6/19/2025 - From Quantum Rap to Angry Baristas: AI Trends

Description

Chapters

Transcript

0:02 [Music]
0:15 [Music]
0:33 Rain
0:35 by
0:37 [Music]
0:41 mountain silence.
0:44 [Music]
0:47 This is the place where the feelings
0:50 aren.
0:52 [Music]
0:55 from an overexposure to balance.
1:00 This is I can slowly face the only one I
1:05 truly can know.
1:09 These are tears from a long time ago.
1:13 Got these tears from a long time ago.
1:17 I need to cry 30 years or so.
1:22 These are tears from a long time
1:25 [Music]
1:28 ago
1:29 [Music]
1:38 Thursday night. Did you do your TPS
1:41 reports?
1:43 Did you do your TPS reports?
1:47 They're still due. I know it's Jun June
1:50 Junth. Happy Juneenth.
1:54 Is that a happy holiday? Do you say
1:56 happy Junth? I don't think you do. Do
1:58 you?
2:02 Yes, you do. Happy Junth. Good lord.
2:09 Every now and again I realize how white
2:11 I am.
2:14 It's not good people.
2:17 [Music]
2:39 [Music]
2:52 Freedom came my way that night
2:58 just like a jet plane in and out.
3:04 I was hauling ass at a million miles an
3:07 hour,
3:09 wondering how hard I did.
3:13 [Music]
3:15 When they came into the st,
3:22 they said I was happy
3:25 [Music]
3:27 and I got no qualms with my situation.
3:34 Say here I am.
3:37 [Music]
3:45 Oh, it's warm in here. It's warm in
3:48 here. It's It's It's Hopefully,
3:52 it is warm in here.
3:56 Come on, air conditioner. Do your
3:57 flipping thing.
4:13 All right,
4:15 let's see.
4:19 [Music]
4:21 Can't post comments to some
4:23 destinations.
4:25 Okay.
4:27 All right. We got no one on.
4:31 [Music]
4:58 Um,
5:02 I think something I want to talk about
5:04 tonight is
5:06 [Music]
5:11 is something that I'm going to tell you
5:12 about. And because you're you, you're a
5:16 bunch of irregulars.
5:19 You're going to take this, you're going
5:20 to run with it. Someone's going to do
5:21 something really big with it. and then
5:22 I'm going to get jealous of it. So, I'm
5:25 going to make a prediction tonight that
5:26 what I'm about to talk about, what I'm
5:29 about to show you
5:32 is
5:35 someone's going to do it successfully.
5:37 Well, that's the definition of
5:40 successfully is to do it well. You've
5:42 succeeded. It was a well done job. You
5:45 did it well.
5:48 And then I'm gonna see that you did that
5:50 and and I'm probably on the live gonna
5:52 go, congratulations. That's so awesome.
5:55 And then in my head I'm like, I want to
5:57 kick Champy.
6:00 Why didn't I think of that? It's going
6:02 to be one of those. So that should be
6:05 fun.
6:07 So really, what I'm saying is it's up to
6:10 you whether or not I kick Champy. If you
6:13 if you don't take this idea and run with
6:14 it, then there's no risk. champ is safe.
6:18 But if you're like, "Hey, that's a good
6:19 idea, Kyle. Why don't I go do that?"
6:24 And then you win some sort of Emmy or
6:26 Pulitzer
6:28 for for excellence in AI journalism,
6:33 you know, then Champy's at risk. He's an
6:36 at risk dog. He's an at risk dog.
6:44 There's only 10 people in here. So, hey,
6:47 irregulars. What's happening? I don't
6:50 know where all of our strangers are.
6:51 That's a little dire.
6:54 Kicking the dog because I'm jealous of
6:56 someone else's success. I know. I know.
6:58 It's It's Listen, I'm human. We're
7:01 imperfect, us humans. It's one of one of
7:04 the things we love about humans is how
7:06 imperfect we are, you know?
7:13 Oh man,
7:16 I don't quite know what the bigger idea
7:18 is yet, but I know there's a thing
7:20 happening
7:22 and it's a thing happening that's that
7:24 is
7:26 was not possible before.
7:29 It was That's not true. It was possible
7:32 before, but it took a lot of work.
7:36 So, the thing I'm going to talk about,
7:39 can't wait to hear the new idea.
7:41 What are you drinking? It's vodka. I
7:44 just have a shot. I just happen to have
7:46 a 32 oz shot glass.
7:50 Oh, man. Um, I was supposed to be at
7:53 Whiskey Club tonight, but I had to I I
7:55 was late at work. I I It was one of
7:58 those days. You know those days where
8:00 you have meetings where you either have
8:01 an hour or half hour between every every
8:04 meeting, but there's meetings for the
8:06 whole day. So I could never really get
8:08 up from my desk to do anything.
8:11 So So
8:13 I'm crispy. I'm crispy.
8:16 [Music]
8:51 [Music]
8:59 I didn't mean to cause you any trouble.
9:05 Didn't mean to cause her to pain.
9:09 [Music]
9:14 Only one time to see you laughing.
9:18 [Music]
9:20 All I want to see you laughing in that
9:23 purple rain. Purple rain. Purple rain.
9:31 Purple rain. Purple rain.
9:38 Purple rain. Purple rain.
9:41 [Music]
9:47 I only want to see you laughing in that
9:51 bubble ring.
9:56 Um
9:58 I uh I've been in a fair amount of uh
10:01 you know Prince sing alike contests. I
10:04 don't like to brag, but you know I'm
10:05 pretty good. You know, pretty good. What
10:09 was what was your dog kicking idea?
10:12 The Oh, I I'll get to the dog kicking
10:14 idea. I'm gonna get to I I actually need
10:17 to get in my head,
10:19 [Music]
10:22 what's happening,
10:25 and what's different about it. Because
10:26 when I'm going to show it to you, you're
10:28 going to go like, "Yeah, that's we've
10:29 seen that before. Yeah, we know. We we
10:31 we know what's going on there."
10:35 So, as I'm as I'm meandering through my
10:39 crazy straw of a brain,
10:42 I'll get there. I'll get there. Um, I
10:46 think we're gonna play SpaceX.
10:49 Just Just came in on kicking the dog.
10:58 I was in that movie. Remember Kak? Yeah,
11:01 that's right.
11:03 [Music]
11:11 Would you ever do a challenge where you
11:13 upload AI music to Spotify? Wait, would
11:17 you ever do a challenge? I have uploaded
11:20 AI music to Spotify.
11:25 What's the
11:27 runner xx23?
11:29 I don't
11:32 [Music]
11:36 We are We are all self-domemesticated
11:38 apes.
11:41 Um yeah, there's AI songs on Spotify.
11:44 I've got one on there. I've got um you
11:47 know what song I've got on Spotify is uh
11:49 the quantum cipher. If you go look up
11:52 quantum cipher ci her quantum cipher on
11:58 Spotify,
12:01 you'll see you'll see the song that I
12:04 put up there. So, yes. Yes, I would take
12:07 that challenge.
12:09 Do I do I get a cookie? I want a cookie.
12:12 I took the challenge. I It's on Spotify
12:14 right now. I did it. Yay. He did it.
12:17 Yay.
12:20 You have Can you show the results? Now
12:23 I'm interested. Yeah.
12:26 I don't know if I can find it on
12:27 Spotify. I haven't looked in ages, but
12:28 let let me change my sharing settings so
12:31 that we can actually hear it. If I find
12:32 it. Oh, wait. Hang on. I got to go to
12:35 Spotify.
12:37 There's So,
12:39 this is where I made the the music part
12:41 of it. I did the lyrics in chat GPT over
12:44 about a 17p prompt amalgamation.
12:49 Um, let's see. Uh, Spotify.
12:55 Spotify.
12:58 Oh, you should have been their podcast.
13:00 Here's a podcast I used to do. Oh, I'm
13:03 not sharing it. How would you know that?
13:09 [Music]
13:11 Okay.
13:14 So, let's see. Sign Oh, not sign up.
13:29 [Music]
13:33 Oh, thank you, Brandon. Brandon's
13:35 protecting my password. Although, what
13:37 was it? 16 60 billion passwords were
13:41 leaked today. The world's largest data
13:43 breach.
13:46 Wee.
13:50 Um. Huh. I can't seem to log into
13:53 Spotify.
13:58 [Music]
14:01 Am I there?
14:03 Continue.
14:05 Oops. Something went wrong. Please try
14:07 again in our help area.
14:10 The [ __ ]
14:17 And I can't log in with this. Did I not
14:19 have a Spotify account? Maybe I don't
14:21 need to log into Spotify. Hang on. Just
14:23 Spotify.
14:25 I can just listen, right? Without
14:27 logging in.
14:35 What's this? Is this my This my Spotify
14:39 post? Yeah, you posted on uh Twitter
14:42 when you launched it on Spotify. I
14:45 literally searched on Google for quantum
14:48 cipher Spotify and this was the second
14:51 hit. Oh, that's great.
14:53 All right.
14:57 Yeah, Dro Kid.
15:00 All right. Here it is. Spotify.
15:06 So, where am I? I'm at Drok Kid. Wait,
15:08 let me go over to spot. It opened in a
15:10 new tab.
15:12 Yeah, I know.
15:17 Did I get it? I got it. All right, hang
15:19 on. Stop screen. Sorry about this. One
15:22 of the hardest things on the planet is
15:25 sharing tabs, apparently.
15:28 All right, so here we are on Spotify,
15:33 right? And there is Laqu Quantum Cipher.
15:38 And let me you know what? I'll tell you
15:39 the story. I'll tell you this is this is
15:41 like the the behind the behind the
15:43 tracks behind the tracks. So
15:47 So there's a couple of things about this
15:49 track. One is um a shortened version of
15:52 this track is in the musical that I've
15:54 been working on uh for the past year.
15:57 So, if you go to Soundcloud and search
16:00 for Sydney and Artificial Love Story,
16:03 I've got a whole soundtrack there of the
16:06 musical that I've been writing for the
16:07 past year. Um, which you should go
16:10 listen to that. Those are not on
16:11 Spotify, but they're on Soundcloud. Um,
16:16 this song, there's a moment, there's the
16:18 moment in the musical where the tech
16:20 reporter says to the chatbot, "Hey, can
16:24 you write um
16:26 write me an Eminem style rap about
16:29 quantum mechanics?" And this is the song
16:31 that plays or, you know, that the lead
16:33 actress sings in the musical.
16:36 But the lyrics for this, I've been
16:40 trying for more than two years. When
16:41 chat GPT first came out, one of the
16:44 first things I did was I I came up with
16:46 this prompt that I was like, what are
16:47 the sort of the two most disperate
16:49 things that that you could smash
16:52 together that if a large language model
16:55 could do them well, would be really
16:57 impressive. And I thought the
16:59 sophistication of Eminem lyrics, right,
17:01 internal rhyme schemes and double
17:02 entandras and triple antandras and all
17:05 that sort of [ __ ] and quantum mechanics.
17:07 And so for two years and a couple of
17:10 months, large language models would
17:12 write lyrics that were rap lyrics, but
17:15 they just weren't very sophisticated.
17:17 They were good at the physics, they were
17:18 not good at the rap. And then in
17:20 November of 2024, so this past November,
17:24 a creative writing upgrade came to
17:26 ChatGpt 40. And I said, "Oh, let me go
17:29 try that." And so this is the this is
17:33 the resulting lyrics. And then I got the
17:35 lyrics uh through a series of different
17:37 prompts and refinements,
17:40 put them together, and then moved them
17:42 over to Sora and generated I don't know
17:44 40 or 50 songs. And this was the one I
17:47 ended up with.
17:50 It's probably going to have an ad in the
17:51 front of it.
18:01 Seriously.
18:09 Oh my god,
18:12 it hurts. I ache. I ache. This email
18:16 already exists.
18:18 Finish creating your account.
18:23 All right, we're going to go back. We're
18:26 just going to play it here.
18:29 So, anyway, it's on Spotify, but we'll
18:31 play it. We'll play it on Drok Kid,
18:34 which is the distributor that I uploaded
18:36 it from.
18:43 Invite guests. Oh, user error.
18:48 All right, here we go. You ready?
18:54 Wait, why
18:56 preview?
18:58 Oh my god.
19:01 You got to be kidding me.
19:04 Can I just play it?
19:07 I can't just play it, can I?
19:14 All right,
19:16 hang on. I'm gonna go find I'm gonna go
19:18 find it somewhere else.
19:20 I'll play you the video of it.
19:23 Um,
19:28 this unbelievable quantum cipher Kyle
19:31 Shannon
19:34 Champ, shut up. Someone's just putting
19:36 the mail in the slot. Come on, you dude.
19:40 You big dumb dumb.
19:45 All right,
19:46 this I know we'll play. Am I sharing it?
19:50 Oh, no. Here
19:56 we go.
19:58 [Music]
20:07 I'm tangled with the beat. Spooky action
20:09 in the booth. Split the atom with a
20:10 pattern. Every partial burst of truth.
20:12 Superp position in the mission. Fusion
20:14 fusing my intentions. Paralleling every
20:16 line. I'm a prism of dimensions.
20:18 Multi-elabic antics. I'm manic with the
20:20 semantics. My rhymes collide like
20:21 particles dynamic and gigantic
20:23 oscillating and dominate. I'm weaving
20:24 quantum state spin a verse and watch the
20:26 multiverse reciprocate. I'm discreet
20:28 like planks beat energy and tight
20:30 packets when I rap it's math and motion
20:32 flipping flows in the fractals. E equals
20:34 NC squ in every letter I write. My pen
20:36 ignites like antimatter when it clashes
20:38 with light. Quick as a flick of a wrist.
20:40 I'm a lyrical catalyst atomically
20:42 splitting the script. I'm equipped with
20:43 the mattest whip. Flipping the tempo
20:45 while pencils igniting the mental
20:46 crescendo. Bending my syllable, slicing
20:48 through time like a sharp innuendo.
20:50 Corner precision. I'm ripping the
20:51 rhythm. I'm spitting in prisms. Sing
20:52 visions. I'm shifting dimensions with
20:53 lyrical decisions.
20:55 Accelerate leather to elevate thoughts
20:56 to the turbulent spin. I'm the velocity
20:58 ripping apart the facade that you're
20:59 certain you're in. Time dilation and
21:01 cadence. I stretch the seconds. I'm
21:02 pacing. Heisenberg with the mic.
21:04 Uncertaintyy's the foundation. Observer
21:06 effecting effect. The second I start to
21:08 speak. Collapse reality's potential with
21:10 every word that I tweak. Yo, my flows a
21:12 spinning lyrical collisions. Every bar's
21:14 a cork sparking a fusion of my visions.
21:16 I'm a wave in a particle. Unpredictable
21:18 art entangled in the cipher. Splitting
21:20 atoms with my heart. From wormholes to
21:22 portals. I'm immortal with the schemes
21:24 through the fabric of this rap. I sow
21:26 the threads of my dreams. Dark matter in
21:28 my patterns unseen but colossal black
21:30 hole soul pulling rhythms through the
21:32 gospel. Double the trouble. I'm stacking
21:33 the rubble. My bombs are combusting,
21:35 speeding and weaving. Believe it. I'm
21:36 leaving your speakers erupting. I'm the
21:38 equation of pace and erasing the limits
21:40 of sound. Taking the crown, breaking the
21:42 bounds. I'm a force. You can't slow
21:43 down. Rapidly tapping the map of my
21:45 mind. I define every pattern in motion.
21:47 Bars are a bullet. I pull it and shoot
21:49 it with ruthless devotion. I'm the
21:50 ignition of friction depicting a world
21:52 that is vivid and bright. Fire and fury.
21:54 I bury the track and ablaze at the speed
21:56 of light.
21:57 [Music]
21:58 [Laughter]
22:04 Shifting perspectives midline like
22:05 quantum flips in my grind. Infinite
22:07 chances to shine. I'm the master of
22:09 spaceime. Every rhymes a singularity
22:11 collapsing the infinite. I'm the big
22:13 bang slang expanding my imprint
22:14 limitless. So step into the quantum
22:16 where the chaos align my bars of
22:18 entropic logic redefining the divine
22:20 from the smallest of particles to the
22:22 cosmos that we know. This is rap at it.
22:24 Quantum flow let it glow.
22:28 Video was done by Kelly Bosch. If you
22:30 don't know Kelly Bosch, you need to
22:33 follow her on Tik Tok. She is
22:35 staggeringly good at at generative AI
22:38 art. And actually, most of her um videos
22:41 that she puts out now have Suno Music on
22:43 them. So anyway, imagine how many views
22:46 this could get if the video was made
22:47 with Google V3. Actually, it's not a bad
22:50 idea. Um,
22:55 I I don't think there's a way to time it
22:57 to to the uh
23:00 I don't think there's a way to time it,
23:04 but what I could do is I could do a
23:05 combination of VO3 and uh I could do a
23:10 combination of VO3 and Hedra
23:14 because Hedra you can do lip syncing
23:16 really good. So yeah, it's not a bad
23:18 idea. I don't I tend not to go back and
23:20 revisit stuff, but that song is one like
23:23 that song really popped off on my
23:24 TikTok. That was my my first video I had
23:27 that was like multiple hundreds of
23:29 thousands of views. Um, so anyway,
23:34 yeah. Yeah, that would be dope. Yeah, we
23:36 could do that. So, I figure what I'm
23:37 going to do tonight is is play with some
23:39 video stuff. But here's the I So, I
23:41 mentioned before I'm going to give you
23:42 all an idea. And I and I actually I
23:44 really do think
23:48 the idea that I'm going to give you is
23:49 not an original idea. The the idea that
23:51 I'm going to give you is an idea that
23:54 happens all the time. It's been
23:56 happening for the past 20 years with
23:59 social media, but there's something
24:00 different.
24:02 Um, all right. So, I've got to share my
24:07 sound.
24:08 So, let me go to
24:12 What are we going to do? We're going to
24:14 go to Tik Tok
24:24 and then I'm going to search for
24:28 barista.
24:34 Yeah, this is it.
24:37 Okay.
24:53 Okay. All right. So, sharing my screen.
24:57 Stop sharing that. Start sharing this.
24:59 Start sharing that. Start sharing this.
25:02 Okay. So, this Tik Tok channel
25:09 That data breach is wild. I know. It's
25:12 crazy. Always use long passwords and
25:14 two-factor authentication. Yeah, it was
25:17 it's the largest data breach in history.
25:19 I don't I don't know who's I heard that
25:21 it was it was like Apple passwords,
25:24 Google passwords, everything. But I have
25:26 no idea like whose data was breached.
25:28 Like that's I didn't hear that part of
25:30 the story. Um okay. So, this channel is
25:35 called
25:40 [Music]
25:42 Hey, do y'all have It's called Feral
25:45 Coffee Crew.
25:48 This channel was created one day ago.
25:52 Okay. One day ago.
25:55 One day. One day ago,
25:58 this channel was created. Okay.
26:02 Every video is 8 seconds long. Hey, do
26:05 y'all have any specials today? Yeah,
26:07 today's special is me not telling you to
26:08 [ __ ] off back to Yelp. You want caffeine
26:10 or a [ __ ] conversation, [ __ ] Can I
26:13 get a venty half oat milk caramel
26:14 macchiato upside down extra hot? This
26:16 ain't build a [ __ ] bear.
26:26 [Music]
26:28 Latte for Claire.
26:30 This is my
26:32 [Music]
26:37 [Applause]
26:38 [Music]
26:46 [Music]
26:47 This is a real coffee coffee company.
26:50 Anyway, latte for Claire. Wait, this
26:54 isn't my iced matcha. Well, are you
26:55 [ __ ] Claire? Then that isn't your
26:58 [ __ ] drink, [ __ ]
27:00 Um, do y'all have straws? The straws are
27:02 right there on the bar if you'd use your
27:04 latte latte for Claire. Wait, this isn't
27:07 my iced matcha. Um, do y'all have
27:10 straws? The straws are POV barista. Got
27:13 that?
27:16 You know what time it is? Hello.
27:20 You know what time? Oh, these are
27:21 different.
27:25 That's why they're weird. Okay, so let's
27:28 see.
27:29 Okay. So, so, so it's so the channel the
27:32 channel is Feral Feral Coffee. Do you
27:37 have any specials today? Yeah.
27:41 And so, here's the here's the idea.
27:44 Here's the trend. And why don't we we'll
27:46 go make some of these. And I I don't
27:48 think I'm gonna start a new Tik Tok
27:49 channel tonight. But, you know, when
27:53 like we've lived in meme culture for 20
27:55 years, so someone will put out a meme
27:56 and then other people copy it. Well,
27:58 well, generally what they did was
28:00 someone would put out a meme and then
28:02 that meme would get proliferated, you
28:04 know, that that single meme, you know,
28:07 or or the original four, you know,
28:09 things would blast out across social
28:11 media and then at some point people
28:14 would start to copy them and those would
28:16 start to go.
28:18 The new trend is when you have a tool
28:21 like VO3 where you can just prompt
28:24 actors and dialogue and a se scene into
28:28 existence.
28:30 You can spin up a Tik Tok channel with
28:32 some clever little hack, right? So yeah,
28:36 this is Stormtroopers, all the
28:38 Stormtrooper videos, all the Yeti
28:40 videos, all the Bigfoot videos, um all
28:44 the babies, you know, doing sitcom
28:46 scenes. We're we're seeing this again
28:48 and again and again, but this this was
28:51 one where I'm I'm, you know, flicking
28:54 through Tik Tok and the first one I saw,
28:57 I didn't immediately recognize it as AI.
29:00 I was just like, "Oh, that's clever."
29:01 And I was like, "Oh, wait, that's
29:03 probably AI." And I clicked on the
29:04 channel and it was Feral Coffee Crew.
29:06 And like all the videos are eight
29:09 seconds long and they're and they're V3.
29:13 So,
29:15 like I said, I haven't quite processed
29:17 what's different about this, but there's
29:19 something different about this. It's
29:20 obviously the speed with which you can
29:23 do this, but why don't we go Why don't
29:25 we go play?
29:27 We'll start over in chat GPT.
29:30 What did I just do? The [ __ ] did I just
29:32 do? Oh,
29:39 share screen. Entire screen.
29:43 Bang bang bang bang bang.
29:47 [Applause]
29:49 Disney ain't doing nothing about being
29:51 used, but let's sue. Yeah. Oh, exactly.
29:54 Yeah, Gavin Biddy. Yeah, dis Disney's
29:58 suing the [ __ ] out of Midjourney.
30:00 Meanwhile, um the you know these VO3
30:02 commercials are all over, you know,
30:05 [ __ ] network television that Disney
30:07 owns. It's it's also hypocritical.
30:10 Um, but that's okay. We don't We've got
30:14 chat GPT now. We can fight back with
30:16 chat GPT. Okay, new chat. So,
30:22 first thing I'm going to do,
30:27 champ, shut up, Kyle. Irregulars,
30:30 please.
30:32 Yes, ma'am. Ma'am, yes, ma'am. Heading
30:34 to irregulars, ma'am. Reload.
30:38 [Music]
30:46 sea shell ASMR. Yeah, the ASMR. That was
30:48 another one I saw today. Yeah,
30:53 yeah,
30:56 yeah,
30:58 yeah,
31:00 yep, yep. ASMR is one. The the glassm
31:05 ASMR.
31:07 Uh, let's see. Um, Glass
31:11 ASMR
31:13 because VO3 has been trained on all of
31:15 YouTube. What's YouTube full of? ASMR
31:19 videos, right? So, let me change my uh
31:21 let me change my settings again so you
31:23 can hear it
31:34 again. Someone here edited some things
31:36 together. are genius.
32:08 So any one of these now you can
32:10 trivially
32:12 duplicate.
32:17 So I'm going to go to I'll just do it
32:19 right here in this tab. flow.google.
32:27 All right. So that's where we're going
32:28 to make our videos.
32:30 And then we'll we'll come back to that
32:32 in a second. Let me stop sharing and
32:34 share my other screen so we can go to
32:36 chat japet.
32:46 All right.
32:52 Yo, Quinn, how's it going? Oh, wait. I'm
32:55 not sharing my audio. Dang it. Oh my
32:58 god, the sharing of the audio
33:07 first. First first I can't log into
33:09 Spotify. Now everything I want to do is
33:12 involves audio because you know what the
33:14 problem is? We're now in a we're now in
33:16 a state where
33:18 AI is going to get completely
33:19 multimodal. It's like
33:23 the blurring of the lines between text
33:25 models, image models, video models, 3D
33:28 world building models, game development
33:30 models, app development models. They're
33:33 all going to start to blur. So, this is
33:35 going to continue to be a problem. But
33:37 anyway,
33:41 we might as well do it right.
33:46 Hey Quinn, how's it going?
33:52 Hey Kyle, I'm doing great. Thanks for
33:54 asking. How about you? How's your day
33:55 going so far? My day day is going okay.
33:57 Okay, so now here's the deal. The
34:00 there's a new video model out called V3.
34:02 Do you know it from Google? VO3?
34:13 Yes, I'm familiar with VO3 from Google.
34:16 It's a pretty cool model that can
34:18 generate videos with synchronized audio,
34:20 including sound effects and dialogue. Y
34:23 it's also being I don't I don't need you
34:25 to describe it to me. Here's here's what
34:27 I want you to do. I want you to um go
34:31 research um best prompting practices for
34:34 V3.
34:44 All right. So, the best way to get great
34:46 results with Vetta 3 is to craft super
34:49 detailed primes. You want to specify
34:51 everything from the subject and setting
34:53 to the camera movement style and even
34:55 the audio elements. Did you write that
34:58 down for me?
35:02 Uh, let's see.
35:04 All right, there it is. Okay, so Okay,
35:07 so let's go back in. Okay. So, now what
35:11 I want you to do is there's a there's a
35:13 meme out right now um of angry baristas
35:18 um and and they're um 8-second videos
35:22 where the customers being a jerk or they
35:24 just ask a normal question and the
35:25 barista cusses at them. Um why don't you
35:29 go see why don't you go do a Google
35:31 search and see if you can find um what
35:34 was it called? Hang on a second. See if
35:36 you can find anyone who's written about
35:38 feral coffee crew videos about these
35:42 angry baristas. I want you to go
35:43 research that.
35:53 So, it looks like the Feral Coffee Crew
35:56 videos are definitely making waves.
35:58 There's one with over 6 million views
36:00 where a barista jokes about not
36:02 squeezing oats in house and another one
36:04 with a ton of likes that pokes fun at
36:06 customers over complicating their coffee
36:08 orders. It's definitely So, here's what
36:11 I want you to do. I want you to come up
36:13 with 20 essentially identical concepts.
36:17 So, we're going to take the basic
36:19 framework of the concept, short 8-second
36:22 videos, um something that um it
36:27 demonstrates
36:29 frustration in the workplace. And like
36:32 one example that pops to my mind is I'm
36:34 sure people that work at Home Depot must
36:37 deal with insufferable know-it-alls all
36:40 day long. So, like a good one might be
36:43 the Home Depot version of this, right?
36:44 Or the big box store. So, I want you to
36:46 come up with 20 new concepts uh that
36:50 that we could kind of knock off this
36:52 idea.
36:55 All right, let's run with that. So, we
36:57 could have a hardware store havoc
36:59 version where employees deal with
37:01 know-it-all customers. Then maybe a fast
37:04 food frenzy where warfers handle
37:06 outrageous orders. All right, now these
37:09 should be written, we hope.
37:16 Uh, let me see. Uh, RIP, show me all 20
37:22 ideas.
37:25 So, we'll do that with with actual
37:27 typing.
37:28 The McDonald ones are a hoot. Yep.
37:31 Workplace parody. Fast food frenzy.
37:34 Okay, we're going to stop it. We're
37:35 going to say, um, I want you to use,
37:41 um, actual brands.
37:44 Um, why, oh, don't we also include
37:50 blockbuster
37:52 video
37:55 video
37:57 um, as if it's still in business
38:02 as one of them.
38:04 That's a good one, actually.
38:07 All right.
38:13 Starbucks snap. We're not going to do
38:14 that. McDonald's meltdown. Trader Joe's
38:16 trauma customer. Do you sell emotional
38:19 support hummus? Only if you promise not
38:21 to cry into it again. Best Buy
38:24 breakdown. I want a TV that's smart but
38:25 not smarter than me. Then you want a
38:27 mirror, sir.
38:36 Oh, Apple store would be a good one. Why
38:39 can't I plug my old charger in because
38:41 we decided love is conditional? That's
38:43 good. CVS chaos. Whole Foods War Zone.
38:47 Is the Celery organic like emotionally?
38:49 It cried when it was picked. Is that
38:51 emotional enough? All right. These are
38:53 good. Good. Lululemon
38:56 REI.
39:00 Uh, okay. So, so now we're going to do
39:04 um
39:08 give me
39:11 your
39:13 your two favoriteites
39:19 uh as well as Blockbuster
39:27 and why Here
39:31 are my two favorite picks. Starbucks
39:33 snap. No, sorry.
39:35 Sorry. No Starbucks
39:40 already been done. And no
39:44 McDonald's.
39:50 Perfect. Let's clear the field.
39:53 GameStop grudge. Can I return this open
39:56 game? Sure. You want me to refund your
39:58 dignity, too? All right.
40:02 Whole Food Fors Whole Foods War Zone
40:05 Blockbuster Blowout. Do you guys stream
40:08 now? No. We suffer. Would you like V VHS
40:12 or pain?
40:18 Okay. Now. Okay. Let's see. Okay. Now
40:23 write
40:27 three
40:32 VO3 prompts
40:36 for each concept
40:41 following
40:43 the best
40:46 practices
40:48 you showed before.
40:59 Why is it putting it in a code block?
41:03 Stop. Wait.
41:06 Just write these as
41:11 formatted
41:13 text in a canvas.
41:17 I don't want these as code.
41:21 You ninny.
41:48 Um,
41:51 we can't use quote marks
41:56 around
41:58 the
42:00 dialogue
42:02 since VO3 Three
42:05 turns
42:08 the words into crappy
42:12 subtitles.
42:15 Rewrite these.
42:18 Rewrite. Rewrite
42:20 these.
42:22 Okay.
42:24 Is it rewriting? What are we doing? We
42:26 doing something. Did we just hang? Did
42:27 we just crash the browser? Oh, no. We're
42:30 in a canvas. That's why. Okay.
42:35 Okay. Whole Foods,
42:38 Blockbuster.
42:45 Um,
42:50 let's see. Let's close this canvas.
42:54 I'm going to say,
42:58 do me a favor
43:01 and go research
43:04 VO3
43:07 best prompt
43:10 practices.
43:12 Your prompts suck.
43:16 Please
43:18 go learn how to do them right.
43:24 If you're new here, if you're new to AI,
43:26 if you
43:29 if you think AI is a genius, AI is not a
43:32 genius. AI is an amplifier. So, what
43:35 it's going to do is it's going to give
43:36 you back what you it's going to reflect
43:38 back at you what you give it. So, if you
43:40 give it something shitty, it's going to
43:41 give you something shitty back. Um,
43:47 so let's have it go off and learn how to
43:49 do a VO3 prompt.
43:56 Okay. What I discovered? Okay. Right.
43:59 There you go.
44:01 Rewritten blockbuster prompt. A scene
44:03 camera audience. There we go. Now we're
44:06 talking. Dialogue
44:09 clerk. No, just suffering.
44:26 Isn't
44:29 the
44:31 dialogue in that scene incomplete?
44:42 Yeah, missed the setup line. Yeah, I
44:44 know. That's why I said it.
44:46 [Laughter]
44:48 Okay. So, now we have a decent prompt.
44:54 So, I'm going to copy this. I'm going to
44:55 say, okay, now go rewrite the canvas
45:03 um
45:05 prompts
45:08 and dialogue. Okay. So, it's going to go
45:12 off and it's going to fix our [ __ ]
45:14 Great. Go do that. So, now we're gonna
45:16 go to VO3.
45:20 Is this it? No.
45:23 Where did I put it? Maybe if you didn't
45:26 have so many tabs open. Shut up. Um,
45:30 flow.google.
45:35 [Music]
45:40 Flow.
45:43 What is going on with Chrome flowgoogle?
45:48 Okay. And I'm not sharing the right
45:50 thing. I know. I know. Uh oh. You know
45:52 what I did, Brandon?
45:56 I I had a different uh I had a different
46:00 uh browser up where where the the
46:02 Post-it note used to be. So, if you're
46:04 wondering why I'm ignoring your Post-it
46:05 notes, because I wasn't looking at any,
46:08 but I am now I am now one with the
46:10 Post-it note. Two nights in a row.
46:11 Shush. Shush.
46:14 Calm down, everyone. Just calm down.
46:16 Wait, what are we doing? We're doing
46:18 flow. You You're going to need to hear
46:20 this.
46:22 So, that's good. Okay.
46:25 All right. Right. So, we got flow. Let's
46:28 jump in. Create with flow.
46:32 Now, if you've never used flow before,
46:35 flow be cool. I think I I'm in this one.
46:39 Yes.
46:42 Flo be cool.
46:45 Um, congrats whoever just gave me a
46:47 donut. Thank you very much. Starf
46:49 fighter kicking out the donuts and the
46:52 cowboy hats. Very nice. Lit up hat and
46:55 mustache. Very, very nice. Thank you so
46:57 much. Okay, so we're going to go new
46:59 project.
47:02 And if you haven't seen this stuff,
47:04 wait, did this do another Oh my god.
47:10 Thank you for the heart. Hang on, I
47:12 gotta change my sharing again because it
47:14 popped open a new window because why put
47:17 something in the same window when you
47:18 put it in a different window? Especially
47:20 if your name's Google and you require
47:22 people to log out and log back in
47:25 because you've got some proprietary beef
47:28 with Apple.
47:29 I have an idea. Why don't you dump your
47:32 corporate battle on your consumers?
47:37 Isn't that a fantastic idea? Because
47:40 maybe if enough consumers [ __ ] about it
47:42 like I do five nights a week,
47:46 Apple will all of a sudden concede to
47:49 whatever the [ __ ] they didn't concede to
47:51 23 years ago, right?
47:55 [ __ ] the customer. All right. Fantastic.
47:59 Bob, tell him what he's won. Oh, you've
48:01 not won anything.
48:03 They took all your data and guess what?
48:05 They leaked it all today.
48:08 I don't I don't I don't know who leaked
48:10 it. I don't know if it was Google, but
48:12 it was someone. They leaked all the
48:14 Apple passwords, all the Google
48:15 passwords.
48:19 Okay. Uh uh uh uh. Let's see if this is
48:23 going to do someone else's brand. So,
48:26 when you're in V3, there's our prompt.
48:28 We dumped our prompt in there.
48:31 We're going to flip over to
48:34 V3 fast
48:38 because it's it's a fifth of the price
48:40 of V3 not fast and it's still pretty
48:42 good. And so we're going to generate
48:44 this. So this should in theory we're
48:48 going to get a Blockbuster video
48:52 in 2025
48:56 that's reminiscent of that other thing.
49:00 Viral viral meme-based video V3 video
49:03 snippets. What are you doing today?
49:05 Yeah, that's it.
49:07 Uh, what is it? Horos
49:10 Crinker or or Georges? I assume it's
49:14 Horhas.
49:17 AI is about to kick you and insult you.
49:19 I know. You know what's going to happen,
49:21 Rick, is because chat GBT has memory and
49:23 all that [ __ ] It's going to be fine.
49:27 It's not It's just going to ignore me.
49:29 until three years from now when when uh
49:33 when I bring a a humanoid robot home to
49:36 help with family chores and clean [ __ ]
49:38 up. Uh
49:40 I'm going to give it some lip and just
49:42 from the back of my head I'm gonna it's
49:44 going to be like you know when your
49:45 mother used to flick the back of your
49:46 head in the grocery store when you were
49:48 being an [ __ ] Flick. I'm gonna get
49:50 one of those from [ __ ] robot. She's
49:53 gonna slap me around a little.
49:56 How did you generate the prompt? So,
49:58 it's pronounced George. Ah, okay. Cool.
50:00 Um, how we generated the prompt was we
50:04 went to um
50:07 we went to to chat GPT. I basically just
50:10 had chat GPT go look for any articles
50:13 that were written about these
50:16 saucy barista videos that came out today
50:19 and someone put like spun up a channel
50:22 overnight basically and then like one of
50:24 the videos has six million views
50:26 already.
50:28 Um, and so I just basically said, "Take
50:30 that concept and give me 20 different
50:32 variations of it." And so it did. And
50:35 then I had it go research. Oh, this is
50:37 fantastic. This is fantastic. Um, let's
50:41 go full screen.
50:45 I hope the dialogue is good.
50:50 Do you guys stream now? No, just
50:52 suffering VHS or pain.
50:58 Do you guys stream now? No, just
51:00 suffering VHS or pain.
51:03 Serving? Isn't it serving? Just serving
51:06 VHS or pain.
51:10 Wait, where?
51:18 Let's see. Hang on.
51:23 We'll do a different one.
51:26 Oh, just suffering. Oh, I see. No, just
51:29 suffering. VHS or pain? That's bad
51:33 dialogue.
51:34 Can I stream something? We suffer. VHS
51:37 or pain.
51:40 Oh, these are all shitty. Hey. Hey. All
51:44 of your dialogue sucks and is the same
51:50 for each
51:52 prompt.
51:54 Get creative.
51:57 Write funny
51:59 jokes
52:03 and make each one unique.
52:14 Um, Louie
52:17 CK is watching and will fire me if your
52:23 jokes suck as bad as they do.
52:31 Are you in the right model? Also, tab if
52:33 you want to. Yeah. No, I was in the
52:35 right model. It was just bad writing.
52:39 Uh, I'll come back to flow in a second
52:41 here. I just need to get me some.
52:46 Let's do something else. We'll do
52:48 GameStop grudge. Can I return this
52:50 opened game? Sure. Let me grab grab the
52:53 dignity refund form. Oh, wait. We're all
52:55 out. That's pretty funny. All right.
52:58 Setting dialogue. Okay.
53:06 Oh, the other thing is you'll notice
53:09 you'll notice when
53:11 I make content on this channel,
53:14 I don't really give too many shits about
53:16 it.
53:19 It's just like if it if it's generated,
53:22 then it's good enough.
53:24 If I were Do you guys stream now? No,
53:27 just suffer. If if I were doing this for
53:31 real, like if I wanted to spin up one of
53:32 these channels, I'd actually pay
53:33 attention to the scripts being written.
53:36 And what I'd probably do is go write my
53:38 own prompt or like have Chat GBT help me
53:41 write it, but then edit it, make it
53:42 really good, and then tell it, hey, this
53:45 is now your base prompt. Do all the
53:47 other prompts like this one. Um, but I'm
53:49 just doing it quick. I don't really we
53:51 don't it doesn't really matter what we
53:52 do here. But anyway, like
53:58 I think it is actually a viable
54:02 business opportunity for probably the
54:05 next year
54:08 to
54:09 scour Twitter and scour whatever for
54:13 something funny
54:15 immediately knock off some variation of
54:17 it, right? Like the ASMR cutting through
54:19 glass fruit. You could do ASMR, cutting
54:23 through glass animals, whatever, right?
54:26 Um,
54:28 and then go spin up a channel. Now, is
54:32 it stupid and mindless and contributing
54:36 to the shitification of the internet? It
54:39 is,
54:42 but if you're creative,
54:46 so there's there's actually two ways to
54:47 do this. One is to be a derivative hack
54:50 and just steal other people's concepts.
54:52 The other one is to come up with
54:54 interesting ideas on your own like this.
54:56 Yeah, meme money. Exactly. I did some of
55:00 those today. There you go.
55:03 Hang on a sec.
55:05 Um, who's that? Gavin Biddy Honeycomb.
55:09 What anti- AI people call slop. Well,
55:12 you can do it as slop, but I So, so
55:15 here's the thing, Gavin. I don't think
55:20 I don't
55:21 I I don't think just because you do
55:24 something fast it necessarily has to be
55:27 bad.
55:29 I think if you're a clever prompter,
55:31 right, if you have your [ __ ] together
55:33 with prompting
55:36 and you can quickly just go grab
55:38 something, copy the style, copy the
55:41 basic idea, iterate 20 different
55:44 concepts, pick the best one, and then
55:46 refine the [ __ ] out of that over the
55:48 course of an hour, and then spin up a
55:50 channel and throw the [ __ ] out there and
55:52 see if it sticks.
55:54 I mean, honest to God, you know what you
55:56 could do? I don't know if Jim Ross is in
55:57 the in the audience tonight.
56:00 But what you could do is every morning
56:04 wake up and in a single hour go find
56:09 something,
56:11 replicate it, spin up a Tik Tok channel
56:14 or spin up an Instagram channel and you
56:17 know put put five videos out.
56:28 memes. Richard Dawkins memes par
56:31 parasitize the human mind in much the
56:34 same way that viruses infect and host
56:37 cells. Um, do you think they paid paid
56:40 to boost the content? It's possible. I I
56:42 don't know. I just thought I thought
56:45 what was
56:49 when I saw the first
56:54 Hang on. We're going to come back to
56:55 this one. When I saw the first um
57:00 meme this morning, the barista meme,
57:04 I thought
57:10 I thought it was this video, not this,
57:13 not [ __ ] Mr. Wonderful. Hang on. Um,
57:16 let's see. Um,
57:19 barista,
57:21 why do you talk like that? Remember this
57:25 video
57:28 with what's his name? Louderm milk.
57:33 What can I get you? I'd like a large
57:36 coffee. Okay. So, hot coffee. Hot
57:39 coffee. Okay. Room for cream. Totally.
57:42 Leave room for cream.
57:44 Why are you talking like that? Why are
57:46 you talking like that? Because this is
57:48 my voice. This is my voice. No, it's
57:50 not. I heard you talking a minute ago. I
57:52 know you don't talk like that. Neither
57:53 do you. Because nobody actually talks
57:54 like this. You choose to talk like this.
57:57 And today I chose to talk like this.
57:59 It's pretty [ __ ] annoying, isn't it?
58:01 Why you so rude, man? Just stop doing
58:03 that. I can't help it. This is my voice.
58:05 No, it's not. It's an affectation that
58:09 annoying teenagers and rich people use
58:11 to sound like they don't give a [ __ ]
58:13 Except you work in a coffee shop, so I
58:15 know you're not rich. So, when I saw the
58:17 when I saw the meme this morning, I
58:19 thought it was this campaign. I mean, I
58:20 thought it was it was like I thought
58:22 someone took this clip and then swapped
58:26 out the people and had them say more
58:28 swear things and then and then I and
58:33 then I saw another one and then I saw
58:34 another one. I'm like, "Oh, no. This is
58:36 like an original campaign." Like what's
58:39 amazing is maybe this is what's amazing
58:41 about it.
58:43 You could create an advertising campaign
58:47 in an hour
58:50 and if it's clever enough and funny
58:52 enough,
58:54 you could get millions of views for a
58:56 product, right? like the the one the one
58:58 of the Bigfoot videos, the Bigfoot's
59:01 talking about a specific product like a
59:03 sunscreen product
59:05 and that video's gotten a [ __ ] ton of
59:08 views and they've sun sold a ton of that
59:11 con that that product. So, so there's
59:14 just there's something here. All right,
59:16 let's go. Let's go back and and uh
59:22 see if we got any action.
59:29 Where?
59:31 Oh, here we are. This is it, right?
59:33 Yeah,
59:39 sure. Let me grab the dignity refund
59:41 form.
59:43 Oh, wait. We're all out.
59:46 That's it. No. No.
59:50 Damn it.
59:53 Dialogue. Customer. Can I return this
59:55 game? No subtitles. Employee. All right,
59:59 we got to try it again. Um, setting
1:00:06 visual cue. I'm going to put real
1:00:09 GameStop. Real GameStop logo blurred in
1:00:12 the background.
1:00:13 Maybe they're pull Maybe they've pulled
1:00:15 out brands already
1:00:28 [Music]
1:00:35 echo chamber regurgitation. Now that's
1:00:38 funny or ironic. Uh someone was saying
1:00:41 something about
1:00:43 Oh, people. Something about creativity.
1:00:47 Gavin.
1:00:49 God damn it.
1:00:52 What anti- AAI people called slop. I
1:00:54 know I'm just saying
1:00:56 you talk about the shitification of the
1:00:58 internet and the antiA people call it
1:01:01 slop and its creativity. People claim AI
1:01:04 images look horrible and ugly, but that
1:01:08 started from the presupposition that
1:01:10 they already have hating AI. So anything
1:01:12 it generates is ugly. It doesn't matter
1:01:15 that what we create with it is, "Oh,
1:01:17 that's not AI slop slop, but a shoe
1:01:20 stuck on the wall is art." Well, so
1:01:24 here's the deal.
1:01:26 A lot of the tropes that you hear anti-
1:01:29 ai people say are the tropes that they
1:01:33 were saying two years ago.
1:01:36 Oh, the images all look like [ __ ]
1:01:38 That's when the we had pretzel hands and
1:01:40 all the skin looked like claymation
1:01:43 and everything looked like [ __ ]
1:01:46 Stuff doesn't look like that anymore. In
1:01:48 fact, after this, we're going to flip
1:01:49 over to MidJourney and I'm going to just
1:01:50 play play with MidJourney for a little
1:01:52 bit more. Um, let's see if we actually
1:01:55 got some dialogue here. That would be
1:01:57 nice. Come on. Give me some dialogue.
1:01:59 Give me some dialogue. Oh, we're still
1:02:01 we're still we're still building. Can I
1:02:04 return this open game? Come on. Come on.
1:02:09 Sure. Let me grab the dignity. Can I
1:02:11 return this open game?
1:02:13 Sure. Let me grab the dignity refund
1:02:15 form. Oh, wait. We're all out.
1:02:18 Can I return this open game? All right.
1:02:21 So, I've got bad prompts here because I
1:02:23 didn't pay attention
1:02:26 because I didn't pay attention to my
1:02:27 prompts. I had shitty prompts. And it
1:02:30 does not look like they're doing um
1:02:32 really good branding because that's not
1:02:33 the real GameStop logo, but whatever.
1:02:37 All right. Um,
1:02:40 and to them still.
1:02:42 Wait. And to them it still does. To them
1:02:45 it's telling. Yeah, I know. Well,
1:02:57 yeah. I I don't know that
1:03:03 there's only so much we can do to bring
1:03:05 people along, right?
1:03:08 Like it's my contention that
1:03:12 whatever opinion you have about AI,
1:03:15 I completely respect
1:03:18 unless you've never tried it.
1:03:22 That's the [ __ ] that pisses me off is
1:03:24 that people say, "Oh, it's evil. It does
1:03:26 this. it does that. They haven't learned
1:03:28 what it is. They haven't played with it.
1:03:30 They haven't learned all the 362 other
1:03:33 things it can do other than steel art,
1:03:36 right? They're looking at it through
1:03:37 this microscopic lens because the
1:03:40 arguments that I've been in around this
1:03:42 stuff, they're like, "Well, you know,
1:03:44 AI, what what AI should be being used
1:03:46 for is to cure diseases,
1:03:49 you And my response to that was, "Oh, so
1:03:51 you're you're okay you're okay with
1:03:55 these large language models being
1:03:57 trained on all of these scientists work
1:04:01 without attributing it to them, but
1:04:04 you're not okay with art." Is that is
1:04:06 that what I'm understand? They're like,
1:04:07 "Oh, well that's different. That's
1:04:09 science." Oh, okay. Right.
1:04:12 So there is some some uh there there are
1:04:15 some intellectual uh fallacies and and
1:04:18 logic loopholes in in a lot of the
1:04:21 arguments. Your phone should be your
1:04:23 smartphone should be curing diseases.
1:04:25 Exactly. Yeah. It's like arguing with a
1:04:27 Trump supporter. It really is. Um okay.
1:04:31 Um
1:04:33 let's go to let's go to let's go to
1:04:36 midjourney. Let's go to midjourney. So
1:04:38 here's the thing about midjourney.
1:04:46 On the surface,
1:04:48 I should not give a [ __ ] about this. And
1:04:50 I'll tell you why.
1:04:54 Ever since VO3 came out,
1:04:57 the fact that we've got video models
1:04:59 that
1:05:01 don't have sound or don't have dialogue,
1:05:03 I think is a massive waste of time. I
1:05:06 think I think it's [ __ ] stupid. I
1:05:08 like I hate the state-of-the-art now
1:05:11 because I want everything to be at least
1:05:13 as good as VO3.
1:05:15 Um, and they're not. And so yesterday,
1:05:18 MidJourney comes out with their with
1:05:20 their video model. And it's integrated
1:05:23 into the platform in this really subtle
1:05:25 way and in this really powerful way in
1:05:27 my opinion.
1:05:30 And on the surface, I should hate it
1:05:32 because it doesn't have any sound. You
1:05:33 can't make sound effects with this. You
1:05:35 can't do dialogue.
1:05:37 But it's really [ __ ] good. All right,
1:05:39 so
1:05:41 let's go play with it. And and one of
1:05:44 the things that that you should know
1:05:48 is that
1:05:52 um let me can I move this up a little
1:05:54 bit and this up a little bit to give a
1:05:57 bit more tick tock black screen. There
1:06:00 you go. All right.
1:06:07 Um,
1:06:11 so yesterday, Midjourney put out a post
1:06:15 that said over the next year,
1:06:19 they they they basically said, listen,
1:06:22 the whole point of MidJourney was never
1:06:25 about image generation.
1:06:28 The ultimate end goal of MidJourney is
1:06:30 to be able to generate 3D worlds with
1:06:34 objects and characters in real time.
1:06:39 And so what they said is in order to
1:06:41 build worlds with characters and objects
1:06:44 and continuity and all that stuff,
1:06:47 there's this lattering that has to
1:06:49 happen. And at the bottom of the ladder
1:06:51 is image generation. and and with the
1:06:54 exception of text generation, midjourney
1:06:57 is simply the best
1:07:00 image fidelity, like just artistic image
1:07:03 quality,
1:07:05 and then they're going to layer on um
1:07:08 video on top of that, 3D objects and
1:07:12 worlds on top of that, and then real
1:07:14 time rendering on top of that. And so
1:07:17 they said over the course of the next
1:07:19 year they're going to drop these as
1:07:20 discrete features
1:07:23 and then probably sometime next year
1:07:25 they're going to roll them all into a
1:07:27 single platform where we're going to be
1:07:29 able to talk worlds into existence.
1:07:31 Yeah. Question.
1:07:33 Have you tried seed dance? I have not.
1:07:39 I don't even know what it is.
1:07:42 Never even heard of it. Seed dance.
1:07:46 Seed dance 1.0. Oh. Oh, the bite dance
1:07:49 thing. Is it out yet? I don't think it's
1:07:51 out yet. Is it
1:07:57 full technical report?
1:08:01 Oh, wait. Why am I Did I share this
1:08:04 wrong? Oh, I shared it wrong. God damn
1:08:06 it.
1:08:09 Oh, but this is a video. I don't think
1:08:12 this is this is out.
1:08:17 vibe coding.
1:08:20 Uh, is there a Tik Tok pin I missed?
1:08:24 Yeah, I don't think this is out. The
1:08:25 seed dance thing. It looks good.
1:08:29 Um, let's go watch the video. Hang on.
1:08:40 Mid Journeys videos are better than
1:08:42 Sora. They are. Okay, let's see. Hey
1:08:45 Kyle, just after we watch this, there's
1:08:47 a model arena. You have to do a quick
1:08:49 authentication check, but then they show
1:08:51 you some more examples and let you play
1:08:53 with the models.
1:08:55 Where? Here it was. Seed dance.
1:08:58 Yeah, it's right next to the full
1:09:00 technical report. Okay.
1:09:04 [Music]
1:09:41 [Music]
1:09:49 [Music]
1:09:59 Model arena.
1:10:10 Why did that not switch? Oh, because I
1:10:12 did sound.
1:10:14 The I'm telling you, I I I I'm gonna
1:10:17 lose my ever flipping mind with this tab
1:10:20 sharing nonsense.
1:10:23 What kind of work do you do? Oh, AI
1:10:24 enthusiast. Oh, I see what you're
1:10:26 saying. AI enthusiast. Uh
1:10:31 AI video.
1:10:34 All right.
1:10:48 If you click back, um I didn't have to
1:10:50 do that. Click the X.
1:10:53 Yeah. Um, just say click the check box
1:10:56 for terms and conditions and say get
1:10:58 started. And it took me right in.
1:11:01 Got it.
1:11:03 Oh, I see. And here's a a prompt.
1:11:06 Actually, let's just paste in what? No,
1:11:10 give it a second to generate. And you're
1:11:12 adding a you're adding a comment. So,
1:11:15 give give it a second to generate. If
1:11:17 you scroll up, it's going to tell you
1:11:18 what it's generating. Uh
1:11:25 uh
1:11:31 on the left hand side you've got uh text
1:11:33 to video, image to video, a couple
1:11:36 different things.
1:11:38 Ah
1:11:41 text to image, text to video, add a
1:11:44 comment.
1:11:45 Yeah, they're going to make you create
1:11:47 an account if you want to actually do
1:11:48 anything with it, but all right. Well,
1:11:50 whatever. All right, I'm gonna go away
1:11:52 from this. All right, let me let me
1:11:53 change my sharing again.
1:11:57 Uh, let's get back to MidJourney
1:12:01 because I because I've been thinking
1:12:02 about MidJourney today.
1:12:05 Um Vicky Vicky shared.
1:12:14 Let's see. Where did she share it?
1:12:18 Yeah, here.
1:12:22 Can I make this bigger? I can. You can
1:12:25 see that, right?
1:12:29 Look at the animation on this
1:12:34 black bar. There you go. Watch this.
1:12:44 Can I loop this? Probably not. More
1:12:48 options.
1:12:52 Let's play it again.
1:12:57 So
1:12:59 this is a midjourney
1:13:01 animation
1:13:04 from that image.
1:13:08 And what's amazing about it is
1:13:13 if we now understand that so so what
1:13:16 what OpenAI talked about a year and a
1:13:19 half ago when they announced Sora and
1:13:22 everyone lost their mind was they were
1:13:23 talking about
1:13:26 a model that understands and simulates
1:13:29 the real world.
1:13:31 What I've noticed about midjourney
1:13:34 animating images is that it seems to
1:13:37 have
1:13:40 just a deeper understanding of the base
1:13:42 image than other imagetovideo models.
1:13:46 The other imagetovideo models isolate
1:13:48 objects and move them. But there's
1:13:51 there's a sophistication in what
1:13:53 midjourney is doing that I I find really
1:13:56 compelling. So, let's go find
1:14:00 let's go find some interesting images
1:14:03 that have got geometry in them. Like
1:14:06 here's an interesting one.
1:14:09 Oh, and I might not be able to do too
1:14:10 many of these because I lost my fast
1:14:12 hours last night. So, we'll do low
1:14:14 animation.
1:14:16 All right. I guess I'm still good for
1:14:18 some some fast. Dynamic animation. Oops.
1:14:22 You've run out of fast hours. Buy one
1:14:25 fast hour for
1:14:27 four bucks.
1:14:36 Thank you.
1:14:39 Purchase confirming.
1:14:44 Confirmed. Okay. Did it go? Did it
1:14:47 generate?
1:14:51 No. Okay.
1:14:53 organize.
1:14:55 So, let's do this one
1:14:58 motion.
1:15:04 Oh, I've got some really cool ones to
1:15:06 show
1:15:08 from last night that we didn't get to.
1:15:11 Actually, these are probably really good
1:15:13 because these have a lot going on in
1:15:15 them.
1:15:20 Let's do that one.
1:15:22 We'll do it with low motion,
1:15:25 by the way. So, what I'm doing right
1:15:27 now,
1:15:31 if you're in midjourney, once you create
1:15:33 an image, you've always had these
1:15:35 options to to like zoom out, zoom in,
1:15:38 upscale, do variations of it. And the
1:15:41 video generation is this lower right
1:15:43 hand quadrant. So you can it there's an
1:15:47 auto version low motion high motion and
1:15:50 then there's a manual version low motion
1:15:52 high mo motion manual allows you to
1:15:56 prompt the motion auto just sort of
1:15:59 figures it out from the image and then
1:16:01 low and high are lower or higher you
1:16:04 know dynamic movement. Um so it's really
1:16:08 simple but you can do things you can
1:16:11 extend videos. So if I turn this into a
1:16:13 video, which I just did, I can then go
1:16:18 watch what it generated and if I like
1:16:20 it, I can then extend it for another 8
1:16:22 seconds and then I can extend it or 4
1:16:24 seconds or eight sec. Wait, four
1:16:27 seconds. It does eight seconds initially
1:16:30 and then you do 4 second extensions four
1:16:32 times or something like that.
1:16:35 I thought you could do unlimited with
1:16:37 slow. I don't think
1:16:40 you can do any video with slow.
1:16:44 I'm struggling. Okay, this is good. Good
1:16:46 question. I'm struggling to trust AI
1:16:48 overall
1:16:50 and with generative AI,
1:16:53 I find it mediocre and it always
1:16:56 requires cleanup. Great, great, great
1:16:59 question. And in fact, I'm in the middle
1:17:02 of writing a book called um I'm at the
1:17:06 beginning of writing a book called Feed
1:17:08 Your Prompt.
1:17:09 Um
1:17:11 that talks a lot about this and I I'll
1:17:13 give you a little bit of a little bit of
1:17:15 insight here.
1:17:17 One of the biggest unlocks for AI,
1:17:21 there's a couple of them.
1:17:23 One is you should never assume that what
1:17:26 AI is going to give you is finished or
1:17:30 even close to finished. You should you
1:17:32 should assume with AI that if you get to
1:17:34 80% of what you wanted, that's a really
1:17:37 good start.
1:17:39 And then you can refine it further from
1:17:41 there. But here's the here's the unlock.
1:17:43 Here's the thing that I'm talking about
1:17:44 in my book.
1:17:47 A lot of people ask the question,"How do
1:17:49 I make the most out of AI?" or "How do I
1:17:51 get the most out of AI?"
1:17:53 That's not what you should be asking.
1:17:54 What you should be asking is, "How do I
1:17:56 get the most out of myself using AI?"
1:18:01 The reason AI is giving you crap is
1:18:04 because AI is not a genius. And and
1:18:07 listen, the the the tech bros in San
1:18:10 Francisco
1:18:12 are douchy douche knuckles. douchy
1:18:16 McDouch knuckles and they are absolutely
1:18:19 promising that these tools are geniuses
1:18:22 and these tools are going to replace
1:18:23 everyone in their work and these tools
1:18:25 are everything that they're cracked up
1:18:28 to be.
1:18:29 AI is not a genius, but what AI is, what
1:18:32 it's a really good what it what it does
1:18:35 really well is it's an amplifier
1:18:38 and a reflector. So it will reflect back
1:18:41 to you whatever you give it and it'll
1:18:44 amplify it, right? So if you give it
1:18:47 generic cold crap like just write me a
1:18:50 LinkedIn post, it will give you back a
1:18:52 LinkedIn post and it will write it
1:18:55 really quick and you'll be like well
1:18:56 that was fast. And then you read it and
1:18:58 it's a piece of [ __ ] because it's it's
1:19:01 just amplifying this kind of dead input.
1:19:03 It's giving you dead output. The key to
1:19:06 unlock AI
1:19:08 is put yourself at the center of it
1:19:11 rather than think of AI as this tool off
1:19:13 to the side that's going to do the work
1:19:14 for you. Think of AI like a
1:19:16 collaborator. So the first thing you do
1:19:19 when you sit down is say is say, "Okay,
1:19:22 what do I want out of this interaction?
1:19:24 Who's the audience I'm writing for? Like
1:19:26 what am I trying am I trying to write a
1:19:27 LinkedIn article or trying to do a
1:19:29 Twitter post or am I trying to write a
1:19:31 poem for my girlfriend?"
1:19:33 Take a moment before you prompt and and
1:19:35 really think about well what do I want
1:19:37 here and what would good look like and
1:19:40 who's the audience and and what would
1:19:41 they like and what's my tone of voice
1:19:44 and then start to think about okay what
1:19:45 are the raw ingredients of the prompt
1:19:48 well okay it's I I want the output's
1:19:50 going to be a poem and the input's going
1:19:52 to be you put yourself in the situation
1:19:55 so here's what my girlfriend looks like
1:19:57 here's what she means to me she's got a
1:19:59 dog named Snuffles
1:20:02 Right. And here's what Snuffles looks
1:20:04 like.
1:20:07 And and that's the first step is really
1:20:09 think about what it is you want. Craft a
1:20:12 prompt that has all those elements in it
1:20:14 that are personal to you and then put
1:20:16 that in there. And what it's going to
1:20:18 then do is it's going to reflect that
1:20:19 back at you. It's going to take your raw
1:20:21 unformed ideas and it's going to amplify
1:20:24 them. And you're going to be like, "Oh
1:20:26 my god, that's actually pretty good."
1:20:28 And that's when you you've started the
1:20:31 process.
1:20:33 You're not done. This is not about
1:20:35 oneshot prompting. You can
1:20:39 use these tools like a like eating
1:20:41 Skittles, right? Just give it a prompt.
1:20:44 It squirts out [ __ ] You share that with
1:20:45 the world. That's like eating Skittles.
1:20:48 But if you want it to be a meal, you
1:20:51 craft your prompt. You put it in there.
1:20:52 Gives you something back. And now the
1:20:54 process has begun. Now it becomes an
1:20:56 iteration where like you're the chef and
1:20:58 and AI is the sue chef and it's
1:21:00 constantly going yes chef yes chef and
1:21:03 you're like no I don't like that part I
1:21:04 don't like that part I want this I want
1:21:06 that you refine it you mold it and then
1:21:09 at the end of that process I promise you
1:21:12 something will come out of it that will
1:21:14 blow you the [ __ ] away
1:21:18 because it's not about the tool it's
1:21:20 about you
1:21:21 and if you can get clear about what you
1:21:23 want and then you get into this
1:21:25 collaborative dialogue with it. It it
1:21:28 does it it is
1:21:32 it is impossible to describe the feeling
1:21:36 when you have your first what I call
1:21:38 Kevin Mallister moment, right? The
1:21:40 moment from Home Alone, that moment.
1:21:44 So anyway, hope that helps. All right,
1:21:47 let's go back to midjourney.
1:21:52 [Music]
1:21:56 the security eagle on Tik Tok. How do we
1:21:58 build our own model like a small
1:22:00 language model? Um I don't know.
1:22:11 Yeah, possibly. Um
1:22:15 there are people within the AI salon
1:22:16 that know that stuff. Um
1:22:20 they're they're so so go to the
1:22:22 salon.ai, click join our community
1:22:26 um and just start intro introduce
1:22:28 yourself and start connecting with
1:22:30 people there. There are people in there
1:22:31 that that do that. I'm not one of them.
1:22:34 Um all right, let's go look at these
1:22:36 images we created.
1:22:45 Okay, that one's good up until some
1:22:47 point.
1:22:51 Where is salon.ai on Tik Tok? Uh, it's
1:22:56 not right now, but if you go to the AI
1:22:58 learning lab, which you're on, um, I'm
1:23:01 the founder of both this and the AI
1:23:03 salon. So, this there's a lot of overlap
1:23:05 in this community and the AI salon
1:23:07 community.
1:23:09 Um,
1:23:12 okay.
1:23:18 So, it's interesting. These are all
1:23:19 zoomins,
1:23:23 huh?
1:23:27 Fascinating.
1:23:33 Kind of cool.
1:23:42 Huh.
1:23:47 Wow. All right. Um, let me show you some
1:23:50 of the things we created last night
1:23:53 because there's some really good ones in
1:23:54 here. All right. This is where we
1:23:56 stopped last night, right? We were
1:23:58 looking at the the smoke. The smoke's
1:23:59 not quite right, but it's better than
1:24:01 most smoke I've seen. Most AI smoke is
1:24:03 really, really bad.
1:24:08 I thought this one was beautiful. Like
1:24:10 the the frosted glass,
1:24:13 the images reflecting and seeing through
1:24:15 it. Jim Ross made some cells today of
1:24:19 generic storage promo video I made in in
1:24:22 V3. So fun. That's awesome. That's
1:24:25 really cool.
1:24:36 This is pretty gorgeous.
1:24:57 I mean, again, if you look at the
1:25:01 if you look at the details of the
1:25:03 strands and all the all the hairs,
1:25:07 most AI video models really [ __ ] up
1:25:10 those little details. And this one
1:25:12 doesn't. This one doesn't seem to [ __ ]
1:25:15 up those details.
1:25:20 There's still weirdness and like double
1:25:23 moves and
1:25:25 but but none of these I did I did any I
1:25:28 didn't do any prompting with these.
1:25:31 These were these were all literally just
1:25:33 images that I said low or high movement.
1:25:36 This this one I think is stunning.
1:25:42 Like just the the masking alone, if you
1:25:44 wanted to take
1:25:47 first of all, if you wanted to create
1:25:48 this image, you'd have to be a really
1:25:51 good illustrator or a really good
1:25:52 compositor, Photoshopper, right?
1:25:56 And then if you wanted to animate it,
1:25:58 right, you got to mask everything and
1:26:00 layer it and all that. Or you just say
1:26:02 low low animation.
1:26:11 This one I love where it looks like it's
1:26:13 carousel.
1:26:17 Crazy.
1:26:20 Oh, look.
1:26:23 We got a puppet.
1:26:24 [Laughter]
1:26:27 I didn't see the puppet earlier. That's
1:26:29 really funny.
1:26:31 Oh, this is the I think this is the high
1:26:33 movement one.
1:26:36 That's really cool.
1:26:39 Wow.
1:26:42 Love that one.
1:26:47 [Laughter]
1:26:51 This one's gorgeous.
1:26:56 What software site is this? So, so we're
1:26:59 on So, as of yesterday, Midjourney added
1:27:03 animation video to their site. What's
1:27:07 fascinating is
1:27:12 what most sites have done that have both
1:27:15 image and video is that they launch them
1:27:18 as kind of two separate tools. You've
1:27:20 got an image tool and then you've got
1:27:22 this video tool and you can create a
1:27:25 bunch of images over here and then you
1:27:26 kind of move them over to the video side
1:27:28 and you make videos.
1:27:30 Midjourneys is weird. It's much more
1:27:32 subtle than that. Midjourney is still an
1:27:36 image generation site and then once you
1:27:39 have an image generated then you
1:27:40 basically just hit animate and it makes
1:27:43 these these really remarkable videos.
1:27:45 It's it's quite a stunning model because
1:27:48 again I think that it is a it is
1:27:51 understanding it is inferring out of
1:27:54 this image kind of a profound
1:27:57 understanding of the elements and depth
1:27:59 and meaning of what's in this image to
1:28:03 be able to pull these things out with
1:28:04 such sophistication.
1:28:08 I don't know which one has it, but
1:28:10 apparently he can't generate.
1:28:13 Apparently he has. Who's he?
1:28:18 I was going to say you can if you have
1:28:21 pro or mega plan with unlimited gym
1:28:25 generations.
1:28:26 Oh yeah, that's right. They they did
1:28:28 relax mode. You can do videos in relax
1:28:31 mode if you pay. If you're one of the
1:28:32 high tiers. Look at this one. This this
1:28:36 im first of all, this image I love, but
1:28:39 look at this animation.
1:28:42 I put this one up on the Tik Tok today.
1:28:49 Absolutely crazy.
1:29:01 There. It's like they're on a boat.
1:29:07 But I mean, if you take that animation
1:29:09 style, just that animation style alone
1:29:15 and just do,
1:29:18 you know, go make 50 or 100 images and
1:29:22 then go make, you know, 30 or 40 videos
1:29:25 based on the best of those images. This
1:29:28 model's really, really good. It's
1:29:29 insane.
1:29:31 Like, there's weird jankiness. Like,
1:29:32 look at
1:29:35 Not that one.
1:29:38 This one, I think. Yeah. So, this one,
1:29:41 if you watch this guy's hat when this
1:29:43 video starts, it's part of his hat and
1:29:46 then all of a sudden he walks out of it
1:29:47 and it's part of someone else's hat. But
1:29:49 the other versions of it don't have that
1:29:51 mistake in them.
1:29:54 And because it's giving you four
1:29:56 different videos, every time you hit
1:29:58 give me a video, you get four variations
1:30:01 of the video.
1:30:05 Like here's another one where look like
1:30:07 it
1:30:09 it's keeping the cityscape intact,
1:30:13 right? It's generating the cityscape
1:30:15 that wasn't there before. It recognizes
1:30:18 that that's way in the distance.
1:30:22 Just bonkers.
1:30:34 Oh, yeah. Here's my uh my 70s muscle car
1:30:37 in an abandoned factory. It's my little
1:30:39 70 Chevy Impala.
1:30:42 Now, the prompt here was for this thing
1:30:44 to roll away, which for most of them it
1:30:46 didn't. But if you look at that one, the
1:30:49 wheels are not horrible. The the physics
1:30:51 on the wheels are not horrible.
1:30:55 All
1:31:11 right, let's get some more images. All
1:31:13 right. Oh, it's 10 o'clock, so I'm gonna
1:31:15 I'm gonna wrap it up here pretty soon,
1:31:17 but let's go let's go find a couple of
1:31:18 more cool images. Go to organize. Any
1:31:21 any thoughts, questions on this stuff?
1:31:25 Where's the smoke coming from the
1:31:26 rubber? Yeah, that's not there.
1:31:30 Oh, I bet these are going to be cool.
1:31:36 Let's do this one. We'll do high motion
1:31:39 on that one. And we'll do low motion.
1:31:42 Low motion on this one.
1:31:45 All right.
1:31:54 I mean, these are really cool animation
1:31:56 styles.
1:32:00 That one's probably pretty cool. Let's
1:32:02 do
1:32:05 Let's do low on that one.
1:32:10 Let's do high on this one.
1:32:18 Let's do low on this one.
1:32:22 Let's do high on this one.
1:32:28 High on that one.
1:32:34 All right.
1:32:51 We'll do low on that.
1:33:15 Oh, that's this is one like
1:33:18 Vicky showed. Let's do low animation on
1:33:21 that.
1:33:30 These things were so beautiful
1:33:34 animated.
1:33:36 So cool.
1:33:42 Oh, this is a good one. Yeah, let's do
1:33:45 this one. We'll do high motion on this
1:33:47 one.
1:33:49 We'll do Let's do low motion on that
1:33:51 one.
1:33:56 Yeah. Hey, Peter. Yeah, we're going to
1:33:58 need you to go on ahead and come on in
1:34:00 on Saturday. Yeah. Oh, and about that.
1:34:13 Wow.
1:34:20 I know you're getting seasick. That's
1:34:21 not my problem. Right.
1:34:25 Question. What's going on with the
1:34:27 mastermind this month? We've got Leah
1:34:32 just started her um her midjourney.
1:34:36 She's doing three days a week midjourney
1:34:39 where if you if you do it live, it's
1:34:41 open to everyone, but if you want the
1:34:43 archives of the recordings and if you
1:34:45 want to join the group, um that's part
1:34:47 of the mastermind. Um Claire Jacobs is
1:34:51 doing a four-part
1:34:54 series on um
1:34:59 it's like self-improvement through
1:35:01 prompting or something like that. It's
1:35:03 it's it's we just we just approved it
1:35:06 today. It's going to be a four-week
1:35:08 series that's going to be in the
1:35:10 mastermind. So that's going on. Um
1:35:15 there's a bunch of meetings in there.
1:35:16 Tomorrow we've got the founder meeting
1:35:18 with Leah and myself and other
1:35:20 mastermind members. And the meeting last
1:35:22 week was really good. So, so just being
1:35:24 being in that group's good. We're also
1:35:27 adding um a chat area just for
1:35:30 mastermind members. Right now you can
1:35:32 chat within different spaces in
1:35:33 mastermind but um we're creating just
1:35:37 kind of a general chat area which by the
1:35:38 way if you don't know what I'm talking
1:35:39 about in the AI salon
1:35:42 um
1:35:44 there's a subscription
1:35:46 uh sub community called mastermind where
1:35:49 the kind of basically where the deeper
1:35:51 dive stuff is where you can build
1:35:53 relationships and things like that. Um,
1:35:55 and for the rest of 2025, it's only 20
1:35:57 bucks a month. So, it's going to go up
1:35:59 to, I think, $49.95,
1:36:01 uh, starting in 2026, but if you if you
1:36:04 join in 2025, you become a founding
1:36:07 mastermind member. Um, and the whole
1:36:09 idea is that the early the early members
1:36:13 of the mastermind actually get to shape
1:36:15 what it's going to become, right?
1:36:17 Because it's a new thing. And and you
1:36:18 know, the the AI salon has been free for
1:36:21 two and a half years. And so the the
1:36:24 core AI salon remains free. This is this
1:36:27 sort of deeper dive for people that want
1:36:29 to like really commit and get to know
1:36:30 one another and things like that. So
1:36:32 it's super exciting. Um I made some
1:36:35 videos out of these images before. So
1:36:37 anyway, go go check out the mastermind.
1:36:39 I made some videos out of these images
1:36:42 before and
1:36:44 um they came out really good. So let's
1:36:47 try them
1:36:49 here.
1:36:53 do that low.
1:36:56 Um, I'm already out of fast hours. How
1:36:59 does that work?
1:37:04 An hour? I thought it was an hour of
1:37:06 render time. I guess it's an hour of
1:37:08 play time.
1:37:10 Oh my god.
1:37:12 This is I got to get in the midjourney
1:37:14 creator program. This is [ __ ] for the
1:37:16 birds. Don't do it.
1:37:20 It's time to go. All right, let's go
1:37:22 look at what we uh what we did.
1:37:30 Yeah, I had a feeling these were going
1:37:32 to be cool.
1:37:52 Whoa,
1:37:59 that's crazy.
1:38:08 These are really good.
1:38:11 These are really good. I'm telling you,
1:38:13 there's something special about this
1:38:14 model.
1:38:16 There really is something special about
1:38:18 it.
1:38:23 Like, it just gets it.
1:38:26 I guess it didn't get that he was dead.
1:39:04 Okay, that's weird. Oh, no, it's not.
1:39:07 It's got some continuity,
1:39:10 huh?
1:39:44 Wow, that's cool.
1:40:05 Yeah. Look at the animation on this
1:40:07 thing on the left. The the it's tracing
1:40:09 the electrical circuits.
1:40:21 Oh yeah. Look at this. Come on. You
1:40:26 [ __ ] kidding me?
1:40:37 Wow.
1:40:43 the fact that it makes the blueprint
1:40:45 that was behind the drawing. Like that's
1:40:48 the that's the [ __ ] to me that says
1:40:50 these guys are going to absolutely be
1:40:52 able to pull off consistent world
1:40:54 building.
1:41:02 [Laughter]
1:41:05 Wow.
1:41:14 I think I'm out. That's it. I think
1:41:16 that's all we get, people. I bought an
1:41:19 hour and that's all we got. That is
1:41:22 [ __ ]
1:41:24 [ __ ]
1:41:26 All right. Well, I'm probably going to
1:41:28 have to I'm gonna have to probably
1:41:29 change my subscription.
1:41:32 Oh, that's the other thing. the the the
1:41:34 uh the front page of Midjourney is all
1:41:37 videos now.
1:41:39 Like it's all videos, right? This used
1:41:41 to just be all images.
1:41:44 They're crazy.
1:42:02 Danielle, that is nuts. I know. Gonna
1:42:04 have to get pro. I know.
1:42:08 That one's cool. God damn, these things
1:42:12 are gorgeous.
1:42:17 You know what's funny about Midjourney
1:42:19 is all these different image models came
1:42:22 out and every time an image model comes
1:42:25 out, they're like, "Oh, it's a
1:42:28 midjourney killer." And nothing yet has
1:42:32 been a midjourney killer. And all these
1:42:34 video models are out and there's just
1:42:36 something about
1:42:38 the midjourney video model that's just
1:42:42 different. It's aesthetically better.
1:42:46 It's like artistically.
1:42:49 Look at that.
1:42:52 That [ __ ] looks like cinema. All the
1:42:55 videos are 8 seconds long. You can
1:42:56 extend them up to 16, I think. Or wait,
1:42:59 you can do it four times. So, it's eight
1:43:02 bass plus three fours.
1:43:07 20 seconds. I think you're going to send
1:43:09 them to something like that.
1:43:18 Look at that.
1:43:36 All right.
1:43:38 Out of here. Okay. Tomorrow it's So
1:43:40 tomorrow's Friday. So office hours, if
1:43:43 if you enjoy this, come to office hours
1:43:45 tomorrow morning at 11:00 a.m. Mountain
1:43:47 time on LinkedIn. You go to my profile
1:43:49 on LinkedIn, Kyle Shannon, and uh just
1:43:54 go to the events, any of the events in
1:43:57 there. Um they all have the same URL.
1:44:00 We've used the same URL for
1:44:05 um so come to that.
1:44:08 Office hours is a really special group
1:44:11 of people. Um, and then right after
1:44:14 office hours, we've got the AI salon
1:44:17 mastermind um, founder hangout with Leah
1:44:21 Fast and myself and Vicki probably and I
1:44:24 don't know if Brandon will be there,
1:44:25 whoever's going to be there from uh,
1:44:27 from Salon Leadership, but come hang out
1:44:29 with the the leadership of the salon and
1:44:31 other mastermind members. It was really
1:44:33 good last week. So, let's let's keep
1:44:35 that rolling. And then tomorrow night
1:44:37 will be Friday night date night. It's
1:44:38 not an art night, so I think it should
1:44:40 just be uh should be normal time
1:44:43 tomorrow night. Um yeah, and I think
1:44:47 that's it. Um I hope you had fun
1:44:48 tonight. We did some good stuff and
1:44:51 yeah, I continue to be kind of blown
1:44:53 away by the the midjourney thing. I am
1:44:55 really cranky about the cost of it, but
1:44:58 I'll figure it out and I'll let you know
1:45:00 what I come up with. like cuz I think at
1:45:04 the next tier of subscription which I
1:45:06 don't know what that jump is with
1:45:08 midjourney
1:45:10 because I think I'm paying 10 bucks now
1:45:11 and I think it jumps to 40 which I hope
1:45:14 it that's not the case.
1:45:17 Manage subscription
1:45:20 standard plan
1:45:27 change plan.
1:45:29 All right goes from 10 to 30. That's not
1:45:32 bad.
1:45:34 15 hours of fast generations.
1:45:38 Unlimited relaxed image generations. But
1:45:41 I wonder what that means for video. Oh,
1:45:44 $60 for unlimited
1:45:47 relaxed image and video generations. Oh,
1:45:50 this sucks.
1:45:58 30 to 60. Yeah. All right. Well,
1:46:00 whatever.
1:46:03 So, come to um yeah, come to the [ __ ]
1:46:07 tomorrow. All right. Hope you had fun
1:46:09 tonight and uh I'm going to get on out
1:46:12 of here. You have yourselves a fantastic
1:46:14 evening and we'll see you on the
1:46:16 tomorrow.