Let's get a car.
56:28 That's nice. Oh, it's a video. [ __ ] Do
56:30 I have a still of that?
56:39 No. There's a good one. Okay. So, we'll
56:41 upload that. Okay. So, we'll say
56:46 um
56:48 blonde man drives.
56:52 woman is in
56:55 passenger seat of the
57:02 car. Um, and we'll say, um, it's a sunny
57:08 day and the car is in
57:14 the parking lot of a
57:20 um, city park.
57:22 in
57:25 Cleveland. We We'll do a We'll do a nice
57:28 uh homage to our missing
57:32 producer who lives in Cleveland.
57:35 Okay, so this is now going to take these
57:38 three images and pull them into
57:40 something. I
57:41 think energy cannot be wasted. Don't
57:45 worry. I've spent the last two months
57:47 learning to train reliable Lauras for
57:51 flux and it feels like a waste now. So,
57:53 here's the deal. Pizza
57:57 later. None of this is a waste. I just
57:59 So, the talk that I gave
58:02 tonight trying to think which which step
58:06 it's
58:08 in. Oh, it's not in there. Basically
58:15 um for the next 3 to 5
58:19 years, nothing that we do, build or
58:23 learn is going to have longevity more
58:26 than probably 3 to 6 months. And that's
58:29 probably too broad. None of the specific
58:31 tools that we learn or workflows that we
58:34 develop are going to survive very long.
58:38 But it doesn't mean you shouldn't learn
58:40 them. Because here's the
58:43 deal. Because
58:45 you've spent the time building and
58:48 understanding
58:49 Lauras, you understood what the
58:51 limitations were. So you can look at
58:53 something like this. Oh, just to
58:57 confirm, do you want to create an
59:00 image? Uh, yes, an image. and keep the
59:06 visual
59:08 style of
59:11 style of the car
59:14 picture.
59:17 Um, you're going to be able to look at
59:19 this and know immediately what it does
59:21 good, what it does bad, why it's
59:23 significant.
59:25 So all of us the the again the reason
59:27 for this channel is is for all of us to
59:29 just be in this conversation and and to
59:31 literally to be saying what you're
59:33 saying and asking questions about what
59:35 you're asking and people going wait
59:37 how's this different than that and what
59:38 does multimodal mean? All of us right
59:40 now are just trying to get our [ __ ]
59:42 heads around what's even
59:45 possible. And if you have any semblance
59:48 of what's possible or in your case where
59:50 you've gone, you know, sort of deep down
59:52 a deep understanding of one piece of
59:54 this, you're going to be way ahead like
59:57 way ahead of anyone else trying to use
1:00:00 this tool. And so none of it's wasted
1:00:03 energy right now, but it's it
1:00:05 adaptability is probably the single
1:00:07 greatest and patience and
1:00:14 um what do you call it when you just let
1:00:17 [ __ ]
1:00:18 go?
1:00:20 Grace. All right, I learned that thing.
1:00:22 I took six months to learn that thing
1:00:23 and now I don't need to know that
1:00:24 anymore.
1:00:25 Well,
1:00:27 but you you now know that you don't need
1:00:30 to know that anymore better than anyone
1:00:32 else
1:00:34 does. All
1:00:35 right. I think Adobe Express is Adobe's
1:00:38 Fast and Furious platform. It It is, but
1:00:42 it's also a red-headed stepchild of that
1:00:46 organization. No waste. What we know now
1:00:48 will put you ahead tomorrow. Yep.
1:00:50 Building building is the roots.
1:00:52 Absolutely. It's important to have
1:00:53 strong roots. Look, there's Liam
1:00:56 Lawson. And there's there's girl me.
1:01:00 There's
1:01:02 Kylie. Liam Lesson has more time on his
1:01:05 hands
1:01:10 now. And there we are at a city park.
1:01:12 And I assume that might be some
1:01:14 building, famous building in Cleveland,
1:01:16 if anyone knows.
1:01:22 Oh, although Oh, there's some there's
1:01:24 some physics challenges
1:01:28 here. I'm just going to say his his
1:01:32 arms look
1:01:36 weird. Um,
1:01:40 she should be
1:01:43 wearing
1:01:44 sunglasses and a pink dress.
1:01:55 And the
1:01:57 door should be
1:01:59 closed. All
1:02:01 right. Yeah, this is this is bonkers,
1:02:04 man. This is [ __ ] bonkers. Oh, the
1:02:08 terminal tower. Chemical farmer knows it
1:02:11 did do Cleveland, Ohio. There you
1:02:14 go. That's [ __ ] awesome.
1:02:20 There you go, people. There's Liam
1:02:22 Lawson. Look how good we look. Don't we
1:02:24 look good
1:02:26 together? Kylie and Kylie and
1:02:30 Liam. He's He's still a little cranky
1:02:32 about losing his ride. Hey, actually
1:02:36 Liam's in the wrong outfit now. He needs
1:02:39 to be in the He needs to be in the
1:02:42 Racing Bulls outfit. Ouch. Yeah, that's
1:02:45 not going to be fun this weekend. Yeah,
1:02:48 you can click on the image to edit it.
1:02:50 Oh, that's true. Oh, but you know what I
1:02:52 noticed, Vicki, is that the the editing
1:02:53 is weird. It doesn't work like it it
1:02:56 used
1:02:56 to. It redo the whole image and it
1:03:00 doesn't keep it the same. So,
1:03:02 something's weird with image editing. I
1:03:04 don't think they got to that
1:03:07 yet. All right, here we are. Let's Let's
1:03:10 see if we got better hands now. His
1:03:13 right hand is holding the left side of
1:03:14 the steering wheel. I know. Yeah, there
1:03:16 was some there was some physics
1:03:17 challenges
1:03:19 there. Tape modder pizza and open-
1:03:22 source models that you can tweak in
1:03:24 general are purposeful.
1:03:27 Yep. Wrong hand on the wrong arm. Yep.
1:03:31 Exactly.
1:03:32 Um there should be a third tall
1:03:34 building, but it's
1:03:36 [Laughter]
1:03:38 okay. What up? What's happening? Marson
1:03:52 AI now Kylie's now now Liam looks like a
1:03:58 bobblehead. It's got the word Cleveland
1:04:00 in the
1:04:01 background. This is pretty
1:04:05 funny. I look like Katy Perry all of a
1:04:08 sudden.
1:04:21 Ah, that's good. But
1:04:24 anyway, you know, still not perfect, but
1:04:26 still pretty [ __ ]
1:04:28 amazing.
1:04:33 Um, all right. What else we gonna do?
1:04:35 What else did I want to show you? Um.
1:04:38 Oh, the one thing I wanted to
1:04:41 try. I know what I want to
1:04:44 try.
1:04:46 Okay.
1:04:51 Um.
1:04:55 Okay. We're going to do a new prompt.
1:05:00 or
1:05:04 say I want to make a
1:05:08 game
1:05:11 where every page of the
1:05:15 book is a photograph or illustration
1:05:18 that has
1:05:29 50
1:05:35 objects that belong
1:05:46 together and
1:05:57 labels that the reader
1:06:06 will use to combine
1:06:13 them. Period.
1:06:28 I can't I can't articulate what I'm
1:06:30 thinking of. I'm thinking of like like
1:06:32 you've got two pair of words like like
1:06:35 um like fishnetss. So you'd have a fish
1:06:38 and nets and then you put put them
1:06:40 together for fish nets. What would that
1:06:42 be
1:06:45 called? Anybody? I redid an old snapshot
1:06:49 of mine.
1:06:51 Um
1:06:53 um for
1:06:55 example, the final word is
1:07:01 fishnetss. And
1:07:03 there is a
1:07:07 picture of a fish and a
1:07:11 net
1:07:15 fishnet. and the
1:07:19 reader pairs them
1:07:22 up. Does this even make any
1:07:30 sense? Compound words. Yes, that
1:07:36 yes. Oh, I tell you it's fun colorizing
1:07:39 black and white. Oh, yeah. That's really
1:07:41 good. Well, this this is a brain teaser,
1:07:43 but it's the compound words that I was
1:07:45 looking
1:07:46 for. Start with and improve the prompt.
1:07:49 Oh, it totally makes sense. It's a very
1:07:51 cool idea. You're describing a kind of
1:07:53 visual word riddle game or a mashup of
1:07:56 the I spy book hidden object game.
1:07:59 Yes. Example page. Fishnet
1:08:04 lighor. Oh, lighthouse. Right.
1:08:07 Okay. Uh correct. And
1:08:11 the
1:08:14 paired
1:08:15 images shouldn't be near each
1:08:20 other. So make
1:08:25 me the items for
1:08:29 image number one or wait number one.
1:08:34 Number one.
1:08:37 It should have the description
1:08:42 of the
1:08:47 object in
1:08:49 detail and
1:08:54 the text for the label below
1:09:00 it. All
1:09:02 right. And then what we're going to do
1:09:04 is we're going to have it make this
1:09:05 [ __ ]
1:09:07 image. So we're gonna have a 50 a 50
1:09:10 object image with Oh, wait. We can build
1:09:14 up to 50 in
1:09:17 batches. Oh, come
1:09:20 on.
1:09:22 Key, lace, cup, horse, hook, net, fish,
1:09:27 glass,
1:09:29 bear. Fine.
1:09:37 Yes. Build the
1:09:44 rest. Yep. Compound word game. Compound
1:09:47 words. All right, we're on it. We're on
1:09:50 it. My brain My brain failed me so hard
1:09:54 there. I'm like midway through
1:09:56 describing the sentence and I'm like I
1:09:57 can't remember the the what it's called.
1:10:00 I can't remember. It just it shut down.
1:10:02 It was bad bad
1:10:09 people. Ah, I see we have not a bot in
1:10:13 the live. It's exactly what a bot would
1:10:16 name themselves, isn't it?
1:10:20 Uhhuh. I see how this is working. All
1:10:23 right. Do we have 50
1:10:25 things? Great. Uh, I'm going to say
1:10:29 now give me No, I don't want it yet.
1:10:32 Okay. I'm going to say
1:10:37 now give
1:10:39 me
1:10:41 three
1:10:44 unique visual style
1:10:49 concepts that might be
1:10:52 [Music]
1:10:54 good for these objects.
1:11:04 Photo realistic grit plus
1:11:08 hyperfocus.
1:11:11 Cool. Collage surrealism. Kind of like
1:11:16 that. Miniature diarama world. Oh, I
1:11:20 like that a
1:11:24 lot. I like number three.
1:11:28 Can you make that image with all the
1:11:35 objects you
1:11:38 listed with all the
1:11:47 labels
1:11:50 below the
1:11:53 item? All right, let's see what we get.
1:11:59 Let's see how good it is at prompt
1:12:04 [Music]
1:12:06 coherence. Fork frisbee cactus collar
1:12:10 letter wheel tape clip
1:12:13 mitten. All right. And then after it
1:12:16 finishes making our little
1:12:18 pitch, then we'll have it make have it
1:12:21 make a key for the for
1:12:23 the for the
1:12:34 thing. I fear it's going to tell you to
1:12:36 come back later for the result. No, it
1:12:38 won't. It'll it'll do it. I mean, it
1:12:41 might not get it right. We've we've
1:12:42 tried to do this.
1:12:44 Um, Ideog is actually pretty good at
1:12:47 this. It's not great at it, but it's
1:12:50 pretty good at it. Um, by the way, Idog
1:12:52 3.0 came out today, so that's another
1:12:56 thing to go play with. I played with it
1:12:58 a little bit and it wasn't super
1:13:05 great. This looks like we're way way
1:13:09 under. Way under. I bet it did the first
1:13:19 15. Peel
1:13:22 book. Oh, these. Oh, wait.
1:13:27 Lace
1:13:30 note rocket. The the labels have nothing
1:13:34 to do Well, they have something to do.
1:13:40 Okay. So, we're going to go um there are
1:13:47 way less
1:13:50 than 50
1:13:52 items and most of the labels are
1:13:57 wrong. Look at your
1:14:02 lists and try again.
1:14:06 I will get fired if you don't get this
1:14:11 right.
1:14:14 I
1:14:16 promised my
1:14:18 boss that this is my
1:14:23 specialty and
1:14:26 told him it would take me a month to
1:14:36 complete.
1:14:38 That was 29 days ago.
1:14:49 Maybe you ran out of credits
1:14:54 probably. I count approximately 36. Did
1:14:57 you really count that many? Let's see. 1
1:14:59 2 3 wait. 1 2 3 4 5 6 7 8 9 10 11 12 13
1:15:06 14 15 16 17 18 19 20 21 22 45 26 27 28
1:15:12 Somewhere between 28 and 36. Yes, but it
1:15:17 ain't
1:15:21 50. Sorry. Choking out.
1:15:30 Good lord. Good people. All
1:15:34 [Music]
1:15:38 right. I ordered new glasses I'm excited
1:15:41 about cuz my eye the these are not the
1:15:44 right glasses for computer work and my
1:15:47 eyes are not
1:15:49 happy. I can't
1:15:52 count. I don't try. Yeah, this isn't
1:15:55 going to get it either.
1:15:59 Lace key. Actually, these look better.
1:16:04 Match roar. It got fish
1:16:07 twice. The cup is a
1:16:10 donut. The pin is a pin, but the pin is
1:16:12 also a fish hook. The bear is a
1:16:18 boar. A jar of marbles is
1:16:24 dominoes. All right.
1:16:26 So, it gives it the good gipper try
1:16:29 here, but it's not great. All right, so
1:16:31 we can move on. I always use the I'm in
1:16:34 trouble unless you do this right
1:16:36 prompting eventually.
1:16:38 Exactly. Don't let me
1:16:41 down. I'm not mad. I'm disappointed.
1:16:49 [Laughter]
1:16:50 [Music]
1:16:51 Um, let's do
1:16:55 um, let's do a new one. Let's go find a
1:17:00 picture. So, we'll do
1:17:04 Dreams. Kyle Shannon
1:17:07 Dreams. Yeah, we'll do that,
1:17:10 dude. All right.
1:17:16 Um, I want you to act as an
1:17:24 animation
1:17:28 supervisor and
1:17:32 recommend five
1:17:42 animation
1:17:44 styles that
1:17:47 are very
1:17:51 popular in
1:17:54 blockbuster movies.
1:17:58 Eg claimmation
1:18:03 um 3D
1:18:08 um rigging I don't know
1:18:13 etc. You don't have to
1:18:17 [Music]
1:18:18 use those
1:18:21 two, but rather give me the top
1:18:27 five styles
1:18:30 and three
1:18:33 movies that use
1:18:36 them each.
1:18:46 That's a cowboy me. Cowboy me coming in
1:18:53 hot. Okay. 3D computer animation,
1:18:56 Pixar's Inside Out, DreamWorks, How to
1:18:58 Train Your Dragon, Disney's Frozen, Stop
1:19:00 Motion, Claymation, and
1:19:03 Puppet. Oh, GMO del Toro's Pinocchio
1:19:07 2022. That's kind of fun.
1:19:10 Motion Capture Plus CGI,
1:19:13 Avatar, Endgame, and Planet of the Apes.
1:19:16 2D
1:19:17 animation, The Mitchells versus the
1:19:20 Machines. I don't know that
1:19:21 one. Studio Gibli's The Wind Rises.
1:19:24 That's Everyone's doing that, so we're
1:19:26 going to avoid that. And then Stylized
1:19:29 Hybrid
1:19:31 Spiderverse Intergalactic. Um, let's do
1:19:38 um I
1:19:40 really really like
1:19:44 the
1:19:46 Pinocchio
1:19:48 style idea. Make the
1:19:54 cowboy. Reimagine the
1:19:56 cowboy.
1:19:59 Re-imagine the
1:20:01 cowboy. I uploaded in this
1:20:07 style
1:20:09 and yeah, let's just do that. Let's just
1:20:11 see see what it
1:20:13 does. You should try Wallace and
1:20:15 Grommit. Oh, yeah. That's that's a good
1:20:17 one, too. All right. Why don't you use
1:20:20 the voice to text? Oh, because I'm just
1:20:22 lazy. Some sometimes I use it, sometimes
1:20:24 I
1:20:26 don't be because because frankly it
1:20:31 gives me great joy to know that it is
1:20:34 torturing people going just turn on
1:20:35 voice to text. Just turn on voice to
1:20:41 text. You OCDers out
1:20:44 there, I'll type I'll type really slowly
1:20:48 and talk each letter while I type it.
1:20:57 I'm an [ __ ] I'm an [ __ ] I'm an
1:21:01 [ __ ] I just forget
1:21:10 sometimes. Well, that hat doesn't look
1:21:12 very
1:21:13 claymationy. Let's see what it does.
1:21:20 I type better when I talk out loud. I
1:21:22 know
1:21:25 exactly.
1:21:27 Um, all right. The the little band of
1:21:30 the hat looks pretty
1:21:35 good. Oh, it made me old and crusty. I
1:21:38 like
1:21:40 it. Weird. But did you
1:21:42 know it's been shown to give you better
1:21:45 results if you're polite to the AI?
1:21:47 Yeah, it's actually it's actually not
1:21:49 weird to me.
1:21:52 Um, every input that you give
1:21:56 um Oh, that's kind of fun. I like that
1:21:59 style. It's very warm. Oh, right. This
1:22:02 is the This is the Pinocchio
1:22:04 style from the GMO GMO del
1:22:08 Toro
1:22:11 Pinocchio. That's really cool. That's
1:22:13 really nice.
1:22:16 Um
1:22:19 um can you
1:22:22 make his
1:22:25 daughter I don't want to call her Belle.
1:22:27 That that would
1:22:29 be
1:22:30 Marge. Can you make his daughter Marge
1:22:33 for me
1:22:39 too? I like that one. That's really
1:22:42 good.
1:22:45 All right. So, now we're going to put
1:22:46 him on a horse. Well, let's We'll see
1:22:59 Marge. Shake
1:23:05 it. Shake
1:23:10 it. Does he always sing like this? It's
1:23:12 very annoying. It It's just the shankity
1:23:16 dank dank. That's not That's not
1:23:18 scatting. I like jazz. I like singers
1:23:21 that scat, but that's
1:23:25 just it's it's like Tourette's with a
1:23:34 melody. Tourette's with a melody.
1:23:42 Hey, apparently Tik Tok's going away in
1:23:44 10 days. So, what are you going to
1:23:47 do? Maybe I'll get kicked off before
1:23:50 they they end
1:23:51 [Laughter]
1:24:06 it. Let's see.
1:24:10 Um, let me stop
1:24:13 it. Um, that looks like an animation.
1:24:18 Uh, an
1:24:21 illustration,
1:24:23 not
1:24:24 claimation. Try
1:24:30 [Music]
1:24:32 again. Why is it so slow? because this
1:24:35 is their new fancy fancy ass model and
1:24:39 everyone's using it. Um, he said today
1:24:43 he's going to extend Tik Tok.
1:24:46 Eh, he said he'd keep our national
1:24:49 secrets safe,
1:24:50 [Laughter]
1:24:58 too. I got your group chat right here.
1:25:07 [Laughter]
1:25:10 Oh my god, the the [ __ ] hearings are
1:25:12 so painful. It's like, come on, just
1:25:15 answer the [ __ ] questions.
1:25:33 G. All right. It made her She's got a
1:25:36 skin
1:25:38 [Laughter]
1:25:45 condition. It gave her a different style
1:25:47 of eyes. Yeah. See, this is this is
1:25:50 rough. This is rough. All right.
1:25:55 Um, let's I want to take
1:25:59 him. Let's download
1:26:02 him and then let's go
1:26:05 to let's go to 11 Labs.
1:26:22 We're going to try a voice
1:26:27 switcher. What's he gonna say? We got to
1:26:30 have him say something, people. What's
1:26:31 he gonna
1:26:33 say? Grandpa Spuds Oxley.
1:26:38 Oh, perfect. All right. So, that's going
1:26:41 to be it. Grandpa Spuds
1:26:46 Oxley. There he's There's that. We're
1:26:48 going to record some
1:26:50 audio. Let's see. What are we going to
1:26:52 say? We're going to say
1:26:59 [Music]
1:27:05 uh I got it.
1:27:23 Time
1:27:24 was when you had a
1:27:26 farm, you needed horses to work the
1:27:29 fields.
1:27:31 Now, I just don't know what to say. They
1:27:34 got these fancy steam machines. They
1:27:38 fill them full of water and coal and
1:27:40 they light them on fire and they turn a
1:27:43 big wheel and it pulls on pulleys and it
1:27:46 pulls on gears and these giant
1:27:49 machines squeak their way through the
1:27:52 fields. And they're telling me that the
1:27:55 the way of the horses are are the way of
1:27:58 the past. And if that's the way of the
1:28:00 past,
1:28:02 then I'm just not here for it. I'm out
1:28:06 of here. They can have my horses. They
1:28:09 can have my
1:28:10 land. I'm heading
1:28:13 west. It's a sad
1:28:19 story. Wait, it cannot be used for voice
1:28:22 conversation.
1:28:25 Why, you [ __ ]
1:28:28 [ __ ]
1:28:31 Um, you used all your slots. Oh, maybe
1:28:34 that's what it is.
1:28:38 [Music]
1:28:42 Uh
1:28:51 no,
1:28:52 Bill.
1:28:54 No. Where Where's
1:28:57 my Do I need to get rid of one of
1:29:03 these? Use
1:29:08 Manage
1:29:10 voices. Oh, I think I just lost my
1:29:12 recording. Damn
1:29:13 it. I [ __ ] hate bad
1:29:18 [Laughter]
1:29:21 interface. It just sucks. Uh, Angry
1:29:43 Alu, Can I get rid of
1:29:46 her?
1:29:48 Delete. Remove
1:29:50 voice. Go back.
1:29:53 Go back. I can't go back. See? You see
1:29:56 what's happening here? Let's see. Can I
1:30:00 choose him
1:30:02 now? Uh, yeah. I lost it. God damn
1:30:06 it. What if I go back one? Is it still
1:30:09 there? No. [ __ ]
1:30:18 time was if you wanted
1:30:20 to. The [ __ ] is
1:30:27 that? Come on, you
1:30:31 [ __ ] This is Okay, I agree. I agree
1:30:36 with you. This got not fun. It was fun
1:30:39 for a while. It was fun. We made a fancy
1:30:41 cowboy and then it's like, "Oh, Kyle's
1:30:44 going to go do something and it's going
1:30:45 to be fun." And then now it's not fun.
1:30:48 Why is it not fun? Shitty software,
1:30:50 unclear interface. Things like not
1:30:53 remembering the [ __ ] I just recorded
1:30:55 because you're making me get rid of a
1:30:57 voice because I don't have enough slots
1:30:59 in the voice to use the thing because
1:31:01 you told me I couldn't use the thing
1:31:02 because I haven't paid enough [ __ ]
1:31:05 money. And now it's not fun
1:31:10 anymore. And then I go to record the
1:31:12 thing and you pop up a [ __ ] box that
1:31:14 says, "Are you sure you got rights to
1:31:16 that?" It's my
1:31:19 voice. Good
1:31:31 lord. Time was if you wanted to plow a
1:31:34 field, you had
1:31:36 to work it with a horse. You had to race
1:31:38 horses. You had to ride horses. You had
1:31:40 to you had to tame horses. You had to
1:31:42 break
1:31:43 them. My daughter Marge and
1:31:45 me, we made a way of life
1:31:49 here making horses right for the
1:31:52 farmers. Along come these big machines,
1:31:54 these steam machines, but they're big
1:31:56 wheels and they're they're they're sound
1:31:58 and they're they're hissing and they're
1:32:00 the heat and the
1:32:03 danger. I got nothing to do with it. I'm
1:32:06 out of here. I'm going to New York City.
1:32:10 I always wanted to be an
1:32:12 actor. All
1:32:14 right, there we go. Can we make Spuds
1:32:18 now?
1:32:20 Yes.
1:32:25 Perfect. That's not
1:32:32 Spuds. Oh, that's Rachel. Stop it,
1:32:35 Rachel.
1:32:38 Uh
1:32:41 uh
1:32:42 Spuds, can we do
1:32:54 him? This is bad interface and it's not
1:32:59 working. Amazing. Grate
1:33:16 All right, fine. Here. Well, I'll I'll
1:33:18 let you listen to it in a bit. Let me
1:33:20 download
1:33:21 this and then we're going to go over to
1:33:26 Hedra. Hedra, if you haven't seen Hedra,
1:33:30 Hedra is [ __ ] cool.
1:33:37 That was 11 Labs where we did the swap
1:33:39 the voice. I recorded my silly
1:33:41 voice. It recorded in the uh whatever
1:33:44 you call
1:33:46 it. Switch the voice
1:33:49 over. Hedra now does the animation. So
1:33:53 good. So good. This is the best one I
1:33:56 think right
1:33:59 now. And so we're going to upload our
1:34:02 dude. Try
1:34:10 now. He does know he can't sing, right?
1:34:13 He knows
1:34:20 that. Time was if you wanted to plow a
1:34:23 field, you had
1:34:25 to work it with a horse. You had to
1:34:27 raise horses. You had to ride horses.
1:34:29 You had to You had to tame horses. You
1:34:31 had to break them.
1:34:33 My daughter Marge and me, we we made a
1:34:36 way of life
1:34:37 here. Making horses right for the
1:34:44 farmers. I like bitter people. Bitter
1:34:47 people are
1:34:49 [Laughter]
1:34:51 fun. All right, that's going to cost us
1:34:54 290 credits.
1:34:57 Fine. Nothing but the best for the AI
1:35:00 learning lab.
1:35:02 All
1:35:04 right, Lord Digital gods, thank you for
1:35:06 the super sticker. I don't have a a a uh
1:35:10 whatever you call it, a producer here.
1:35:12 I've got some mods over on TikTok, but I
1:35:14 don't have a producer here, so thank you
1:35:16 for that. I don't know quite what that
1:35:18 looks like. Error. Could not Wait, what?
1:35:21 Could not perform the action for the
1:35:23 image asset.
1:35:27 Why? Wait, did I not pick Kedra
1:35:30 character 3? I
1:35:34 did.
1:35:36 Um,
1:35:38 cowboy
1:35:44 talks. I I I swear to God,
1:35:47 everyone, I I'm so [ __ ] done with
1:35:52 [ __ ] just not
1:35:56 working. All right, now it's going. I
1:36:00 don't know what was up with
1:36:06 that. And now let's go to
1:36:13 [Music]
1:36:18 Suno. We'll go create. We'll go sign in.
1:36:21 We'll go
1:36:23 Google. Ye old Kyle. Ye of little faith.
1:36:27 Ye of little faith. How dare they?
1:36:31 Instrumental. Then we'll put in olden
1:36:39 timy.
1:36:40 Olden
1:36:42 tiny cowboy
1:36:45 [Laughter]
1:36:51 music. Fantastic.
1:36:55 Bob,
1:36:57 this No one's No one's talking. No one's
1:37:00 saying anything. Are you just watching
1:37:02 me spiral out of control without not
1:37:05 without knowing what I'm doing? That
1:37:07 makes
1:37:08 sense. Um, there's
1:37:11 him. Network error occurred.
1:37:14 Um, show me the
1:37:18 cowboy and
1:37:21 Marge out by the
1:37:24 barn in the
1:37:28 claymation
1:37:30 style. Let's go listen to our song. I
1:37:33 got to do this. I gotta do this. I gotta
1:37:37 go. Uhoh. Oh, I'm in the wrong Oh, good
1:37:41 lord. I can't even share the audio. All
1:37:44 right. Don't you make No, never
1:37:50 mind.
1:37:58 No, no. Olden Tommy Cowboy music. A
1:38:03 sad
1:38:06 lament on an out of tune guitar.
1:38:18 Oh, I'm not sharing anything now, am I?
1:38:21 No, you're
1:38:22 not. You big dumb dumb. Listen, you big
1:38:26 dumb dumb. We're trying to learn here
1:38:28 and we can't learn if we can't see
1:38:30 anything. We're visual learners, Kyle. I
1:38:32 thought you knew
1:38:34 that. I thought you said this channel
1:38:37 was neurodeivergent friendly.
1:38:40 I'm a visual learner. I can't learn by
1:38:42 just seeing you yapping your jaws,
1:38:45 Kyle. Ding ding ding
1:38:58 ding. Okay, good. Perfect. Let's
1:39:02 download
1:39:03 that. Download. Okay, we got to wait for
1:39:05 that to load.
1:39:09 [Music]
1:39:15 I'm blown away by this. I'm pretty
1:39:17 proficient with chat GPT, but the voice
1:39:19 and animation are legit. Yeah. No, it's
1:39:22 crazy, isn't it? Did you see the chat
1:39:25 GPT update? Crazy, right? Yeah, that's
1:39:27 what we're working on. Yeah, the the the
1:39:30 image update is [ __ ]
1:39:34 bonkers. It's just bonkers.
1:39:38 They made her out of felt. That's what
1:39:41 they
1:39:49 [Music]
1:39:55 did. Dang, dang. That's a really cool
1:40:00 It's a really cool
1:40:02 character. I wish I could have used that
1:40:04 old crusty
1:40:09 dude. Okay, so now we got a
1:40:14 song. Oh, and you know what else we
1:40:16 should do? We can go back to 11 Labs.
1:40:18 Let's do sound effects. But wait, I
1:40:25 need 8 seconds. Oh, you can do it
1:40:28 longer. 13. Oh, you can do 20
1:40:46 seconds. All right, here we go. We got
1:40:48 We got
1:40:50 Okay, now wait a
1:40:53 minute. Hang on a sec. I need to grab
1:40:58 this. Move this over
1:41:01 here. Go.
1:41:08 Hedra
1:41:11 library. Oh, library.
1:41:18 [Music]
1:41:29 19 days ago, 14 days ago, six days ago.
1:41:32 Oh, here we go. Okay. All right. So, let
1:41:35 me stop sharing this. Start
1:41:39 sharing this. This
1:41:43 this. Okay.
1:41:48 So, here's our
1:41:56 dude. You
1:42:06 ready? Time was if you wanted to plow a
1:42:09 field, you had
1:42:11 to work it with a horse. You had to race
1:42:13 horses. You had to ride horses. You had
1:42:15 to you had to tame horses. You had to
1:42:17 break
1:42:18 them. My daughter Marge and me, we we
1:42:22 made a way of life
1:42:24 here making horses right for the
1:42:27 farmers. Along come these big machines,
1:42:29 these steam machines with their big
1:42:32 wheels and they're they're they're sound
1:42:33 and they're they're hissing and they're
1:42:35 the heat and the
1:42:38 danger. I got nothing to do with it. I'm
1:42:41 out of here. I'm going to New York City.
1:42:45 I always wanted to be an
1:42:53 actor. All right. So, so we so so we
1:42:59 have ourselves some
1:43:01 [Laughter]
1:43:07 comedy. Oh, good lord. Now we go throw
1:43:11 some
1:43:13 uh
1:43:19 iMovie. We'll go throw some audio in
1:43:22 there. Some background music as they
1:43:25 say.
1:43:29 [Music]
1:43:34 [Music]
1:43:38 Lady Fantastic. Bob. Bob. Hey, Bob. Bob.
1:43:43 Bob.
1:43:51 Bob. That's pretty funny. Oh, you can't
1:43:53 see anything. See, this is If I had a
1:43:55 producer here, he'd be yelling at me,
1:43:58 but I don't. So, you're gonna have to
1:44:00 deal with my [ __ ] ADD.
1:44:03 Frank was more
1:44:05 fun. Yeah, Frank was more fun. All
1:44:08 right, so down
1:44:09 goes He doesn't have a name, this dude.
1:44:17 [Music]
1:44:29 We'll go grab the end of
1:44:33 [Music]
1:44:37 this. And we'll fade it on
1:44:41 out. And we'll fade this out.
1:44:48 Oops. I always wanted to be an actor.
1:44:51 [Laughter]
1:45:02 I like it. I like it a
1:45:10 lot. I think we'll just do
1:45:13 black.
1:45:15 Go and then we'll
1:45:17 go Oh, no. We'll go uh Oh, wait. We'll
1:45:21 go
1:45:22 and we'll
1:45:27 go nothing to do with it. I'm out of
1:45:30 here. Going to New York
1:45:33 City. I always wanted to be an
1:45:36 actor. Time was if you want
1:45:41 to go to New York
1:45:44 City. I always wanted to be an actor.
1:45:47 [Laughter]
1:45:53 Oh, what did I do? All right. Save share
1:45:58 a
1:46:01 file
1:46:05 720p. Will it save it just as a little
1:46:08 thing? Cowboy. He needs a
1:46:13 name. Desktop.
1:46:16 [Music]
1:46:21 Carl. Oh yeah, Carl's good. That's what
1:46:24 everyone when I grew up in York,
1:46:25 Pennsylvania, they always called me Carl
1:46:27 because Kyle was too complicated for
1:46:29 their tongues. Hey, Carl. It's Kyle.
1:46:32 Yeah,
1:46:37 Carl.
1:46:41 Greg. All right, let's see if this
1:46:43 worked. So, I go here.
1:46:47 Is that Carl? No, that's the raw
1:46:50 one. Cowboy
1:46:53 time was if you wanted to plow a field,
1:46:55 you had
1:46:56 to work it with a horse. You had to race
1:46:59 horses. You had to ride horses. You had
1:47:01 to You had to tame horses. You had to
1:47:03 break
1:47:04 them. My daughter Marge and me, we made
1:47:08 a way of life
1:47:09 here, making horses right for the
1:47:12 farmers. Along come these big machines,
1:47:15 these steam machines. They're big
1:47:17 wheels. They're their sound and they're
1:47:20 they're hissing. The heat, the
1:47:23 [Music]
1:47:25 danger. No, it's Carl. It's Carl. We're
1:47:28 going to We're going to call him Carl
1:47:31 the
1:47:33 cowboy.
1:47:38 Carl. Carl Cowboy. All right. And now
1:47:41 we're gonna go to the
1:47:47 [Music]
1:47:51 Twittering ding ding
1:47:55 ding. Carl the
1:48:01 cowboy
1:48:04 resisted steam
1:48:06 engines at first.
1:48:12 He was real
1:48:30 mad. He
1:48:35 learned resisted
1:48:39 He
1:48:41 learned don't be like Carl with
1:48:49 AI. Everything's a [ __ ]
1:48:57 lesson. Carl
1:48:59 cowboy. Carl the cowboy resisted steam
1:49:02 engines at first. He was real mad. He
1:49:06 learned. Don't be like Carl with AI.
1:49:09 This is good. I like this. This is a
1:49:10 public service announcement. We'll do
1:49:17 PSA. PSA. Carl the
1:49:19 cowboy and then we'll just
1:49:25 do He resisted steam engines at first.
1:49:28 He was real mad.
1:49:37 He learned. All right. Don't be like,
1:49:39 "Carl, did it upload?" No, of course
1:49:41 not. Unfucking believable. Hey, Elon,
1:49:45 why don't you fix your [ __ ]
1:49:53 [Music]
1:49:57 website? That's okay, Elon. You know
1:49:59 what? I'll go save this in a different
1:50:01 format.
1:50:04 I'll do I'll do the work for you. Not
1:50:06 it's not a problem. I got it. I got it.
1:50:09 You're good. You just keep doing the
1:50:12 awesome stuff you're doing in the
1:50:17 [Music]
1:50:24 [Music]
1:50:26 country. Shang dang dang dang dang.
1:50:31 If this doesn't upload, I'm gonna be a
1:50:33 cranky
1:50:37 boy. Carl Dirk Diggler because we all
1:50:41 want to be
1:50:42 actors. Honest to God, it just
1:50:46 like How about if I hit post? Will we
1:50:49 upload it?
1:50:51 Please,
1:50:53 please. Did it upload it? No. Error.
1:50:56 Post.
1:51:01 Come
1:51:06 on.
1:51:08 30 19
1:51:11 21. Does anyone know how to upload
1:51:14 something to X? Am I just being an
1:51:17 idiot? Is this like a boomer moment? Oh
1:51:20 yeah. I was watching this AI learning
1:51:22 lab and he was trying to upload
1:51:24 something to Twitter and he was like he
1:51:26 doesn't even know how to upload a file.
1:51:28 Like, oh my god, that's so
1:51:30 ridiculous. Like, why doesn't he learn
1:51:32 to upload a file if he's going to take
1:51:34 all this time to make a file? Why didn't
1:51:35 he learn to upload the file? Stupid dumb
1:51:38 dumb. He's a dumb dumb. This is a dumb
1:51:40 dumb channel. So dumb. So dumb.
1:51:49 [Music]
1:52:06 Anyone? Do you have to shrink the file
1:52:08 first? Maybe. No. Like people are
1:52:11 uploading hour-long videos to this
1:52:13 stupid [ __ ]
1:52:15 site. Oh, you know what? I could maybe
1:52:18 try. No, Kyle, why don't you tell us?
1:52:21 Shut
1:52:22 up. Let me try an incognito window
1:52:26 because maybe, just
1:52:29 maybe, there's some stupid [ __ ]
1:52:32 extension that's incompatible with that
1:52:35 stupid [ __ ]
1:52:38 site.
1:52:39 [Music]
1:52:42 Uh, oh, because I'm in incognito. Let's
1:52:46 see.
1:52:56 I don't remember my
1:53:08 password. Wrong password. Yeah, I know.
1:53:27 Anyone? Any
1:53:30 ideas? Maybe a smaller file will work. I
1:53:33 don't know. I This is so [ __ ]
1:53:36 annoying. Export as
1:53:41 480p to the desktop. We'll replace that
1:53:44 one. Yes, we'll replace it. All right,
1:53:47 that's good. We'll go here. We'll get
1:53:49 rid of this. We'll go here. We'll go
1:53:51 here. We'll click
1:54:00 [Music]
1:54:06 that.
1:54:08 50 60 100. Yay.
1:54:21 post. All right. Oh, and I can't share
1:54:24 this with you because I'm in a wrong
1:54:27 window.
1:54:48 All right, did it. I lost all my verve
1:54:52 for life in that process.
1:54:59 But if you'd be so kind, go to Twitter
1:55:01 and like that tweet, that Zeit, the
1:55:04 yeet, whatever the [ __ ] you do on X now.
1:55:07 And uh quote it, quote tweet it and like
1:55:10 it and send it to Sam Alman. And I
1:55:12 should Oh, I should tag Hedra, shouldn't
1:55:14 I? I should hang on. I should tag edit
1:55:21 post at
1:55:24 Hedra Hedra Labs and I should tag at
1:55:31 Sununo and I should tag at
1:55:36 OpenAI and I should tag at
1:55:39 OpenAI
1:55:44 developers and I'll do at Globalizer.
1:55:58 [Music]
1:56:03 Sir, look, someone else doing chat GPT
1:56:07 plus Hedra
1:56:08 Labs. Car sync available on Hedra.
1:56:19 Well, I just did Carl the cowboy. You
1:56:22 stupid [ __ ] Scooter the
1:56:26 Whimo. Carl the cowboy is better than
1:56:28 you dumb
1:56:30 car. Not that I'm competitive,
1:56:33 jealous. You big dumb dumbs.
1:56:43 [Music]
1:56:45 just had an
1:56:47 idea. Or even better yet, ask
1:56:51 [Laughter]
1:56:57 Grock. Oh, lordy, lordy, lordy. All
1:57:01 right, I'm going to get out of here. Uh,
1:57:04 it's Wednesday. Tomorrow's Thursday.
1:57:06 I'll be back at the normal time
1:57:07 tomorrow, 8:00
1:57:09 PM. And
1:57:11 uh yeah, we'll we'll keep figuring out
1:57:14 what's going on here. I don't I don't I
1:57:15 don't know what else we're going to play
1:57:16 with, but there's I feel like there's I
1:57:19 I feel like we've barely scratched the
1:57:21 surface with this image model. Like I
1:57:23 don't know why there's my my gut is
1:57:25 telling we we don't know what this is
1:57:27 really about yet. So it maybe it could
1:57:31 just be like yeah, it's just another
1:57:32 picture thing. I don't think it is. I
1:57:34 think it's bigger than that, but we'll
1:57:36 see. We'll see. We'll see. All right.
1:57:40 Peace out, everyone. Hope you had fun
1:57:43 tonight. Later. Later.