AI Learning Lab

5/7/2025 - AI Learning Lab Celebrates a Birthday With Tequila and Tech

vXW1hBSaYMY
Live Stream2025-05-081:25:4084 views

Description

It's my day. Come hang and let's chat AI Join Kyle on his 60th birthday as he explores the wild world of AI video and image generation, fueled by tequila aged in a Blanton's barrel. He dives into LTX Studio, a modular movie-making tool, generating a short film from the logline "Swipe Right for the Apocalypse." While impressed with LTX's story assembly, Kyle finds its image generation lacking, leading to humorous inconsistencies and anatomical anomalies. He highlights the importance of structured workflows for LTX, comparing it to assembling a complex Lego set. The AI adventure continues with Midjourney and RunwayML. Kyle demonstrates Midjourney's image prompting, style referencing, and omni referencing, showcasing how to incorporate specific objects and aesthetics into AI-generated images. He then shifts to RunwayML, praising its superior image compositing capabilities and reference tagging system. A comedic demonstration ensues, combining disparate images into a surprisingly realistic, yet deeply unsettling, video. Kyle concludes with a serious message about the ethical implications of deepfakes and the importance of responsible AI usage, encouraging viewers to explore and learn AI tools. Learn more about AI on TikTok: https://tiktok.com/@aiLearningLab. #AI #Deepfakes #Midjourney #RunwayML #LTXStudio #VideoGeneration #ImageGeneration #ArtificialIntelligence Chapters: 00:00:00 Birthday Wishes 00:02:18 Downsides Of Going Live 00:03:38 Blanton's Tequila 00:04:41 Tequila Tasting 00:07:33 Losing His Dog 00:09:25 Career Chapters 00:11:01 Producer Brandon Joins 00:13:22 Ai Learning Lab Highlights 00:14:05 Ai Learning Lab Intro 00:18:20 Ai Tools Are Janky 00:20:15 Ai-Generated Birthday Greetings 00:20:22 Playing With LTX 00:22:16 Atomic Level Assembly 00:25:11 More Tequila? 00:25:50 Exploring LTX 00:27:22 LTX Explained 00:29:25 LTX Projects 00:32:18 Blockbuster Log Lines 00:34:09 Back To LTX 00:36:52 LTX Settings And Cast 00:43:46 LTX Scene Breakdown 00:46:07 LTX Timeline 00:48:01 LTX Sound Effects 00:51:44 LTX Conclusion 00:52:56 LTX Recap 00:55:36 Midjourney Intro 00:58:29 Midjourney Omni Reference 01:01:49 Midjourney With People 01:07:12 Runway ML Intro 01:09:14 Runway ML Assets 01:11:29 Runway ML Image Generation 01:13:19 Runway ML Video Generation 01:16:02 Runway ML Overview 01:19:26 Deepfakes Are Dangerous 01:21:27 Final Thoughts 01:23:11 Ai And The Future Of Work

Chapters

Transcript

0:00 Let's
0:02 see. Let's see. There's that. Start a
0:06 new project from
0:09 scratch. Start with content. Start with
0:12 a concept. Start from
0:15 scratch. Scratch.
0:23 [Music]
0:32 [Music]
0:44 Freedom came my way that
0:50 night just like a jet plane even
0:56 side. I was hauling ass at a million
1:00 miles an
1:01 hour. Wonder how hard I
1:08 hit. When they came
1:11 into
1:14 town, they said I was bad beyond
1:20 repair. But I got no qualms with my
1:26 situation here
1:30 [Music]
1:35 right
1:38 say won't you
1:41 dare say Sheree Sheree Sheree won't you
1:45 dare
1:47 to say Sheree Sheree
1:52 Sheree happy
1:54 birthday.
1:56 Yeah. Take a
2:00 [Music]
2:02 time. Take a time to want to satisfy
2:08 me. Take all these old fantasies and
2:12 send them care of
2:15 [Music]
2:18 me. Oh, good day everyone. Thank you for
2:21 the well wishes. One of the
2:23 downsides of going live five nights a
2:26 week for multiple years in a row when
2:29 you got irregulars involved is that
2:31 they'll remember
2:33 [ __ ] Like
2:35 say when you [ __ ] turn
2:43 60. Thank you TBlock Tom. Thank you
2:46 everybody. Dr. J, all the happy wishes.
2:50 Appreciate it. Appreciate it.
2:56 No, exactly. Spin. I know. I
3:02 know. I can no longer
3:06 pretend, you know, I'm just this side of
3:11 Young. I can't do it anymore.
3:14 But the other thing that's nice about
3:16 having a group like the Irregulars is,
3:18 you
3:20 know, and look, oh wait
3:24 here, hang on. Hang on, people. Just
3:27 calm down. Just calm
3:28 down. All the aggressiveness. It's a
3:31 very aggressive
3:32 bunch. Oh no. What did I do? There we
3:36 go.
3:39 Look. Aged in a Blandon's barrel. single
3:43 barrel
3:44 tequila. Now, I'm not a massive tequila
3:47 fan, but it'll be really interesting to
3:50 see if Bllandon's is the salvation. It
3:54 looks like a beautiful, beautiful color.
3:58 So, we'll You know what? Let's just do
4:00 it now. Let's Let's Sorry, Tik Tok.
4:04 Yeah, I'm embibiving. You know what else
4:05 I like?
4:08 Gambling. Oh, yeah.
4:11 All
4:14 right. See? See? Tik Tok. What are you
4:17 going to do? You [ __ ] shut me down.
4:19 You banned me for life. You going to do
4:21 it? Do it. Do
4:24 it. All
4:26 right. Pinkies
4:28 out. Why do we have pinkies
4:31 out? Cuz we're classy.
4:41 Wow. I can honestly say, and maybe this
4:45 isn't tequila, but it No, it is tequila
4:49 corazone de argave single barrel
4:54 anjo exclusive single estate tequila
4:57 aged. I went over to So, in my in my
5:00 bourbon in my bourbon drinking
5:03 club, we went over to this guy's house
5:05 once and he wasn't a bourbon guy. He was
5:07 a tequila guy and he's like, "I'm going
5:09 to convince you that you like tequila."
5:10 And I'm like, "I don't but I don't think
5:12 you can." No. And so, he gave me all
5:16 these different tequilas. And every one
5:18 of them I was
5:22 like there was one that was like a $300
5:25 bottle and it was semi- tolerable but it
5:27 was still a little like you know
5:29 remembering my 21st birthday where my
5:31 friends got me drunk on tequila and I
5:32 can't ever drink it again. But this is
5:35 actually [ __ ]
5:36 good, huh?
5:43 So, you're telling me that Bllandon's is
5:46 so
5:47 good that it makes something that I
5:50 normally find undrinkable actually
5:52 pleasant? I It's unbelievable. I'm a
5:56 tequila guy all the way. Ah, Gareth, I'm
5:59 so sorry. And And listen, very sweet.
6:03 The shot glasses are very sweet. I'm not
6:06 doing [ __ ] shots. I'm a sipper. I'm a
6:09 sipper. I just like sipping, you know. I
6:11 put them lots of rocks. Lots of rocks. I
6:14 like to get little like icicle on the
6:16 outside. Lots of rocks. And then I just
6:18 sip. I just sip away. Let me just sip
6:21 it. This is really
6:25 good. I am blown away. I am blown
6:29 away. I was expecting to give you all a
6:32 bunch of
6:33 [ __ ] Huh? All right. Consider myself
6:37 stunned.
6:38 It's the It's the 31st anniversary of
6:41 your 29th birthday. He's a Sin City
6:44 sipper. I am a Sin City sipper. I just
6:47 zip it up like I'm in Sin City. And you
6:50 know what we do in Sin City, right? No,
6:52 you don't because we leave it
6:54 there. It's f It's just tasty. Oh, it's
7:00 tasty. How I'm behaving right now, by
7:03 the way, is why people on LinkedIn ask
7:06 me if I'm doing all right.
7:16 Yeah, exactly. Vicki understands Sin
7:18 City sipper, not Sin City shooter.
7:23 Shotter. Shooter.
7:27 [Music]
7:33 Shotter. Oh, I lost my dog.
7:37 You know it's bad when your dog's like,
7:39 "Yeah, this is too much for me. I'm just
7:41 going to
7:42 [Laughter]
7:45 [Music]
7:55 go. Is this a place I can rest my
8:01 head?"
8:03 Um, all right. Well, listen. No need to
8:06 stand on ceremony. We'll just do this.
8:08 Happy birthday. Thank you, Silver Fox.
8:10 Very sweet of
8:11 [Music]
8:14 you. Dogs is like, "Yeah, I can see
8:16 where this night is going."
8:18 Exactly. I'm just going to go lick my
8:21 paws and pass
8:22 [Laughter]
8:28 out. God damn, that's
8:30 good. I
8:35 It just doesn't make
8:37 sense. It just doesn't make
8:40 sense. It's got like a Bllandon
8:43 smoothness to
8:45 it. Single
8:47 barrel.
8:49 Huh. All right. All right, people. I got
8:53 nothing for you. I would I would have
8:55 bet money. I would have bet money that
8:58 that wouldn't have been palatable.
9:05 [Music]
9:11 in a westerly
9:14 [Music]
9:17 direction. Car is my
9:20 [Music]
9:21 train. Been driving. I've been
9:25 wondering. Happiest of birthdays. Hope
9:28 this is the beginning of the best
9:29 chapter of your life. Thank you, Source
9:31 Camp. I would I would love for that to
9:34 be the case. I think we got a good shot.
9:37 I was talking to someone today. They're
9:40 like, maybe this is your third and final
9:43 chapter of of of like your
9:46 career. And I was like, I resented the
9:49 final part. Why Why would this be my
9:51 third and final chapter? Why not like my
9:54 third of
9:56 seven? That's that's just it's just it's
9:58 just you
10:00 know it's as they say in the program
10:04 that's stinking
10:08 thinking you know I'm going
10:11 [Music]
10:13 for we just we just got to hang out for
10:16 five more years and we hit lev life
10:19 escape velocity. You know that one,
10:21 right? Where for every year you live,
10:24 they they
10:25 advance survival by more than a
10:30 year. Five
10:32 years. But knowing our health system, 10
10:35 years. So got to survive 10. If you're
10:39 loaded, five. If you're a human being,
10:44 [Music]
10:49 10 60 years of life, love, and countless
10:53 memories, actually. Yeah. And 1100 hours
10:55 of those memories are on YouTube.
11:01 Hey, Kyle. Wednesday. Oh, producer
11:04 Brandon. Yes, sir. Hello, sir. First of
11:06 all, I don't think you have mirror open.
11:08 Issue.
11:09 What's that? Sorry, I said I don't think
11:11 you have Miro open. Oh, I don't. No. Oh,
11:15 I do. But it was in a I'm post I'm
11:18 posting to no one. Yes. What What were
11:21 you saying? I I was asking if Champy got
11:24 his cheese. There should have been a
11:26 little cheese. Oh, there was There was a
11:27 lovely cheese in there. Champy's not
11:29 getting the nice cheese. That's the
11:31 people are going to eat that. I didn't
11:33 realize the cheese was for Champy. No,
11:35 no, the people are going to eat that.
11:37 Well, happy birthday from the
11:39 irregulars. Oh, that's very nice. Yes,
11:41 Champy got his cheese. Thank you for the
11:43 I'm glad I'm glad the selection doesn't
11:45 taste like crap.
11:47 No, it's it's like dude, it is so good.
11:51 I figure if the Bllandon's name's on it,
11:53 it's got to have, you know, some
11:56 credence. Yeah, but normally that's just
11:57 marketing [ __ ] you know, and No,
12:01 no, it's actually really, really good.
12:03 So, so you mentioned you have um 1100
12:07 odd hours of
12:10 uh of uh content on yield YouTube. Yes.
12:15 Yes. And uh Lord Digital Gods has done
12:19 an incredible job building out some
12:22 content
12:24 uh around that and capturing that and
12:27 archiving that in a way that makes it
12:29 very searchable. Really, he does. And
12:33 so Lord Digital God's absolutely amazing
12:36 because I because the way these these
12:39 streams work now is they're just saved
12:41 to YouTube, right? Like I don't have to
12:44 manually do them. And when I was
12:45 manually doing them, you know, I would
12:47 go what what was it six nine months
12:49 without updating a title. And then I
12:52 just I just noticed that every day the
12:54 titles are updated. So first of all,
12:56 thank you to Lord Digital Gods for just
12:58 regularly doing that. That kind of
13:00 consistency is not something I have in
13:02 my life. But I do. So I I reached out to
13:06 him and I asked him if he how hard it
13:09 would be to put together some
13:12 clips, a highlight reel if you will of
13:16 the AI learning lab from the past couple
13:18 of years. Oh boy. And uh
13:22 so you know I'm the producer so I could
13:26 just kind of take over the show. So,
13:27 we're we're gonna make a 80 add left
13:29 turn here. All right. For the next few
13:32 minutes. Okay.
13:34 I wanted to make sure the alcohol got
13:36 there before you we got this to you. So,
13:38 yeah. I'm already a little
13:39 uncomfortable. So, this is this is
13:41 taking it to a no new level. So, thank
13:43 you both or all of you, whoever was
13:45 involved with this. This sounds awful,
13:47 but let's go. And hopefully the uh
13:50 hopefully the sound will come through.
13:52 Um and uh enjoy. Happy birthday. It's
13:55 not a Friday night unless you're
13:57 inducing vomit. You know what I
13:59 [Laughter]
14:04 [Music]
14:06 mean? Embrace the jagging. Embrace the
14:08 jag. People, people come here, they're
14:10 like, I come for the hair. I stay for
14:13 the madness. Some people, quite the
14:15 opposite. I I come for the madness and
14:18 the hair I just tolerate. A masterful
14:21 collaboration of four different
14:24 toupes, cleverly stitched together to
14:27 seem
14:28 like natural hair with an extremely
14:32 large forehead. The real singer in the
14:34 room's name is Champ. Full
14:38 disclosure, he's never been a champion
14:40 of anything. We took him for a walk and
14:43 he was sweet and every like, I want him.
14:44 We want him. I'm like, I don't think so.
14:47 But no.
14:51 No. So, we kept him because, you know,
14:54 I'm just the husband. Oh, yeah. We do AI
14:57 stuff here,
14:59 too. We went from 102 down to 71. People
15:03 are like, "Uh, what? Who is what? Why is
15:06 he talking about his dog like this?
15:08 Seems very
15:10 aggressive. Haven't felt pleasure until
15:13 you've huffed some
15:15 neon." I am not an IT scientist.
15:19 an ex actor. The hell do you want from
15:21 me? I've got video proof of it. A wiener
15:24 popped up on screen. I was just
15:26 acknowledging the wiener that was in my
15:29 face. My other channel's a twerking
15:31 channel. If this were about money, I'd
15:33 be in a manini selling Madagascar
15:36 vanilla beans, you know. Fantastic.
15:39 Fantastic.
15:42 Thinking you in a manini, for the love
15:46 of God, please say no. Got to give the
15:48 people what they want. I can answer
15:50 anything.
15:52 We like to call it mansplaining as a
15:54 service. He's a little
15:56 moist. Little Did he put on foundation
15:59 tonight? He looks blotchy and he looks
16:01 moist.
16:02 Couple of filters, a little bit of
16:04 makeup, a crew of seven. You too can
16:08 look like this. I have a degree in
16:09 acting. And I have no qualifications to
16:12 be running this channel. So, what are
16:13 you going to do?
16:15 Tomato sauce is not made from tomatoes.
16:18 Tomato sauce is made from cooked and
16:19 pureeed tomatoes, but it doesn't come
16:21 from the same thing.
16:24 But isn't tomato sauce made from
16:26 tomatoes? No. Again, tomato sauce is not
16:29 made from tomatoes. Listen, I live in
16:31 Denver, but it sounds like you've been
16:33 smoking that wacky weed. What the hell
16:34 are you talking about? What the hell are
16:37 you talking about? What? They're always
16:39 out to get us. Who's out to get you? Who
16:42 are? They're not our friends. Who are
16:44 not your friends? They're the enemy. The
16:46 police of the enemy. Out to get us. Oh
16:48 my. They out to harm us. Moshi.
16:52 Moshi. That sounds very aggressive.
16:56 I'm ready for the game. They're like,
16:57 "Frank, Frank, you look [ __ ]
17:00 delicious." I'm like, "Hey, hey, hey.
17:02 Back up. Hey, whoa, whoa, whoa, whoa.
17:03 What is it, fella? Get your hands off my
17:05 butt."
17:07 Urgent care. It's like 7-Eleven but with
17:10 needles. I'm looking for a man in
17:12 finance. 65 just fine.
17:17 [Music]
17:20 65. That's That's That's good. That's
17:23 good. The the giant penis nose. Am I
17:26 allowed to say penis on here? I'm going
17:28 to get cancelled for [ __ ] ever.
17:30 Perman permanently banned. You big
17:32 stupid [ __ ] Tik Tok channel. You big
17:35 fat dum dum. I I can hear you. I know
17:38 what you say about me. Anyone who tells
17:40 you they're an expert is full of [ __ ]
17:43 My role in life is to give you
17:45 permission to feel like an idiot, too.
17:48 Oh,
17:50 hello. Let's breathe heavy into the
17:52 microphone, shall we?
17:56 Hi. Hi. It's
17:59 Kyle. So, listen. AI tools are janky. AI
18:04 learning lab
18:05 professors equally janky. Probably more
18:08 janky. Maybe tomorrow we wake up to
18:12 AGI or probably
18:15 not. Peace.
18:20 Oh, that was [ __ ] awesome. Boy, do I
18:23 say some stupid [ __ ] When you put it
18:26 all together, I just seem like an idiot.
18:33 Oh, that was awesome. Well, thank you.
18:38 You're welcome. So, we'll memorialize
18:40 that in irregulars for you. And there's
18:41 also some images in there now, too, from
18:44 uh the uh the irregulars that are
18:47 playing along from the peanut gallery.
18:48 All right, let's let's go let's go
18:51 around.
18:52 If you're new here, yeah, if you're new
18:55 here, jump over to the AI salon. You go
18:58 to the salon.ai, click on join our
19:01 community, and that'll take you into
19:03 this wonderful place where you get
19:07 welcomed. M. All right. Have a great
19:09 show. Happy birthday. Wow. Thank you so
19:11 much. That was awesome. Thank you. All
19:14 right. Yeah, if you want to do uh Oh,
19:16 this Oh, this is cool. Wow. Oh, very
19:21 cool. Loading. Come on. Why is it not
19:27 loading? That's
19:29 awesome. Silver
19:33 Fox 70s muscle car birthday cake.
19:36 Fitting. Lovely. I'm 60. That's just
19:39 rude. That is just
19:44 rude. Serena, happy birthday, Kyle. 69
19:47 GTO judge. She got it. She nailed it.
19:51 Nailed it from Brandon. Very nice with
19:55 the llamas and the turtles. That's
19:58 actually a really nice nice uh likeness
20:01 drawing. Lord digital gods,
20:04 man. Frigin awesome. Well, thank you all
20:08 so much.
20:10 Um, all right, let's get going. Let's
20:13 let's let's get this show on the
20:15 road. What do we want to talk about? We
20:18 can talk
20:19 about We could go play with uh
20:23 LTX.
20:25 Um, I also was thinking that it might be
20:27 fun to play with omni reference in
20:31 midjourney
20:32 and references in runway. Um, I don't
20:37 know how to use either very well, but I
20:39 thought it might be nice to compare
20:41 them.
20:44 Um, yeah, Lord Digital Gods killed it
20:47 with that edit. My god, that was
20:49 [Music]
20:53 funny. I'm actually just joining a
20:56 speakers bureau right now and I'm like,
20:58 if I never want to get a speaking job, I
21:00 should just send them that.
21:03 [Music]
21:16 Uh uhuh. Please refresh your regulars.
21:20 Right.
21:25 [Music]
21:27 Fine. Danielle. Ah, the new baby. Kyle
21:31 baby at 60. That's so cute.
21:35 Have you seen like Tik Tok is now just
21:37 littered with there's a channel that's
21:39 just they're basically taking Joe Rogan
21:42 and Theo Vaughn clips and famous movie
21:44 clips and doing them as babies. It's
21:46 very cute. So I like it. I dig it. I
21:50 approve of me as a toddler. That's not
21:52 even a toddler, is it? As a baby. A
21:55 little baby. He's a little baby.
21:59 [Music]
22:17 If I had known, I would have baked the
22:19 cake. It's not too late. You just go
22:22 make a cake on the on AI.
22:25 I heard I heard some guy
22:27 today I think it was on Tik Tok and he
22:31 was some old dude like I don't know the
22:33 context of the clip. He was like an old
22:35 dude. And I know you're like well you're
22:38 old but he was like older than me old.
22:40 And you know how there's like there's
22:43 there's like old people but then there's
22:44 people that act old. Like he was
22:47 probably one of those people that is
22:49 about my age, maybe a little more, but
22:50 acts old, right? I act like, you know, a
22:54 sophomore in high
22:57 school and I'm this
23:00 age. He acted like a scientist. But
23:06 anyway, he was talking about when we get
23:11 um atomic level
23:13 assembly basically anyone can have
23:16 anything. He goes once we have that
23:18 anyone can have anything.
23:21 And the interviewer is asking him
23:22 like,"Well, but what do you mean like
23:24 like wouldn't just a handful of people
23:27 have those things?" And he goes, "No,
23:28 because once you have a replicator, once
23:31 you have an atomic level
23:33 assembler, the first thing that you have
23:35 it build is another atomic level
23:37 assembler. And then you have those two
23:40 make two more." And then they just keep
23:42 making, you know, assemblers. And then
23:45 you just hand them people and say, "You
23:46 can just go make anything now. You need
23:49 a steak, go make it. You need a oil
23:52 filter for your 69 judge, just go make
23:57 [Laughter]
23:59 it. He said I think he said five years
24:02 or 10 years for that. I don't buy that,
24:04 but you know, whatever. Here we are.
24:08 Like at this point, I we we are already
24:11 far beyond where I thought we'd be
24:13 science fiction wise. So,
24:18 oh man, that's why I made a baby version
24:20 of you. Very nice. Very
24:23 [Music]
24:26 nice. Oh, lordy lordy lordy lordy lordy
24:30 lordy lordy lordy lordy lordy lordy
24:33 lordy lordy lordy lordy lordy. Oh, you
24:36 know what else we could do?
24:39 [Music]
25:12 Hey, Champy. You want some tequila?
25:14 Champ, you want tequila? You going to
25:16 have a little drinky poo with daddy? No.
25:19 Oh, he's
25:24 coming. Brandon's like, "You're not
25:27 going to give your dog alcohol, are
25:28 you?" No, I said I give it to my cats. I
25:32 don't have
25:33 [Laughter]
25:36 cats. Could we 3D print the cake? We
25:39 could. Oh
25:41 man, Star Trek kind of undersold the
25:44 ship concept. Yeah, I know. Or the ship
25:46 computer,
25:48 right? Okay.
25:51 Um, another thing that I was kind of
25:53 blown away by the other day was
25:57 SNO. And we've also got lovable. Like
26:00 there's so much stuff we could play
26:01 with. Why don't you tell me what you
26:03 what you have a hankering to to dig into
26:06 because I don't I don't care. Like I
26:08 feel like at this point we could play
26:11 with chat GPT. We're waiting on Grock 3
26:13 to come out which is apparently going to
26:15 change
26:20 everything. Midjourney. Midjourney
26:22 versus Runway.
26:24 Manis. All right. Dip champy's cheese
26:27 and tequila. There you go. Just relax
26:30 him a little bit. He can sing it down a
26:33 [Laughter]
26:35 key. What is the
26:37 LTX? I kind of want to see the LTX that
26:40 you suggested. All
26:43 right, let's go there. Because for one
26:46 thing, they've they've they've improved
26:48 their
26:49 um their interface. Well, I don't know
26:52 if they've improved it. They've made it
26:53 more visual. It is. It is more It is
26:57 more
26:59 prettier. Create an
27:01 actor. Here's all my here's all my
27:04 shitty LTX
27:11 [Music]
27:14 projects. What's LTX?
27:18 Okay.
27:22 So, back in the olden timey days, you
27:26 know, three years
27:29 ago, back when I was young,
27:36 um there were these programs like
27:39 um one was called
27:42 Contour. I think there was a software
27:45 companion to the Save the Cat book.
27:47 There there were these programs that you
27:49 would you would sort of give it an idea
27:51 and it would make you like a screenplay
27:54 outline. LTX is kind of like that except
27:58 that it makes the whole movie.
28:01 So, think of it as a modular
28:05 um movie making tool that starts with
28:08 the story and scenes and shots and then
28:13 you define a look and feel for the
28:15 movie. You define characters for the
28:17 movie. You define dialogue for the movie
28:20 and then it breaks those scenes into
28:23 shots and then it creates images for
28:26 those shots and then it turns those
28:27 images into videos and then it stitches
28:30 all that [ __ ]
28:31 together. So you can make full-on movies
28:34 with it.
28:35 Now like any AI, it is jankier than
28:40 [ __ ] Like it's just bad. like the what
28:43 what was cool about LTX is how they
28:46 assembled a story. What was shitty about
28:48 LTX is they used all the cheapest
28:52 freest models and image models so that
28:54 it was lots of six finger pretzel hands
28:58 and it was lots of people turning their
29:00 head and their head would turn into a
29:01 pumpkin or you know what it was a lot of
29:04 that stuff.
29:05 In fact, we could probably go look at
29:07 some things I've made and
29:10 see see how bad it is.
29:14 Um, I think I'm going to need to share
29:16 differently. Hold,
29:20 [Music]
29:25 please. Start with a concept of a plan.
29:28 Yeah,
29:29 exactly. Winston, hi people. Missing
29:33 irregulars and Kyle and Champy. Well,
29:35 you're here now. Damn it. I like it. I
29:39 like it a lot. I like it. I like it. I
29:42 like it. Let's see which one of these
29:45 did we do. View
29:48 more. Um, I don't remember a lot of
29:53 these. Office
29:55 shenanigans. Yeah, this was one. See, I
29:58 didn't make a lot of these into movies.
29:59 These are just the images. But it's
30:01 like, so these are like all the shots
30:02 that it made. And you can turn any one
30:04 of these shots into a movie. And then,
30:08 you know, in each one of those shots, it
30:10 has things like Mia Johnson, right? So,
30:12 here's a character that you define. You
30:15 have a
30:16 cast,
30:18 right? I guess it's rendering these
30:21 rather than caching
30:23 them. They need a Gen Xer in the corner
30:26 office.
30:29 So you have all these characters and
30:31 then the scenes it makes it puts the
30:32 characters in and then they're supposed
30:34 to have character consistency anyway. So
30:36 that's kind of it. It's in a it's a it's
30:40 a an
30:42 assemblage of components of a story that
30:46 the promise is you push a button and out
30:49 comes a movie. The reality is it's
30:52 jankier than [ __ ] and you really got to
30:53 work hard to get something good out of
30:54 it. And with ADD I can't ever finish a
30:57 project in it. So, if you have ADHD,
30:59 it'll seem good on the surface and about
31:01 a third of the way through your project,
31:03 you'll be like, I'm [ __ ] exhausted.
31:04 I'm going to take a nap. Um, it could be
31:08 the tequila. Oh, it could be. All
31:11 right.
31:13 Now, they have added V2 to this. Um, V2
31:18 is Gemini's new video model. That's
31:21 really, really good. But they have
31:25 also added their
31:28 own, they don't have it explicitly
31:30 called out here. They've added their own
31:34 homegrown video model that's supposed to
31:36 be 30 times faster and better and do
31:38 consistency better.
31:42 So, do we want to start with a concept
31:44 or do we want to do a blank project? I
31:49 think turn any idea or script into a
31:52 vivid project. Develop a concept with
31:55 AI. Stick to the script. Visualize your
31:58 idea or script as written. Okay, let's
32:01 do
32:05 that. No, [ __ ] it. Let's see. Let's see
32:07 how good the AI is. Examples. All right.
32:10 So So I'm just going to run over and
32:12 grab 10 log lines. Nah, 20 log lines.
32:16 I'm in chat
32:18 GPT. Uh, okay. So act like
32:24 a
32:28 Hollywood pitch person and
32:33 come up with
32:37 20 Blockbuster
32:41 log
32:42 lines
32:44 for a
32:46 threeinut
32:48 movie. Um, some funny, some
32:55 sad, all
32:58 awesome.
32:59 Okay. Can't see chat GBT. Doesn't
33:02 matter. It's fine. I'll just read them.
33:05 Last call at the end of the world. A
33:07 grizzled bartender serves his final
33:09 drink before Earth's destruction until a
33:12 mysterious stranger orders a second
33:15 round for tomorrow.
33:19 Ah, that's intriguing. It's pretty good.
33:21 Three minutes to dad. A man races across
33:24 a crowded airport with one shot to tell
33:27 his aranged father to forgive him before
33:29 the plane takes off. That's depressing.
33:31 Shark tank wedding. Two marine
33:33 biologists tied the knot, literally
33:35 while scuba diving in a shark
33:38 tank. One thing they forgot, the shark
33:42 loves veils.
33:43 [Laughter]
33:51 Swipe right for the apocalypse. Two lo
33:54 two two lonely doomsday preppers match
33:56 on a dating app five minutes before the
33:58 missile sirens go off. That's [ __ ]
34:01 depressing. Although that one could be
34:02 cool visually. Let's do that one. All
34:05 right. So, we got our log line. All
34:07 right. Back to
34:10 LTX. Now I went to chat GPT and had it
34:14 come up with a movie concept for
34:17 us. If you still have
34:20 ego and think that your brain is
34:23 important, you can also use your brain.
34:26 You can come up with ideas like
34:28 whiteboard it, do post-it notes, like
34:30 write a poem, turn it into a
34:33 thing. You can do that too. Full totally
34:36 optional now. But like if you like your
34:40 brain, congrats on on on your
34:43 functional
34:46 mass. Okay, next. So now it's coming up
34:50 with a story line. Let's see if it does
34:51 any
34:52 good.
34:56 [Music]
35:00 Uhuh. YouTube comment from real estate.
35:03 Happy birthday, Kyle. I also celebrated
35:05 the 39th anniversary of my 21st birthday
35:08 this year. This year as well.
35:10 Congratulations. It's It's fantastic
35:12 when you when you when you factor your
35:16 [Laughter]
35:20 decades. All right. It's called Love in
35:22 the Bunker. Good name. A pair of
35:25 doomsday preppers find love amidst chaos
35:28 in as the missile sirens signal the end
35:30 of their first date. In a world where
35:33 anxiety over impending doom is
35:35 commonplace, Jack and Mia, two extreme
35:38 doomsday preppers find themselves
35:40 swiping on a dating app looking for
35:42 someone who shares their passions for
35:44 survival and conspiracy theories. When
35:47 they match, their excitement turns into
35:49 tension as they engage in a rapidfire
35:51 conversation. Oh, see, this is not good.
35:54 Be Well, you know what? Let's let
35:58 um what's it called? LTX used to be
36:01 really bad at dialogue, but I think it's
36:02 better now. Unbeknownst to them, the
36:04 situation outside is about to escalate
36:06 dramatically. As they exchange witty
36:08 banter and flirt
36:11 flirtatious exchanges, Jack invites Mia
36:13 for a quick meet up in his fortified
36:15 bunker, hoping to impress her with his
36:17 extensive collection of survival gear.
36:23 Just as they begin to connect
36:24 personally, the ominous sound of missile
36:27 sirens blare through the city, sending
36:29 them both into a frenzy, caught off
36:31 guard. They realize their romantic
36:33 connection may be their last chance at
36:35 love in an uncertain world. With minutes
36:39 to spare, they scramble to gather their
36:41 supplies for the bunker while sharing
36:43 their fears and dreams, deepening their
36:45 bond. Um, can I can you modify this? Oh,
36:49 wait. Custom Doomsday Preppers romantic
36:53 comedy. Okay. Unexpected love. Okay.
36:55 Well, [ __ ] it. Whatever. Next. Settings
36:58 and
37:01 cast. So, I mean, so this is basically
37:04 it. It's kind of cool, right? I mean,
37:05 it's it's
37:09 like like again, you don't need to use
37:12 your brain. Now, will this produce
37:14 something good and usable? I don't know.
37:16 I haven't done this in probably a year,
37:18 so this has probably gotten much better.
37:20 I'm digging this idea. I am, too. This
37:22 is kind of a good idea. There's Jax
37:24 Thompson. He's a bit of a hottie.
37:26 They're too hot. They're too hot. No,
37:29 no, no. Sorry. Face switch. We can put
37:33 my face on it. That'd be
37:35 funny. Doomsday prepper. Caucasian male.
37:38 Square jawline with light stubble. No.
37:41 Let's say um
37:46 um a few too many bags of
37:53 Doritos. Green eyes with a determined
37:56 gaze. Short brown hair styled in a
37:59 practical cut. Should No. Uh let's see.
38:03 Unckempt. Unckempt brown hair.
38:09 unckempt brown hair
38:17 um that hasn't seen a brush in a
38:22 while. Yeah, we're not making a
38:24 Hollywood blockbuster with [ __ ]
38:27 square jawed model actors. It's doomsday
38:31 preers. Athletic build with broad
38:32 shoulders. No.
38:35 Um,
38:37 computer. Let's see.
38:41 Computer technician
38:47 build with
38:51 uh drooping
38:57 shoulders.
38:59 Oops. Magic Mike Apocalypse Now. Yeah,
39:02 exactly. with drooping shoulders. Okay,
39:05 let's see what he looks like
39:12 now. Come
39:17 on. He's still too skinny. Let's say uh
39:20 let's say stocky
39:23 build with drooping
39:26 shoulders. Computer dadbod. Oh yeah,
39:28 that's good. Stocky dad
39:32 Bob with drooping shoulders. How? Oh, 30
39:35 years old. Yeah. No, he's got to be
39:37 older than this. He's got to be He lived
39:39 He lived with his mother until he was
39:41 38. He's finally on a He finally saved
39:43 up a known to
39:45 get Jesus Christ. Does everyone have to
39:48 look like a goddamn model? Um, we're
39:51 going to make him 40
39:56 43. Fitted olive green shirt. No, we're
39:59 going to just say we're going to say
40:03 um
40:05 um faded rock
40:08 t-shirt with a rough neckline, rugged
40:11 dark cargo pants. Rugged. See, you got
40:14 to take out words like rugged or you're
40:16 going to get
40:19 hotties. All right, let's see what we
40:21 got now. 43 is the new 23. Yeah.
40:26 Jesus, he's still a dude. Caucasian made
40:30 male uncampented brown hairs hasn't seen
40:33 a brush in a while.
40:37 Huh? Why can I not?
40:42 Okay, whatever. Fine. That's that's
40:46 him. And
40:48 then it's
40:53 exhausting. Hispanic female. She doesn't
40:56 look Hispanic at
40:59 all. Heart-shaped face. Let's say round
41:04 face.
41:06 Um long wavy black hair. Let's go. Long
41:11 wavy.
41:19 Um
41:21 brownish brownish hair.
41:26 I know. Listen, if you're like, "This is
41:29 boring." I know. And she's going to be
41:33 45. We're not having We're not This is
41:35 We're not doing
41:37 hotties. Older people know how to swipe.
41:43 Okay. Stupid AI
41:50 writers. Oh.
41:53 [Music]
41:57 Uh, I've noticed AI has difficulty with
42:00 any gradients between 32 and an eight
42:04 pack and 50 plus. Yeah, it's a good
42:06 point. AI doesn't seem to understand
42:10 that anyone who's going to be in a
42:12 movie, you know, isn't isn't Hollywood
42:15 hot. Why is this taking so
42:19 long? Oh, actor library. Pick or create
42:23 a consistent
42:31 actor. Okay. And then what style do we
42:36 want? View all. Let's go look at our
42:39 styles. This is
42:42 it. 3D cartoon anime scribble.
42:47 Boost. Okay. I guess we'll do boost.
42:52 Loving the bunker. Consistent is
42:57 good. These do not look
43:09 like they don't look like what I
43:11 described, but whatever. Whatever.
43:14 Whatever.
43:23 It's cuz all the programmers are 32 with
43:25 an eight
43:27 pack. Looks like Ben and JLo. Yeah,
43:30 exactly. All right. Well, we're going
43:32 with Hollywood hot. Even though listen,
43:35 you saw it here. I tried to describe not
43:37 Hollywood hot and it didn't like
43:40 it. All right. So, now it's going to
43:42 give us a
43:46 breakdown. Scene one. Match on the
43:49 dating app. Dimly lit room cluttered
43:51 with survival
43:53 gear. All right. Their exchange is
43:55 fitted with
43:57 enthusiasm. Scene two. Frantic
44:00 preparations. Jack's fortified bump
44:02 bunker characterized by reinforced walls
44:05 stacked of canned goods and an array of
44:07 survival
44:08 tools. Jack invites Mia to meet in his
44:11 fortified bunker. Excitement palpable in
44:15 the
44:16 air. Moment of intimacy in the
44:19 bunker. All right, start. I think this
44:23 is more than a three scene thing, but I
44:24 did say a threeminute movie, so we'll
44:31 [Music]
44:38 see. How was the podcast today? Podcast
44:41 was great. podcast was great.
44:45 Um, we had a a lovely artist
44:50 um on and it was just it was just good
44:53 to be back and we had fun. We had a good
44:55 conversation and then it was very very
44:58 good. Very very
45:01 nice. Yeah. There's Jack looking hot in
45:04 his in his little love chair.
45:14 Look at his
45:18 belly. He went
45:23 he Look at him. Look at him in scene
45:28 one. Okay, there he is in scene one. And
45:31 and three minutes
45:38 later. Oh my god. Loving the time of
45:43 catastrophe. Oh, that is joyous. That
45:46 belly happened in one sitting. That's a
45:48 30 pack. Well, as I like to call it, keg
45:51 abs. You know, some people have
45:53 six-packs. I've got keg
45:55 [Laughter]
45:57 abs. Oh, lordy, lordy, lordy. All right.
46:01 Uh let's see now. So do we have
46:05 character
46:08 dialogue voice
46:10 over sounds has moved to the timeline?
46:13 Try it. Oh, they've got a timeline now.
46:17 Oh, interesting.
46:19 [Music]
46:29 A, she's so cute. I like Mia. Oh, he's
46:32 got an extra arm.
46:39 [Music]
46:53 Wow. All right. Uh, how do we do voice
46:57 over videos for
46:59 sound? Soundtrack sound
47:02 effects text. Melissa voice over. What
47:06 if I want to record it?
47:09 Insert. I want to record it.
47:13 How do I record
47:18 it? You're
47:21 stupid. Why can't I record it? This is
47:25 dumb. Under
47:28 voices, choose your own.
47:38 Oh, no.
47:44 Guess
47:45 not. Produc producer Brandon faking me
47:49 out with
47:50 the thinking I missed something
47:53 here. I guess he can't record [ __ ]
47:56 Well, let's go back to the Let's go back
47:58 to the story. Motion
48:02 editor. Well, anyway, this is LTX. All
48:06 right, here's Franic Preparations. Why?
48:09 Why did it not generate all this
48:13 [ __ ] Generate
48:15 scene. Oh, I see. They're They're trying
48:17 to save their their their processors
48:20 from melting. I get it. All right, let's
48:23 go add some sound effects. We'll create
48:26 create
48:28 composition. Tik Tok pin. Kyle, how
48:32 often do we all usually say we need an
48:35 extra armor hand a lot? I know.
48:40 Wait, Jack has a Jesus.
48:43 Jack wide shot, medium
48:50 shot. Oh, that's Mia now. She's got
48:53 slump shoulders. Wait, she went from
48:55 being a hottie to She can't stand up
48:58 straight
49:04 now. All right, she's hot. It's okay.
49:07 She's hot again. He And he took off the
49:10 30 pounds before she got
49:16 there. Retry. Oh my god, this is
49:19 hilarious. Let's do Let's do audio
49:22 effects. Sound
49:24 effects. How do you generate
49:27 them? Uh, generate motion.
49:32 [Music]
49:40 The thing about this is if you didn't
49:43 have ADHD, like if you're one of those
49:45 people who can do step by step, like
49:47 let's say let's say when you buy the
49:51 4500piece Lego
49:53 set, you start at page one in the
49:56 instruction manual and you end on page
50:01 472 with a model of Tit Titanic. with no
50:04 pieces left over. If if if you're like
50:08 that, you'll love
50:14 LTX. If you're like me where you open
50:18 the Lego box, you open all of the
50:20 packets at once, and then you mix them
50:22 with all the other Legos that you have
50:24 and try to make a race
50:28 car and realize you don't have wheels
50:31 after you've made the body.
50:33 Then LTX probably ain't going to be your
50:41 thing. I just got bored. Then then don't
50:45 finish. I know. That's exactly. I just
50:47 get bored then don't finish. Exactly.
50:49 That's where I am. Um do we have
50:51 something that we can look at
50:54 here?
50:57 No. Still
51:00 going. Uh uh
51:05 uh. Well, it's still janky. Like what I
51:09 would do if I really wanted to do
51:11 something good here now that chat GPT
51:15 has decent image generation, I would
51:18 probably generate all my images in chat
51:20 GPT and then either do all of the
51:23 animations outside of this thing and
51:24 just assemble them in here. Um, or maybe
51:27 have this do the the images or the
51:30 animations. But I don't know. I don't
51:32 know. I don't know. I don't know what to
51:33 tell you. But I'm I'm tired of this now.
51:37 I've grown
51:41 weary. Okay, that was that.
51:45 Um, let me come back out and just chat
51:47 for a
51:49 bit. That exhausted me because you know
51:53 what part of it is? Is
51:56 That's one of those things that to do it
51:58 really well, like if I
52:01 were if this channel were like me doing
52:04 a twoeek deep dive on something, that
52:07 would be the thing to do.
52:12 [Music]
52:28 Yeah, exactly. What day of the 30 days
52:30 of deep dives are we on? Yeah. How How
52:32 long did I make it into 30 days of deep
52:34 dives?
52:35 14, I think. No. No, it wasn't even 14.
52:38 It was 10. It was two weeks. And I don't
52:41 think I finished the second week. And
52:42 then I never came back to it.
52:45 See, me and consistency are not your
52:48 friend.
52:53 If it starts feeling like work, I'm no
52:54 longer interested. Just arrived. Mind a
52:57 recap? Yes, I will absolutely recap. We
53:00 just played with a tool called LTX
53:02 Studio, which has a new video model.
53:04 It's supposed to have better character
53:06 consistency, but it doesn't.
53:09 Um, well, at least it doesn't have
53:12 character consistency that's good enough
53:15 that it's not hilariously bad.
53:21 Um, but what it is, if you haven't
53:24 played with LTX Studio, it will
53:27 generate, you know, multi-shot scenes
53:30 and and it will kind of design and
53:33 assemble a movie for you with all sorts
53:35 of granular control for each of the
53:37 shots and things like that. Um, it's
53:39 just that it's it's really janky. And
53:42 you can use things within it like V2,
53:45 which I think it's 60
53:47 cents. I think in here V2 is 60 cents
53:51 per shot. So, you could rack up some
53:54 money pretty quickly. But if you wanted
53:55 to do something that looked good and put
53:57 together a story, you could do that. And
53:59 then I just got bored. It's my birthday.
54:01 I'm drinking tequila, which I normally
54:03 don't like tequila, but this tequila was
54:06 aged in a barrel of Bllandens. And it's
54:08 pretty
54:12 good. All right, Mike Harding, enjoy the
54:16 adventure, Kyle. Thank you so much. I
54:18 appreciate that. I appreciate I
54:23 I like box
54:26 chocolates. Tik Tok
54:28 question. And you can access access LTX
54:32 directly from chat GPT. No. Um, LTX
54:36 Studio is I think it's just Yeah,
54:43 app.ltx.studio. That's how you access
54:46 it. It's a standalone
54:49 service. And then if you're like, is it
54:51 going to cost you money?
54:54 Probably to do anything good. They they
54:56 might give you a free, you know,
54:58 assembly or two. Uh, but there's going
55:00 to be a point at which it's going to
55:02 it's going to cost you. Anything video
55:04 related is going to cost you money for
55:05 the most
55:06 part.
55:10 Um, all right.
55:13 Let's What do I want to play with? What
55:16 do I want to play
55:18 with?
55:19 [Music]
55:24 Uh, lots of
55:27 requests for image stuff. All right.
55:31 Yeah, let's let's
55:34 let's hop into
55:37 midjourney. So, here we are in
55:42 midjourney. Hang on. I can I can do it,
55:45 Brandon. I got
55:49 it. There's
55:51 midjourney. New image for you in a
55:53 regulars. Okay. Um
55:57 [Music]
56:02 These are some images I made today for
56:04 the for the AI readiness
56:08 project. And then I was designing some
56:12 people for video avatars at Story Vine.
56:17 Some of them are actually really good.
56:19 That one's really quite
56:21 good. Very
56:24 natural. New image for you. And look
56:26 what I made. Okay, hang on.
56:30 [Music]
56:37 Two hours into the live. Three tequilas
56:47 down. That's pretty good. I like
56:51 it. Oh, that's cute.
56:53 Danielle, a little fishy teacup on a
56:57 thimble thimble of tea on the finger.
57:00 That's nice.
57:02 Lovely. Um, let's see. Here's more
57:05 inventing
57:10 people. Mediterranean dude. Asian dude.
57:25 Um,
57:30 okay. So, we're in
57:34 Midjourney. And if you're new to
57:38 Midjourney, they've they've changed the
57:41 the the the interface for MidJourney
57:43 changes all the time. Oh, thank you very
57:45 much, Side Hustle Mimi. That's very
57:47 sweet.
57:52 Um, there's there's three areas where
57:55 you can augment your prop pro prompts.
57:57 Well, well, there's three areas where
58:00 you can augment your prompts with
58:04 images. And one's one's called an image
58:07 prompt, and it will retain the sort of
58:09 structure of the image. One's called a
58:11 style reference, and it'll pull in the
58:13 style. And then one's called omni
58:15 reference. And omni reference is
58:17 interesting because it can be you can
58:19 put a person in that or you can put like
58:22 a bottle of perfume or like this car. So
58:25 if I put this car in omni
58:30 reference. So it's over
58:34 here. And then I say,
58:39 um, muscle
58:42 car rolls down the
58:49 highway. And actually, why don't I put
58:51 I'll put something else into style
58:54 reference. Oh, I'll grab one of my
59:01 cool Yeah, here's one.
59:05 I've got a cool
59:09 uh that shot like sunset sunset with
59:12 clouds,
59:14 right? And then we'll just try that.
59:16 Let's make
59:18 that. And so omniresence should pull
59:21 that car in. So, we should have this
59:28 green impala looking
59:32 thing in our shot with the cool sunset.
59:52 [Music]
1:00:00 [Music]
1:00:06 close. It was interesting. It brought
1:00:09 the I guess because we did style
1:00:12 reference. It had all those golden hues
1:00:14 in it. It turned the car gold, which is
1:00:16 kind of interesting. I wonder if we said
1:00:20 green muscle
1:00:23 car if it'll pull that olive green
1:00:26 in. And then we do that. Wait, we
1:00:32 do we do this
1:00:34 [Music]
1:00:35 as that. And that is omni reference. And
1:00:39 let's turn up the omni
1:00:43 strength. And
1:01:01 Isn't watching the way these things make
1:01:02 images of a mind trip? Yeah, it made it
1:01:05 green
1:01:06 now. That's kind of cool.
1:01:17 77%
1:01:21 complete. Midjourney be a little slow
1:01:23 tonight. It's a little slow tonight.
1:01:26 Winston, happy birthday, Kyle. Sent you
1:01:28 a doggy gift for champion, too. Oh,
1:01:30 thank you very
1:01:31 much. I appreciate that. Thank you.
1:01:36 Um, yeah, this looks pretty that looks
1:01:39 pretty on
1:01:44 point. And now you can do the same with
1:01:46 people, right? So you could
1:01:50 say, let me go grab a person here. We'll
1:01:53 grab Ann Murphy. We got Ann Murphy in
1:01:55 here. So we'll drag her over to Omni
1:01:59 reference. And then we'll say, let's put
1:02:02 her
1:02:04 in we'll put her in front of the in
1:02:08 front of the Big Bunny Motel. So, we'll
1:02:10 use that for the image prompt. And
1:02:15 then I hope she's in here. And then
1:02:19 we'll do some sort of crazy ass cloudy
1:02:22 sunset. Let's see.
1:02:27 Um, yeah, that's a fun one. That for a
1:02:30 style reference. Okay. So, so we got we
1:02:34 got Big Bunny Motel for an image
1:02:36 reference. So, we'll go
1:02:38 woman
1:02:39 stands in the motel parking
1:02:44 lot
1:02:46 smoking a cigarette.
1:02:52 She She's gonna She's going to sue me.
1:02:55 She's going to sue me for
1:03:02 defamation. Let's see what we
1:03:10 get. Oh man. Wondering how Ann feels
1:03:13 about being dragged in front of the Big
1:03:15 Bunny Motel. Exactly.
1:03:23 This could go horribly
1:03:31 wrong. Oh
1:03:35 no.
1:03:37 Oh bug honey. Bye
1:03:42 Mundisy. Yeah, this is these are all bad
1:03:44 but she is smoking a cigarette so this
1:03:46 is good.
1:03:54 Oh, this is good. This is solid. This is
1:03:57 what
1:03:59 Okay, there's a lot of people that talk
1:04:03 about AI is going to
1:04:06 solve disease and it's going to cure
1:04:09 hunger and it's going to we're going to
1:04:11 live in a hyper prosperous
1:04:13 society. That's not what AI was designed
1:04:15 for. AI was designed to make fun of your
1:04:19 friends in increasingly sophisticated
1:04:21 ways.
1:04:23 Okay? Like, take it in for what it
1:04:25 actually is. There's all this hype about
1:04:28 it's going to cure
1:04:29 cancer. Yeah, but what if you can make
1:04:31 fun of your friends
1:04:33 easily? That's what it's for. Do you
1:04:36 see? Do you understand why I'm excited
1:04:38 about it now?
1:04:54 Look, she got an extra arm. Everybody
1:04:56 needs an extra
1:05:01 arm. I think I think we got to send this
1:05:04 to her, don't we? I'm pretty sure that's
1:05:07 how the law works.
1:05:14 Let's see where we got. Maybe Ann, I
1:05:17 think this is her her little Is that it?
1:05:20 We'll head over in five. Uh,
1:05:23 sorry. This just
1:05:35 happened. Oh my god, that's bad. you um
1:05:43 probably
1:05:46 shouldn't join the
1:05:56 live. Oh
1:05:59 god. All
1:06:01 [Laughter]
1:06:05 right, that was good. All right. So,
1:06:09 yes, this is this is why we have
1:06:13 midjourney. Okay. So, let's
1:06:16 see. Tik Tok
1:06:18 pin. After ask her if she remembers this
1:06:23 night. Remember that night outside of
1:06:26 [Laughter]
1:06:33 Reno? Oh, you're going to get it.
1:06:40 Listen, the people just asked me to show
1:06:43 them how to use MidJourney. I just
1:06:45 happened to have your image in there. It
1:06:47 wasn't a It wasn't on
1:06:49 purpose.
1:06:51 [Laughter]
1:06:56 Yeah. Oh, good lord. Oh, man. Um, okay.
1:07:03 Let's see.
1:07:05 Let's go to
1:07:12 runway
1:07:17 runwayml.com. Let's see. Now, the thing
1:07:20 about runway, they've got this new thing
1:07:22 called
1:07:24 references. Gen 4 images and references
1:07:28 for free. Wait,
1:07:30 what? Reference testing.
1:07:36 So I think you can only make references
1:07:39 consistent scenes with
1:07:41 references. Use one to three character
1:07:44 or location images to build your scene.
1:07:48 So even
1:07:58 though even though this
1:08:01 is a video creation tool, I think with
1:08:05 references, you start out by making an
1:08:07 image first and then turn it into a
1:08:12 video. All right, so let's let's see if
1:08:15 we can grab the same. Let's let's go
1:08:18 into
1:08:19 here.
1:08:22 Uh, okay. This is I feel I feel guilty
1:08:25 now, but that's okay. Ann
1:08:29 Murphy, the it is not is it is not ours
1:08:33 to
1:08:34 judge what we do in life. It's exactly
1:08:38 ours to judge what to do in life. Shut
1:08:41 up. All right, there's my picture. Save
1:08:44 image as. Oh, this will be interesting
1:08:46 to see. So, we'll call this clouds. Um,
1:08:48 it'll be interesting to see how this
1:08:50 thing does. And here's the big bunny
1:08:52 motel. Okay, good. And it's got like the
1:08:54 creepy dude outside of it. This This was
1:08:57 the actual
1:09:00 photo. All right. So, we'll do save
1:09:02 images and we'll do motel. This is on
1:09:05 KFax, by the
1:09:07 way. Um, okay. So, we're going to come
1:09:11 here. We're going to do assets. And then
1:09:13 we're going to do drag and drop. Okay.
1:09:15 So, let's go
1:09:16 grab those three
1:09:25 assets. Okay. We're going to drag those
1:09:28 there, didn't
1:09:32 we?
1:09:34 No.
1:09:36 Unbelievable. Bad interface. Bad
1:09:38 interface. And we can only upload one
1:09:41 image at a time cuz they don't have a
1:09:44 Gen Xer sitting in the corner. They
1:09:47 don't have a Gen Xer sitting in the
1:09:49 corner. They don't have a Gen Xer
1:09:52 sitting in the corner bitching about all
1:09:54 their shitty interface. Yeah, it's a new
1:09:58 song I just wrote. It's a hit coming out
1:10:01 soon. It's called You Need a [ __ ] Gen
1:10:04 Xer sitting in the corner.
1:10:07 Okay. Hover over the reference to tag
1:10:10 and
1:10:13 save
1:10:16 image
1:10:21 motel. You can add reference by using
1:10:25 at Oh, motel. Oh, I
1:10:29 see. That's cool. Okay. I'm kind of
1:10:33 digging this now. So, this is clouds.
1:10:37 And this is
1:10:39 an. All right. This I like. So I can say
1:10:43 at an
1:10:46 um smoking a
1:10:53 cigarette.
1:10:55 Um looking up at the at
1:11:01 clouds, right? Oh, and then you at it
1:11:03 gives you a little pull down menu. This
1:11:05 is pretty slick. Okay, right now for
1:11:08 interface, even though they pissed me
1:11:10 off with their non multiple file upload
1:11:13 thing, this is kind of cool. Looking up
1:11:15 at the
1:11:17 clouds, standing in the parking lot of
1:11:22 at
1:11:24 Motel. I like
1:11:27 it. I like it.
1:11:30 Generate. I have a feeling this is going
1:11:32 to be good.
1:11:34 Welcome to the Hotel, California. This
1:11:37 could go horribly
1:11:40 wrong. I still have a runway sub. Yeah,
1:11:43 runway. Runway's pretty
1:11:47 good. I'll just have a little more
1:11:53 tequila. I wonder if she got back to me.
1:11:56 She's probably
1:11:58 pissed now.
1:12:01 She She's She She's probably out to
1:12:04 dinner. She'll see that when she gets
1:12:06 home. What did you
1:12:16 do? Knock knock knock. Who's there? And
1:12:19 run, Kyle. Run. Oh my god.
1:12:31 Oh my god. Oh, this absolutely needs to
1:12:35 be illegal. Okay, I used to be like, you
1:12:40 know, I'm like, yeah, deep fakes are
1:12:42 deep fakes, but like, you know, how bad
1:12:44 could they possibly be? Well, look at
1:12:49 [Music]
1:12:52 this. The Big Bunny
1:12:55 Motel. Oh my god.
1:13:03 Yes, this becomes a video now. That's
1:13:06 exactly what this is. I think we're
1:13:08 going to go with this one. The dreamy
1:13:11 one. Use. Yeah, I would I would say use
1:13:15 is is is the appropriate thing. Okay,
1:13:19 we're gonna say um woman
1:13:24 um takes a
1:13:28 drag of her
1:13:34 cigarette and
1:13:38 blows smoke at the
1:13:43 camera. I'm sorry. I'm sorry for what
1:13:46 I'm about to do, an
1:13:56 It's so bad. It's so bad. I really I
1:14:00 really do think today might have been my
1:14:02 last podcast with Ann because when she
1:14:05 sees this, it's not going to go well.
1:14:08 It's just not going to go well.
1:14:14 [Laughter]
1:14:19 Um, by the way, the the AI learning lab
1:14:23 is sponsored by She Leads AI. It's a
1:14:27 fantastic community of women in AI led
1:14:31 by the estimable Ann
1:14:35 Murphy. What website is this, please?
1:14:38 This is runway
1:14:41 runwayml.com. And as much as we're
1:14:45 laughing here, like one of the reasons
1:14:47 we're laughing is that it looks so much
1:14:49 like the source materials, right? Like
1:14:52 that's definitely an with that cloud
1:14:54 background I gave it standing in front
1:14:56 of the Big Bunny Motel and now she's
1:14:58 going to be smoking a cigarette.
1:15:01 Um, I think I think that
1:15:03 this this assembly model is better than
1:15:07 MidJourney
1:15:09 is. This is this is quite good. Like if
1:15:12 we were take like I'm I'm literally
1:15:14 putting random pieces of content
1:15:16 together here and like that are
1:15:18 aesthetically different and just
1:15:20 completely different. I got chills.
1:15:22 They're
1:15:24 multiplying. Oh
1:15:26 man, I hope she doesn't come in. And I'm
1:15:29 going to be I'm going to I'm going to
1:15:30 get a talking
1:15:35 too. Oh, good
1:15:50 lord. I want her to come
1:15:54 in. What's the name of the AI tool? This
1:15:56 AI tool is runway
1:15:59 ML. Runway ML. So the so
1:16:03 runway there
1:16:07 are well there are probably about a
1:16:09 dozen decent video makers out there
1:16:12 right
1:16:13 now. There are kind of four four or five
1:16:16 major ones. Runway is kind of the
1:16:19 granddaddy of
1:16:20 them. U and they're starting to add some
1:16:23 interesting features. And this this
1:16:27 ability
1:16:28 to put multiple images into a scene and
1:16:32 then tag them in your prompt is is
1:16:34 really quite
1:16:36 clever. It really is. And what the [ __ ]
1:16:40 are you? Okay.
1:16:41 [Laughter]
1:16:51 Exactly. That's for showing us what not
1:16:54 to do with AI.
1:17:00 This is this is why people are afraid of
1:17:02 AI. Exactly. Where can we get runway?
1:17:05 Just go to um
1:17:08 runwayml as in machine learning.com. All
1:17:11 right. Here we go, people. This is this
1:17:15 is uh now's when she needs to come in,
1:17:17 isn't
1:17:20 it? Oh, here we go.
1:17:27 [Laughter]
1:17:43 Oh
1:17:51 no. Vicky comment. You may not live long
1:17:55 enough to see the next version of Runway
1:18:07 ML. Should I send this to her? Sure.
1:18:21 Oh my god. Oh, this is just
1:18:29 bad. I'm going to say you've been
1:18:34 [Laughter]
1:18:43 violated. Oh my
1:18:49 god. Wow.
1:18:51 Wow. Tik Tok pin for the win. New pope.
1:18:54 Yeah, exactly. New
1:19:02 Pope. Kyle was a bad boy on his
1:19:12 birthday.
1:19:14 Jesus.
1:19:16 Oh. Wow. That's uh that's just not
1:19:20 right. Okay, people listen very
1:19:24 seriously. Deep fakes are dangerous.
1:19:27 Okay. Okay. You don't want to take
1:19:30 someone's likeness without their
1:19:31 permission and in front of a bunch of
1:19:34 people do bad things with them. Right?
1:19:38 You don't want to do that. And I'm just
1:19:40 sharing this is an example of what not
1:19:42 to
1:19:43 do. Right?
1:19:47 It may seem like I'm laughing on the
1:19:49 outside, I'm weeping
1:19:51 inside, right? Because things have
1:19:54 changed. Things are not
1:19:57 okay.
1:20:01 Okay. I always knew this was a teaching
1:20:04 [Laughter]
1:20:06 moment. This is one of those do as we
1:20:09 say, not as do as we do moments.
1:20:11 Producer Brandon. Oh, lordy.
1:20:17 Wow, my gut hurts on the inside. I'm
1:20:21 laughing. I should probably stop at
1:20:23 looping, shouldn't I? No, I shouldn't.
1:20:26 It's [ __ ]
1:20:29 hilarious. Let's see if she
1:20:32 responded. I have officially become my
1:20:34 mother is what she responded in
1:20:41 text. Oh my god.
1:20:45 Oh
1:20:53 god. Okay. The more you know. All right,
1:20:57 folks. You know, listen, we take our
1:21:00 education here at the AI learning lab
1:21:01 quite seriously. What we've learned here
1:21:04 tonight is references. And you take one
1:21:07 image of say a photograph you took of a
1:21:10 lovely sunset, another photograph you
1:21:13 took of a sleazy motel, and then say a
1:21:15 photograph of someone you don't really
1:21:17 know. You just combine them all
1:21:22 together. I'm getting a contact high.
1:21:26 Oh. All
1:21:27 right. Let's Let's wind this up because
1:21:31 I'm about to pour another tequila and
1:21:33 that's I think we're I think I'm I'm
1:21:37 drunk enough on on this
1:21:43 exercise. You know what's going to be
1:21:45 the funniest thing? Because I have
1:21:47 ADHD. I'll like not use Runway for two
1:21:50 or three weeks and the next time I go in
1:21:52 to use it, this will be the video that's
1:21:54 sitting there.
1:21:58 Yeah. Tequila. M.
1:22:01 Tasty. Wow. All right. Okay. I'm I'm a
1:22:05 little I'm a little hurting
1:22:07 [Laughter]
1:22:14 inside. Let's see. Wait. Also, what did
1:22:17 she say? Also, everyone knows I'm a
1:22:20 gummy girly, so this isn't me. But
1:22:21 nicely
1:22:24 done. Oh, she's a good sport. That's
1:22:27 awesome. Okay, cool. Um,
1:22:30 Groovy, my gut hurts. Oh, Marina, it's
1:22:37 funny. All right, what what day is
1:22:39 today? Today's Wednesday.
1:22:42 Yeah, Wednesday, May the 7th. Yes, my
1:22:46 birthday. My birthday. Thank you
1:22:48 everyone for for hanging out and
1:22:50 allowing me to act ridiculous night
1:22:52 after night. Um the uh the little the
1:22:56 little super cut edit was awesome. Thank
1:22:58 you. Thank you, Winston. Thank you,
1:22:59 Brandon. Whoever had a hand in that.
1:23:01 Thank you. That was really
1:23:03 incredible. Um and just thanks for all
1:23:06 your good wishes. This is this is really
1:23:07 sweet of you. So, hope you had fun
1:23:09 tonight. Maybe you learned a little
1:23:10 something.
1:23:12 Um yeah, if you've got the brain fork,
1:23:15 go play with LTX. If you don't, don't.
1:23:18 Um just go play with anything. I I will
1:23:23 go I want to return I read a thing last
1:23:27 night from the CEO of
1:23:31 uh of Fiverr to his staff about you know
1:23:36 AI is coming for your jobs. He he wrote
1:23:38 this really blunt
1:23:40 email to his staff and the first thing
1:23:44 he says is e you know AI is coming for
1:23:47 your job and naturally what most people
1:23:49 are talking about is AI is coming for
1:23:51 your
1:23:52 job. Where that email ultimately lands
1:23:57 is the thing that you need to do right
1:23:59 now is turn yourself into like give
1:24:03 yourself superpowers with AI.
1:24:06 Learn all the tools. Hang out with
1:24:08 people that know the tools and use the
1:24:10 tools. Hang out at places like the AI
1:24:12 learning lab and the AI
1:24:14 salon
1:24:16 because over the next couple of
1:24:19 months, these tools are getting so good
1:24:21 that work is absolutely going to start
1:24:24 being disrupted. And if you're like
1:24:25 paying attention to some of the videos
1:24:27 on TikTok these days, there's there's a
1:24:30 fair amount of videos going around of
1:24:32 like I'm a Yale graduate with a 4.5 G,
1:24:36 you know, GPA or 3 point or whatever 3.5
1:24:39 GPA and I've applied for 250 jobs and I
1:24:44 haven't gotten, you know, a single
1:24:46 interview or a single offer, whatever it
1:24:47 might be. And the guy looks like he's
1:24:49 chiseled out a rock.
1:24:52 um it's going to start getting brutal.
1:24:54 And who it's going to not get brutal
1:24:56 with, I think, at least initially, is
1:24:59 people that know AI. So, I know I'm
1:25:01 preaching to the choir here, but if
1:25:02 you're new here and if you're just
1:25:04 trying to figure this AI [ __ ] out, get
1:25:07 aggressively
1:25:09 curious. Get aggressively curious.
1:25:12 Really start to dig into this stuff.
1:25:14 Keep coming back here, okay? and we will
1:25:16 keep
1:25:18 edimocating even though we're not
1:25:20 qualified to do so. Um, you will not get
1:25:23 a s certificate from this
1:25:28 channel. You might get a lawsuit from
1:25:30 this channel, but you won't get a
1:25:33 certificate. All right. All right.
1:25:35 Thanks for the kind wishes, everybody.
1:25:37 And uh, peace out. I will see you
1:25:39 tomorrow. Bye.