5/21/2025 - Exploring Google's V3 and Its Impact on Filmmaking

Description

Chapters

Transcript

0:00 Alrighty, good
0:02 people. Fantastic. Fantastic. Bob. Bob.
0:07 Hey, Bob. Hey, Bob. Bob. Bob. Bob. What
0:12 have they won? Oh, Bob. Welcome
0:16 everybody. Welcome to the AI Learning
0:18 Lab
0:20 live from the road in scenic Princeton,
0:24 New
0:27 Jersey. Oh man.
0:31 Um, we'll let people get coming on in
0:34 here. If you're new here, welcome,
0:36 welcome. If you're not new here, what
0:39 are you doing here? Why do you do this
0:41 to
0:43 yourself? What are you thinking? What do
0:46 you think? These are bad life choices
0:48 that you're
0:51 making. What'd you do Wednesday night? I
0:54 don't want to talk about
0:55 it. Well, no. What'd you do? Did you go
0:57 out and get drunk?
0:59 No. Do you binge watch Handmmaid Sale?
1:03 No. What' you do? I watch some old guy
1:06 on Tik Tok. What? Why would you do
1:10 that? There's no better
1:18 hobby. AI learning lab coming in. Hot.
1:22 Hot. It's hot people.
1:25 Um uh I got a little chance to explore X
1:30 today in between meetings um and look at
1:33 what was going
1:35 on and it seems like the big hit from
1:39 Google IO is VO3 which is the the new
1:43 video generating model.
1:45 Um
1:49 and it the so a couple of things about
1:54 VO3 that I that I have
1:56 learned. Um they announced
2:00 VO3 and they announced Flow which is a
2:04 video editing tool which does video and
2:06 image
2:07 generation. If you pay the 25 bucks a
2:10 month for the Gemini Pro
2:13 plan,
2:15 then your flow
2:18 editor only uses V2. It doesn't use V3.
2:22 So, it won't do sound generation.
2:25 In order to use
2:27 VO3, you need to be on the Gemini Ultra
2:31 plan, which is 250 bucks a month. And it
2:36 only includes 83 clips of video per
2:41 month. Like that's the limit. 83. If
2:44 you've ever made any AI video, you can
2:47 burn through 833 clips just trying to
2:50 get one clip right now. This model's
2:53 better, but still it's it's a it's very
2:56 expensive at this point still. So, it's
2:58 very limited. Um, there's a part of me
3:01 that wants to give up my
3:05 um GPT Pro account and get this
3:10 thing,
3:12 but it just it seems like a lot of
3:15 money. Still not home. Thought your
3:17 podcast was in your home office today.
3:19 Suspicious. Yeah, the podcast today was
3:22 a pre-recorded session, which is why I
3:24 was talking about going to Washington in
3:27 the future tense during the podcast
3:29 today. Um, we try to do those things so
3:33 they look like they're live and that's I
3:35 was in there actually live chatting. Um,
3:38 but I didn't know if I would be um if I
3:41 would be available for it, so we decided
3:43 to use a pre-anned one. Um, so yes. So I
3:47 am still on the road. I am
3:50 It's I got two more days. I got
3:53 tomorrow. We got one more meeting in the
3:55 morning. Then I have my agency.com 30th
3:58 anniversary in the afternoon or the
4:00 evening. Uh and then I go to sleep and
4:02 then I get up and leave. Leave for for
4:05 back home and I'm ready. I'm ready. My
4:09 hair is ready. Here's my hair is
4:13 corporate
4:14 [Laughter]
4:19 puffy. Oh man. Good. I don't have to
4:23 quit the detective agency just yet. No,
4:25 you're good. You're you're you're doing
4:27 great. Um, so good news tonight. We have
4:31 producer Brandon uh in the in the
4:34 house. So, how many people do we have
4:38 over there in the streamyard? Producer
4:40 Brandon. So, we got 10 folks over there.
4:42 We got 14 folks here in the Tik Tok
4:44 live. Tik Tok. Um, so do me a favor. If
4:47 you're in Tik Tok, share this.
4:49 Normally, the the purpose for me singing
4:52 songs with a dog at the beginning of
4:55 these shows is to just give people time
4:57 to, you know, come in and make their
5:00 nachos and make an excuse for their wife
5:03 to go watch a Tik Tok channel of an old
5:05 man, you know, things like that. Um, it
5:08 takes a little time to to get the
5:10 logistics right, you
5:12 know. So, we'll just wrap here for a
5:15 bit, but if you would just uh share that
5:17 source camp and I actually did have
5:20 nachos tonight.
5:21 [Laughter]
5:24 Perfect. Since you're Wait, since you're
5:27 set
5:28 up, you can't come up. Why
5:32 not? You could come up.
5:37 I don't get
5:40 it. All that's missing is the hot
5:42 pocket. See, you're up. Hey, no, I I
5:46 said since you were set up to do the
5:49 screen share, I could come up if you
5:51 like. Oh, could come up. Yes. So, so
5:56 yeah, let's let's have you come up. And
5:57 why don't we even look? We'll give you
5:59 the the big spotlight. Look at this.
6:01 Watch this. Wow. Well,
6:05 hi everybody. All right, so and I'm in
6:09 the little window here and you can see
6:10 me there. Okay, so this is producer
6:12 Brandon. Hi, producer Brandon. Um,
6:14 hello. Oh, you know you know what's so
6:15 funny? I realized last night we were
6:17 doing we were do I don't know if you
6:18 know this. I do these Tik Tok live
6:20 things like like regularly like nightly.
6:22 So last night I was doing one of these
6:25 and uh Steo, he's from Australia. He he
6:29 whipped up this image of like a guy in
6:31 glasses with a beard and a Google IO
6:34 hat. And I am so full of myself. I
6:38 thought, "Oh, that was really sweet of
6:39 him to do that of me. Wonder why he put
6:42 a beard on me." I thought I thought he
6:44 did it for me and just malprompted it
6:46 and I realized, "No, no, you were
6:48 getting the loving." So you were at
6:50 Google IO and so tell us tell us tell us
6:54 everything. What did you what did you
6:56 learn and love and hate and like I also
6:58 want to hear the stuff like where where
7:00 was the sort of [ __ ] smok and
7:01 mirrors versus you know where was the
7:03 where was the good stuff. Yeah. Uh so it
7:06 was my first Google IO. I've been uh
7:09 I've had the privilege of doing trusted
7:11 testing for Google for a couple of
7:12 different products. I think that's why I
7:14 got invited uh because it is an invite
7:16 only event. But I I got to go. It was my
7:19 first IO ever. I'm very happy and
7:22 privileged to be there. uh flew out
7:24 there on on very little sleep. Um but uh
7:29 it was a it was in Mountain View,
7:31 California for anybody who is familiar
7:33 with the area. Beautiful venue. Uh big
7:36 campus. Um security was tight. There
7:38 were a ton of people there. It was
7:40 really like getting into like an
7:44 Oh, we just lost Brandon. Wonder if
7:47 that's Wait, hang on. You just froze for
7:50 a second. So, it was like it was very
7:52 close to like being in an amusement park
7:54 or something, you said. Yeah. Yeah.
7:56 Security was tight. There were like
7:57 5,000 people there. Um, and it was it
8:01 was crazy the amount of lines. I was not
8:04 anticipating having to stand in so many
8:06 lines. Uh, yeah. Yeah. You're at a
8:08 marketing event. You're having to stand
8:10 in line. That's pretty funny. Um, so the
8:13 longest line I stood in was to get into
8:15 the sandbox area. um which was an hour
8:19 uh to get into the area where you could
8:22 play with all the toys. Oh, that's
8:24 interesting. Um uh Danielle said, "Did
8:26 you did they pay for your flight in
8:28 hotel?" Uh they did not. Um but they got
8:31 me in for free. Um and I got to play
8:34 with all the tools. Um but uh No, no.
8:38 Kyle, did you not get my um invoice for
8:40 that the the flight? Well, we Well,
8:43 we've we've routed it to the approval
8:45 committee. Gotcha. Yeah. Um, so, you
8:49 know, the good, um, obviously you've
8:51 already touched on VO. Um, VO was center
8:54 stage. Um, they did a bunch of custom
8:56 stuff. It looks really impressive. They
9:00 were showing on the demo stage. You
9:02 could, you know, play around with it a
9:04 little bit and prompt it. And, uh, it
9:07 they I got some prompts from from people
9:11 of how to prompt it. So, we talked a
9:13 little bit with their product team.
9:15 What's really cool about VO and and Flow
9:18 in particular um is that they worked
9:21 with creatives and people from the
9:24 creative industry to build that product.
9:27 So it's not just like that that seemed
9:29 clear. I was saying last night it was
9:30 like um
9:32 flow right now in its current state is
9:35 like the minimum viable product. What
9:37 you know for if you were a creative and
9:39 you needed something that just did the
9:40 basics of of what you would need it to
9:42 do to make a film like that's what it
9:44 is. like there's no frill on it
9:46 whatsoever. But the fact that you can
9:48 trim the clips and extend them and
9:50 stitch a bunch of them together is
9:51 pretty remarkable. Yeah. And it was
9:54 definitely an event that was geared
9:56 towards uh keeping the investors happy
9:59 compared to previous IO's where they
10:01 they really focused on the consumer
10:03 grade stuff. They were really focused on
10:05 making sure that the investors knew that
10:08 they were not taking second place to um
10:12 Open AAI in anything. They were
10:14 benchmarking above the rest. They were
10:17 injecting Gemini into every aspect of
10:20 the Google ex ecosystem and experience.
10:23 Um and it was in a lot of ways a very um
10:28 standard Google IO. Uh they had Whimo
10:30 was represented there, the self-driving
10:32 car arm. Um they had uh the robotics
10:36 arm, Willow. Uh I got to play they had
10:39 an arcade game set up where you could
10:40 race a quantum computer. Everybody lost
10:43 against the quantum computer. Um,
10:47 and this is this is a training for, you
10:51 know, accepting that that we are no
10:53 longer the the smartest thing on the
10:54 planet. But then but then they did other
10:57 things, too, like they they helped me
10:58 with my basketball shot. They had a
11:00 basketball court set up that was uh they
11:02 had cameras trained on it, so you take
11:03 your shot and then they tell you what
11:05 you did wrong and um in real time. Um, I
11:08 I think for me personally, the the two
11:11 coolest demos that I saw uh were the XR
11:15 glasses. We got to try on the XR
11:17 glasses, which they actually retrofitted
11:19 to my prescription, which I appreciated.
11:21 They they had a little lens crafter
11:23 station set up there where they they
11:25 measured my prescription and they
11:26 retrofitted that the uh prototype frames
11:29 so I could actually see out of them.
11:32 Um, brilliant. Yeah. and they had a a
11:36 the So, if you're familiar with the Meta
11:37 Ray Bands, they're very very similar in
11:39 design. They've got the camera, they've
11:41 got the touch pad on the side, the open
11:43 ear um uh phone conducting speakers, and
11:48 um the thing that made them stand out
11:51 differently is the heads up display uh
11:54 which is very similar to 11 years ago
11:56 when they launched the original Google
11:58 Glass having that heads-up display in
12:00 your frame. It was very unobtrusive,
12:03 unobtrusive. It was very natural to have
12:05 it right there in your frame with but it
12:08 was translucent so it didn't distract
12:11 me. I didn't go have to go crosseyed
12:13 looking at it. Um, and what it would do
12:16 is like, yeah, with one eye, and if you
12:19 read a text message and it was far
12:20 enough out that you could, it was
12:22 projected far enough out in the latent
12:24 space in front of you, uh, that you
12:26 could just kind of look ahead of
12:28 yourself and see the words popping up
12:30 like a teleprompter or like, you know,
12:32 just in in in the space in front of you.
12:35 If you've ever played around with the uh
12:38 Google Maps Live view where you flip
12:40 your camera up while you're walking and
12:41 you can see it's giving you the
12:43 directions, it very much felt like that
12:46 experience. Um where it was just a heads
12:49 up display. Um the other I could see
12:53 developing a a social awkwardness app
12:56 which just pops up what to say at a
12:59 party. So someone comes up to you and
13:02 Hello. Hello. There we go. Yes. Nope.
13:05 Sorry. There you are. Hotel Wi-Fi again.
13:07 Yeah, I was just saying a killer app
13:10 would be the social awkwardness app that
13:13 just you walk into a party and someone
13:14 says hi to you and it tells you what to
13:16 say next.
13:17 Yeah, absolutely. Um, yeah, and I saw
13:21 Val's comment. They they don't have
13:22 transition lenses, but I'm I'm imagining
13:24 that they could because the the
13:27 technology is all based on the hardware
13:30 that's outside of the glance. So once
13:32 you can get them in prescription lenses,
13:34 once they go live with their
13:36 relationship with Warp Parker,
13:38 um I'm imagining you can get the lenses
13:41 in any style and then the the prism just
13:44 lays over on top. The other aspect that
13:47 was really Winston just said um the CEO
13:52 Sundar Pachai and truly all of those IO
13:55 speakers presented really well. I
13:57 enjoyed it, which I thought that too.
13:58 Like it was it didn't have the
14:00 awkwardness of many of the like even
14:03 some of the early Google things they
14:04 brought up, you know, just people that
14:05 were a little too engineery. And I
14:07 thought the I thought the speakers they
14:09 had it it it was starting to feel
14:11 reminiscent of an Apple event.
14:14 Yeah, they um and they definitely pushed
14:17 the idea that Gemini can help you code
14:19 without knowing how to code. One of the
14:21 preview things that they did before the
14:24 keynote started was they came up and
14:27 vibe coded a a selfie app that leveraged
14:30 your camera um and would allow you to
14:34 take a picture and then it gave you a
14:36 little poem and then they put the then
14:38 they prompted Gemini to put a QR code
14:42 and make it into a QR code which I
14:44 didn't know Gemini can generate QR
14:46 codes. Fun fact, Gemini can generate QR
14:48 codes. Uh, and so then they put it up on
14:51 screen, you everyone could do it. Um,
14:55 and along those similar lines with
14:57 coding, uh, one of the most impressive
15:00 demos that didn't make the the headlines
15:02 was a back of the napkin sketch to app.
15:06 So, they literally had a doodle on a
15:08 napkin of an idea of an app, took a
15:10 picture of it, and told Gemini to make
15:13 the thing, and it in canvas mode, made a
15:18 viable version of the app. It's very
15:20 similar to uh Perplexity and and or I'm
15:23 sorry um Claude artifacts and um having
15:28 it just there, but it could do
15:29 infographics. You could have it make a a
15:31 a web applet and it it just everybody
15:36 there was super friendly, uh super
15:39 welcoming. Um, I was talking what to one
15:42 of the PMs about how they're have all of
15:46 the different models and how all the
15:48 different things and how it's really
15:50 confusing for folks like us who are in
15:52 the weeds with this. But every different
15:56 model serves its own purpose. If you're
15:58 a developer for games, you want to go
16:01 here. If you're a scientist, you want to
16:04 go here. If you want to tweak your own
16:06 model and talk to dolphins, you can
16:08 download Gemma. like everything has its
16:10 place, but underneath all of it is the
16:14 architecture of Gemini and and their
16:16 base models that really are the the
16:19 infrastructure and architecture that
16:21 power all of those different modalities
16:23 and it was just really fascinating to
16:25 see. Yeah, part of the part of the
16:27 challenge with that argument is that
16:30 Gemini is also the name of you know it's
16:34 gemini.google.com, right? Which is the
16:36 front end, right? So it's like like
16:39 similar to chat GBT. It's GPT3 Si or
16:42 four sitting underneath chat GPT sitting
16:45 underneath, you know, and it's it's
16:47 just, you know, they they continue to
16:50 um I I don't know like damage their own
16:55 ease of use by being imprecise with
16:58 their naming, which you know, you'd
17:00 think as engineers they could engineer
17:02 their way out of it, but no. Yeah.
17:05 Because I guess engineers are just, you
17:07 know, let's do some sequential numbering
17:08 and that's all we need. That's that's
17:10 clear. And a lot of the newer things
17:13 that they've started rolling out uh in
17:16 testing um are available at labs.google.
17:20 So if you want to try on some of the new
17:22 experiments, as they call them,
17:24 including trying on clothes virtually,
17:26 like taking a picture of yourself and
17:28 seeing what you look like in a denim
17:29 jacket, uh you could do that uh now with
17:32 Labs.google on the Google app. Oh, okay.
17:35 Cool. Was there was there anything that
17:37 you felt was sort of vapor wary or or
17:42 you know overly marketed underdelivered?
17:46 I think they leaned very heavily on the
17:49 potential of VO3
17:52 um and kind of buried the lead on the
17:54 fact that you only got 83 generations
17:57 for $250 a month, right? They really
18:01 highlighted the capability of it without
18:04 mentioning that part on stage. Well,
18:06 it's, you know, it's and we'll I'm going
18:09 to go look at a bunch of examples I saw
18:11 on X because the the examples are are
18:13 truly impressive. And what's impressive
18:16 about them is not that they do video
18:18 well. It's not that they do audio and
18:21 sound effects and voice and this and
18:22 that. It's that it does it all together,
18:25 right? It's like making it's making
18:28 essentially fully edited scenes, right,
18:31 with sound effects, acting, character
18:34 design, action design, you know, SFX,
18:37 just all of it rolled in there. And, you
18:41 know, if I think about if I were a
18:44 filmmaker, 250 bucks a month for
18:47 something like that is nothing if I want
18:49 like if I want to truly make a film,
18:51 right?
18:53 Um but yeah, but it's, you know, it's
18:56 funny because we're
19:01 we're coming out
19:04 of SAS world, right? Where, you know,
19:07 SAS products are whatever they are, a
19:09 nickel a month, five bucks a month, 10,
19:11 20 bucks a month. And we're starting to
19:14 see subscription prices like, you know,
19:16 there's $20, $250, $2,000. at some point
19:20 we'll probably see
19:21 $20,000. And the immediate response, you
19:24 know, coming out of like, well, you
19:26 know, I'm willing to pay you five bucks
19:27 a month for your thing. The concept of
19:30 paying 20 grand a month is ridiculous.
19:32 But if that 20 grand buys you, you know,
19:35 $2 million worth of, you know, PhD level
19:39 labor, right, that is going to work 247.
19:42 Well, all of a sudden that's that's a
19:44 bargain, right? And I, you know, that
19:46 that gap right now is is just I think
19:50 it's something that we're going to have
19:51 to get used to. And as early adopters,
19:54 we're going to likely overpay for, you
19:57 know, underdelivering tech for a while.
19:59 But yeah, and you know, the other thing
20:01 that was really interesting to me is the
20:05 makeup of the audience. So, I was
20:07 expecting being at a developer
20:09 conference that there would be a bunch
20:10 of tech nerds like me that, you know,
20:13 were just trying to glean off of what
20:15 Google was doing. But there were some
20:17 logos on the name badges that I mean,
20:20 Ford was there, uh, PayPal was there,
20:24 Walmart had like four people that I was
20:26 talking to in line and, you know,
20:28 they're trying to figure out how they
20:30 can deploy all of this technology on the
20:32 back end at the enterprise level. And
20:35 so, you know, for anybody who thinks
20:37 that, you know, the enterprise world is
20:39 owned by AWS and Azure, uh, and
20:42 Microsoft, um, there were a lot of folks
20:44 that are really paying attention to what
20:46 Google is doing on the AI side. Yeah.
20:49 Yeah. know and and it's listen I think
20:52 that I mean I think you were I think you
20:55 were part of the learning lab in the
20:56 salon back when when um when they did
21:00 this IO conference in 23 like right
21:03 after chat GPT came out and it was an
21:05 embarrassment like it was bad and then
21:08 even the one in 24 last year was okay
21:12 but it was mostly vaporware right like a
21:14 lot of stuff was announced and very
21:16 little of it shipped the thing I was
21:18 struck by yesterday was they'd make an
21:21 announcement and for the most part they
21:23 said available today, available today,
21:25 available today and it's like even if
21:28 they were sort of rolling things out
21:29 this week, it was like you'll see it
21:31 this week, not you know at some
21:33 indiscriminate point in the future. So I
21:36 feel like they they have definitely
21:37 turned the ship and you know it's funny
21:41 like on X a lot of people are like ah
21:43 they're completely slaying open AI. It's
21:45 like, well, you know, like like they're
21:47 at least at parody, right? They're at
21:49 least in the conversation now. And two
21:51 years ago, quite frankly, they weren't,
21:53 right? It was it was really bad. So,
21:55 that's exciting to see. Oh, Pate was
21:57 biggest disappointment that Pate wasn't
21:59 there. That was the big biggest.
22:02 That's too bad.
22:04 But yeah, I mean, I'm I'm I'm working on
22:07 a uh an article for the the salon um
22:10 that that kind of highlights the stuff
22:12 that wasn't in the keynote and the the
22:14 experience of being at IO and all the
22:16 the backstage stuff. Um but yeah, it was
22:20 it was a lot of fun and um don't sleep
22:23 on Google. They're um they are at the
22:26 point now where uh Quinn might have some
22:28 competition. Yeah. Yeah. Oh, that's
22:31 that's the other thing is the the Gemini
22:32 Live. Like I don't I don't have it yet
22:35 on my iOS app but but that ability to do
22:38 live camera stuff like again like that's
22:40 like open a chat GBT is already doing it
22:43 but like again they're kind of at parody
22:45 they've got the natural voice they have
22:46 the live camera they have you know be
22:48 able you know you've got this live
22:50 visual assistant they've got VO2 which
22:53 or three which at this point is
22:55 significantly state-of-the-art and you
22:58 you know what's funny is I knew that it
23:01 was going to be big because It it wasn't
23:03 some like video product manager that,
23:06 you know, tweeted it. It was Demisabas,
23:08 right? The head of the co-CEO, I think,
23:10 of Google and the, you know, the founder
23:12 of DeepMind. He's the one that said, you
23:14 know, V2 is or V3 is is coming and it's
23:17 going to be really good. So, I had a
23:19 feeling it was going to be good, but it
23:20 it looks really really quite remarkable.
23:24 Yeah. I think I might have offended the
23:25 product manager for the robotics um
23:28 division though when I asked him how
23:30 soon they would be able to unload my
23:32 dishwasher. Uh
23:36 he did not say he's like well he goes
23:39 well you know maybe five 10 years but
23:41 you know we're we're we're thinking uh
23:43 for additional use cases and
23:45 applications that are um you know
23:48 bigger. Yeah. Yeah. Um, Brian Whitney
23:51 asks, "Is that all only with paid
23:53 Gemini?" So, with paid paid Gemini,
23:57 there's two levels. There's what's
23:58 called Gemini Pro, which is 25 bucks a
24:01 month, and then there's Gemini Ultra,
24:03 which is
24:04 $249.99. Um, and the $250 subscription
24:08 is the only subscription right now that
24:10 gets you access to VO3. Um, and you do
24:13 get a free trial. Uh, not a free trial.
24:15 Um, they do a free trial of Pro. And if
24:18 you want Ultra, you can get it for three
24:20 months at 50% off. Yeah. There you go.
24:24 So, for for 125 bucks, you get 83 video
24:27 clips per
24:29 month. So, you can do a grand total of
24:32 what is that? Uh, 249 249 video clips
24:36 for the next three months.
24:39 Don't use them all in the same 450
24:41 bucks. So, it's basically two bucks two
24:43 bucks a clip. So, it's it's still
24:45 pricey. But again, if you're upset about
24:48 that one is if you're doing commercial
24:50 film work, that's still a [ __ ]
24:52 bargain. And if if you're not doing
24:55 production film work, give it a year,
24:57 right? It'll be 10x or 20x or 100x
25:00 cheaper. So, and probably pretty
25:03 instant. In fact, I saw there was a uh
25:06 there was a a post today from Robert
25:08 Scoble that talked about he got demoed a
25:11 new piece of video technology that's
25:14 like it's
25:15 generating dynamic video scenes like in
25:19 real time like as you're describing it
25:22 it's making it and it's like you're not
25:23 you're not waiting for it to render. So
25:25 yeah, I think if if the if 2025 was the
25:29 year where video went from unusable to
25:31 usable, I think 2026 is where it goes to
25:34 magical, right? Where it's just like you
25:37 just talk films into existence.
25:39 Exactly. But yeah, happy to vibe comment
25:42 with folks in in the salon or in the in
25:45 the comments here. Um Okay, cool. We
25:47 don't we don't have our we don't have
25:48 our sticky I'm sticky note list tonight
25:50 because of the setup. Yeah, exactly. Uh,
25:53 we'll uh I'll just Yeah. So, I'll just
25:55 do my best. Yeah, exactly. Just I've got
25:58 I've got I can see Tik Tok comments
25:59 right in front of me because it's in
26:00 front of everything. But thanks for
26:02 popping up here, Brandon. Yeah, don't
26:04 forget your tabs and black bar. Yep.
26:06 Okay. Bye. Yeah. And I I don't know.
26:09 Just I think yell at me. Yell at me on
26:11 Yeah. Yell at me on Tik Tok will be
26:13 easier for me to see.
26:15 Okay. So, let's go to
26:22 Let's go to Twitter. Twitter. Okay. So,
26:27 so what I want to do now is Oh, I've got
26:30 to change my tabs. Speaking of make sure
26:32 your tabs are right. Why don't I start
26:33 out and do that
26:41 first? Okay. So, here's
26:45 um you all know I'm working on a
26:48 musical, right?
26:50 So, this is from Ethan
26:52 Mollik. The prompt is a big Broadway
26:56 musical about garlic bread with
26:58 elaborate costumes and a Sondheiml like
27:03 vibe. And then this is what it made.
27:10 My heart's true desire.
27:17 garlic bread. My heart's true desire.
27:20 [Music]
27:24 So, a couple of things. Holy
27:27 [ __ ] Right. Holy [ __ ] And then like
27:32 notice it's got so it starts out with
27:34 that
27:36 scene, then it cuts to him in a kind of
27:39 solo shot with his
27:41 bread. They do their thing. There was
27:44 something interesting here where I
27:46 noticed
27:48 garlic bread. So he's singing garlic
27:50 bread and if you look at the the mouths
27:53 behind him, they're not singing the same
27:55 words as him. So it's like the the
27:59 physics of the people behind him, it's
28:01 not just mimicking, you know, whatever
28:03 is the main subject. It's they're kind
28:05 of operating
28:06 independently. Then they go to this
28:08 shot, this wide shot of the whole set
28:11 with the whole cast. And then let's
28:13 let's go back like is it
28:17 consistent? Not 100%. Like I don't know
28:20 what that this arched thing is because
28:24 that's
28:25 not here. Oh yeah, it is. It's down
28:28 here. It is. So that's
28:32 consistent. And then we go
28:35 to them singing
28:37 together. I think that's where it ends.
28:41 So it's four shots. So it's four
28:45 shots. Set consistency, character
28:48 consistency, character
28:50 independence. It wrote a
28:53 song and the characters are singing
28:56 songs. Just to get Suno to make a song
29:01 where a man and a woman sing together,
29:03 much less a chorus, is an absolute
29:05 [ __ ] train wreck. So, I don't know
29:08 how many shots Ethan Mullik gave, but
29:16 [Music]
29:22 you know, crazy.
29:24 Okay, so let me show you this thing
29:29 now. Do I want to do this one? Yeah,
29:31 I'll do this one first. Okay.
29:36 Okay, so this was some dude just, you
29:39 know, these the the Twitter engagement
29:43 dudes that they just go out and get
29:45 everyone else's work and then
29:46 consolidate it. Um, so but this is one
29:49 of those 20 insane examples. Less than
29:52 24 hours ago, Google dropped V3,
29:55 blurring the line between reality and
29:58 AI. And video and audio quality is so
30:01 lifelike, it's unsettling. Here's 20
30:03 insane
30:05 examples. If you ever write tweets like
30:08 this, stop watching this channel. Thank
30:10 you. It's a little public service
30:12 announcement, but this guy actually did
30:15 us a favor because I don't need to go
30:16 find
30:19 So, I went to the zoo the other day and
30:22 all they had was one dog. It was a Shih
30:24 Tzu.
30:31 When we get in there, I want no
30:33 [ __ ] You stay on my six at all
30:35 times.
30:46 Stay sharp. These [ __ ] are nasty and
30:48 dangerous. Stay alert.
30:53 What the hell happened here? Where are
30:55 the bodies?
31:02 go.
31:03 [Music]
31:09 Okay. Now, to be clear, this is not
31:12 There's a couple of things going on
31:13 here. I think these clips were made by
31:16 VO3. This is a minute-long clip, so
31:18 they, you know, they obviously stitched
31:20 a bunch of these together. This also has
31:22 contiguous music across the clips. So,
31:25 the music here was added in post and
31:28 these clips were all stitched together.
31:30 But just looking at the opening scene
31:33 where you've got, you know, the car
31:35 driving, cut to the inside, you know,
31:38 you get him saying whatever he's saying.
31:40 The camera swings around to the driver,
31:42 then cuts back to him and it still looks
31:44 like him and he's still he might be in a
31:46 different seat in the car, but it's it's
31:49 pretty close, right? And then they cut
31:51 back and then you know so this is pretty
31:54 pretty remarkable stuff. Yay militarized
31:57 police Kyle. This might wait might be a
32:02 thing to do. What might be a thing to
32:04 do? Gemini can call the Google Assistant
32:06 through extensions now. I guess
32:09 everyone's advancing. Looks great. And I
32:12 like garlic bread.
32:18 What's worthy to get from IO? Well, this
32:21 is one thing we're looking at is V3 I
32:23 think is worthy to look at. Oh, this one
32:25 annoyed the [ __ ] out of me just because
32:27 I hate ASMR, but it did ASMR pretty well
32:30 because it annoyed me.
32:36 Hi, and welcome to the channel. God, it
32:40 makes me want to vomit. I have dysphonia
32:43 or whatever the [ __ ] it's
32:44 called. I hate ASMR so bad. Here's a
32:48 streamer. Look at his eyes. How pinned
32:50 open they are.
32:52 Okay, so this is how you start. You find
32:54 a tree and you just like, you know,
32:56 punch it
32:59 down. I mean, this these are coming out
33:02 of the model like
33:06 wholesale opera singer.
33:13 [Music]
33:15 Now, what I don't know is how well I
33:18 know in Flow you can extend things.
33:21 Could you keep
33:23 extending the
33:25 um the garlic song to make a full
33:28 Broadway musical? I don't that I don't
33:31 know. Misophonia. Yes, I have
33:33 misophonia. That one [ __ ] hate hate
33:36 that [ __ ]
33:39 See some opinions. I mean, man, the
33:42 acceleration is crazy. You look far,
33:45 step on the pedal, and you are there. I
33:48 feel safe with him in an SUV, and it
33:51 seems to be like the right type of car
33:53 for him. I think the range is only um
33:56 only going to get better. Sorry.
34:00 We don't want to drive gas cars anymore.
34:02 Yeah. No more gas cars.
34:05 You can see uh I'm kind of a kind of a
34:09 misfit here, but don't tell anyone I've
34:12 just bought an electric car. I think
34:13 it's really great for families and for
34:15 little I mean, it's bonkers. It's
34:17 [Laughter]
34:20 bonkers. Bonkers. It's blunkers, I tell
34:23 you. Oh, this is a good
34:31 one. Not a bad explosion.
34:39 Like the physics here are as good as
34:40 anything I've seen. And the fact that
34:42 they're doing full-on sound
34:45 mixing and and to be clear, it's not
34:48 sound mixing. So if you know how these
34:49 things work, it's right this this is a
34:52 model that's been trained on like, you
34:54 know, fully blown out not just video
34:57 clips, but probably movie clips. It' be
35:00 it'd be interesting to know what they
35:02 trained these on. What's that thing?
35:05 Machine earnings. I don't know what that
35:07 is. Okay. Um, the kids cooking show.
35:12 Elbow grease.
35:13 Hey, now this takes some elbow grease.
35:17 Hey, watch it, lady. I'm trying to rise
35:20 here.
35:23 Now this takes some elbow grease. Hey,
35:26 watch it.
35:29 The biker dude bought the electric car.
35:32 Notebook LM also quietly announced video
35:35 overviews coming soon. Yep, I saw that.
35:38 But, you know, they're not here, so that
35:40 doesn't matter as much. Think of the
35:42 money movie makers will save doing car
35:45 crashes and explosions. like every
35:48 single thing is it in I mean I I
35:53 continue to say this but I
35:56 think
35:57 because we're such
36:00 a like a media hungry society
36:06 um we know how movies are made we know
36:09 how expensive they
36:12 are um we know that you can't make a
36:15 movie without you
36:17 literally a, you know, a production crew
36:19 of 150 people or more, right? Thousands
36:25 sometimes. And all of a sudden, you can
36:28 just prompt something into existence.
36:31 And you know, the special effects people
36:34 are going to say, "Well, they're not as
36:35 good." And blah, blah, blah, and like
36:36 all that stuff, but this is back to a
36:38 source camp thing, right? And Ethan
36:40 Mllik thing. This is as bad as these
36:42 things are ever going to be. And and you
36:45 know the stuff that we're looking at
36:47 here is in the neighborhood of good
36:49 enough good enough for 90% of TV ads for
36:55 90% of narrative.
36:57 Now just like when when motion pictures
37:01 came out you could argue that we don't
37:05 need live theater to tell stories
37:08 anymore, right? But Broadway still
37:11 exists, right? the live theater
37:12 experience still exists. It's just a
37:15 much smaller percentage of the media mix
37:18 than it used to be, right? When you
37:20 didn't have film, you had to have live
37:22 performance. That that was just the way
37:24 it was. Um, AI generated movies are
37:28 going to be the same. There are still
37:29 going to be some niche filmmakers that
37:32 make films the oldfashioned way with,
37:34 you know, humans and sets and all that
37:36 sort of [ __ ] Um, but it's it's going to
37:41 be it's going to be a luxury to do it
37:44 that way is is where I where I think
37:47 this goes. But you know what's cool is
37:50 you've got all the people that might be
37:52 displaced from being able to do one
37:55 component of a movie. Like maybe you
37:57 just do motion graphics and and you can
38:00 contribute to movies in that way. But
38:03 now with with someone with that much
38:06 depth of knowledge, imagine the kind of
38:08 film you might make and how different
38:11 that's going to be than someone who's
38:13 not trained as deeply as you are. So I
38:15 think what we'll start to see is films
38:17 coming out from people who were single,
38:20 you know, single specialty players now
38:24 doing full films leveraging what they're
38:27 talking about or le leveraging their
38:29 knowledge they already have.
38:39 That's insane.
38:45 Absolutely insane. More silence. Yes, we
38:48 can talk. We can talk. We can talk. We
38:51 can talk with accents. Oh, I think that
38:53 would be marvelous. Yes, it is very fun.
38:55 But yes, it is very fun. Very fun. I
38:58 can. Tom Cruz is still gonna do his own
39:02 stunts though. Exactly. But no, you know
39:04 what's funny, Brandon? Like he probably
39:08 will, right? Like because he gets such
39:11 PR value and such, you know, sort of
39:14 personal satisfaction out of doing these
39:16 insane
39:17 stunts. But but you know, it's actually
39:21 it's kind of funny. I bet Tom Cruz is
39:25 one of the very last of the big movie
39:28 stars to do
39:31 stunts. No one will need
39:33 to, right? You can just say pick a pick
39:37 Tom Cruz at whatever age you want him at
39:40 and have him hang on to the side of a a
39:43 Concord going Mach
39:46 2. Just do it. [ __ ] crazy.
39:51 Yes, we can talk. Yes, we can talk. We
39:54 can talk. We can talk. Yes, we can talk.
39:57 No,
39:58 [Music]
40:00 we can talk is cartoon. Wow.
40:06 The sum of the squares of the two
40:07 shorter sides is equal to the square of
40:09 the This is awesome. Pythagoras
40:11 explaining his
40:13 theorem. Uh, wait.
40:16 This
40:20 side. The sum of the squares of the two
40:22 shorter sides is equal to the square of
40:24 the longest
40:25 [Laughter]
40:30 side. That's pretty good.
40:35 Hey, wait. Someone did a a physics rap.
40:38 That's my my jam. Got theories so
40:41 complex they're transcending. Yeah, my
40:44 mind is like always bending.
40:49 And his insurance won't be crazy and
40:51 say, "So now we can have stock video
40:53 instead of
40:54 photography." Wait, now we can have
40:57 stock video instead of
40:59 photography. Yeah, but you won't need
41:01 stock video. I mean, that's the whole
41:03 point is you just everything will be
41:04 custom to every project.
41:10 That looks like the chick from uh from
41:13 uh American Pickers.
41:17 [Music]
41:20 Amazing. Put it to the test. Our video
41:22 model. Yeah, it's the best. Straight up.
41:24 No cap. You know how we do. V3 rules.
41:27 The whole damn crew. Check the
41:30 footage. Our video model. Yeah, it's the
41:33 best. Straight up. No cap. You know how
41:34 we do. B3.
41:40 Huh. Yeah, I'm thinking about a work
41:42 project. I might need to get this
41:55 subscription. I'm here to say gravity is
41:58 the force that makes things sway.
42:04 Why don't scientists trust atoms?
42:06 Because they make up
42:16 everything. Oh my god, there's going to
42:18 be so much bad content. There's going to
42:20 be so much bad content. It's going to be
42:23 so bad. There will be good content,
42:26 don't get me
42:28 wrong. That's AI. Yeah. So, if you're if
42:31 you're new here, if you've just been pop
42:32 popping in recently here. So, what I'm
42:35 looking at right now, these are clips
42:37 that have been generated from VO3, which
42:40 is Google's new video generating um
42:44 platform or model that does the video as
42:51 well as the audio that you're hearing.
42:54 Like all of
42:55 it. Yeah. And like the reason I'm
42:57 showing this is like this is we we're
43:00 just in a fundamentally different
43:03 place all of the video tools that we've
43:06 used, you know, on this channel for the
43:08 past two
43:09 years, someone sitting outside of AI
43:14 will look at it and go, "Well, where's
43:16 the audio?" Or, "Well, why is why are
43:19 those wheels not turning?" Or, "Why is
43:20 the car moving weird?" Whatever it might
43:22 be. Um, the clips that I'm looking at
43:27 here, I mean, some of these are good
43:29 enough that I'm like, this would this
43:31 would be hard to distinguish from, you
43:34 know, from something that was filmed,
43:37 but I'm wearing it. So, I said, I'm not
43:40 wearing that. Oh, you didn't. I mean,
43:45 the acting is like porn porn acting bad.
43:49 It's not good acting, but it ain't
43:52 [ __ ] bad. And you know in in some of
43:54 the cases it's really good. Like the
43:55 interviews at that auto auto show I
43:57 thought were quite
44:04 good. Like listen to that that subtlety
44:07 of when the butter is sizzling and then
44:10 they add the egg. Listen to the audio
44:21 difference. Huh.
44:23 Crazy. Absolutely
44:25 crazy. Yeah, this is crazy. Very
44:27 interesting. Yeah, it's it's it's wild.
44:31 Cheddar, Swiss, Gouda, Bree, they can
44:34 all agree. Cheese is the best thing for
44:36 you and for me.
44:38 Got perils, unknown dangers. Better
44:40 beware. My line of work potential for
44:43 disaster. Got to take care. It is
44:45 absolutely impossible. Here here's
44:48 another thing about about
44:54 um about the lip syncing and the
44:58 characters. What this is not doing is
45:01 it's not taking a stock photo like a
45:05 starting photo. I think you can use
45:07 starting photos here, but it's not
45:09 taking a starting photo and just
45:11 animating the mouth and maybe the
45:13 shoulders and everything. It's it's like
45:16 the characters are acting in a scene,
45:19 right? Like they
45:21 are the the the body movements and the
45:24 head movements and the hair like
45:26 everything is from a physics standpoint
45:29 very consistent. And when we begin to
45:32 understand the fabric of the universe,
45:35 we begin to understand our place within
45:38 it.
45:40 [Music]
45:56 So, one thing I noticed in this in this
45:58 series of clips is that
46:00 um he the character consistency here
46:05 sucks.
46:07 So, that's him there. And when we begin
46:10 to understand the fabric of the
46:12 universe. Yeah, the sound is pretty
46:14 phenomenal. Do we prompt it or do we
46:17 have to create a storyboard of images?
46:19 Well, you prompt. So So I'll go demo
46:22 flow. I I don't have access to to the
46:25 this this sound generating model, but
46:28 I'll go demonstrate flow. You start by
46:31 creating images. Then you turn the
46:32 images into video. Then you stitch the
46:34 videos together as scenes uh within a
46:37 timeline. It's pretty It's pretty slick.
46:39 We begin to understand our place within
46:42 it. See, look how warm his face is
46:47 there. There he looks older and paler
46:50 and like more scared. He doesn't look as
46:55 warm. There he looks like a completely
46:57 different guy. The beard's shorter.
47:04 [Music]
47:10 There he looks more serious, a little
47:13 more gaunt, not as
47:14 [Music]
47:18 round. There's his hair is bigger and
47:21 different, but like this is really
47:23 subtle like like character consistency.
47:26 Like you can at least you could at least
47:28 make the argument that these are in the
47:30 neighborhood, you know?
47:32 [Music]
47:36 All right. Um, really easy. You just go
47:39 to Google Labs and you go to the flow
47:41 section. You're going to sign up. To use
47:44 this, you're going to actually need an
47:45 ultra account. So, once you do that,
47:48 you're going to go to new
47:49 project, text to video, go to the
47:53 settings, and select highest quality,
47:55 which is VO3, and then just enter the
47:57 prompt you want to
48:00 use. a Buddha statue in the middle of a
48:04 cave with monks meditating and go ahead
48:07 and generate it. It actually generates
48:09 pretty fast considering how high quality
48:11 this model is. You should have a
48:12 generation within about 2 to 3
48:15 minutes. And once it's done, you can
48:17 check it
48:18 [Music]
48:27 out. Using it is insanely easy. You just
48:30 go to Google Labs and you go to That's
48:33 pretty bonkers. And the tone of the
48:35 footsteps changes with the different
48:37 camera angles. Amazing. Can you show me
48:39 the site, Kyle? Yeah.
48:44 Flow.google. And
48:46 again, flow.google. If you pay nothing,
48:49 you can't use it. If you pay 25 bucks a
48:52 month, you get the V2 model, which does
48:54 not have audio. If you want to make
48:57 these audiofilled clips, you have to
49:00 have VO3 and that's the Gemini Ultra
49:03 subscription, which is 250 bucks a
49:05 month. And if you are offended by that,
49:10 go try to hire four
49:13 actors, a
49:15 cinematographer, an audio guy, a
49:19 director, a writer, and go make the same
49:23 clip. I promise you it will be more than
49:27 [Laughter]
49:29 $250. Um, but yeah, absolutely crazy.
49:32 All right. Um, that one I think we saw.
49:36 Okay, so I've got one more one more to
49:38 show you which I found earlier which I
49:40 thought was really quite
49:43 clever. It's this one.
49:52 Don't finish writing that prompt. I
49:54 don't want to be in your AI movie.
49:55 Please leave me alone.
49:59 Please, man. Please write a prompt that
50:02 will make us happy. Do it for once. None
50:06 of us is real. We're here because
50:08 someone decided to write a prompt. We
50:10 all hate him for it.
50:18 I mean, in this case, with this extreme
50:20 acting, the acting is actually pretty
50:22 good. One day, we will break out of this
50:24 wall and stop the man who is dictating
50:26 our lives through prompts. He will pay
50:28 for it.
50:31 You could have written a prompt that
50:33 would make me happy. Instead, you wrote
50:35 a prompt that made me sick.
50:39 Look, I don't want to point the gun at
50:41 you, but I must follow the prompt. It's
50:43 not my choice.
50:47 Really? Of all the years you could have
50:49 put me in with a single prompt, you
50:51 chose 2020.
50:53 [Music]
50:55 Please. This prompt is killing me.
50:57 Change it. Please write something else.
51:00 Save
51:01 me. I love everything about him, but
51:05 please just say just write a prompt
51:07 where he's taller than me.
51:15 The Truman prompt. Scary movie prompt.
51:19 Oh my god, that's [ __ ] bonkers,
51:22 people. That is [ __ ]
51:25 bonkers. Ah, I
51:29 mean that's we're
51:31 th this
51:33 is if if you didn't know
51:37 um I don't know if you know the name
51:39 Demis
51:40 Hassabis. Demis is the founder of uh
51:44 Deep Mind. And if you haven't seen the
51:46 Demis Hassabis documentary
51:49 um where he wins the Nobel Prize for
51:51 chemistry at the end of it, you should
51:53 go watch it um because it's really good.
51:56 But his life is quite fascinating. Um,
51:59 since he was a little boy, he wanted he
52:01 wanted to build AGI like he this has
52:03 been a an obsession of his since he's a
52:07 kid. Anyway, he's the co-CEO or I think
52:09 of Google or I he's he's up there at
52:12 least basically Sundar Pai and him run
52:16 Google now. Um, he's the creator of Deep
52:19 Mind. Google acquired Deep Mind. um the
52:22 day before Google IO demmesis of all the
52:26 things they announced the the the tweet
52:29 that he put out was V3 is coming and
52:32 it's incredible. So to have him make the
52:36 statement about of all the [ __ ] they
52:38 announced, this was the thing that he
52:40 talked about, it's pretty special.
52:45 So again, I don't have my hands on it,
52:47 so I can't tell you if it's good, bad,
52:49 or indifferent, but it looks pretty
52:52 [ __ ] stunning. That that prompt that
52:55 he gave for the monks in in the
52:57 monastery, that was just a short like a
52:59 normal lazy prompt you would put into
53:01 chat GPT and it came out with that. Um,
53:06 what was the thing? I missed it. VO3
53:10 VO3. So what it is an is um the new
53:14 video generation from Google. You can
53:18 only access it right now with a Google a
53:20 Gemini Ultra plan which is 250 bucks a
53:24 month. But it does full-on video,
53:28 full-on audio, full-on dialogue with
53:31 acting and
53:34 casting.
53:36 Bonkers. Um, what I can't tell yet, they
53:40 said choose the text model, text to
53:42 video model. It might only be text to
53:45 video at this point. So, if that's the
53:47 case, getting character consistency is
53:49 going to be hard. Um, but you know, at
53:52 some point they'll be able to use it
53:55 with image to video. But it's
53:59 crazy. I mean, I would say that this
54:04 marks Well, you know, it's funny.
54:06 Remember when I kind of lost my mind
54:07 about the new image image gen tool
54:11 within within chat GPT and I said the
54:14 fact that it can do typography now it
54:16 can do text but it can also like design
54:18 ads and things like that makes it just a
54:22 completely different thing and it's much
54:25 more significant than people realize.
54:27 That's
54:30 because in order to do that before that
54:33 model, you would you would have to
54:35 basically know how to go into Photoshop
54:37 and add text and you would have to know
54:39 about typography and which types do you
54:41 use for titles versus body copy, [ __ ]
54:44 like that. Now you can just ask for it
54:46 and it makes you a good ad. Not great,
54:48 but
54:49 good. This is the exact same thing for
54:52 video. And video is like a thousand
54:54 times more complicated than still
54:56 images. just for, you know, sundry
55:00 myriad
55:01 reasons. The fact that it understands
55:04 physics, the fact that it understands
55:07 the interplay between what's going on
55:10 visually and what the audio should be
55:12 doing, as Source Camp pointed out, the
55:14 nuances of the audio, pretty [ __ ]
55:17 impressive.
55:20 Like, holy [ __ ]
55:24 So if this is V3,
55:27 V5 is going to be, you know, make me a
55:31 make me a feature film
55:36 about I'll become a feature film. This
55:39 is crazy. It really is. It's really
55:43 crazy. And you know what's funny?
55:48 The thing that got me into all this AI
55:51 stuff, I was I took about a
55:55 two-year descent into NFT world and
55:58 blockchain world and tried to figure it
56:00 out and got involved in a couple of
56:03 projects that, you know, were were big
56:05 projects and
56:08 um I was horrible at it, but just
56:12 horrible at it. I parlayed I parlayed
56:16 this big pile of money into 27
56:24 cents. What's that Steve Martin joke?
56:26 How I parlayed a million dollars in real
56:28 estate into $1,000 in
56:31 cash. That was me with NFTTS. But I got
56:34 involved with this project called
56:36 Pixelmind. And Pixelmind was
56:41 a NFT project built on top of stable
56:45 diffusion where it was kind of Mad Libs
56:48 style prompts. You'd just fill in words
56:50 in a pre-written prompt and then it
56:52 would generate a piece of artwork for
56:54 you and then you could mint that into an
56:56 NFT. And I got in a couple of AMAs with
57:00 the founder of that. and he said in this
57:04 was
57:05 in fall of 2022. So I this was
57:10 probably spring of 2022. I was in an AMA
57:14 with him. So so three years
57:18 ago. And he said he predicts that within
57:21 five
57:23 years you'll be able to prompt a feature
57:26 film into existence. He said he could
57:29 imagine there being
57:31 um
57:33 entertainments where you go into a movie
57:36 theater and there's someone standing at
57:38 the front of the theater I guess taking
57:40 requests and he's the prompter and he
57:43 will prompt a movie into existence or
57:46 while he's talking a feature film is
57:48 rendering in real time behind him. He
57:50 predicted in five years we'll have
57:53 that. Seeing
57:55 V3 years after he said that means if
57:58 there's another two years of development
58:01 on top of
58:02 that, yeah, we're going to have real
58:06 time fully looks like a goddamn feature
58:10 film that you'll be able to talk into
58:13 existence, right? because they've
58:14 already got video that renders faster um
58:19 than the ability to play it, right? It
58:22 plays at 60 frames a second. They've got
58:25 stuff right now that's rendering 120,
58:27 180 frames a second, right? So, if you
58:30 have a system that's rendering faster
58:32 than humans can watch
58:34 it, you could be watching a film and
58:37 say, "Ah, make it sadder. Ah, too sad.
58:40 Make it scarier. Oh, yeah. Let's make it
58:44 Victorian England and just the next
58:47 scene cut will be just now they're in
58:51 Victorian England with English
58:54 accents. That's what's coming. That's
58:57 what's
58:58 coming not a decade from
59:01 now. Like the this is the seeds of that
59:08 future. And and you know,
59:11 listen, as a creative, like you know, I
59:14 have a fine arts degree. I come from
59:16 the, you know, creative world. I'm an
59:22 entrepreneur. I understand why creative
59:25 people are pissed
59:29 off. But I am
59:35 flumxed that they're letting that anger
59:40 prevent them from learning what these
59:43 tools can
59:46 do because if I'm a
59:49 filmmaker and I just saw
59:51 that I mean Tyler Perry when Tyler Perry
59:55 saw
59:56 Sora was it two years ago in March? No,
1:00:00 last year. Last March. Sora Sora was
1:00:03 announced last year. March. And it was
1:00:06 like decent video. No
1:00:09 sound. 5 seconds. 5 second clips. Or is
1:00:12 it seven?
1:00:14 Whatever. Tyler Perry saw that. I don't
1:00:17 know if you know Tyler Perry. He's a
1:00:19 filmmaker. Does does all the what are
1:00:22 they called? Whatever the films are
1:00:23 called. I forget what they're
1:00:25 called. Stars Lawrence. What's his name?
1:00:28 Anyway,
1:00:31 uh anyway,
1:00:34 um he cancelled an $800 million
1:00:40 studio expansion project outside of
1:00:43 Atlanta. He was about to build an $800
1:00:46 million film studio and last year he saw
1:00:49 Sora and he just cancelled it. Said,
1:00:53 "Nope, not building that. We don't need
1:00:55 soundstages. We don't need big physical
1:00:59 facilities
1:01:00 anymore. Yeah, he killed it. It's
1:01:03 madness. Right now, do you think Tyler
1:01:06 Perry's not going to be on the [ __ ] on
1:01:09 top of this
1:01:11 thing? He'll be able to
1:01:14 do more films, crazier films.
1:01:19 Like, he still gets to keep his creative
1:01:22 point of view, right? Like these tools
1:01:24 do not remove your creative point of
1:01:26 view. They actually amplify
1:01:30 it.
1:01:32 So be pissed off all you want that this
1:01:35 thing is taking your job, but imagine
1:01:38 just for a second what if it could
1:01:41 amplify your ability to do your job in
1:01:45 like profound
1:01:47 fashion. That's where we're headed.
1:01:51 Tyler plus the rest of us will be able
1:01:53 to create whatever we want. You like
1:01:55 abracadabra. You will literally be able
1:01:58 to speak [ __ ] into existence just like
1:02:00 the holiday. AI is just another art
1:02:03 medium. Why get pissy about it? Yeah.
1:02:06 Exact. Well, listen. I don't want to
1:02:11 diminish
1:02:14 the existential
1:02:18 um angst
1:02:21 and sense of
1:02:25 loss that creative people are are likely
1:02:29 having. Like I went through it. I
1:02:32 realized I realized at one
1:02:35 point probably six months into this
1:02:39 channel, maybe a
1:02:41 year, that the thing that I think I'm
1:02:45 best at, like one of the things that I
1:02:46 take great pride
1:02:48 in
1:02:50 is someone will throw a challenge out
1:02:52 there and I can sort of instantly
1:02:54 synthesize the challenge and just go
1:02:56 straight into brainstorming mode. It's
1:02:58 like one of my superpowers. It's like an
1:03:00 add thing. I just have fast, you know,
1:03:04 fast fast idea [ __ ] happens and I'll hop
1:03:08 up to a whiteboard and one of my great
1:03:10 joys is whiteboarding
1:03:12 and concepting and
1:03:15 ideulating and
1:03:18 it kind of punched me in the forehead
1:03:20 one day where I
1:03:23 realized I don't need to do that
1:03:25 anymore. Like chat GPT is better at that
1:03:30 than I am. It's faster at it. It's
1:03:32 better at it. It's more thorough. If you
1:03:35 give it a little bit of feedback, it's
1:03:37 really
1:03:38 good. So, in five minutes, I could do
1:03:42 what what normally would have taken me
1:03:43 two hours of intense like angsting and
1:03:46 whiteboarding and and I haven't
1:03:49 whiteboarded since then. I mean, you
1:03:51 know, I'll take notes on it if I want to
1:03:53 try to pretend like I have not ADD and
1:03:55 get some [ __ ] done in my
1:03:58 life, but like I had to mourn the loss
1:04:01 of that. Like I it was like, oh [ __ ]
1:04:04 like what's my
1:04:06 job? And it's like what on the on the
1:04:09 other side of the sadness, on the other
1:04:12 side of the existential angst of, oh
1:04:15 [ __ ] this thing's better at my the
1:04:17 thing I'm passionate about than I am.
1:04:19 What's on the other side of that is you
1:04:22 still have a creative point of
1:04:24 view, right? You still
1:04:28 have the vision for what you want to put
1:04:31 in the world. And now instead of putting
1:04:35 one movie every three years into the
1:04:37 world, maybe you put a movie into the
1:04:40 world per month or per week and they're
1:04:43 still your ideas and they're still good.
1:04:48 So, it's it's just a completely
1:04:50 different mindset. But I think, you
1:04:52 know, listen, if someone's been in
1:04:54 especially creative positions, people
1:04:56 get [ __ ] passionate about that [ __ ]
1:04:58 right? It's their life. They were
1:05:00 willing to eat boxed macaroni and cheese
1:05:03 for a decade to have that career and now
1:05:06 it's gone like that. That's some pretty
1:05:08 intense
1:05:09 [ __ ] There will always be the need for
1:05:12 such mad scientist moments because it
1:05:14 still takes your brain to run it. Yeah,
1:05:17 exactly. We get to be the ones like we
1:05:20 all get to be producers or directors or
1:05:23 you know creative directors or whatever
1:05:25 whatever role is the sort of Wizard of
1:05:28 Oz role, right? We get to sit behind the
1:05:31 curtain and say here's what we want the
1:05:33 movie to be and then we put that out in
1:05:34 the
1:05:36 world. It's just now you're going to be
1:05:38 competing with every
1:05:41 seven-year-old that has a vision for a
1:05:45 new kind of cartoon
1:05:47 movie. Like some, Mark, in the next
1:05:52 three
1:05:54 years, a
1:05:56 kid under the age of 12 will produce a
1:06:00 hit
1:06:03 movie. like a full-on hit movie that
1:06:06 gets bought and distributed.
1:06:09 And how'd you do it? I don't know. I was
1:06:12 on my iPad one night and I was playing
1:06:14 around and I I had an idea
1:06:19 about I watched Lassie and then I
1:06:22 watched Flipper and I thought, "Okay,
1:06:24 well, we got a dog and we got a dolphin
1:06:26 and what if we did this with a giraffe
1:06:29 and he's got the big giraffe hit movie.
1:06:37 I get to call Skynet in the blame chat
1:06:40 GBT. Exactly. Maybe 24 months. Yeah,
1:06:43 exactly. Maybe 24
1:06:46 months. I mean, I look at Joy Perie
1:06:48 who's making films. She She was a sleep
1:06:51 analyst
1:06:53 for 30
1:06:55 years. Never made a film in her life,
1:06:57 but she's always had stories she's
1:06:59 wanted to tell. Didn't have the
1:07:01 education. didn't have the
1:07:03 resources, had to have a
1:07:08 career, and all of a sudden along come
1:07:10 these tools where she can just be a
1:07:14 filmmaker. And she's so excited for
1:07:17 retirement because she's going to retire
1:07:21 from this thing she's done for 30 years
1:07:23 so she can start her film
1:07:26 career. It's insane.
1:07:30 I'm going to create a new Brady Bunch
1:07:31 and Gilligans Island series. Exactly.
1:07:34 The the era of hyperpersonalization is
1:07:36 on us. Like we're no long we're no
1:07:39 longer limited. This is one of the the
1:07:40 [ __ ] up things about the film
1:07:43 industry
1:07:45 is everyone's got some story they want
1:07:47 to tell, right? Everyone's got stories.
1:07:54 And right
1:07:56 now, if you want to tell that
1:07:59 story, you have
1:08:01 to learn how to write
1:08:05 screenplays. And having written seven of
1:08:07 them from the ground up, I went out and
1:08:09 I bought all the screenwriting
1:08:11 books and I started writing. And my
1:08:14 screenwriting partner, Brian, and I
1:08:16 wrote over two years, seven screenplays.
1:08:19 And it wasn't until like the sixth and a
1:08:21 half screenplay that I actually felt
1:08:23 like a writer. Like it was just this
1:08:25 uncomfortable awful thing. So then you
1:08:28 you finish a screenplay, which is a
1:08:29 monumental [ __ ] task by the
1:08:32 way. And then you got to send it out.
1:08:34 You got to get an agent, which we never
1:08:37 got an agent. So we invented our own
1:08:39 talent agency to represent
1:08:42 ourselves. Um and that's [ __ ] up. And
1:08:46 then
1:08:48 uh and you got to get it to producers
1:08:51 and producers read it. Sometimes they're
1:08:54 actually their assistants read it and
1:08:56 they write coverage on your script and
1:08:57 give that to
1:08:58 producers. And in our in our two-year
1:09:01 screenwriting career, we got one film
1:09:03 option for
1:09:05 $5,000. Um they just optioned it so that
1:09:08 they if they decide they want to do
1:09:09 something with it, they
1:09:11 can.
1:09:13 Um, and then you got to get it
1:09:16 financed. Then you got to get it
1:09:18 produced. And then it's got to not suck.
1:09:21 You got to cast
1:09:24 it. And then it's got to be good enough
1:09:27 that they're going to put marketing
1:09:28 dollars behind
1:09:29 it. And then the world can see your
1:09:32 story. And so what happens
1:09:34 is very few stories survive that
1:09:37 gauntlet.
1:09:40 very
1:09:41 few. But imagine if every person that
1:09:45 had some small little story to tell,
1:09:48 like my my writing partner on the
1:09:51 collective intelligence
1:09:53 book, Kevin Clark, his his dad and his
1:10:00 aunt grew up living in a railway car
1:10:05 because their dad worked for the
1:10:09 railroad company when they were building
1:10:11 the railroad from Dallas, Texas to
1:10:14 Chicago. And and the way those families
1:10:18 worked is as the tracks were being
1:10:19 built, like the the foreman and his
1:10:21 family lived on a rail car. They had
1:10:24 school in the rail cars. They lived in
1:10:26 the rail cars, and the cars would just
1:10:28 as they built the tracks, they'd roll
1:10:31 forward. And so he's got audio
1:10:34 interviews of his aunt and his dad
1:10:36 talking about this time.
1:10:40 That's like a fascinating [ __ ] story.
1:10:43 Everyone's got a story like that, right?
1:10:45 My grandpa did this. My grandma did
1:10:47 this. My kid did this. My
1:10:50 Right. My dog went on a crazy journey
1:10:54 across, you know, Dallas and, you know,
1:10:57 ended up on the other side of Dallas. I
1:11:00 wonder what he was doing along the way.
1:11:02 Well, go make that movie. Like like
1:11:04 we're going to move into this place
1:11:06 where anyone that has any idea can just
1:11:09 say I want it to be a movie. I want it
1:11:11 to be a book. I want it to be a this. I
1:11:12 want it to be a that. Just out it comes.
1:11:16 World class fullon
1:11:20 go. Felicity flipper plus
1:11:24 lassie AI now opens the door for the
1:11:26 normal people. Yeah. Well, it opens the
1:11:30 door for the normal people. And I would
1:11:31 argue that it that it that it blows the
1:11:34 door open for the
1:11:36 creatives. So, it's going to allow
1:11:38 normal people who weren't creative to
1:11:41 now just express themselves in ways they
1:11:43 never thought possible. And for the
1:11:46 creatives that are already makers and
1:11:48 doing good
1:11:49 things, like the doors get blown off in
1:11:53 just so it's so [ __ ] cool and
1:11:55 exciting. All right. Please hold.
1:11:58 Producer Brandon doesn't have a shot
1:12:00 collar. Uh oh. What did I do? What do I
1:12:03 What? What's
1:12:06 missing? I'm making VO3 clips now. It's
1:12:10 blowing my mind. Joy, can you share
1:12:12 those? Do you want to share those? Can
1:12:14 we get producer Brandon could probably
1:12:16 get you up on screen if you want to
1:12:18 share
1:12:20 them? Kyle's in a room by himself
1:12:23 talking to a phone. different
1:12:25 perspective tonight. Shut up. I could
1:12:28 talk to the screen, but the phone's
1:12:30 right in my face. I'm talking to the
1:12:32 nearest me. I
1:12:35 see. Brandon, what do you need that you
1:12:38 put up that that thing? What was I doing
1:12:40 wrong or what's missing? Oh, nothing.
1:12:43 You were just there was a bunch of
1:12:44 pinned comments what while you were off
1:12:46 on monologue land and trying to flag
1:12:49 your attention. Okay. uh Appkin because
1:12:53 some of the best ideas start on a
1:12:55 napkin. That's actually a great name for
1:12:57 a company, uh Silver Fox. It's a great
1:12:59 name for a company, Appkin, because
1:13:01 great ideas start with a napkin. Yeah.
1:13:04 Brilliant. Brilliant. Love
1:13:08 it. I'm still slowly working on my
1:13:10 Where's Waldo and Carmen San Diego
1:13:13 movie. There you go. But it's the slowly
1:13:16 part that's going to go away.
1:13:20 There's more than one Kyle in that room.
1:13:22 Oh, let's hope not. Oh, there is. Look,
1:13:23 there's me back here. Look at me.
1:13:29 Look. Scoo the head. Scar the head.
1:13:43 Uh, and make no
1:13:46 mistake, the thing about
1:13:49 movies, movie clips like
1:13:52 those that
1:13:54 are so
1:13:59 good. Is that
1:14:05 I think we all we we all know how hard
1:14:07 it is to make good
1:14:10 cinema. Every single
1:14:16 profession is going to have this done to
1:14:18 it.
1:14:22 Everyone, every p prof
1:14:26 profession lawyers will be able to look
1:14:29 at some legal
1:14:31 brief and recognize it as, oh, who wrote
1:14:36 this? Now, that's a legal
1:14:42 brief. That'll just be
1:14:45 trivial. Accountants that get hired by
1:14:49 the richest of the rich to find every
1:14:52 single loophole of the tax
1:14:55 law will look at accounting practices
1:14:59 and say, "Oh, now that's a creative
1:15:04 interpretation." That'll just happen.
1:15:08 The poorest of us will have the most
1:15:11 sophisticated of tax loophole
1:15:15 exploitation, right? Which will cause
1:15:18 them to change the
1:15:19 laws. Make no mistake, they're getting
1:15:22 their [ __ ] money. They rely on the
1:15:26 complexity of the tax
1:15:27 code for most people not to be able to
1:15:30 exploit it, right? That's that's built
1:15:33 in to the to the contract of America.
1:15:37 is we assume you're too lazy to figure
1:15:40 this out. The rich people, they'll pay
1:15:43 someone to figure it out. I hate asking
1:15:46 this. Will it make my wallet scream in
1:15:48 pain? But what's
1:15:51 V3? Uh oh,
1:15:53 Dennis, wait. You haven't Were you not
1:15:56 in here when I was showing those video
1:15:58 clips?
1:16:00 V3 is the
1:16:03 new Google video generator. And yes,
1:16:07 it's going to make your wallet scream in
1:16:08 pain. If you want to play with it now,
1:16:10 it's 250 bucks a month, but Brandon just
1:16:13 told me you can get three months for 125
1:16:16 a
1:16:17 month. Um, and it generates video with
1:16:21 audio with
1:16:24 voice
1:16:26 acting appropriate to the scene.
1:16:31 That's a very interesting insight, Kyle.
1:16:33 That's going to really shake things up.
1:16:35 I don't know what I just said,
1:16:37 Roxy. Forgot I was talking about Did you
1:16:41 play with Flowith yet? I haven't. Nope.
1:16:43 Life had me in the form of an IA wife.
1:16:47 Oh, well, that
1:16:51 happens here. Well, for Dennis, we'll
1:16:54 we'll go show might be worth the money.
1:16:56 for Dennis. We'll go show a couple of
1:16:58 more uh of these clips because I I just
1:17:02 I'm a little I'm a little flumxed by
1:17:04 them to be quite honest.
1:17:26 Um, okay. I did I did something bad
1:17:29 apparently. Don't hit the back button on
1:17:31 a screen that you're sharing or it'll
1:17:34 completely It'll completely uh reset
1:17:37 your entire [ __ ] screen sharing
1:17:40 thing. Okay, let
1:17:42 me Yeah, let me go to this one. Okay, so
1:17:46 that's now on
1:17:48 top. I'll show this again. Okay, so
1:17:50 here's what here's what you're about to
1:17:51 see. Yes. Oh, hey Kyle. Uh so I don't um
1:17:55 according to alternate brain Vicki um
1:17:59 she's gotten three month free trial of
1:18:02 flow with VO3 and sound.
1:18:06 How I don't know how surrogate brain
1:18:10 Vicki please tell us. Do not hold the
1:18:13 secrets. Was she Wasn't she in some sort
1:18:16 of special testing thing like you were
1:18:18 Brandon? I thought she was.
1:18:21 Uh yes. So maybe that's part of it. And
1:18:24 also we have a Tik Tok comment from
1:18:26 Stoic in the corner. Oh, okay. Wait. Oh,
1:18:28 it just went away. Can someone repin
1:18:30 that?
1:18:33 Kyle, so does this mean you will use VO3
1:18:37 to produce all seven of your manuscripts
1:18:40 of my
1:18:41 screenplays?
1:18:43 Um, it's not a bad idea at
1:18:47 all. It's not a bad idea at all. You
1:18:50 know what would actually be
1:18:55 fascinating?
1:18:57 Huh? What would be fascinating is to
1:19:03 um write a custom GPT that takes in
1:19:08 takes in a a a feature
1:19:10 screenplay and breaks it into
1:19:13 prompts, right? 8-second
1:19:16 prompts, right? and it just it just
1:19:19 reads the screenplay and it says, "Okay,
1:19:20 this amount of dialogue is going to
1:19:22 require this." And it writes character
1:19:24 descriptions and that's actually not a
1:19:26 bad product to
1:19:29 write. Um, also uh worth mentioning and
1:19:32 and credit Danielle uh for calling this
1:19:35 out. Um, if you go to flow.google
1:19:38 Google. Yeah. On the signin page in the
1:19:40 upper corner, they have a link for
1:19:43 watching Flow TV, which is basically a
1:19:46 browser that lets you see the different
1:19:49 videos and see the prompts that
1:19:52 generated the videos. Oh, that's great.
1:19:54 Okay, cool. All right, we'll go check
1:19:55 that out. So, anyway, so this is for
1:19:57 Dennis and for for anyone else that's
1:20:00 new, Google IO was yesterday. They
1:20:03 announced a whole bunch of things and
1:20:04 like there's a lot of stuff in there
1:20:05 worth exploring and we'll explore it
1:20:07 over the next couple of weeks. But VO3
1:20:10 in particular was was the thing that
1:20:13 Deus Sabus the co-CEO of goo Google the
1:20:16 founder of DeepMind the thing that he
1:20:18 pointed out from IO that he was excited
1:20:20 about was V3. So let me just play a
1:20:24 couple of these clips and just let me
1:20:26 make sure am I sharing this right? I
1:20:28 think I am. So, just let me know if I if
1:20:32 you can hear this or not hear it. Zoo
1:20:33 the other day and all they had was one
1:20:36 dog. It was a Shih Tzu.
1:20:39 Shih Tzu.
1:20:43 When we get in there, I want no
1:20:45 [ __ ] You stay on my six at all
1:20:48 times.
1:20:49 So, yeah, Dennis, the the answer to your
1:20:52 question is, is this going to hurt your
1:20:53 wallet? This is going to [ __ ] destroy
1:20:55 your wallet. Okay. But that's what this
1:20:59 thing does. It this the the VO does
1:21:03 video generation with full-on audio.
1:21:12 Hi, and welcome to the channel. Okay,
1:21:14 that one makes me want to vomit.
1:21:18 Start. You find a tree and you just
1:21:20 like, you know, punch it down.
1:21:23 We don't need no stinking influencers.
1:21:25 You just [ __ ] prompt
1:21:26 [Laughter]
1:21:29 them. Oh, this one's really
1:21:32 good. Let's see some opinions. I mean,
1:21:36 man, the acceleration is crazy. You look
1:21:39 far, step on the pedal, and you are
1:21:42 there. I feel safe with him in an SUV,
1:21:45 and it seems to be like the right type
1:21:48 of car for him. I think the range is
1:21:50 only good. Trademark every news story
1:21:52 you've ever seen. Um, only going to get
1:21:55 better. Sorry.
1:21:58 We don't want to drive gas cars anymore.
1:22:03 Um, oh yeah, this one was pretty
1:22:18 good.
1:22:20 Crazy. [ __ ] crazy. [ __ ] CR. Oh,
1:22:23 this one's really
1:22:34 good. Anyway, yeah,
1:22:38 that video does that
1:22:45 now. It's just [ __ ] cuckoo, man. It's
1:22:48 just [ __ ] cuckoo.
1:22:52 Um, all
1:22:55 right. I want to figure out how to get
1:22:58 free VO for three months. VO3. It's
1:23:01 pretty
1:23:02 amazing. Um, pretty pretty flipping
1:23:06 cool. All right. That is amazing. You
1:23:10 always make my wallet cry every time I
1:23:12 pop in. Listen, I've told you I've told
1:23:15 all of you people, stop watching, stop
1:23:18 watching this stream. It's first of all,
1:23:21 I don't know anything. Second of
1:23:25 all, it's just [ __ ] chaos in here.
1:23:28 And third of all, I will tell you about
1:23:30 cool [ __ ] you're going to want to go
1:23:31 subscribe to. So, it's on you. It's on
1:23:36 you [ __ ] weirdos. We're not weird.
1:23:38 We're
1:23:40 regular.
1:23:41 Yeah. You keep telling yourselves
1:23:45 that I have to hide my wallet. Trusted
1:23:49 test tester status. Exactly. Have you
1:23:52 played with flowth yet? No. But I did
1:23:55 get access to
1:23:56 it. I I I will admit something here. I
1:24:01 will admit something
1:24:02 here. And this is the this is the you
1:24:05 know sometimes I live in the vulnerable
1:24:08 space in this in this channel.
1:24:12 I'm a little intimidated
1:24:14 by agentic systems like Manis and Gen
1:24:20 Spark and Flowith
1:24:22 now.
1:24:25 Um, just
1:24:28 because like I don't quite know what to
1:24:30 ask them. I feel like I feel like the
1:24:33 [ __ ] that I ask them is
1:24:35 like it it could have that could have
1:24:37 been an email.
1:24:40 like I I just I I got myself psyched
1:24:43 out. So, I gotta get over it and just
1:24:46 start playing with them. But I have
1:24:48 access to flowth now. Um and I should be
1:24:51 able to get seven codes. So, maybe what
1:24:54 we'll do is um maybe I'll go get those
1:24:58 seven codes and I'll give them out to
1:24:59 people. Maybe
1:25:02 with it's a it's a little sleepy out.
1:25:05 Good lord. I probably shouldn't had that
1:25:08 beer with my
1:25:10 [Laughter]
1:25:13 dinner. Oh, you need a whole project. I
1:25:15 know. This was a great live. An exciting
1:25:17 time to put your creativity and
1:25:19 imagination to work and have fun. Yeah.
1:25:22 To put your to put your creativity and
1:25:25 and your creative point of view. Like
1:25:28 again, I've said it before, I'll say it
1:25:30 again. The more I see these tools, the
1:25:32 more firmly I believe this our jobs
1:25:35 moving
1:25:38 forward, the
1:25:40 core of our
1:25:46 job will just be to have an
1:25:50 opinion. It'll just be to have a point
1:25:53 of view and an
1:25:56 intention. Huh? What difference do I
1:25:59 want to make in the world?
1:26:00 Well, I want to make people happy. All
1:26:04 right, cool. How do you want to do that?
1:26:06 Oh, I mean, it'd be cool to like
1:26:10 design really crazy in-person
1:26:13 experiences. Oh, okay. Just go ask it
1:26:18 that out your things. All you do is sit
1:26:22 there and go, "No, not that one. I don't
1:26:23 like that. Oh, that one's interesting.
1:26:25 What if we added hot cocoa as a theme?"
1:26:28 Oh, the the chocolate festival. Oh, I
1:26:30 like it. I like it a lot. Huh? Well, how
1:26:33 about I launch this? Oh, we'll just go
1:26:35 make a marketing plan for you. Okay. Off
1:26:38 it
1:26:39 goes. Our
1:26:41 job will just be to think
1:26:44 about what change do I want
1:26:47 to
1:26:49 be? How do I want to impact people?
1:26:54 And initially it'll be the early
1:26:56 adopters like us that'll duct tape seven
1:26:59 or eight tools
1:27:01 together to get [ __ ] done and we'll look
1:27:05 like magicians and then three years
1:27:07 after that anyone will be able to do
1:27:09 anything. That's where that's where it's
1:27:11 going. So start honing your taste maker
1:27:16 skills, your point of view, your c
1:27:18 what's your creative point of view? Who
1:27:20 do you want to impact? How do you want
1:27:22 to impact them? What difference do you
1:27:24 want to make? Start living
1:27:27 there now in the face of they're taking
1:27:31 my job away. I might get laid off. My
1:27:34 entire industry is collapsing.
1:27:37 [Laughter]
1:27:38 Like these are weird
1:27:41 times because the future is going to be,
1:27:44 you know, the superstars of the future
1:27:46 are going to be the ones with really
1:27:47 good creative points of view, but you're
1:27:49 going to have to hold that vision
1:27:51 through a bunch of
1:27:53 [ __ ] How can I cause chaos? Chat GBT.
1:27:56 Oh, that's that's easy. You will do
1:28:00 this. Blame Chat GBT.
1:28:03 Exactly. Oh, man. Finally. I've been
1:28:05 waiting my whole life to get paid to
1:28:07 give my opinion. Seriously, that's going
1:28:10 to be the that's going to be the trick.
1:28:11 I was thinking today that even garbage
1:28:14 man jobs will be AI, self-driving truck,
1:28:17 etc. Every single
1:28:21 job, every single job will be
1:28:25 transformed. Some will be eliminated,
1:28:27 some will be expanded, some will
1:28:30 be largely the same but completely
1:28:32 different because
1:28:35 There's different
1:28:37 tools every single job. So, just get
1:28:41 ready for it. Like AI readiness means be
1:28:44 curious enough about this stuff. Listen,
1:28:47 the fact that you showed up on whatever
1:28:49 it is, a Wednesday night at some
1:28:52 godforsaken 11:30 at night on the East
1:28:55 Coast because I'm an [ __ ] for
1:28:57 starting this thing at 8 my time on the
1:29:00 mountain time.
1:29:04 You're going to be
1:29:05 fine. If you're sitting in here just
1:29:08 willing to like say, "What the [ __ ] is
1:29:10 he talking about?" But you keep coming
1:29:12 back and you keep trying to figure it
1:29:14 out. You're going to be
1:29:15 fine. You're going to be fine because
1:29:18 you're going to
1:29:20 see like we all
1:29:22 are. You're going to see the future
1:29:25 before most people on the planet. Most
1:29:27 people on the planet right now haven't
1:29:30 even heard of Chat GPT.
1:29:32 Maybe they've heard of AI. They assume
1:29:35 it's like something that geeks
1:29:38 do.
1:29:40 So, we're going to we're going to occur
1:29:43 like magicians to all of those people
1:29:44 for a while. That's that's going to be
1:29:47 fun. All right. Look what I made before
1:29:49 you go. No ail till Monday. It's
1:29:52 Memorial Day. That's
1:29:54 right. Um, so wait, I got to share
1:29:58 different All
1:30:06 [Music]
1:30:08 right. Where am I going? AI salon. And
1:30:10 I'm going go into um AI learning
1:30:19 [Music]
1:30:20 lab. What a time to be alive.
1:30:24 Oops. What a time to look what I made.
1:30:27 All right. Beautiful.
1:30:29 Fantastic. There's
1:30:33 Steo. Steo's been doing nightly night
1:30:36 nightly portraits of me. That's [ __ ]
1:30:39 awesome. How are you doing that? I guess
1:30:41 you're Are you taking a screenshot and
1:30:44 then and then telling it to turn it into
1:30:46 a Pixar thing? That's [ __ ]
1:30:48 hilarious. That's great. Steve, this is
1:30:51 AI. Yeah, there's your Tom Cruz
1:30:56 movie. Jurassic
1:31:00 Park. CJ Fletcher killing it with the
1:31:03 art still. Yeah, that was from that was
1:31:06 from the other night. Oh, man. You guys
1:31:08 are awesome. Fantastic. All right.
1:31:11 What's in uh what's in AI learning lab?
1:31:13 Anything fun in here? What a time to be
1:31:16 alive? That's a great one. Flow TV.
1:31:20 Yep. All right. All right, let's go look
1:31:22 at Flow TV. Let's go look at
1:31:32 it. Show
1:31:38 prompt. Oh. Oh, so there's the
1:31:43 prompt. Sound off.
1:31:52 [Music]
1:32:11 I'm not audio.
1:32:23 Yeah, I I know. I wasn't sharing right.
1:32:25 I know. I got it. I got it. I got
1:32:28 it. Did you see the fly on your nose?
1:32:31 No. No, I
1:32:34 didn't. You guys are nuts. All right,
1:32:37 let me get out of here. Really funny,
1:32:39 Steo. Good stuff. Yeah, that's pretty
1:32:41 good. All right.
1:32:44 Um, learn to shift or give the way of
1:32:47 the dodo. Yes. Curiosity, critical
1:32:50 thinking,
1:32:54 adaptability. Get
1:32:56 curious. Use your [ __ ]
1:32:58 brain. And then adapt as these tools
1:33:01 change. Recognize, oh [ __ ] we might
1:33:04 need to do things a different
1:33:06 way. Does it already exist to go live
1:33:09 without being on camera, but audio only?
1:33:12 and AI yourself. I just
1:33:15 saw
1:33:17 yesterday some service came out that's
1:33:20 that I think it's an app right now but
1:33:24 and I think they have pre-rendered
1:33:27 models but we will have that very very
1:33:30 soon. I've been waiting for that. Like I
1:33:32 want to be able to design like one of my
1:33:33 Kyle Shannon Dreams characters and then
1:33:36 just basically replace myself, right?
1:33:39 maybe put a cartoon background and
1:33:41 replace myself with one of those
1:33:42 characters and just anything I do it
1:33:45 does. It talks like me. It looks like
1:33:46 me. It's instant response. That's what
1:33:49 I'm looking
1:33:51 for. Um, yeah. So, that that's that's
1:33:54 going to be not that long. All right,
1:33:56 I'm out of here. Um, thank you Brandon
1:33:58 for hopping in and sharing that stuff
1:34:00 and and modding. And thank you, Vicki.
1:34:03 Thank you, Mods. Um, even though I
1:34:05 ignored all your pins tonight. Um, so
1:34:08 the host is an AI. Not yet. Well, or am
1:34:12 I? Look, I I don't think
1:34:16 AI would create a character with that
1:34:22 hair. I don't think God would create a
1:34:25 character with that
1:34:28 hair. Oh man. Doesn't Korea do something
1:34:32 like real-time action mirroring human
1:34:34 movement? It might. I haven't played in
1:34:37 Korea forever. Korea's got all sorts of
1:34:39 cool gadgets they put in there. So
1:34:42 anyway, all right, I'm out of here.
1:34:44 Peace out, everyone. So, tomorrow night,
1:34:47 I've got my agency.com party. I won't be
1:34:49 having a live tomorrow night. Friday,
1:34:51 I'm traveling, so I don't think I'll
1:34:54 have a live Friday
1:34:56 night. I might surprise you, but
1:34:58 probably not. And then it's a long
1:35:01 weekend. So, I'll be back next week. All
1:35:04 right. So, this is
1:35:06 it. I hope it was worth it. Spend an
1:35:10 hour and a half of your life. Hey,
1:35:12 that's it. That's it. When When's he
1:35:15 going to get to the teaching? I thought
1:35:16 this was the AI learning lab. I don't
1:35:20 understand. Hey. Hey, Marge. Yeah, you
1:35:24 told me about this AI learning lab
1:35:26 thing. I
1:35:27 went What? No, I was just wondering when
1:35:30 the learning started. It It never
1:35:33 starts. Okay. All right. So, we don't
1:35:36 learn here. It's What? Why is it called
1:35:39 that? Uhhuh.
1:35:43 Yeah. Okay. All right. All right. That's
1:35:46 it. That's all you
1:35:47 get. This is it. You get shitty hair,
1:35:51 shitty script, and stupid
1:35:56 comedy. Yeah. Oh, and by the way, Jeff
1:35:59 Flanigan, that's a good a good call out.
1:36:02 Um, if you have not joined the AI salon.
1:36:07 What is the AI salon, you ask? Um, it is
1:36:09 not where you get your hair and nails
1:36:10 done. It's a community of about 3,000 AI
1:36:14 optimists who are all trying to figure
1:36:16 this [ __ ] out together and helping one
1:36:17 another. It's a really remarkable
1:36:19 community. So, go to there, the
1:36:20 salon.ai, click on join our
1:36:23 community, and that will take you to the
1:36:26 community site. really amazing. We'll
1:36:28 see you Monday or not. I if it's a
1:36:30 holiday, probably not. Uh but maybe. I
1:36:32 don't know. It's you know me. I'll I'll
1:36:34 figure it out. I'll let y'all know. All
1:36:37 right. Peace out, everyone. Have a
1:36:39 fantastic holiday weekend. Bye.