5/12/2025 - AI's Creative Leap: From Text to Coherent Character Animation and Scene Control

Description

Chapters

Transcript

0:02 [Music]
0:06 [Applause]
0:10 [Music]
0:13 [Applause]
0:14 [Music]
0:23 The boys were sing
0:26 shingling the summer that we
0:32 You were 10 and
0:35 17. How could I
0:39 forget? All the stars from near and far
0:43 were watching from
0:45 [Applause]
0:46 a watching two
0:49 teenagers fall in love.
0:54 The way we danced was not a dance, but
0:58 more a long
1:01 embrace. We held on to each
1:05 other and floated there in
1:09 space. I was shy to kiss you while the
1:14 whole wild world could
1:16 see. So
1:18 shingling said everything for me.
1:22 [Music]
1:24 And all the poor old folks, they thought
1:29 we'd lost our
1:31 minds. They could not make heads or
1:34 tails of the young folks funny rhymes.
1:40 Uh, you and I knew all the words and we
1:43 always sang
1:45 along to old jam
1:50 dong
1:52 jam bang.
1:55 [Music]
2:01 Well, after years and after days and
2:04 after summer's
2:07 passed, we held on like we held on to
2:10 each
2:14 other. I don't know the
2:16 [Music]
2:23 words. Happy Monday, everybody. That is
2:27 Champ singing along. He knows the words
2:28 better than I do at this point. producer
2:31 Brandon checking in saying hi. Good day.
2:33 Good day. Good day. What's
2:37 happening?
2:40 Um, we, you know, sometimes these
2:43 Mondays, it might be a meltdown Monday,
2:46 but I don't think it's going to be a
2:47 meltdown
2:48 Monday. But I I
2:52 I've right now that is similar to a
2:55 feeling I had back in December, January.
3:02 And you know what came after that was,
3:05 you know, some of these really fancy
3:06 models and some of the agentic stuff. I
3:10 have a feeling we're on the cusp of
3:12 something big
3:13 [Music]
3:17 again. Kevin Stewart, happy belated
3:20 birthday. Thank you very much. I
3:22 appreciate that. You got the number
3:24 wrong. It's not 60. That was a rumor.
3:27 That was, you know what it was? There
3:29 were there were internet trolls and the
3:32 internet trolls were like, "Yeah, he's
3:33 [ __ ] old. He's [ __ ]
3:36 60." And you know what's amazing about
3:40 because of AI now? If you go back and
3:43 you look up my birth certificate, the
3:45 trolls got to that, too. Like, my birth
3:48 certificate would agree with that
3:50 assessment. But we I mean, look at me.
3:54 This is not 60. Come on.
3:57 So anyway, still can't tune my guitar
3:59 for [ __ ]
4:05 [Music]
4:14 [Music]
4:31 [Music]
4:38 KKE. Hey Kyle, can you show us how you
4:40 aggregate content about AI for post
4:44 ideas?
4:50 I listen there are no bad questions.
4:53 That's just a question that built into
4:56 your question are some assumptions that
4:57 I have a [ __ ]
4:59 plan. So I will absolutely tell you how
5:02 I I put my thoughts together for
5:05 things and what I pay attention to and
5:08 how how I pay attention to it. I'm happy
5:11 to do that.
5:15 Um, I mean to be quite honest, the the
5:20 biggest the biggest thing Okay, Steve
5:23 has something for me in the regulars.
5:25 The biggest
5:26 thing that allows me
5:30 to follow this stuff is just a massive
5:34 sense of
5:35 curiosity. Like I just want to know like
5:38 what's coming, like what's next? What's
5:39 coming? How does it work? What's it do?
5:41 What's it make
5:43 [Music]
5:50 possible? I'm getting yelled at by my
5:53 producer. It's very rude. He's saying,
5:55 "Switch your cameras. Do the black bar.
5:57 Everyone's annoyed with your
6:00 life. I just, you know, do do I deserve
6:03 this?" No. I'm here. I'm here. I'm just
6:06 I I've got a guitar. I've got 17 chords,
6:10 15
6:11 songs, and a
6:16 [Music]
6:19 dog. Wow, this is so out of tune. It's
6:22 not playable.
6:27 [Music]
6:42 [Music]
6:48 That's better. A little bit better.
6:51 Thank you, Kyle. I'm thinking of blog
6:53 content topics. I appreciate it. I have
6:56 seen you do it before and like the
6:58 thing.
7:01 Um, you know what we'll do? So, I'll
7:04 tell you what we'll do, Kake. I'll I'll
7:05 tell you my general how I do it.
7:09 Like, I I'll share my methodology with
7:12 you. If if if you're expecting my
7:14 methodology to have anything resembling
7:16 an actual
7:19 methodology, you'll be gravely
7:21 disappointed,
7:24 right? Um, no, I don't use Nomatic. I
7:27 created New Zomatic. I don't use New
7:29 Zomatic.
7:36 Um, I now have a couple of uh like
7:38 Brandon will send me stuff. I've got
7:40 people that send me stuff
7:41 occasionally here. I'll show you what I
7:43 do. Let me just I I'll share my screen.
7:45 We'll go look at some
7:50 stuff. I've got um Chat GBT doing some
7:54 deep research on how to get Sydney
7:56 produced. So, we'll come back to that.
8:02 Um, this is mainly what I do.
8:05 So, okay. So, so I'm on I'm on the
8:09 Twitter. So, it's mostly Twitter and
8:12 then I'll I'll
8:13 occasionally if it's something big, I'll
8:16 I'll occasionally look at LinkedIn, but
8:19 it's mostly Twitter. And the key with it
8:24 is you need to groom your you need to
8:28 prune or groom or garden your who you
8:32 follow. And so what I did when I first
8:35 started getting into this is, you know,
8:38 I found people I liked. I know Robert
8:40 Scoble was way into this stuff. I knew
8:43 Ethan Malik was way into this stuff.
8:45 Paul Ritzer. Um I'm trying to think who
8:48 else were some of the early ones.
8:50 you know, Sam Alman, Greg Brockman,
8:52 like, you know, players at the major at
8:54 the major labs. And so I would follow
8:57 them and then I forget who told it to
9:00 me. I don't know, someone probably
9:04 younger said the people that you most
9:07 respect, go figure out and and Twitter
9:10 will tell you who they follow. And then
9:13 so I just spent like a week kind of
9:16 following people that are followed by
9:19 um people I liked and then I would pay
9:22 attention to who posted [ __ ] and then I
9:24 would go look at who they followed. And
9:26 so I did that for I don't know maybe a
9:28 month probably back in 2022. I don't
9:31 think I've updated it since
9:32 then. Um so that's part of it. So part
9:36 of it is just having and you can do
9:38 lists. There's there's things you can do
9:40 where you can do searches and save those
9:42 as like I've got I've got groupings of
9:45 people I follow like here's
9:47 genai list but I just I just basically
9:51 go to my for you page for the most part
9:53 there's Riley
9:56 um and then what I do is this and then
9:59 this is here's the big if if there's a
10:02 secret that I have again like it's not
10:05 that much of a methodology I'll go down
10:08 and I'll I'll like there are so many
10:11 people now doing sensational headlines
10:14 that I I'm starting to recognize if it
10:17 says um you know such and such is dead I
10:23 just ignore that if it says um if it
10:27 says um Google just released right and
10:32 then I look at what they're talking
10:34 about what Google just released and if
10:37 Google released it two months ago. They
10:40 didn't just release it. That's just some
10:42 [ __ ] right? Because Google just
10:45 released X followed by 10
10:49 crazy 10 crazy examples. Th those are
10:54 like everyone's copying everyone. So, so
10:56 those I just completely ignore. Then
10:59 what I do if I see someone like hater, I
11:01 like um slow developer. I like um here's
11:04 retrunk. I like him. he's an irregular.
11:07 Um Timmy is the guy from Runway. Like
11:11 you just start to recognize people that
11:13 post regularly and that you like. And
11:16 then what I do is if I see one person
11:19 post about something, I'll look at it if
11:22 it's someone I respect and I'll check it
11:24 out. If I see two or three, I'm like,
11:26 "Huh, there's something here." If I see
11:29 10, you know, just in in like a single
11:32 scroll pass, then I'm like, "Something's
11:34 up.
11:35 So most of what I pay attention to, like
11:38 I tend to ignore um open
11:41 source just because I'm not technically
11:46 comfortable enough to
11:49 just be filling up hard drives and
11:52 switching things and maintaining
11:54 versions and just all that [ __ ] just
11:57 exhausts me. So I just kind of ignore
12:00 that. And and I figure at some point
12:02 people are going to use those open
12:04 source tools to build some cool thing
12:05 and a lot of the tools that we use today
12:07 are are that, right? So so I figure the
12:10 open source will take care of
12:12 itself.
12:14 Um I tend to ignore things like iPhone
12:19 apps and things that are like little
12:22 small things that are built on top of
12:23 other people's tech.
12:28 Um, and then I just and then I just pay
12:30 attention and
12:32 and you know often what I'll do like
12:36 there are a couple of people that I
12:37 watch. Ethan Mllik's one of them. David
12:39 Shapiro is one of
12:41 them. Um, occasionally Robert Scoville.
12:44 His are usually one oneoff hits, but
12:47 like Ethan Malik and Robert Sco or uh
12:50 Ethan Malik and
12:53 uh [ __ ] who the hell was I did I
12:56 say? Uh who was the other one I said?
12:59 Ethan Mllik and oh David
13:02 Shapiro. They tend to go on topic
13:07 benders, right? So like David Shapiro
13:09 will do 10 things about universal basic
13:12 income or you know post post labor
13:15 economics. Um and when he's on a bender
13:18 something's something's simmering right
13:20 and Ethan Malik will often go on a
13:22 bender or or Ethan Malik will do three
13:24 pieces of research in a row that are all
13:27 related and that from those two in
13:30 particular. They've got early access to
13:32 stuff. They're they're up on this way
13:34 more than I am. that tends to point me
13:37 in a direction and then I'll start
13:39 looking for what they're talking
13:42 about. Share the
13:49 live. Which which live? Oh yeah, we've
13:54 got shitty numbers here. So do me a
13:57 favor. If you're
13:59 here, if you're on YouTube, if you're
14:01 watching this anywhere, can you send
14:03 this to your friends? We only have a
14:05 handful of people here. Danielle
14:07 Rockstar sharing like like a banshee.
14:10 Love it. Thank you. Thanks for all the
14:13 cool little What are all the cool little
14:16 heart hearty hearts? We're getting
14:19 hearts and lipsticks and party favors.
14:23 What is all that [ __ ] Did I miss a memo
14:25 from Tik
14:28 Tok? Of course you did, Silver Fox. You
14:31 always do.
14:35 Wow, we got all sorts of cool new
14:37 goodies instead of just hearts tonight.
14:40 Um, so anyway, so that's
14:43 basically that's basically what I do.
14:45 And then, you know, it's funny. I go
14:48 through these waves of I won't post
14:50 anything and I'll be like, uh, I don't
14:52 have any ideas. And then someone will
14:55 post something and and all of a sudden
14:57 just a really clear concept will pop in
15:01 my head. And sometimes they're small and
15:03 sometimes they're big. And that's
15:04 usually when I write an article on
15:07 LinkedIn. Um, I used to post a lot more
15:10 about this launched and that launched
15:12 and that launched. Now I'm much
15:14 more interested
15:16 in did something launch that's actually
15:19 going to change the
15:27 game. So I don't know if that helps.
15:34 [Music]
15:49 [Music]
15:52 irregulars video. So I got to change
15:56 tabs. What am I looking for here?
16:00 Get out of here with your stupid pole.
16:04 What am I looking
16:14 for?
16:26 Steo. I don't. Maybe I gotta
16:31 [Music]
16:37 reload. Oh, my first attempt at anything
16:41 other than an image on Dollali. It was
16:44 fun to make. Hope it brings a smile. Oh,
16:47 Champy singing. Okay. All right. Here.
16:50 Let me let me get it
16:52 rolling. So, it's probably going to
16:53 launch a new tab.
17:04 What? Oh, it downloaded
17:09 it. Oh, I think I can do it without
17:11 sharing tabs. Here,
17:17 find a night.
17:21 Every fella's got you
17:24 [Laughter]
17:27 in. What you doing with a clown like
17:34 me? Surely one of life's
17:37 little
17:39 mystery. So,
17:43 [Applause]
17:44 that's
17:47 you. You want to sing with
17:49 you? Listen.
17:53 [Music]
18:00 What did I
18:04 say? That's
18:08 awesome. The the uh the backup singer
18:11 there was not great, but the lead dog
18:12 was good. Pika dropped a new thing. Even
18:15 Champion
18:16 is lip-syncing is adorable. Yeah, I
18:19 know,
18:21 right? That's
18:24 awesome. Thank you, Steo, for that. That
18:27 was very nice. Very nice. Very nice.
18:30 Maybe I'll share that on the uh on the
18:33 Twitter. Oh, man. Um
18:37 [Music]
18:49 [Music]
19:00 [Music]
19:06 Is this place I can rest my
19:09 [Music]
19:12 forehead gather my thoughts in sweet
19:16 silence.
19:19 [Music]
19:20 There's this place where the feelings
19:23 are
19:24 [Music]
19:26 dead. From an
19:30 over this place I can slowly face. The
19:35 only one I truly can
19:38 know. These are tears from a long time
19:42 ago. Got these tears from a long time
19:45 ago. I need to cry. 30 years or
19:50 so. These are tears from a long
19:55 time
20:00 [Music]
20:08 ago. You done? You good?
20:13 Um, okay. I think Champy's
20:15 good. So, there's two things that I
20:18 think I want to do tonight. Um, and you
20:20 could derail me if you ask a question
20:22 and I lose my [ __ ] on it. Cool. We can
20:24 do that, too. Um, or if you come in and
20:27 troll me in an uncreative way. It's
20:29 meltdown Mondays. We We just might get
20:31 some fireworks. Who knows? Oh, Winston.
20:34 Oh, that's gorgeous. Thank you very much
20:36 for that. Very generous. Starmap
20:39 Polaris. Love it. Damn.
20:42 Dag.
20:44 Um, so I don't know if you remember Mr.
20:47 Strawberry. He he was he came out I
20:51 don't know a year or so ago and was
20:53 hyping up hyping up you know feel the
20:56 AGI and hyping up Open AI and then Sam
20:59 Alman retweeted him
21:01 and he annoys the [ __ ] out of me because
21:05 the stuff that he posts is just hypy
21:07 [ __ ] and it like it rarely like it
21:11 often doesn't come true and and it's
21:15 like it's just it's just hypy [ __ ]
21:17 But he posted something today that looks
21:20 like he didn't take his meds or he's
21:22 seen
21:23 something. And so I want to read it, not
21:26 because I particularly believe him, but
21:30 there's something about how it's written
21:32 that I think is what we're about to
21:35 experience and and it's kind of nuts. So
21:39 even if the timeline's different from
21:40 what he says, I think this is about what
21:43 we're all about to experience. So that's
21:45 that. And then the other thing that I
21:47 thought might be fun to try and we'll
21:49 see how fun it is because it this could
21:51 also be not fun. Um I got access to
21:55 dscripts agent where you can just talk
21:59 video editing into
22:02 existence. Um and I tried it once and
22:05 sort of it kind of failed on me and I
22:06 didn't I never really dug deeper. So, I
22:09 thought it might be fun to, you know,
22:11 sort of talk about the point that he
22:14 made or prediction that he
22:17 made and and what that's going to look
22:20 like because I I think we need to get
22:22 ready. And again, I think, you know, the
22:24 the folks that hang out on this channel
22:26 regularly,
22:28 um, we're as prepared for what's coming
22:31 as anyone. So, I, you know, I think
22:33 we're fine. Um, but I think shit's going
22:36 to start getting weird.
22:38 Um, and then I thought we'd just go play
22:41 with Dcript and just like I'll just sort
22:44 of, you know, learn and and fail
22:46 publicly or learn and succeed
22:49 publicly. Um, damn, you and the dog and
22:52 the guitar make a great trio. Oh, thank
22:54 you very much. I appreciate that. Yeah,
22:56 you know, the doggy the doggy's the lead
22:58 singer. I'm just here backing up the uh
23:01 the the diva. The diva we call champ.
23:05 [Music]
23:34 All right, Let's let me go. Let me go
23:36 find this thing and then and then once I
23:39 find it, we can we can share it. Let's
23:42 see.
23:45 Uh I
23:51 [Music]
23:58 [Music]
24:00 rule. Hold please. Oh, wait. I responded
24:03 to him. So, let me I can find it on my
24:06 thing. Did I respond or did I did I
24:09 reply? I don't think I
24:11 replied. Hey, I I'll be with you. Just
24:15 calm down. Here it is.
24:21 Okay. All right.
24:26 So
24:28 again, the context here is
24:32 that the guy whoever's behind this
24:36 account is mostly full of
24:38 [ __ ] and it's mostly hypy [ __ ] and I
24:42 I you know it's it's got a crypto bro
24:47 kind of feel to it. So I have a feeling
24:48 this is someone that's kind of connected
24:50 with Worldcoin or you know World ID or
24:54 something like that.
24:56 But I'm just gonna I'm just gonna read
24:59 what he said and then so take it with a
25:02 massive grain of
25:04 salt.
25:07 Um
25:09 and and then I'll sort of unpack at
25:12 least like this is one of those things
25:15 where this has me now starting to
25:17 starting to pay attention to what's
25:19 coming. So he goes, "It feels quiet,
25:23 doesn't it?
25:24 You are not prepared for the next four
25:27 weeks. Not at all. And by the way, if
25:28 you want to see this account, it's I
25:30 rule the world mo as in master
25:35 organization mo.
25:38 Um, just spent six hours with Deepseek
25:43 R2 and I literally can't process what I
25:46 witnessed. The cognitive architecture is
25:48 [ __ ] alien. No one is ready to admit
25:51 the intelligence gap between the last
25:53 gen models and what's happening now.
25:55 We're not seeing 2x or 10x jumps
25:57 anymore. We're seeing whole order of
25:59 magnitude leaps in
26:01 capabilities capability landscapes that
26:04 didn't even exist before.
26:10 Um, there's a good possibility this is
26:18 horseshit. Jim Ross is now officially in
26:21 love with Lovable. I'm telling you,
26:23 Lovable, if you have not played with
26:25 lovable.dev, go there right now and
26:27 instead of listening to me, just go tell
26:29 it to make you an application of any
26:31 kind and watch what it does. If you
26:34 haven't played with it yet, it's insane.
26:36 Uh, as in Strawberry. Yes, this is Mr.
26:39 Strawberry is is what what he was called
26:41 when he was doing all that. I would
26:43 snark about Nicks on social media sites,
26:47 but I'm a sea
26:49 slug. Okay. Um, I watched it solve three
26:53 impossible math conjectures while
26:55 simultaneously building a theoretical
26:57 framework of
27:02 consciousness. Like, as I'm reading
27:04 this, I'm thinking, is this just
27:05 horseshit? This could just be horseshit.
27:12 Let me keep reading
27:13 it and then just assume that everything
27:17 I'm reading is
27:18 horseshit because I I I do want to talk
27:20 about it at the end because I do think
27:23 that what he's describing here is going
27:25 to happen. I just don't think it's going
27:27 to happen in the next four weeks.
27:29 Um, I wa I watched it solve three
27:32 impossible mathematical conjectures
27:33 while simultaneously building a
27:35 theoretical framework for consciousness
27:38 transfer. Uh, consciousness, yeah,
27:40 consciousness transfer that made actual
27:42 sense, not metaphorical sense, not
27:44 philosophical sense, engineering
27:46 specification sense. Researchers in the
27:48 room were physically shaking. One kept
27:50 laughing
27:51 uncontrollably uncontrollably saying,
27:53 "This isn't supposed to be possible
27:55 yet." 03 Pro and Gemini Ultra are
27:59 running similar architectures now and
28:02 the emergent behaviors between
28:04 interconnected systems are producing
28:06 capability spikes that nobody predicted
28:08 on any timeline. forget AGI debates.
28:15 Um, uh, Grock 3, Grock 3 is the only one
28:20 being somewhat transparent about what's
28:22 happening under the hood. And even
28:24 they're downplaying the weirdness by at
28:25 least 60%. That's a weird number, but
28:28 like why is that number there? So again,
28:31 like I I feel like he might be off his
28:33 meds. When when these systems start
28:36 self-modifying beyond our ability to
28:39 trace execution paths, the acceleration
28:42 curves be become actual mathematical
28:45 singularities, not
28:47 metaphorical literal computational
28:50 explosions that create capabilities no
28:52 one can
28:53 predict. Um, and so he goes on, he says
28:56 this is going to happen in the next four
28:59 weeks. So whether this happens in the
29:01 next four weeks or the next four months
29:03 or even the next four years,
29:06 um I do think there's going to be kind
29:10 of a snap um in
29:13 capability. And I also agree with him
29:16 that all of the from from what I've
29:19 experienced over the past year, all of
29:22 the frontier model
29:26 companies, they're either copying one
29:29 another or they're starting to use the
29:31 AI to improve itself and it's
29:34 improving the models in similar ways.
29:39 But but I think we're going to keep
29:40 seeing this thing where you're going to
29:42 get one of them release something
29:46 remarkable and then you'll have a bunch
29:48 of very quick follow on. So next week is
29:52 Google IO and Google's supposed to be
29:54 releasing new [ __ ] and everyone else is
29:58 probably going to launch new [ __ ] to
30:01 counteract that. So what we don't know
30:04 is is anyone going to launch new [ __ ]
30:06 that is actually really gamechanging and
30:08 is kind of a snap function that is way
30:11 beyond what we have right now. And I
30:13 don't know but
30:16 um I just have this simmering feeling
30:19 like something something is coming that
30:22 is not going to
30:25 feel we're not going to know what to do
30:27 with it is I guess what I'm saying.
30:31 Um, but you know, [ __ ] I feel like I
30:34 don't know what to do with the current
30:35 tech we have. So, so there's that. Uh,
30:39 thoughts,
30:41 questions. As now that I've now that
30:44 I've read that, don't know lovable. If
30:46 you need to go play with
30:47 lovable. Hi, H. Hello, Stacia. What's
30:51 happening?
30:54 [Music]
30:58 I have ranch dressing on my word
31:01 [Music]
31:14 salad. Co-pilot lied to my face repeated
31:18 repeatedly
31:19 today. Co-pilot.
31:23 Yeah, all of these things lie
31:25 repeatedly. They just do. They're just
31:28 bad. They're just bad. We're going to so
31:31 look back on this time and be like, "Oh,
31:33 remember when they used to
31:36 lie and now they just flat out
31:40 manipulate." All right. Um, what are we
31:44 going to do? Maybe we'll go make
31:45 something
31:46 lovable. I have a couple of use cases
31:49 lined up for lovable.
31:52 What should I make in lovable? You guys
31:53 that are doing it. Jim Russ, what's
31:55 something I should make? Hey, Mr. It,
31:57 what's happening? I'm doing
32:00 well.
32:02 [Music]
32:05 Uh, the more I read of that Mr.
32:08 Strawberry thing, the more I think it's
32:11 [ __ ] So, that's why I kind of
32:14 cooled down on not digging deep on that.
32:15 I do think that we are going to have at
32:18 some point one of these frontier model
32:20 companies is going to launch something
32:22 that's just going to be very different.
32:25 Um and then and then and then stuff
32:27 starts to get crazy. All right.
32:32 Um maybe I should tell it to be
32:35 impatient. Oh, I'm impatient. B Yeah,
32:38 you can listen. If you're not yelling at
32:40 your if you're not yelling at your uh
32:43 your your large language models, you're
32:45 not doing it right. I yell at them all
32:47 the time. Tell them they're being
32:48 idiots. Tell them to dig deeper. I like
32:51 all the time. I'm like it
32:54 just they're exhausting.
33:00 They really do give you sort of sort of
33:02 like
33:03 the the average answer, the average non
33:07 shitty answer to every question unless
33:09 you, you know, trick it into thinking
33:13 deeper. So,
33:16 um, all right. Here we are on lovable.
33:19 Let's
33:20 see. What do I want to make here that
33:22 Oh, you know what would be fun that that
33:24 I haven't done before? Okay, we'll we'll
33:26 do this.
33:30 Um, I want to
33:34 create a
33:39 synthesizer.
33:43 Synthosizer. How do you spell that? S I
33:46 Z er. Yes. I want to create a
33:50 [Music]
33:54 synthesizer musical
34:00 instrument
34:04 that
34:06 plays like a video game. You know what?
34:09 We're not going to we're not going to
34:11 architect this here. We're going to
34:12 architect this in chat
34:15 GPT because you only have so many hits
34:19 with
34:22 lovable. And so we're going to do this.
34:24 We're going to do this new. We're going
34:25 to do this with 04 mini high, which is
34:29 great at coding and visual
34:32 reasoning. Or do we want to do
34:34 03? Let's do 03. We'll do the big
34:37 MacDaddy model. All right.
34:40 So I want to create a synthesizer
34:43 musical instrument that plays like a
34:45 video game.
34:48 I want it to be very visual.
35:02 Incorporate music
35:05 theory in a
35:08 way that the user just
35:14 benefits from its
35:17 knowledge but doesn't need to
35:22 know theory.
35:35 Um, the video game
35:45 component
35:46 should score you higher as you layer on.
36:00 [Music]
36:02 sounds and
36:05 beats. Act like a
36:10 senior
36:13 product visionary and manager.
36:19 And let's
36:22 uh create a spec for this bad
36:29 boy or
36:34 girl. How are you all? My father-in-law.
36:38 I got my father-in-law to stop start
36:40 watching Kyle over the weekend. Sorry
36:43 about that, father-in-law. It but it
36:45 doesn't get better than this. This is
36:48 what you're experiencing right now. This
36:49 is the pinnacle of what this channel is.
36:52 So, if you're expecting more, ju
36:56 just don't because what where it very
37:01 quickly can
37:02 devolve is things like this. You know,
37:06 you know what I'm
37:09 saying? And I'll just talk about [ __ ]
37:11 with a pink bow on for no reason.
37:15 Well, mainly the reason is I was told by
37:18 the woman that gave me this, I'm wearing
37:20 it wrong. Every headband I've ever worn,
37:23 this is how you wear it.
37:26 Apparently, I got I got woman explained
37:29 how to wear a pink bow. Okay. Okay. He
37:33 and Kyle are the same. Well, that's
37:40 bad. If you're expecting more, just
37:43 [Laughter]
37:47 don't. Oh, my daughter who writes grants
37:50 for a nonprofit had chatbt write an
37:53 entire grant first time. She's pretty
37:55 excited. Oh, that's cool. Yeah, when you
37:57 when you do stuff like
37:59 uh when when someone hasn't had their
38:02 first Kevin Mallister moment and if
38:04 she's in grant writing and this thing
38:05 just knocked out a [ __ ] grant. Yeah.
38:07 It's, you know, because you got to
38:10 figure large language models have been
38:12 trained on all the [ __ ] that's been put
38:14 on the internet. Well, what [ __ ] that's
38:16 been put on the internet? Every [ __ ]
38:18 grant application, you know, grant
38:21 writing course, every grant, right?
38:23 Every failed grant, every succeeded
38:25 grant, every analyzed grant, every
38:27 business school
38:29 grant, they're all in there. So, it
38:31 knows everything. So, like that's a
38:33 great use case where it's just going to
38:34 be really good at that [ __ ] All right.
38:37 03 thought for three seconds. It's
38:39 called Synth Quest, a gamified visual
38:41 synthesizer experience. Product vision.
38:45 Invent the first Play It Like Mario,
38:48 sound like Brian Eno instrument. That's
38:50 brilliant. Play it like Mario, sound
38:53 like Brian Eno instrument. Players
38:56 sculpt beats and melodies by jumping,
38:58 sliding, powerup, stacking audio layers.
39:01 Under the hood, we enforce solid voice
39:04 leading harmonic rhythm and modal
39:07 interchange. I don't even know what the
39:09 [ __ ] that is. Tik Tok pin. Yeah, if
39:12 you're expecting more,
39:15 don't. Well, welcome to the AI learning
39:18 lab where ain't no one learning and
39:20 ain't much
39:22 AI. If you're expecting more, don't.
39:31 All
39:32 right. Invent the first. Play it like
39:34 Mario. Sound like Brian Eno instrument.
39:36 Like that's a [ __ ] marketing scheme
39:38 right there, isn't
39:40 it? Player Sculpy. Okay. Under the hood,
39:43 we enforce solid voice leading harmonic
39:47 rhythm and modal interchange. So
39:49 whatever they do sounds legit, but the
39:52 player only sees color, shapes, and
39:54 points
39:55 climbing. Target personas. The bedroom
39:59 producer, streamer Sam, the music music
40:03 ed
40:05 mentor, core gameplay and UX
40:10 loops, invisible invisible theory
40:13 engine, circle of fifths, world map
40:15 ring, warp pipes, modulate keys. Okay,
40:19 voice leading rules. It auto renders
40:21 collect. Okay, it doesn't [ __ ]
40:23 matter. This is already above my head
40:26 just in the product
40:29 description. This is
40:31 unbelievable. Oh, so Oh, yeah. Base
40:34 game. Okay. So, wait. Okay. Okay. This
40:36 is great. You can pull out the
40:43 marketing stuff for now and the
40:49 timeline stuff.
40:53 I want you to write this as
40:58 a
41:00 product
41:04 spec assuming we can just make it
41:12 happen
41:14 order. Uh let's see.
41:17 I'll be using an AI
41:24 vibe coding app
41:28 um to
41:31 execute it. So I just need it very well
41:39 articulated. Uh, make sure that the game
41:46 play and
41:48 visuals are
41:50 world
41:53 class. Don't get lost in the geeky music
42:00 [Laughter]
42:06 [ __ ] All right. Fantastic. Fantastic.
42:10 Bob, tell him what he's won. Uh, he
42:14 hasn't won anything. It's a This is not
42:16 a game
42:17 show,
42:26 right? Create the first game field
42:29 synthesizer.
42:30 Wait,
42:33 where? Okay, we're going to go steal the
42:36 the product vision.
42:40 So, we're going to pop that into
42:43 lovable.
42:45 Boom. Then we're come back over to chat.
42:49 Jetpack. We're going to scroll on down
42:51 to the bottom here. End of
42:54 spec. Oh,
42:57 Jesus. Game play note router. Unity
43:00 render loop theory engine. Good
43:05 lord. I wonder if I can I copy all this
43:08 together? I bet I can. All right, fine.
43:11 So, here we're going to
43:13 go. All right. Purpose and North
43:19 Star audio
43:21 requirements less than 10 milliseconds.
43:24 All right. Here we go, people. Boom.
43:26 This is a big [ __ ] complicated ass
43:30 thing.
43:32 Paste. Make it. Oh, sign in. Damn it.
43:40 Shut
43:41 up. You know what? Does he have any
43:44 qualific What are his qualifications?
43:46 Does he have any
43:48 qualifications? Is he a professor? He he
43:51 seems he seems like a an atypical If
43:55 he's a professor, he seems
43:57 atypical. He He doesn't seem like he he
44:00 knows anything. is that that this is
44:03 what you do
44:06 here.
44:07 Huh. It's a novel approach to
44:11 education. It takes my breath away. It's
44:13 just
44:17 a
44:21 lot. All right, we're
44:25 off. Spinning up the preview. Okay, so
44:28 now it's going to do its [ __ ]
44:30 This is vibe coding. If you haven't seen
44:32 vibe coding before, it's like come up
44:34 with an app idea and tell it to go make
44:36 it and then it does [ __ ] and then it
44:38 works or it
44:42 doesn't. He got a certificate from a
44:45 cereal box. Yeah, but I lost the cereal
44:47 box before I cut the certificate off it.
44:50 That's the problem. Nobody else hear
44:53 that? Hear
44:54 what? Joker, are you hearing those
44:57 voices again?
45:02 You You gota You gotta get off the the
45:05 ether there,
45:07 buddy. We're off like a herd of
45:14 turtles. Oh, man. All right. What's it
45:17 doing? What are we doing
45:25 here? Too many mushrooms. Exactly.
45:29 Exactly. Duty-free designs World Bazaar.
45:33 I like it. Solid name. Solid
45:36 name. Okay. I'm excited to create this
45:40 innovative music game hybrid that
45:42 combines performer mechanics with
45:44 audiovisisual synthesia.
45:48 The concept blends the tactical feel of
45:50 Mario platforming with Brian Eno's
45:53 ambient soundsscapes in a way that makes
45:55 music creation intuitive and
45:59 visceral. Drawing inspiration from games
46:02 like Rez Audio Surf and Hi-Fi Rush, I
46:06 envision a colorful fluid experience
46:08 where movement directly translates to
46:11 music with vibrant visuals responding to
46:13 every note. The cell-shaded neon
46:17 aesthetic will with particle effects
46:21 will
46:22 certainly will create
46:25 sinthetic connection between sound and
46:27 visuals. For the first version, I'll
46:30 implement basic platformer mechanics,
46:33 sound generation, visual
46:35 feedback, theory engine. Cool. I'll use
46:40 vibrant color palette and neon stuff.
46:42 All right.
46:44 It's making [ __ ]
46:47 now. Now, if you haven't played with
46:49 Lovable before, like what it's doing
46:50 right now, I think. Can we go look at
46:54 uh open preview
46:57 tab? Yeah. Here's all the code it's
47:01 writing. So, see this little block over
47:03 here to the
47:05 left? This little thing right here? It's
47:07 got the little code. It says writing the
47:09 code. So if you click on
47:12 that, here's all the code it's writing.
47:14 So here's
47:17 index.tsx game
47:20 context. A lot of game
47:24 context, then audio
47:29 context. If you've never coded before,
47:32 you can just ignore this [ __ ]
47:35 But if you're like curious, like a good
47:37 way to learn coding is watch what these
47:39 things are
47:40 doing. They're doing it way
47:43 fast. Way fast and way
47:48 competent. The intro
47:51 overlay configure. Oh, spinning up
47:53 preview. We just went on to our next
47:55 thing. So, it
47:56 created
47:58 Whoa. Whoa. All right. What do we
48:02 got? Error. Try to fix
48:06 it. Vibe coding
48:16 baby. Kyle, this is Mr. It trapped in my
48:19 tablet in my other Tik Tok
48:24 account. Dutyfree designs is is not
48:28 really a duty-free design shop. Well, it
48:30 might be, but it's Mr.
48:31 it. So, it's an irregular. Ah, okay.
48:35 Harmonic
48:37 jumper. Create beautiful music through
48:40 movement. Jump and collect orbs. Okay.
48:42 Arrow keys to move.
48:45 Okay. Space or up arrow to jump. Collect
48:49 sound orbs to create melodies. Land on
48:53 platforms to trigger beats. Oh, error.
48:55 Try to fix. Damn it.
49:01 Start creating.
49:08 Oh, didn't work.
49:11 Okay, we're facing an issue with the
49:14 context provider router. Use audio
49:16 context must be used with an audio
49:22 provider. All right, it's it's fixing
49:24 it. But look, we got we got neon glowy
49:27 [ __ ]
49:28 [Laughter]
49:34 Harmony. See if we can get something
49:36 working
49:37 here. So, does Lovable make apps? Yes,
49:40 Love Well, Lovable makes web apps. So,
49:43 this is
49:44 important. A web app can only run in a
49:47 in a web browser. I guess you could
49:49 probably do some sort of container where
49:50 you run it on the desktop, but it's
49:52 basically just a it's running in a
49:54 browser.
49:59 um that doesn't mean it's not valuable
50:02 and powerful, right? Because there's all
50:04 sorts of ways um all sorts of apps that
50:07 we use are just web apps. Um but if you
50:10 want to make like a native iPhone app,
50:12 you can't do that on it. Or if you want
50:14 to make a native standalone Mac OS app,
50:17 OSX app, it's not doing that. So, it's
50:20 basically making web apps, which the
50:22 advantage of that is that it can run the
50:25 thing it writes right in your browser
50:27 while you code it. Web apps only. Yep.
50:30 Exactly. Can our main character be a
50:32 turtle? Yes, we'll do that. Start
50:34 creating.
50:35 Okay.
50:41 Um Okay.
50:44 Um,
50:46 the main
50:49 character should be a
50:53 turtle.
50:54 And when I try to move, the main
51:01 character
51:06 disappears.
51:09 Also, the
51:13 Um,
51:15 there are
51:17 only
51:19 dots on a
51:22 grid and it doesn't feel like a video
51:29 game. Make the game
51:33 play
51:39 better. Start jumping, Kyle. Yeah. Oh,
51:42 let me let me reload this and then maybe
51:44 we can reload. Let me jump first. Start
51:48 creating.
51:50 No, I can move right and left, but then
51:53 it
51:55 disappears. The issue seems to be with
51:58 the character disappearing in the
51:59 overall game not feeling engaging
52:01 enough. Let's fix that.
52:06 Let's But anyway, like even if this
52:08 didn't work, like the fact that it like
52:10 just understood the game, spun this
52:13 up, did the splash screen, you know, did
52:17 all the other stuff, it it's it's just
52:20 crazy. And again, what what you need to
52:23 do is do what Jim Ross did with this is
52:26 like, you know, just play with it. Try
52:28 to make an app for your business or try
52:29 to make an app for your I don't know
52:32 recipe
52:33 organizing what whatever you want to
52:35 [ __ ] do your your uh your obsessive
52:38 compulsive uh seed organizing by time of
52:42 season for planting
52:48 whatever. Uh what what champ? What do
52:51 you want? Oh, you want to go out because
52:53 you hear people out there and you want
52:54 cheese now. Is that what's going on?
52:56 Yeah. Yeah, you a cheese eater. Oh,
53:00 please. We'll let that thing keep
53:02 writing. Come
53:22 on. Still writing. All right, let's go.
53:26 Let's go over to decco here while we're
53:29 doing Well, that thing's working. Let's
53:32 do some chat add here, shall
53:35 we? I hear people and I want cheese,
53:40 too. Nice site, by the way.
53:44 Um, you can't do anything while it's
53:47 working. Yeah, I know it's a drag. Okay,
53:49 wait. So, I'm going to go to decode
53:50 here. Not decode here. Descript.
53:55 vscript
54:02 web.dscript.com. And if you haven't seen
54:04 dscript, dscript is a very
54:08 weird though very
54:11 powerful AI video editing tool. What's
54:15 weird about it is you upload a video and
54:17 it transcribes it and you kind of just
54:19 interact with the script of the, you
54:23 know, the transcript of what people
54:26 say, but
54:29 they've but they've now added an agent
54:33 to it that you can just ask it to do
54:34 [ __ ] and it does
54:36 it. Dropping you something in a regular
54:39 is very nice. Why is this taking so long
54:41 to load?
54:42 Let me just go to dscript.com. This was
54:45 web.dscript.com. So something's
54:48 up.
54:51 Um, well, it's [ __ ] up again.
55:18 Reload
55:21 that. Okay. Well, I I don't even know
55:24 where to
55:26 begin. Okay. Turtle is fine.
55:31 The
55:34 background star noise
55:37 is
55:39 distracting
55:41 and not
55:45 musical. So, let's kill
55:49 it. the left and
55:57 right.
55:58 Just jump the
56:05 turtle to
56:08 opposite sides of the
56:14 screen. Up
56:16 arrow doesn't work.
56:20 It doesn't
56:23 scroll. In fact, it
56:28 sucks. Make it
56:34 playable. So, yeah. Uh, tabs for audio
56:37 outside of Tik Tok. Yep. Dropped you
56:39 something in irregulars. All right,
56:41 let's go look at irregulars while we're
56:42 at it. Well, that thing's working.
56:47 [Music]
56:49 new
56:51 activity archetypal
56:53 architect. What are your
56:56 coalifications? Yeah, I see what you did
56:58 there. You see that? That was for
57:06 Steo. Um, certified AI consultant Karen,
57:10 thank you for sharing the live. Much
57:13 appreciated.
57:15 Uh, if you haven't shared the live and
57:17 want to share it and annoy some people
57:18 in your life, that would be a good
57:20 thing. Now's the time to do it. Um, if
57:25 you have people in your life, or if
57:26 you're one of them that are actively
57:29 ignoring this AI [ __ ] you got to knock
57:31 that [ __ ] off. Like, seriously, it's
57:33 getting serious now. If your friends are
57:36 like, "I don't care about AI is evil. I
57:38 don't know why you like that AI." You
57:41 gota they got to get their [ __ ]
57:42 together. They got to [ __ ] snap out
57:44 of it because this stuff's
57:47 coming. And here's the here's you want
57:49 to know the dirty little secret about
57:51 all the bitter people about AI? I
57:54 promise you, they're going to be the
57:56 ones that fall the hardest where they're
57:58 just like, "Oh my god, it's so amazing.
58:00 I didn't realize it could do that. I
58:03 just thought it was evil." But now that
58:05 I use it, wow, it's gonna be a lot of
58:08 that going on. So anyway,
58:11 yeah.
58:14 Yeah, Dennis referred you. Great.
58:17 Awesome. Welcome, welcome, welcome. So,
58:19 okay. So, if there are some new folks
58:21 here, my name's Kyle Shannon. Uh, this
58:24 is a neurospicy friendly channel. So, I
58:28 have lots of uh lots of lots of issues
58:31 uh conditions, ADHD being the primary
58:34 one. Um, I am also hopelessly optimistic
58:38 and um,
58:42 uh, encouraged by generative AI and
58:46 what's coming. Um, most of the people
58:48 that are really angry about this in my
58:50 experience haven't used it. So, there's
58:53 a lot of people on soap boxes that are
58:56 on the sidelines. So, I don't care if
58:59 you're on a soap box, but if you're on a
59:01 soap box and you don't know anything
59:02 about what these tools actually do and
59:04 how they work and what the issues really
59:07 are, then I don't want to [ __ ] hear
59:09 it
59:12 because not for
59:14 nothing, this is the the kind of purpose
59:16 of the channel, and I'll get to the AI
59:18 salon in a second. The purpose of the
59:20 channel
59:22 is not to judge AI one way or the other.
59:28 My opinion about whether AI is good or
59:31 bad or your opinion about whether AI is
59:34 good or bad is completely independent of
59:38 the fact that it's here and it's not
59:39 going away. So what this channel is
59:42 about is is the basic premise of the
59:44 channel is if AI is not going away and
59:47 if its capabilities are accelerating
59:50 which they are, what are we going to do
59:52 about
59:53 it? How do you learn? How do you keep up
59:56 with it? I've been trying to keep up
59:58 with it for two years, two and a half
1:00:00 years. It's absolutely impossible. You
1:00:02 cannot keep up with it. Okay, so what do
1:00:05 you do with
1:00:06 that? How do you build? What do you do?
1:00:08 How do you incorporate this in your
1:00:10 business? How do you incorporate this in
1:00:11 your job? How do you incorporate it in
1:00:13 your
1:00:14 family? I don't know. So, what we do
1:00:17 here is we just hang out five nights a
1:00:19 week. So, I come in generally around
1:00:21 8:00 p.m. Mountain
1:00:23 time and I just go for about two hours
1:00:26 and just ramble. So, it's like chat
1:00:29 add or chat TMZ when we're doing
1:00:32 speculation about what's
1:00:34 coming. Um, and then sometimes I'll play
1:00:37 with stuff, sometimes I'll teach stuff
1:00:38 if I know it, sometimes I'll experiment
1:00:40 with things and just watch it fail over
1:00:43 and over
1:00:44 again. It's so sad that it's the 20somes
1:00:47 that are the most opposed to it. Well,
1:00:49 so here's the thing, archetypal
1:00:51 architect. The [ __ ] education system
1:00:54 told them that if they used it, they'd
1:00:56 be kicked out of
1:00:59 school. So So the the Gen Zers and the
1:01:03 Gen Alphaers have basically been
1:01:05 educated that AI is
1:01:09 cheating and so they're not using
1:01:13 it. It's It's [ __ ] inexcusable. It's
1:01:17 [ __ ]
1:01:18 inexcusable. Um,
1:01:22 anyway, good for us Gen Xers. Gen Xers
1:01:25 are kicking ass with uh and elder
1:01:27 millennials are kicking ass with AI
1:01:28 because we got to experience what
1:01:31 happened with the worldwide web and how
1:01:33 that changed everything, right? And now
1:01:36 with AI, we're like, "Holy [ __ ] this is
1:01:40 a [ __ ] thousand times more powerful
1:01:42 than what the worldwide web was." And
1:01:44 look what that did.
1:01:46 Um, plus we're also just ornery
1:01:51 pricks where it's like we just want to
1:01:53 [ __ ] some stuff up and AI is the perfect
1:01:55 thing to do that, right? It is
1:01:58 shattering, you know, it's going to
1:02:01 shatter every everything we do. Every
1:02:03 way of working is going to be completely
1:02:05 upended and
1:02:07 transformed. And who gets to benefit
1:02:09 from that are the
1:02:11 curious. That's what this channel's
1:02:13 about. Okay. So,
1:02:16 um, go to the AI salon. Uh, Brandon had
1:02:19 the address up earlier, the
1:02:22 salon.ai. Scroll down a little bit and
1:02:24 click on join our
1:02:26 community. And that'll take you over to
1:02:28 the active community. There's about
1:02:30 3,000 people that are there that are AI
1:02:33 optimists and adventurers and people
1:02:36 trying to figure this stuff out. And
1:02:40 it's a remarkably like remarkably
1:02:43 generous
1:02:44 community. And so if you go there, the
1:02:48 second step, we ask you to basically,
1:02:50 you know, come in, take a look at what
1:02:51 we're about. Step two, introduce
1:02:53 yourself. So I would strongly encourage
1:02:55 you go to the salon and introduce
1:02:57 yourself. Um, and if you're like, "Yeah,
1:03:01 but I don't know anything about AI and
1:03:03 just someone told me to come here and I
1:03:05 think you're kind of weird and I don't
1:03:06 even know why I'm here, but I'm still
1:03:08 here and this is stupid and you don't
1:03:10 have any qualifications." Yeah. Yeah.
1:03:12 All that. Go in there anyway and
1:03:16 introduce yourself and just say that.
1:03:18 Just say, "I don't know what I'm doing
1:03:19 with AI, but I it seems like something I
1:03:22 should do."
1:03:24 Great. It's a very welcoming community.
1:03:28 Um, and if you know what's going on with
1:03:31 AI, get in there and share what you're
1:03:33 learning. As you start figuring this
1:03:36 stuff out, start sharing it. And in
1:03:39 fact, one of the things we talk about in
1:03:41 the salon, actually, the first thing we
1:03:43 talk about in the salon is what we call
1:03:45 the AI readiness
1:03:47 cycle, which is play first, mindfully
1:03:51 create, generously lead, play, play with
1:03:55 AI. Why do I act like an idiot in here?
1:03:59 Because everyone's taking AI so [ __ ]
1:04:01 seriously. Lighten up,
1:04:04 people. Sure, it's gonna decimate the
1:04:07 job market, but let's have fun on the
1:04:10 [Laughter]
1:04:13 way. There's actually some seriousness
1:04:16 behind the play first though, and that
1:04:18 is this.
1:04:19 Everything is moving so fast and there
1:04:22 is so much and it is so insanely
1:04:28 powerful that you can't right now if you
1:04:31 haven't been playing with the this AI
1:04:32 stuff. You can't right now even imagine
1:04:36 what's now possible that wasn't possible
1:04:39 two years ago. Like literally wasn't
1:04:42 couldn't do it. It is mindshattering
1:04:47 what is now possible. And so one of the
1:04:50 best ways to do it is get in a community
1:04:52 like the salon and just start playing.
1:04:54 And as you're playing, you'll learn some
1:04:56 [ __ ] You'll learn, oh, I can make like
1:04:59 a really cool song on Sunno and I can
1:05:01 make really cool images on MidJourney or
1:05:03 Chat GPT. And oh, I can write a
1:05:05 marketing plan. I didn't think I could
1:05:06 write a marketing plan. It does a whole
1:05:08 marketing plan. Well, I bet it can't
1:05:10 write social media copy. Oh my god, it
1:05:12 wrote social media cut. Like you you
1:05:15 will just what we call them your Kevin
1:05:17 Mallister
1:05:18 moments right from home
1:05:21 alone. Have some of
1:05:24 those and then mindfully create is once
1:05:27 you've learned a couple of things now go
1:05:29 oh okay let me try to do this thing for
1:05:32 my business or do this thing for my mom
1:05:33 or do this thing for my uncle who wants
1:05:36 to start a business. Maybe I can do
1:05:39 that. Right? That's the mindfully
1:05:40 create. try to actually use AI to do
1:05:43 something once you play with a pile of
1:05:46 tools and then generously
1:05:49 lead is about teaching what you're
1:05:52 learning. So here's the the constant
1:05:55 that I see in the salon is that people
1:05:58 consistently
1:06:00 undervalue their
1:06:04 contribution. So they're like, "Well, I
1:06:06 just I I just I'm like really into
1:06:08 gardening and like all I did was use
1:06:11 chat GPT to come up with an automatic
1:06:13 gardening thing and it's not that big a
1:06:15 deal." Well, to people that are not into
1:06:18 gardening, it's a pretty big [ __ ]
1:06:19 deal. It's gardening season. Maybe they
1:06:22 want to get out and garden, right? So
1:06:24 you teach people what you're learning
1:06:26 and then someone else will go, "Oh, I
1:06:28 never would have thought about it like
1:06:29 that." That's what this community is all
1:06:33 about. And as you then share that,
1:06:37 people give you feedback. Oh, here's a
1:06:39 thing I did. And then you're like, oh,
1:06:41 that's inspiring. Let me go try that.
1:06:42 You get back to
1:06:43 play, right? You just keep doing that.
1:06:46 If you keep doing that, you will win.
1:06:49 You will win. Whatever the next 5, 10,
1:06:52 20 years is is going to bring is going
1:06:55 to
1:06:58 require nimleness, curiosity, adventure,
1:07:03 adaptability, judgment,
1:07:07 curation. It's not going to require
1:07:09 taking a class in prompt engineering.
1:07:12 It's not going to require I've got to
1:07:13 learn
1:07:14 Python. It's not going to require I've
1:07:17 got to go back to school for music
1:07:18 theory. All of that's out the
1:07:22 window. Everything that you thought you
1:07:25 couldn't do in your
1:07:27 life, you either now can do or you will
1:07:30 very shortly be able to do
1:07:33 anything. Well, I always did want to be
1:07:36 a quantum physicist. Okay, you can
1:07:38 probably do that now.
1:07:44 And like a year from now, you'll have
1:07:46 like a research lab of like a thousand
1:07:49 PhD level quantum physicists working for
1:07:54 you. So you want to prove that the Earth
1:07:57 is flat and it actually lives on the
1:08:00 head of a
1:08:01 pin in some 1870s Taylor shop. Cool. Go
1:08:06 prove it.
1:08:13 We all get to reinvent who we are and
1:08:16 what we do. We all get to We all get to
1:08:18 ask ourselves the question. This is the
1:08:20 thing I'm most excited about with
1:08:23 AI. If I could do
1:08:29 anything, like period, if I could do
1:08:33 anything, if I could accomplish
1:08:35 anything, if I had the skills to do
1:08:39 anything, what would I choose to do?
1:08:42 what difference do I want to make in the
1:08:43 world? And then you can go do that. Now,
1:08:46 right now, because we're so early, you
1:08:48 have to go play with all these tools and
1:08:49 figure all this [ __ ] out. That's where
1:08:51 all the opportunity is. So, welcome. I
1:08:55 love philosophical nights. Doubletapping
1:08:58 on my tablet and my
1:09:01 phone. Thank you very much. Um, all
1:09:05 right. Where were we? We were at
1:09:07 Lovable. Let's see if Harmonic Jumper is
1:09:10 back.
1:09:12 It already wants to refactor code. I've
1:09:14 made several improvements. Start the
1:09:17 adventure. Error. Try to fix it. All
1:09:20 right, that's broken. Back to dcript.
1:09:23 Loading is taking longer than normal.
1:09:25 You
1:09:28 think I don't know what's up with
1:09:30 Dcript. Maybe they're redoing their
1:09:32 servers.
1:09:34 [Music]
1:09:41 [Music]
1:09:48 Um
1:09:51 descript sign
1:09:56 in. Yeah, it seems to be
1:09:58 broken. It seems to be broken.
1:10:08 Oh, the other thing that I thought would
1:10:10 be cool to experiment with, I saved in
1:10:13 my
1:10:14 little my mind, my little this
1:10:19 thing. Yeah, let's go. Let's go look at
1:10:22 this
1:10:23 thing. Go back to lovable and refactor
1:10:26 files. It suggested you didn't read the
1:10:28 bottom. Oh, fine. What did it say?
1:10:33 The issue was with the
1:10:36 function. Oh, okay.
1:10:40 Refactor. Can I do both of
1:10:43 these? No. All right, it's off. Thank
1:10:47 you, Vicki. Vicki Vicky serves as my
1:10:50 second
1:10:54 brain. Brandon Brandon serves as my um
1:10:59 Put the black tab there for the Tik Tok
1:11:00 people.
1:11:01 [Laughter]
1:11:04 You'll have to learn to create video for
1:11:10 [Music]
1:11:12 starters. What's new in the AI space?
1:11:15 Well, good lord.
1:11:21 Um, let's see.
1:11:24 deep research in OpenAI right now can
1:11:27 now export um PDFs, fully formatted
1:11:32 PDFs. So, did I is that what this is?
1:11:35 No, that was that one. Hang on.
1:11:38 Innovative Pathways to
1:11:40 Broadway. So if I click
1:11:44 share, download as
1:12:00 PDF. Let's see what it does if I if it
1:12:03 formats it nicely.
1:12:08 Launching a new musical and shephering
1:12:10 it to Broadway has traditionally been a
1:12:13 lengthy, high-cost
1:12:15 endeavor. The traditional path versus
1:12:18 recent
1:12:20 changes.
1:12:22 Um, you know what I'm going to tell this
1:12:24 thing? I'm going to
1:12:26 say, listen, [ __ ]
1:12:30 Uh, I'm gonna say, um,
1:12:34 um, have you met
1:12:38 me? I have ADHD and a wall of
1:12:44 text makes my eyes
1:12:48 cross. It feels like homework.
1:12:55 Use bullets and
1:12:58 [Music]
1:13:01 uh and
1:13:03 uh catchy
1:13:07 uh
1:13:09 uh
1:13:12 techniques to keep me
1:13:17 interested. Again, one of the best
1:13:20 things you can do with AI is just tell
1:13:22 it what you don't like. If it gives you
1:13:23 something shitty, just tell
1:13:25 it. If Oh, this is a new thing. If
1:13:28 you're if you're not paying attention to
1:13:29 this stuff, about a month and a half
1:13:31 ago, six weeks ago, eight eight weeks
1:13:33 ago, um, chat GPT now has long-term
1:13:36 memory. So, it now can remember all of
1:13:39 your previous chats. So, you can say
1:13:42 based on what you know about me, what
1:13:44 are some ideas that we've talked about
1:13:46 that I never, you know, that I only
1:13:48 mentioned once and it'll go find them
1:13:50 and bring them back. It'll make some up
1:13:52 too, which is frustrating, but you know,
1:13:58 um, Runway now has a thing called
1:14:01 references, which is pretty flipping
1:14:04 cool. Um, Dcript now has an agent that's
1:14:07 in beta testing.
1:14:10 Um, Gen Spark is [ __ ] crazy. Um, it's
1:14:16 kind of like Manis but but
1:14:19 slicker. Um, I don't know if it's as
1:14:21 robust, but it's pretty slick. Pretty
1:14:24 slick. Look, I got tables and bullets
1:14:27 for my marketing
1:14:29 plan. Yay. Um, now I'm going to say I'm
1:14:33 going to turn on deep research again.
1:14:35 I'm going to
1:14:36 say this looks very top
1:14:40 line.
1:14:42 Now deeply
1:14:45 research exactly what I should
1:14:49 do and who I should talk to.
1:14:56 Make it a world
1:14:59 class plan that will change
1:15:04 Broadway
1:15:07 forever. All right, so AI is off working
1:15:11 for me now. Um, if you haven't played
1:15:14 with the 03 model in chat GPT, it's
1:15:18 [ __ ] insane. Oh, I've got to answer
1:15:20 questions for it. Hang on, please hold,
1:15:22 please.
1:15:23 one. What's the current state of your
1:15:26 musical? Uh script and
1:15:30 soundtrack
1:15:33 finished
1:15:35 two. What kind of support do you already
1:15:38 have?
1:15:40 uh a co-writer and
1:15:44 some
1:15:46 interested
1:15:48 producers and
1:15:51 theaters but nothing
1:15:56 concrete. Um three, are there specific
1:16:00 cities you plan to launch from like
1:16:02 Denver? Are you open to digital first?
1:16:05 prefer New York City, but
1:16:10 um open to digital
1:16:14 first
1:16:17 since
1:16:18 we use AI as a
1:16:22 collaborator
1:16:24 already. Oops, I didn't do the other
1:16:27 questions. Whatever. Now it's going to
1:16:30 go. What were the other questions? Do
1:16:33 you want help finding All
1:16:38 right. So, that's off
1:16:40 now. So, it basically goes into thinking
1:16:43 mode and just starts searching the web
1:16:46 and thinking and reasoning and doing all
1:16:49 of its [ __ ] Speaking of reasoning, how
1:16:52 are we doing here?
1:16:55 Um, error build
1:17:03 unsuccessful. Try to fix
1:17:08 it. I probably, for the purposes of this
1:17:11 channel, shouldn't have created an
1:17:14 ambitious genre
1:17:16 smashing music synthesizer that plays
1:17:19 like a video game.
1:17:21 We we could have gone with a game of
1:17:23 Frogger and and know that we'd have
1:17:25 gotten somewhere with
1:17:28 [Music]
1:17:30 it. All right, let's see. What else do
1:17:35 we have
1:17:38 here? Yes, join the salon.ai. It's free.
1:17:42 Lots of people in there to help and
1:17:43 share. Yep. It's great people. It's
1:17:46 great people. It's also great people in
1:17:48 here. If you're if you're new to this
1:17:51 channel, whether you're watching it on
1:17:52 YouTube or LinkedIn or X or on Tik Tok,
1:17:56 um just keep coming back night after
1:17:58 night. We do this regularly. Although
1:18:01 next week's going to be [ __ ] up. I'm
1:18:02 in New York. In New
1:18:04 Jersey, New York and New
1:18:08 Jersey. I fixed the type mismatch errors
1:18:11 by properly handing the note value
1:18:13 within the collect orb function. Thank
1:18:16 goodness.
1:18:17 [Music]
1:18:21 Um, let me reload
1:18:29 this. Uhoh.
1:18:41 Uh,
1:18:42 everything
1:18:45 flies
1:18:47 immediately off
1:18:53 screen. Ah, vibe
1:18:57 coding. Uh, what kind of video is that?
1:18:59 What kind of video is what?
1:19:03 what I was just looking
1:19:15 at. Yeah, there's there's lots of people
1:19:17 that are good with video. You got the
1:19:19 high score though. Try
1:19:21 preview. Oh, maybe. Yeah, let's try
1:19:24 preview.
1:19:32 [Laughter]
1:19:47 Yeah, she she's done
1:19:49 broke. Can't get D from
1:19:52 here. I fixed the issue where everything
1:19:55 flies off screen. I'll be the judge of
1:19:58 that. Yeah, you fixed it.
1:20:02 Wait. Okay.
1:20:05 Um, you need to
1:20:08 establish
1:20:09 platformer levels and the turtle needs
1:20:14 to start at the
1:20:19 beginning
1:20:21 and go
1:20:23 right. you know, like a
1:20:31 game. Uh, next year I'll get AI to make
1:20:34 my own smartphone OS. It'll be better
1:20:38 than iOS. It can't be worse than iOS
1:20:40 right now. iOS has just turned into the
1:20:42 biggest pile of dog [ __ ] It's so
1:20:45 bad. It really is bad. It's The only
1:20:48 thing worse than it is OSX. like Apple
1:20:51 has
1:20:53 just they apparently Apple needed an
1:20:56 [ __ ] CEO
1:20:59 um that just was relentless with people
1:21:02 like it can't go out until it's
1:21:05 brilliant. Um no one's doing that
1:21:07 anymore
1:21:08 and hence the
1:21:15 result. Not to mention it sounded like
1:21:17 Uncle Bob after Thanksgiving dinner.
1:21:27 run. All
1:21:28 right. Thanks, Tim Apple. Exactly. Um,
1:21:32 let me go to Let's go to runway. Oh, so
1:21:37 this is what I wanted to show
1:21:42 you. Oh, here. This is what I wanted to
1:21:44 show you. Okay.
1:21:50 So there's the image that runway
1:21:56 generated. There's the storyboard scene
1:22:00 lay layout. Tom, Brad, happy sad. Tom,
1:22:04 Brad, happy. I don't know if it's sad,
1:22:07 but at least, you know, like he's trying
1:22:09 to hold in a
1:22:11 fart. Characters Tom and Brad.
1:22:16 It's so funny. Every time I see Brad
1:22:18 Pitt, I'm like, "Wait a minute. Is that
1:22:20 me? Am I looking in the mirror? Shut
1:22:23 up." And then here's what they're
1:22:28 wearing.
1:22:30 So, and then there's what they're
1:22:32 wearing. So, like it can do this
1:22:35 now. Well, this post claims it can do
1:22:41 it. So, let's let's just go ahead and
1:22:44 steal
1:22:46 We'll just steal their storyboard. Copy
1:22:51 image. Oh, no. That's image address.
1:22:53 Copy
1:22:54 image. And let's go to runway.
1:23:00 [Music]
1:23:13 [Laughter]
1:23:19 Oh, it's so funny. All right. References
1:23:22 testing.
1:23:23 Wait,
1:23:25 where? How do I do
1:23:29 new Gen 4 for image and video? Okay,
1:23:32 here's our here's our test. So, let's
1:23:34 get rid of these things. How do we get
1:23:36 rid of them?
1:23:40 with that. No. Oh, I
1:23:43 see.
1:23:45 Okay. Can I
1:23:49 paste?
1:23:50 Back. Paste. No. Damn
1:23:56 it. All right. So, save image as
1:24:03 uh scene.
1:24:06 Save that to the
1:24:11 [Music]
1:24:22 desktop.
1:24:25 Let's I guess we can go I'm just trying
1:24:29 to figure out how much photoshopping I
1:24:32 actually want to do.
1:24:35 Oh, I guess we'll do some [ __ ]
1:24:38 work. It's 9:23. See, this is a this is
1:24:41 a [ __ ] rabbit
1:24:43 hole. Hi, Champy. What's up? You got
1:24:46 your cheese? You back to bug me now? Sit
1:24:50 down. Lay down. Lay down. Do you want to
1:24:54 sing
1:24:55 again? You want to sing
1:24:57 again? You're wagging your tail. You
1:25:00 want to
1:25:01 sing? Ready?
1:25:04 [Music]
1:25:18 Well, I heard there was a secret
1:25:21 call David played and it
1:25:25 alone. You don't really care for music,
1:25:29 do you?
1:25:32 It goes like this. The fourth, the
1:25:35 fifth minor fall in a
1:25:38 major
1:25:40 baffing composing.
1:25:44 Hallelujah.
1:25:48 Hallelujah.
1:25:50 Hallelujah. You're not into it. You were
1:25:53 just faking me out. Go lay down. Go lay
1:25:55 down, you big fat soul. That dog is fat.
1:25:58 I don't know if you knew that. That dog
1:25:59 is large, Marge. Okay. Yes, I'm fat
1:26:04 shaming my dog. Um, I grew up in a time
1:26:08 where that was allowed. Okay. All
1:26:10 [Laughter]
1:26:12 right. You people you people with your
1:26:17 your treating people with dignity. I
1:26:20 don't know how to do that. Okay.
1:26:29 All right. Scene layout's going to be
1:26:31 that. Let's go grab another picture.
1:26:34 We're going to go back over here. We're
1:26:36 going to go
1:26:38 grab and we'll grab this
1:26:45 [Music]
1:26:50 one clothes. Let's get rid of Tom and
1:26:54 Brad. And then maybe we'll go find some
1:26:57 cooler
1:27:03 clothes. All right. Who should we go?
1:27:05 Who should we put in
1:27:07 this? Let's put Ann Murphy. Hang
1:27:11 on. If since we're since Since we're
1:27:15 already bismerching her reputation with
1:27:19 AI, let's keep that tradition up. She's
1:27:22 on vacation, so she can't defend herself
1:27:25 unless they have Wi-Fi at the fancy
1:27:27 resort she's going to. Okay. Um, where
1:27:31 did I have her picture last? I had it
1:27:33 Oh, I had it in in uh in runway, didn't
1:27:36 I? We can go get it
1:27:41 there. Copy image. All right. Paste. Oh,
1:27:46 it's too small. That's all right. We can
1:27:48 scale that [ __ ] up.
1:27:55 All right, Tom. No. All right, Ann
1:27:59 Murphy. Boom. Done. We could put her
1:28:02 with Brad Pit, but that seems that seems
1:28:05 too It's not appropriate. It's just not
1:28:08 appropriate. I think who she needs to be
1:28:11 with. Let's see.
1:28:19 Let's see who would be good. She's
1:28:24 tall. Any
1:28:27 ideas? She's smoking right now. William
1:28:31 Defoe. Okay, that's good. Silver Fox.
1:28:33 Silver Fox is always there with the
1:28:35 brilliant William
1:28:40 Defoe images. And we got to get we got
1:28:43 to get the right William Defoe. It can't
1:28:45 be the when he was handsome.
1:28:48 No, no, that's not the
1:29:01 uh we won't we won't make him that ugly,
1:29:03 but let's see which
1:29:05 which William Defoe is the question.
1:29:10 It's got to be one where he was out on a
1:29:12 bender. Let's see. William
1:29:15 Defoe.
1:29:18 Paparazzi. Let's get him when he looks
1:29:20 like [ __ ] Come
1:29:29 on. Oh, that's is that William Defo?
1:29:32 That looks like Dennis Liry or Dennis.
1:29:35 Yeah, Larry. William
1:29:39 Defoe. All right. Bearded William Defoe.
1:29:41 This is good.
1:29:46 Right. Oh, for [ __ ] sake. Everything's
1:29:49 a [ __ ]
1:29:55 salesfunnel. There we go. All right. So,
1:29:58 here's William
1:30:00 Defoe. This is the one we're going to
1:30:02 use. This is This is who Ann's going to
1:30:05 be in her little scene with.
1:30:13 Hey, if any of you tell her that I'm
1:30:14 doing this, that would be inappropriate
1:30:16 and rude. Would be rude of you to do
1:30:21 [Music]
1:30:22 that. All right. Now, we go back over to
1:30:25 photo
1:30:26 slop. Paste that bad boy in there. Let's
1:30:30 chop off part of
1:30:38 that. All right. So, we cover up
1:30:43 All right. So, now we're going to change
1:30:46 some Wait, I got to go to this layer. I
1:30:50 got to pull this [ __ ] out. Kill
1:30:52 it. Oh my god, this is exhausting. You
1:30:55 see why I didn't want to do this?
1:30:56 Everything's a [ __ ]
1:31:01 production.
1:31:03 And And then this is We're going to call
1:31:06 him
1:31:08 uh Billy.
1:31:13 Okay. So, we got Ann and Billy. Now,
1:31:17 let's So, what else do we need? We need
1:31:19 their clothes. And then, what was the
1:31:20 other thing? We had the
1:31:23 layout. All right. So, we're going to
1:31:25 make Ann sad and Billy
1:31:29 happy. Okay. So, this is going to be
1:31:32 Ann
1:31:34 and this is going to be Billy.
1:31:40 All right. So, we got that. That's good.
1:31:43 Now, we need clothing. All right. Back
1:31:46 to the
1:31:47 Google. Um, let's see. Uh,
1:31:52 co
1:31:55 dress. Let's get something fundamentally
1:31:59 ridiculous for an
1:32:02 um Oh, well, let's see.
1:32:14 There we
1:32:15 go. Is that good? Yeah, that's
1:32:29 good. All right. There's There's Ann. We
1:32:33 got to resize that bad boy.
1:32:36 Is Is anyone still
1:32:40 here? If you're wondering what this
1:32:43 channel's all
1:32:46 about and why anyone watches it, I don't
1:32:50 know either.
1:32:52 Okay. Um, and then what are we going to
1:32:55 put Billy in? I think Billy needs to go
1:32:57 in like a gas station. Uh, let's see.
1:33:00 Um, gas
1:33:02 station.
1:33:04 Uh, jumper
1:33:09 uniform. Yeah. Here we go. That's good
1:33:11 right there. Right. Yeah. That looks
1:33:14 like Billy Defoe.
1:33:16 [Laughter]
1:33:27 Uh, if one of the questions you're
1:33:30 asking yourself is, is this channel
1:33:33 really just an opportunity for Kyle to
1:33:35 entertain himself? Yes, that is exactly
1:33:39 what this channel is.
1:33:42 Okay, so now we need some letters.
1:33:46 So this is
1:33:49 Billy
1:33:51 and this is Ann.
1:33:55 Okay. And then did we need one other
1:33:57 thing? I think we might have needed one
1:33:59 other
1:34:01 thing. Oh no, that's just the output.
1:34:04 Okay. So we've got our three things,
1:34:05 right? Okay. Beautiful. Back to
1:34:07 Photoshop. So we're going to do export
1:34:09 as
1:34:11 PNG. So this is clothes.
1:34:15 to the desktop clothes. Um, if you're
1:34:19 new here, um, pro file management tip.
1:34:23 Save all of your files loose to the
1:34:26 desktop with completely generic names
1:34:28 you'll never be able to find again.
1:34:30 You're welcome. All right. Export quick
1:34:34 PNG. This is
1:34:38 characters. All right. Fantastic.
1:34:42 And then uh and then this is scene
1:34:49 layout
1:34:51 export quick export is so let's call
1:34:54 this scene. Okay, back to
1:35:02 runway. Anybody have any questions? Have
1:35:04 I lost anyone? Is anyone still here?
1:35:08 This is why the irregulars show up night
1:35:09 after night to defend our
1:35:13 reputations. Yeah, exactly. You go on
1:35:16 vacation as an irregular, this shit's
1:35:18 going to happen to you. So, yeah, get
1:35:22 over it. All right, here we go. So,
1:35:24 we're going to
1:35:27 upload
1:35:29 desktop scene character clothes.
1:35:37 Then we're gonna
1:35:39 say
1:35:41 an
1:35:43 and
1:35:45 Billy sit
1:35:48 in a
1:35:53 horserawn carriage in
1:35:58 Central Park, New York City.
1:36:06 Uh, I guess I'll just say
1:36:14 follow the
1:36:17 image
1:36:21 guidance, I
1:36:24 guess. All right, we're going. We're
1:36:27 off. Generative AI. Yes, that's what
1:36:31 we're doing here. So, this channel is
1:36:33 primarily about generative AI, not about
1:36:36 weights and embeddings and
1:36:39 math. Not the guy. Um, there are people
1:36:43 in here who are really good at that. I'm
1:36:45 not the guy though. I talk about
1:36:47 generative AI. How are we using these
1:36:48 tools? Why did that not generate? What's
1:36:51 going on? Oh, generating. Here we go.
1:36:53 Oh, okay. We're gonna have two
1:36:55 sets. And B, what's happening? Good to
1:36:58 see you. Welcome. Welcome. Welcome.
1:37:06 [Music]
1:37:10 All right, these are almost
1:37:12 done. Let's see how lovely they
1:37:26 are.
1:37:28 Damn. Come on.
1:37:31 Come on, runway. You can do
1:37:45 it. Oh my god. It's too perfect. It's
1:37:50 too perfect. Is the processing time
1:37:52 based on their server or yours? It's all
1:37:54 theirs. I'm not doing anything locally.
1:37:56 So, so on this channel, I tend to just
1:38:00 ignore open source. Not because I don't
1:38:02 love it and think it's important. I just
1:38:04 don't have the attention span for it,
1:38:06 the attention to detail for it, the hard
1:38:09 drive space for it, the computer, the
1:38:12 computational resources for it. Um, and
1:38:16 I also just feel like, you know, about
1:38:18 the time I'll get something up and
1:38:19 running, it'll just have moved on to the
1:38:21 next technology. So I just basically
1:38:24 what I do is I focus on the commercial
1:38:26 commercial big players like who are the
1:38:28 ones that are the the least waste of
1:38:31 anyone's time and that's what I focus
1:38:33 on. Um the fact that we have this level
1:38:36 of of coherence right now is is actually
1:38:40 kind of mind-blowing to me.
1:38:43 Um, look at Ann. Oh,
1:38:49 sad.
1:38:51 Frozen. Um, why are they sitting so far
1:38:54 apart? Cuz she wore her dress and he
1:38:56 came out in a
1:39:05 [ __ ] Look, he looks giant. She looks
1:39:08 teeny.
1:39:13 Oh my god, this is good. This is so
1:39:16 good. Okay, there's one where she's
1:39:17 really
1:39:19 pouting. Let's see
1:39:24 that it's sitting really far
1:39:27 apart. Although the carriage looks
1:39:29 [ __ ] up there. Let's see there. She's
1:39:31 pretty
1:39:33 pouty. All
1:39:37 right. Okay. So, now we're gonna make a
1:39:42 movie. Hey, Joy, if Joy, have you How
1:39:46 long have you been here? Have you seen
1:39:47 this? Oh, I'm buffering. Oh, that sucks.
1:39:50 Leave her alone. Is this generated
1:39:52 locally? No, this is generated uh this
1:39:55 is
1:39:55 RunwayML. So,
1:39:58 app.runwayml.com. They have this new
1:40:00 thing called references, and you can put
1:40:02 references in that do things like put
1:40:05 her on the right, him on the left, her
1:40:07 sad, him
1:40:08 happy. Um, just got here. Okay. So, so
1:40:12 here here's what we
1:40:14 did. So, I went I went on Twitter and
1:40:17 someone someone had this with Tom Cruz
1:40:19 and Brad Pitt. Um, and
1:40:23 so here's the scene layout, right? So,
1:40:25 the person on the left is happy, the
1:40:27 person on the right is sad. Um, it's
1:40:30 Billy on the left and on the right. So,
1:40:31 I did that. Then I said, you know,
1:40:35 here's what the people are wearing,
1:40:37 right? Billy's wearing the mobile gas
1:40:40 shirt and she's wearing this big ass
1:40:42 fancy dress. And then, you know, it's
1:40:45 William Defoe and Anne Murphy are the
1:40:48 sources. All right.
1:40:51 And then we
1:40:54 went, it just [ __ ] did it. So Joy,
1:40:58 for you in particular, the character
1:41:01 character coherence here is
1:41:04 insane. Uh oh my god, this is hilarious
1:41:07 and awesome, isn't it? Well, now we're
1:41:08 going to make it into a
1:41:10 movie. Let's see. So we're gonna go All
1:41:14 right.
1:41:17 Um camera tracks
1:41:22 with carriage as
1:41:28 it rolls
1:41:37 through Central Park.
1:41:43 The man is
1:41:46 laughing and the woman is
1:41:50 [Laughter]
1:42:01 pouting. So naughty and
1:42:05 fun. Dragonfly alchemy. What does Ann
1:42:08 have to say about it? She's probably
1:42:10 crying, laughing. She's on vacation. So,
1:42:14 so what I what I inadvertently did, so
1:42:17 when I first started testing
1:42:20 um this this reference thing, um I did
1:42:24 it first
1:42:26 in Midjourney. And in midJourney, it
1:42:30 kind of sucks. It's not really good. And
1:42:33 so then I came to runway and I just I
1:42:36 just had used I had a picture of Ann
1:42:38 from when we were setting up the
1:42:39 podcast. And so I just grabbed it and
1:42:41 threw it into into
1:42:44 um into runway with two two photos I've
1:42:48 taken. One of a photo of clouds that I
1:42:50 took and one of this motel. And so I
1:42:52 put, you know, the clouds, the motel and
1:42:54 her in the picture. And if [ __ ] came
1:42:57 up with that and then we turned it into
1:42:59 a video and it made this. Wait, hang
1:43:03 on. This is [ __ ] hilarious.
1:43:06 [Laughter]
1:43:17 Now, that might seem random, right? And
1:43:20 it might seem like glitchy, like no one
1:43:22 can hold that much smoke in their lungs,
1:43:24 but it's like, you know, have you ever
1:43:26 partied with her? You don't know. Maybe
1:43:28 she maybe she's she's a toker. Um, all
1:43:32 that white smoke. The next
1:43:35 day we had a new
1:43:37 pope. I'm not saying Anne had anything
1:43:40 to do with it. I'm just saying she blew
1:43:44 white smoke from her cigarette, got a
1:43:47 new pope the next day. All right, new
1:43:53 pope. All right, let's see if we have a
1:43:55 video yet. 61%. We're almost there. It's
1:43:58 going to be so good. Did Lovable make
1:44:00 our game yet? our game
1:44:04 thing.
1:44:07 Nope. Oh,
1:44:09 wait. No. Ah, damn
1:44:13 it. Got nothing,
1:44:18 [Music]
1:44:21 people. Why are they sitting so far
1:44:24 apart?
1:44:29 Oh, lordy, lordy,
1:44:37 lordy. I think I'm going to text her
1:44:39 [Laughter]
1:44:45 this. Oh, it's so
1:44:48 good. I'm a bad man.
1:45:04 Uh oh, I'm not charging my phone and my
1:45:06 phone's about to
1:45:08 die. That would be
1:45:11 awkward. He didn't even say
1:45:19 goodbye. All right, here we go. Here we
1:45:23 go, people. Do we have Do we have comedy
1:45:27 gold AI
1:45:29 weirdness? What is
1:45:36 it? It's perfect. It's perfect.
1:45:56 Oh my god. Oh my god. Okay, we got to
1:46:01 download
1:46:08 this. Oh god. Oh, it's so good. It's so
1:46:13 good. All right. Can
1:46:16 we Does it allow Oh, restyle. That's
1:46:19 interesting. Use
1:46:22 frame. Oh,
1:46:25 interesting. All right, let's go here.
1:46:29 We'll go here. Oh, yeah. That's Oh, you
1:46:32 can't see any of that. Oh, could you not
1:46:34 see any of that? Hang on. Billy has a
1:46:37 comment. Yeah, I know. Billy said
1:46:39 something at the end there, didn't he?
1:46:41 Um, hang on a sec. Let's
1:46:44 see. Maybe Ann. Okay.
1:46:57 Okay, this is good. I'm just what you
1:47:00 can't see right now is I
1:47:03 am
1:47:05 texting our
1:47:06 [Laughter]
1:47:16 friend here and I'll say I'll put the
1:47:18 message.
1:47:19 Oops, it happened again.
1:47:25 I didn't know you and William were so
1:47:35 close. Ah, that's comedy right there,
1:47:39 people. I am telling you that is comedy
1:47:43 gold. Comedy go. I know. Oh my god. So
1:47:46 funny. We still can't see. She may not
1:47:49 be your friend for long if you keep this
1:47:50 up.
1:47:58 It's so good. It's so
1:48:01 good. Okay, so listen. I said this
1:48:04 before. We'll just keep watching this
1:48:07 because it's
1:48:08 awesome.
1:48:10 Um, I said this before when I was
1:48:13 talking about the AI salon. If you're
1:48:14 new here, I said play first. Remember
1:48:17 when I said play first? No, I don't
1:48:19 remember that. When did you say play
1:48:21 first? I wasn't here. So, how could I
1:48:22 have heard that if I wasn't here?
1:48:26 Um, as stupid as this was, as silly as
1:48:31 this
1:48:32 was, we learned something really
1:48:34 important tonight. Like, Joy Perie is a
1:48:36 full-on
1:48:38 filmmaker
1:48:40 and now knows that this is possible.
1:48:43 Like, this is a level of control that we
1:48:46 haven't had an AI before. And it's kind
1:48:49 of crazy, right?
1:48:51 Um, I wouldn't have gotten to this if I
1:48:55 wasn't just [ __ ] around. Like, I
1:48:57 wouldn't I wouldn't know how to do this
1:49:00 if it was any good, right? You know,
1:49:03 there's something about having someone
1:49:05 that you know, like I I know enough what
1:49:07 she looks like. She looks She at least
1:49:09 starts out looking like her. She kind of
1:49:10 morphs into someone else by the end.
1:49:15 But that's bonkers.
1:49:20 This is the stage we're
1:49:25 at. I could
1:49:30 probably you could you could probably
1:49:32 put this into another
1:49:38 um lip syncing tool and give him
1:49:43 dialogue. I've had students wanting
1:49:46 consistent characters all year. There
1:49:47 you go. Well, they exist. It It exists
1:49:51 now. The guy following them has has some
1:49:54 rhythm. Oh, yeah. He does, doesn't
1:49:57 he? I think he he looks like he
1:50:00 recognizes William Defo. He's paparazzi.
1:50:02 This guy back here. That's
1:50:04 [Laughter]
1:50:06 great. Billy cheers her up at the end.
1:50:09 Yeah, she does. She gets a little
1:50:10 happier at the end. She can't help him.
1:50:12 He's
1:50:13 lovable. I'm going to do this, Kyle.
1:50:16 Yeah. Like like this is crazy. This is
1:50:19 crazy. Zoom in on the guy following. Can
1:50:22 I zoom in? I don't think I can. Hang on.
1:50:24 Maybe I can
1:50:26 here.
1:50:31 Um, hang on.
1:50:37 There's the guy
1:50:44 following
1:50:45 upscale enhance.
1:50:49 Exactly. Oh, man. All right, good
1:50:52 people. Uh, it is
1:50:55 late. Oh, 9:49. Now, we're we're we're
1:50:58 sort of right on time. You could totally
1:51:00 out add a soundtrack. Yeah, you could do
1:51:02 soundtrack, you can do music, you
1:51:05 can put dialogue in
1:51:07 there. Again, again, people, people,
1:51:13 people. If you
1:51:17 think that any
1:51:21 job is safe from
1:51:24 transformation, you're
1:51:26 wrong. Like, this stuff has gotten so
1:51:29 good in the past six months. like all of
1:51:31 it, large language models, coding tools,
1:51:35 video tools, image tools, like the image
1:51:37 coherence within Chat GPT at this point
1:51:39 is
1:51:40 insane. But this is image generation
1:51:43 within Runway ML, which is a video
1:51:45 model, but this image coherence, we've
1:51:47 never had this before where you could
1:51:49 put in
1:51:51 um like layouts of how you want to
1:51:54 arrange
1:51:55 things. It's just bonkers.
1:52:04 just
1:52:11 bonkers.
1:52:14 Anyway, all
1:52:16 right. Maybe that's Billy's
1:52:19 agent.
1:52:22 Billy, I prefer William. Yeah, whatever.
1:52:25 Billy, listen.
1:52:28 What did you What did you do to piss her
1:52:30 off so
1:52:31 [Laughter]
1:52:34 much? All right, Kak's off to bed. Good
1:52:37 night. I'm gonna get off to bed, too.
1:52:39 Um, tomorrow's Tuesday. I don't think we
1:52:41 have anything tomorrow, do we? Tomorrow
1:52:44 night? No.
1:52:47 Um, so we'll be here tomorrow, same
1:52:51 time, same bat time, 8:00 PM.
1:52:56 Um, if you're on YouTube, like and
1:52:59 subscribe. We're trying to build our
1:53:00 user base over there, our subscription
1:53:02 base over there in case Tik Tok flies
1:53:05 the coupe, which it is, and it isn't,
1:53:08 and then it is, and then it isn't. So,
1:53:10 um, go subscribe over there so you can
1:53:13 not miss You wouldn't want to miss a
1:53:15 minute of one of these, would you? No.
1:53:17 All right, cool, Groovy. All right,
1:53:19 peace out, everyone. Have a good Monday
1:53:21 night and I will see you