5/20/2025 - Google's AI Innovations: Gemini, Flow, and V3

Description

Chapters

Transcript

0:02 Fantastic. Good evening. Good
0:09 evening. Let's see. What are we going to
0:14 do? We're going to go
0:19 here. I think what I'll do is bring this
0:22 here. Yeah, that's what we're going to
0:27 do. All right.
0:30 What's
0:31 happening? Hello, Vicky.
0:35 Brian, I think I need my other
0:37 glasses. Need the ones I can see close
0:40 up. These eyes ain't what they used to
0:43 be. All
0:45 right. The hair she is
0:50 looking That's baffling.
0:54 That
0:55 is that's like some sort of fire hydrant
0:58 waterfall kind of
1:01 situation. Google cooked today. Google
1:05 Holy [ __ ] So Google events are normally
1:11 um you
1:13 know 20 announcements. I mean every one
1:15 of those events has 20 announcements,
1:17 right?
1:18 Google's historically have
1:21 been 10 of them were here's [ __ ] we did
1:24 before that was really
1:27 cool and
1:28 then five of them were here's things
1:32 we're going to show you that you're
1:33 going to get all excited about and
1:34 you'll never [ __ ] see again and then
1:37 five were some sort of thing that they
1:39 would release in some form that usually
1:42 sucked um and occasionally there'd be
1:46 one or two good things Um, they launched
1:49 a bunch of [ __ ] today. Like the first
1:52 pile of announcements were, you know,
1:55 available today. Available today. And
1:58 the stuff they showed was pretty [ __ ]
2:02 bonkers.
2:04 Now, you got to be clear about
2:06 something. Two two things are true with
2:09 Google. Some of the stuff they show
2:11 might only be available for developers.
2:15 Um,
2:17 and as with all of these events, whether
2:19 it's OpenAI or some fly by night, you
2:24 know, AI startup or Google, these are
2:27 marketing events. So, what they
2:31 show is the cherrypicked
2:35 [ __ ] for, you know, for what's best,
2:38 what's best for the brand. Uh, by the
2:41 way, if you're on Tik Tok and you want
2:43 to see this streaming on YouTube, if you
2:47 go to the name of the channel here or
2:49 the no, the name of this episode or
2:51 whatever it is, there's the link to the
2:56 other channel which is at
2:58 YouTube and it's uh learning lab-ai over
3:02 on YouTube. All right, so then you can
3:04 see it there. Hey Corey Sandler, what's
3:07 happening? Good to see you. So feel free
3:10 to you can comment in either place. I
3:11 can see comments in both. So I figure
3:14 what we'll do tonight is we'll just kind
3:16 of go through some stuff.
3:19 Um we'll take a look at
3:25 uh at stuff from the Google event and
3:28 I'll I'll show stuff. Where where do you
3:30 want us to
3:32 list? I don't know what that means. Like
3:35 where do you where do I want you to
3:36 comment? You can comment in either
3:38 place. I don't
3:40 care. I don't care. The YouTube one gets
3:45 recorded and so your comments will live
3:48 in
3:49 perpetuity. Oh, where do you want us to
3:51 watch from? I think watching from
3:52 YouTube makes more sense. I
3:55 mean, I just do the the Tik Tok thing
3:58 cuz we've always done it and there's a
4:00 lot of people that do it, but I think at
4:02 this point, certainly for the
4:04 irregulars, they're usually doing both.
4:06 Like I get comments that people I don't
4:08 [ __ ] know. I don't know. I don't know
4:10 people. I just don't
4:14 know. Okay. So, over on
4:24 to So, how am I going to do this? I'm
4:26 gonna have to share that. I want to
4:27 because we're going to be playing
4:28 videos. So, hang on a sec.
4:32 [Music]
4:34 By the way, if you have not seen I I
4:37 haven't seen the the um finale yet
4:40 myself, but if you have not seen the
4:42 finale of American
4:44 Idol, this the guy that won it, spoiler
4:47 alert,
4:51 um is
4:54 stunning, like remarkable.
4:57 Um, and the song that he sang that's the
5:00 last song that was sung of the
5:02 season is just it's it just gives you
5:06 chills. It's so good. Um, I hate when
5:10 for a demo they prompt things like
5:13 where's the nearest mac and cheese and
5:15 what are Yeah, I know. Yeah, fun things
5:17 in terraott. I I know. Listen, they're
5:20 they're marketing demos. So, what are
5:22 you gonna do? All right, so let's let's
5:24 watch this video. Um,
5:27 24. Do you want to get lunch?
5:31 I'm here with the latest research
5:32 prototype of Project Astra, and we're
5:34 going to see how it can help me fix my
5:35 bike.
5:39 Hey, go online and look for a user
5:41 manual for a Huffy mountain bike.
5:44 I have found a user manual for you. What
5:47 would you like to know about it? Okay,
5:48 cool. Now, scroll until you find the
5:50 section about the brakes. Sure thing.
5:53 Looking for that page now. I think I
5:55 stripped this screw. Can you go on
5:57 YouTube and find a video for how to fix
5:59 that? Of course. I'm opening YouTube
6:02 now.
6:05 This looks like a good video. Place a
6:07 rubber band over the head of the screw
6:09 and press down firmly. Can you go
6:11 through my emails with the bike shop and
6:13 find which size hex nut I need? Sure.
6:16 According to the email from Bicycle
6:17 Habitat, you need a 3/8 in hex nut for
6:20 your sprocket. I've highlighted the
6:22 correct bin for you.
6:23 It seems like I need a spare tension
6:25 screw. Can you call the nearest bike
6:28 shop and see what they have in stock?
6:29 Yep. Calling them now. I'll get back to
6:32 you with what they have in stock. I
6:33 think my brake pads are wearing down.
6:35 Can you check the manual again and see
6:37 if it talks about them anywhere?
6:38 According to the manual, if you look on
6:40 page 24, do you want to get lunch? Yeah.
6:44 Give me 5 minutes. Can you finish what
6:47 you were saying? As I was saying, if you
6:49 look on page 24, section 2, you'll see
6:52 how to replace the brake pads. Hey, uh,
6:55 any updates on that call? Yep. I just
6:58 got off of the bike shop. They confirm
6:59 they have your tension screw in stock.
7:01 Would you like me to place a pickup
7:02 order? Could you show me some examples
7:05 of dog baskets I can put on my bike?
7:07 Sure, I can help you with that. Just
7:09 give me a
7:10 moment. Here are some options. I think
7:12 Zuka would look really great in these.
7:15 [Music]
7:20 Okay. Action intelligence plus action
7:24 intelligence. So, not artificial
7:26 intelligence, action intelligence. And
7:29 let's see, how am I going to I'm trying
7:33 to I've got my computer up on a cooking
7:37 [Laughter]
7:40 pot, so I'm trying to get things squared
7:44 away here. Hang on. Hang on, people.
7:46 Just calm
7:49 down. All
7:52 right, I'm trying to give you
7:53 everything. You're just going to have to
7:55 deal with the shiny keyboard. I don't
7:56 have black paper. Oh, wait. I have blue
8:00 paper. Hang
8:03 on. Look at
8:05 that.
8:10 Fancy. Okay. Can you pay my parking
8:13 ticket out of your account? I think who
8:15 is it? Sea slug of doom said all the all
8:18 these AIs have vocal fry. I know they're
8:21 all trying to sound like Sam Alman. AI
8:23 is gonna get so sick of being
8:25 subservient for our forgetful meat
8:29 brains. Yeah. I mean, you know, that's
8:32 the thing. I think that the con the
8:35 concept of getting sick of something is
8:39 a very human concept. Um, I don't think
8:43 they will unless they're trained to get
8:47 sick of [ __ ] right? You could you could
8:49 prompt a large language model to get
8:52 sick of stuff, but I don't think that's
8:54 not how they've been put together. So,
8:56 let's see. Content retrieval, interface
8:59 control, agent highlighting, call
9:02 assistance, knowledge
9:04 grounding, knowledge, contextaware
9:07 dialogue, which contextaware dialogue is
9:10 where the friend came in and interrupted
9:12 them and he said, "Can you continue what
9:14 you were saying?" And she knew what that
9:16 meant. And then native audio dialogue.
9:19 So that's that's pretty
9:21 slick. Um action intelligence. All
9:24 right.
9:26 escape
9:29 there. Um, is this all working over
9:34 on StreamYard? We are live there, right?
9:38 Brian Whitney, hello. Corey Sandler here
9:40 now. Vicki, that I'll get back to you
9:44 has turned off so many people now. Will
9:47 we ever believe it? Well, I think so,
9:50 Vicki. I think listen as soon
9:53 as the
9:55 the chat GPT telling you that it's made
9:58 you a PDF file when it hasn't made you a
10:01 PDF file or it's made a phone call for
10:03 you when it hasn't. Um that looks like a
10:07 lot of what they uh are trying to fix
10:10 here with this thing. So, I think
10:15 people we that are paying attention to
10:17 this every day are pissed off at that
10:20 stuff and
10:21 and it's it is annoying. But a lot of
10:25 people are going to discover these tools
10:27 when that stuff's all gone. And then I
10:30 think all of us like if the if the tools
10:32 just start working and they actually do
10:34 go off and do some work for you, I think
10:36 we'll kind of dig it.
10:38 So, I mean, I think we'll we'll we'll
10:40 forgive it
10:43 quickly. But it'll give us something
10:45 good to [ __ ]
10:46 [Laughter]
10:51 about.
10:54 Okay. Yeah. This is David Scott
10:56 Patterson. All human jobs will be
10:58 replaced by AI and robots by 2030. I
11:01 don't buy
11:02 that. Everyone will will receive a high
11:05 guaranteed income from the government. I
11:08 don't have faith in
11:10 that. Women will return to their
11:13 traditional status of not needing to
11:15 work. Men will lose their traditional
11:16 role as providers and
11:19 protectors. Huh. What will be the
11:22 impact? Oh, I don't
11:24 know. The erosion of of personal purpose
11:28 and self-worth for some some amount of
11:31 time until we figure this [ __ ]
11:33 out. I think I think that's part of it.
11:36 All right. Now, this one, Google
11:42 V3. So, we had V2, which if you if
11:45 you've been paying attention to the
11:47 video space,
11:49 V2
11:51 was pretty much the
11:53 state-of-the-art. Um, there are some
11:56 features in some of the other tools that
11:58 were better than V2. Um, it looks like
12:00 most of those have been replicated in
12:03 VO3. and this new thing they launched
12:05 called Flow, which looks pretty insane.
12:08 I have access to it to some amount of
12:11 time on it. I don't know how much time.
12:18 Um, but V2, VO3 looks amazing. So, it's
12:22 better at video obviously. And then
12:27 um and
12:30 then it it also does native audio. So
12:35 watch this video. This is pretty
12:38 nuts. So both the audio and the video
12:42 here are done in V3.
12:45 [Music]
12:52 We're going to light up the
12:55 [Music]
13:02 sky. I'm not sure.
13:07 [Music]
13:11 like that camera move from the guitar up
13:13 to the dude in a single
13:17 shot. Like just getting his fingers
13:19 looking like they're actually playing
13:21 chords or, you know, the right notes.
13:25 Pretty impressive because
13:26 normally AI guitar hands are like, you
13:30 know, the fingers are just sort of
13:32 randomly mashed either through the
13:34 fingerboard or around it. Um, and that
13:39 actually looks like he's playing
13:47 that that sounds like AI lyrics. I'm not
13:50 sure I can go on in this
13:53 tapestry we delve
14:00 into. Anyone know what Trump was signing
14:03 the agreement about AI? Is it necessary?
14:06 Hey
14:07 Ocean. Um, I'll give you my two cents on
14:11 it. Um, the AI salon has taken
14:15 a a conditional support of
14:19 it. So, let me give you some context.
14:22 So, right now there's no
14:25 um right now there's
14:31 no Hang on. I
14:33 gotta switch some [ __ ] Oh, I got to
14:36 also get a phone charger here in a
14:38 second. Uh, what am I doing? I am doing
14:41 that.
14:43 Hi. And I am
14:50 [Music]
14:58 doing so. Okay. So, some context for the
15:02 for the Trump executive order about AI.
15:05 Um, let me go get my charger so that my
15:08 phone doesn't die because that would
15:10 suck.
15:12 [Music]
15:27 I know the this channel without Champ is
15:31 a [ __ ] drag, isn't it? It's just like
15:34 me [ __ ] warbling on about
15:38 [ __ ]
15:39 [Music]
15:41 Um, okay.
15:49 So there's been
15:52 no federal position on
15:55 AI. No, no national position. I think
16:00 Biden did some sort of I don't know
16:03 signing statement or some thing that
16:05 basically said we should deal with this
16:07 at some point. That's about as much as
16:10 they had. So in the absence of
16:13 that in the past two years I think there
16:16 have been more than 700
16:19 um state
16:20 level AI laws proposed passed whatever
16:24 to to varying degrees
16:27 of clarity comp competence
16:32 um education most of them most of the
16:35 legislation is only based in fear um and
16:40 They tend to be overly broad and not
16:42 understand the difference between
16:45 traditional AI like facial recognition
16:48 kind of stuff that's been around for 30
16:50 years and generative AI which is the
16:53 stuff that we play with on this channel
16:54 and is is just a completely different
16:57 beast.
16:59 Um, and I don't think most politicians
17:02 understand the positive side of AI.
17:09 um the you know the the the opportunity
17:13 the fact that people are reinventing
17:15 themselves. We've got people in the
17:16 salon coming out of retirement like you
17:19 know rediscovering passions in their
17:21 lives.
17:23 So the the result is going to be this
17:26 patchwork of every state's going to have
17:29 different AI legislation and if you're
17:31 building something like on APIs on top
17:34 of chat GBT or whatever um you're going
17:37 to have to [ __ ] manage regulations
17:40 from 50 states. And so and as the
17:42 technologies change those laws are
17:45 likely going to sort of fall out of
17:47 relevance. they're going to be
17:49 legislating things that might not be
17:51 true even a year later.
17:55 Um, so so there was sort of a pending
17:58 absolute disaster coming and these state
18:02 regulations were coming hot and heavy
18:05 and then Trump got elected and then
18:07 Trump did what Trump did, right? Did all
18:09 the tariffs and did all the just threw
18:11 threw everything into chaos. And so a
18:15 lot of organizations like the ACLU that
18:18 was pushing really hard for AI
18:20 legislation is now dealing with like
18:22 human rights stuff. Um so the focus has
18:26 drifted off of AI. So a lot of these
18:29 bills were initiated and then have just
18:31 sort of you know not happened for
18:33 whatever reason.
18:34 But at some point, you know, the
18:37 government's going to stabilize and all
18:39 these state level legislations are going
18:42 to go through and it's going to be an
18:44 absolute [ __ ] nightmare. So, what
18:46 Trump proposed, which I think is too
18:49 radical, but is kind of what's needed,
18:52 is he he um it the signing statement is
18:55 a 10-year moratorum on state level
18:58 legislation.
19:00 Now 10 years. Are you [ __ ] kidding
19:03 me? Like this channel's been around for
19:05 2 years and it feels like 30 years. Like
19:07 it's like the amount of stuff that just
19:10 was announced today um is insane, right?
19:13 So 10 years is a really long time. And
19:15 there's also not anything in the the
19:18 provision that says anything about
19:20 getting input from people that actually
19:22 know what the [ __ ] they're talking
19:23 about. and there's not, you know, sort
19:25 of check-in times to say, "Hey, let's
19:28 check in every three years and see if
19:30 this still makes sense, right?"
19:33 Because good legislation that will
19:36 protect us from the risks of AI as well
19:38 as protect our ability to access
19:41 powerful AI that can help us make money
19:45 and do all sorts of
19:47 things that I I fully support. So the AI
19:51 salon's position on this is that we
19:53 conditionally support it and say, "Hey,
19:55 listen, if you improve it and do things
19:59 like put in, you know, the following
20:00 things, and Daisy Thomas, who's who who
20:03 heads up the uh the policy and civic
20:05 engagement um hub within the AI salon,
20:10 she's she's been a part of drafting this
20:13 positioning. She really knows her [ __ ]
20:15 Um and I trust her implicitly with this
20:18 stuff. So, um, so yeah, that's kind of
20:21 where we are. That's what it is. All
20:24 right. Yeah, three years or three
20:26 months. Yeah, exactly. Something like
20:28 that. Maybe it's a year. It's a thousand
20:31 years in AI time. It's crazy, right?
20:33 Okay, let's go back to uh let's go back
20:36 to the
20:39 Twitter and just kind of scroll down
20:41 through here.
20:45 [Music]
20:52 Gemini diffusion Google deep
20:58 mind call. I'm knackered with two hours
21:01 of constant posting, commenting, and
21:02 keeping up with Google IO
21:04 updates. Yeah, I couldn't do it today.
21:07 Like, normally I would be way deeper in
21:09 this than I was. I because I I watched
21:11 it at one point today. We were driving
21:14 between meetings and so I was I was
21:16 watching the the live Google IO
21:20 broadcast. Um, and it was pretty good,
21:22 but I wasn't taking notes or anything.
21:24 So, all right. Um, oh, wait. And let me
21:28 let the Tik Tok people see what I'm
21:36 seeing. Hello, Tik Tok people. You're so
21:40 good. You're such good people.
21:43 You're people. People, aren't you
21:46 fantastic, Bob? Tell him what he's won.
21:50 Uh, he hasn't won anything. This is uh
21:53 not a game
21:55 show, right? Cheerio. Rather wondering
21:58 by chance, Prayel.
22:01 Sir, sir, I do
22:04 declare. What? By chance.
22:08 Now, your
22:14 qualifications. I think we all know the
22:16 answer to that by now, don't
22:18 we? Okay. Gemini's Gemini Live's camera
22:23 is free now for everyone. That's the
22:25 other thing. A lot of these things are
22:26 launched. So, if you go to the Gemini
22:28 app now, it should have a lot of these
22:30 new
22:32 features. So, Gemini Live's camera is
22:34 free now for everyone. real time AI
22:36 assistant in your pocket. We've had that
22:39 with chat GPT. Um, but it looks like
22:41 this one is at least as good, probably
22:45 better, but I don't know. Haven't played
22:47 just see marketing things what they're
22:50 doing there. Showing
22:53 um, you know, pointing it at your car
22:56 dashboard. I've done that with Chat GPT.
22:58 That [ __ ] works. It's crazy. Nvidia,
23:01 Intel,
23:03 AMD, what a powerhouse right now. Pretty
23:05 awesome to see. Yeah, it's pretty crazy,
23:07 isn't it? What kind of trouble are we
23:10 cooking up today? I don't know yet.
23:12 We're going to need a bigger table.
23:14 Yeah, that was a I I got to say that
23:18 um I have given Google [ __ ] over the
23:22 years for their um ridiculously
23:26 bad presentations. You remember the the
23:29 first year the first Google IO meeting
23:32 after chat GPT came out? the the woman
23:36 goes and now we're gonna demo our app
23:38 that does some chat GPT like [ __ ] and
23:42 she didn't have the phone and no one
23:44 could find the phone. The demo phone was
23:47 not available. So, uh they they have
23:51 upped their game which is good. They
23:53 needed to um real time speech
23:56 translation. All right, let's watch this
23:58 video.
24:00 It's nice to finally talk to you.
24:03 Hi, Camila. Let me turn on speech
24:08 translation. It's nice to finally talk
24:10 to
24:17 you. You're going to have a lot of fun
24:19 and I think you're going to love
24:21 visiting the city. The house is in a
24:23 very nice neighborhood and overlooks the
24:26 mountains. That sounds wonderful.
24:29 Wow. I didn't see that at at the uh the
24:32 IO event today. That's pretty slick.
24:36 So, the way that does that, I hadn't
24:39 considered that as an option because I
24:41 always wondered like how are you going
24:43 to hear the person's voice in the other
24:46 language and hear it translated. So,
24:50 they basically do real time audio
24:52 mixing, right? person speaks in their
24:54 native language and when it translates
24:57 enough of it, it lowers the volume of
25:00 that and raises the volume of the
25:02 translation with the cloned voice.
25:05 That's pretty [ __ ] slick. Let's watch
25:06 that
25:09 again. That's a very sophisticated
25:13 um that's a very sophisticated
25:17 solution. They tried to do it with the
25:19 glasses, but it didn't work properly.
25:20 Yeah, that one the the dude's glasses
25:23 hung on him. He he goes, "This is
25:25 probably a bad idea to do." And then her
25:28 thing worked and then his thing just
25:30 hung.
25:36 Hi, Camila. Let me turn on speech
25:41 translation. It's nice to finally talk
25:43 to you.
25:50 [Music]
25:51 You're going to have a lot of fun and I
25:53 think you're going to love visiting the
25:55 city. The house is in a very nice
25:57 neighborhood and overlooks the
25:59 mountains. Wow, that's a [ __ ]
26:01 gamecher. Good lord. Good lord. Hang on
26:04 one second on Tik Tok.
26:08 [Music]
26:33 All right. Fantastic. No, no, I just
26:35 want to wait for a second. No, I just
26:38 had to check something. All right, let's
26:41 let's keep let's keep moving on. That's
26:43 that's pretty amazing.
26:45 All right, let's see. Gemini is now
26:48 built into Chrome.
26:50 This seems like a good idea, but I
26:54 just I I find it intrusive, but I think
26:58 that's just because maybe it hasn't been
26:59 implemented well yet. So, let's look at
27:01 what this
27:04 is. Um, ask Gemini anything. Summarize
27:09 the reviews of the sleeping bag. That's
27:11 a pretty cool use case. All right.
27:21 Very
27:22 cool. Very cool. All
27:25 right. Um, let's see. Kyle had more
27:28 important friends
27:30 visiting. Oh, for the for the meeting
27:32 tonight. Yeah, I had I had uh a business
27:35 dinner. A business
27:38 dinner. Uh, where am I going? What am I
27:41 looking at? Oh, I wanted to go look at
27:42 comments over on the YouTube.
27:45 Let me Hang on a sec. Let me This self-p
27:49 prodducing thing is rough. I'm used to
27:52 I'm used to uh producer Brandon helping
27:55 me navigate through all this stuff. I'm
27:57 doing it on my own. We've been burned
27:59 too many times. Yeah, exactly. Oh, hell
28:00 no. Uh oh. We'll be like, is it Kyle or
28:03 V3? Oh, I saw a really cool thing today,
28:06 too. YouTube is really lagging behind
28:09 Tik Tok. Oh, that's fascinating. All
28:12 right. Well, I don't know. I don't know
28:14 what to
28:15 say. I got I'm doing the best I can with
28:19 hotel
28:21 wifi. What was the real time translate?
28:24 So the real time translate I don't know
28:27 where real time translate lives. Let's
28:29 let's see if we can go find it. Hang
28:39 on. Try it here.
28:50 So it's
28:52 in it's in Gemini within
28:58 workspace. So I guess within within
29:00 Google Meet if you have a workspace
29:05 account. I don't know. I don't know.
29:11 Like I said, I haven't been able to play
29:12 with this stuff today. So, I'm, you
29:15 know, I looked at it today, but I didn't
29:18 um play with any of it. YouTube is
29:20 almost is almost always a tad behind Tik
29:24 Tok. In a few weeks, Gemini Live will
29:28 work with your favorite Google apps like
29:30 Calendar, Keep, Task, Tasks, and Maps.
29:34 Yeah. So, they talked about the Google
29:37 AI assistant being proactive where it
29:40 can just go do things like make
29:43 appointments for you and put them in
29:44 your Google calendar. You know, go see
29:47 if I have time available to go do this
29:49 thing. Um, so that's pretty cool. Google
29:52 just introduced Flow, an AI film studio.
29:56 You can use VO, imagine Gemini all
30:00 together to keep your characters and
30:01 visuals consistent across every shot.
30:04 This changes everything for indie
30:06 filmmakers. Watch how we're using it.
30:08 So,
30:10 let's let's watch the video and then
30:13 I'll this I have access to so we can go
30:16 play a little bit with
30:18 it.
30:21 Um, what's interesting today is a lot of
30:24 the stuff that they
30:26 launched, they launched like a stubbed
30:29 version of it. So like I was in this
30:33 thing called flow which is this video
30:34 editor thing and you can generate new
30:37 images but you can't upload images but
30:39 there's a button that says upload coming
30:41 soon. So I think a lot of the features
30:44 they are holding back the things that
30:47 haven't been tested you know so it's
30:49 sort of like the most stable feature got
30:52 launched today. They're like we got to
30:53 launch some [ __ ]
30:56 So fascinating fascinating
31:05 times and all the
31:09 pieces start falling into
31:12 [Music]
31:15 place. It all feels pretty clear in my
31:18 head. I see these like flashes of
31:22 possibilities almost like I'm traveling
31:24 through dimensions.
31:26 I'm looking down at myself and my
31:28 characters in these different worlds and
31:31 it's almost coming to life on their own
31:33 even though I know I'm in control of
31:34 that
31:35 narrative. It feels like it's almost
31:38 building upon itself at some
31:40 point. You know, you could have an
31:44 infinite amount of endings to your
31:45 story.
31:48 So the work isn't built brick by brick
31:51 by brick. It blooms like a spontaneous
31:56 garden. It grows naturally fully vibrant
32:00 and
32:02 complete. I'm not forcing it. I'm I'm
32:06 just finding it and that's when I know
32:09 I'm in the right place.
32:16 Um so so that's pretty cool. So couple
32:18 of couple interesting things from today
32:21 um about the video stuff
32:25 is they kind of positioned
32:31 VO3 they they had I forget uh what's his
32:35 name? Arnowski, the the film director,
32:38 Dennis Dennis David David Arnowski, I
32:41 think his name.
32:44 Um, they had a real filmmaker that's
32:47 been using V3 and then they had a bunch
32:49 of indie filmmakers that were working
32:51 with
32:53 Arnowski. Um,
32:56 and the the one video they cut together
32:59 of this indie
33:01 filmmaker, like it it was
33:03 like, well, I filmed all the all the
33:06 liveaction stuff that needed to be
33:08 filmed, which was like the people and
33:10 the acting and the and me and Right. But
33:12 then when I needed shots that were like
33:14 really hard like things at the molecular
33:17 or cellular level or you know some sort
33:20 of crazy drone shot or me as a baby
33:23 that's when I used AI. So I felt like
33:26 they were very careful from a marketing
33:30 positioning point of
33:32 view to say that this is a tool that
33:36 augments current filmmakers workflow.
33:39 And in fact the flow the the new flow
33:42 app is very much a linear or a nonlinear
33:46 editor right it's a it's a timeline
33:48 editor um where you can generate in this
33:51 timeline and then edit it like you edit
33:53 a film. So they very much presented it
33:55 as this is a tool to augment
33:58 filmmakers which I understand why they
34:01 did it right because they don't want you
34:03 know a revolt from Hollywood but at the
34:05 same time you know I thought it would
34:08 have been nice if they kind of had the
34:10 cojones to say um and we have people
34:14 like Joy
34:16 Perie who have never made a film in
34:19 their life that are now movie makers
34:22 um you creating a a completely new kind
34:25 of narrative language with, you know,
34:27 with these tools. Um, so I know that's
34:31 kind of a subtle thing, but you
34:34 know, I suppose it makes sense because
34:36 there's a lot of vitriol from the
34:39 creative community. Um, but it was like
34:42 it was really well done. The piece was
34:44 really well done. Um, and that that
34:46 piece we just saw was really well done.
34:48 Those videos are very cool.
34:50 Um, I'm eager to see how OpenAI responds
34:53 to this. I don't know if they can keep
34:55 up. Oh, I think they can keep up,
34:58 Gareth. I I think I think Open AAI's got
35:02 a lot of [ __ ] coming. Um, they've
35:05 already got Sora. Um, it's been out now
35:09 for whatever six months. And when it
35:12 came out, they already were demoing Sora
35:15 2. So, Sora 2 is sitting in the wings.
35:17 It's at least in testing. Um, I saw
35:20 today that it's been confirmed that GPT5
35:24 is in red team testing, right? So, it's
35:27 in safety testing right now. So, that
35:30 probably means within less than three
35:32 months, probably more like one and a
35:34 half or two, um, we're going to see
35:36 GPT5. That is going to
35:39 be, I assume, bonkers. I assume we're
35:42 not going to have to pick models that'll
35:43 do all that [ __ ] dynamically. I assume
35:46 things like Sora and operator and agents
35:49 and codecs and reasoning models and deep
35:54 research and just all the [ __ ] will be
35:56 within it. I assume that custom GPTs
35:59 will either be officially flushed down
36:02 the toilet or made
36:04 better.
36:07 Um my challenge with Google and it looks
36:10 like it got better today. My challenge
36:14 with them is that they they have such a
36:16 lack of product strategy that it's
36:18 impossible to know who their customer
36:20 is. Um, but it that looked a little
36:23 better today. They've got, you know,
36:25 they've got a consumer app. They've got
36:26 the Gemini app. They've got things like
36:29 Flow.google, which is the video editing
36:32 tool. Um, anyway, hi Champy too.
36:35 Champy's back in Denver. I am in
36:37 Paripony or no, I am in Whipony, New
36:41 Jersey. whatever the [ __ ] that is. It's
36:44 near
36:46 Parcipany. Uh we sir, we'd like to call
36:50 this a uh a a new
36:52 township. We call it Paripony. Yeah, but
36:55 but I'm over here on the other side. I'd
36:57 like to I'd like to have my own
36:58 township. Uh yes, Mr. Whipple. Uh that's
37:02 fine. Uh what would you propose you call
37:04 your township? Uh well, that's part. How
37:07 about uh Whippony? Yeah, that's that's
37:11 funny, m Mr. Whipple. That's that's we
37:14 don't joke around here. This is this is
37:16 politics. It's a very serious
37:18 game. Any objections to
37:21 Whippony? Whippony it is. The eyes have
37:24 it. All
37:27 right. After this week, I don't know how
37:29 the Worldwide Developer C conference
37:33 doesn't become a depression fest. Shane,
37:35 I listen, I've talked about this now for
37:37 the past year or
37:39 two.
37:42 Apple's complete [ __ ] absence from
37:45 the AI game is one of the most
37:48 shockingly disappointing
37:50 things I've experienced from them in
37:54 like all of their
37:56 history. They They have all the cash.
38:00 They have very smart people.
38:03 Like what the [ __ ] You know this this
38:07 is what happens when you know Steve Jobs
38:09 put in Tim Cook who's a
38:11 CFO. When you put a CFO in to run a
38:14 company, what the CFO will do is make
38:16 sure that the company doesn't go out of
38:17 business, but they won't do things like
38:20 say invest billions of dollars to win
38:23 the AI game. It's amazing. It's amazing
38:25 to me. Just count your blessings. You're
38:27 not in South Jersey. What's going on in
38:29 South
38:30 Jersey? I was just in South Jersey.
38:32 We're going back to South Jersey
38:34 tomorrow. The [ __ ] going on in South
38:37 Jersey? Tell
38:40 me. Don't squeeze the shaman whipple. I
38:44 remember that commercial. That was a
38:45 good one. All right, let's keep moving.
38:48 We'll keep
38:49 roll. All right, let me flip back over
38:54 to this for a second. See the
38:57 comments. Red slider. Brian, you can
39:00 make the slider catch up. Oh, there you
39:03 go. All
39:05 [Music]
39:16 right. Okay. What's this one? Oh,
39:18 imagine four is here.
39:22 gemini.google.com. All right. Oh, so
39:24 okay. So, we need to go play with Flow
39:26 and we need to go play with Gemini. Uh,
39:28 yeah. Gemini uh imagine 4. So imagine is
39:33 their image generation model. It's
39:34 apparently really good at text now. So
39:37 again, one of the things we're seeing
39:39 and we're going to keep seeing this. All
39:41 of these companies are going to catch up
39:43 with each other. You're going to be able
39:46 to do everything that Google announced
39:50 today within six months at all the major
39:55 players. It's bonkers. And and so I
39:59 think an implication of that
40:01 is it doesn't matter which tool you
40:04 pick. It doesn't matter which company
40:06 you back. Maybe you love Google and hate
40:09 Open AI. Maybe you love Open AI, hate
40:11 Google. Maybe you're a Meta fan. Meta's
40:14 got to pick it up. They're falling
40:15 behind. Anthropic's falling
40:18 behind. Um but but you know, we got some
40:22 we got some pretty major [ __ ] tools
40:24 in the works. Um, so let me change my
40:28 sharing my sharing
40:33 [Music]
40:35 settings. Look at you all giving
40:37 yourselves YouTube um, uh, management
40:41 lessons over there in the comments.
40:43 That's super cool. Yeah, if you're if if
40:45 I'm way behind, you might be able to to
40:48 jump to the end of the of the live
40:50 stream, but if if I said something
40:56 poignant, but if I did say something
40:58 poignant, um, you can go back and just
41:01 drag drag your little, you know, the
41:03 little playhead, the red line, drag the
41:06 ball back and you can jump back in time
41:08 and go listen to my brilliance.
41:12 Um that I said that tongue and
41:33 cheek
41:36 futurism. Where where where for art
41:40 thou?
41:43 Okay. So
41:52 reload scene
41:58 builder reuse
42:00 [Music]
42:02 prompt. So I'm at flow.google
42:05 Google, which then takes you
42:08 to labs.google,
42:18 labs.google/fx/tools/flow.
42:21 Um, I'm starting to feel like a
42:24 courtroom artist with my images of Kyle
42:26 every day. Look what I made.
42:35 Well, you know what's funny? You know
42:36 what's funny, Steve, is
42:44 um what what
42:47 what you're participating in, Steve,
42:50 is
42:52 is
42:56 hyperpersonalization. And and you're
42:58 starting to see it like this. So, so if
43:00 you notice on X and probably on other
43:02 platforms as
43:04 well, somebody puts out a cute little
43:08 video of a famous movie scene, but
43:13 instead of adults, it's talking babies.
43:16 Someone does that
43:18 once and then everyone else can just
43:22 copy it, right? like what they did was a
43:26 particular tool or two, a particular
43:28 technique or two, but it's easily
43:31 discoverable. And so we see like instant
43:34 copies of things that rise above the
43:37 sort of meme
43:39 noise and doing things like, you know,
43:42 making cute little images of me while
43:44 this thing is going on. I mean, that's
43:46 effectively fan art, right? But it's fan
43:48 art that's happening in real time. I
43:51 think we're going to increasingly
43:54 see and have the ability
43:57 to just create [ __ ] that we want. If
44:00 you're just like, "Oh, I like that. I
44:02 want to turn that into something." I'm I
44:04 This is one of my new little side um
44:08 distractions is I'll be surfing through
44:12 X and I'll see something cool and I
44:15 literally just screenshot it and put it
44:17 into chat GPT and say reimagine this and
44:19 give it a different prompt and it it'll
44:22 copy the style. It's amazing. I can do
44:25 parodies of your lives. The
44:26 possibilities are insane. Yeah, exactly.
44:29 Like you you could absolutely do
44:31 parodies of my life, right? Take take a
44:33 transcript from one of my lives, tell
44:36 Chat GPT to write it into a you know a
44:39 90 second script that is that is a you
44:43 know a a comedy of it a uh a parody of
44:48 itself and it'll write it and then yeah
44:51 turn me into a baby or a talking muscle
44:54 car or something like that
44:56 and clone my voice, clone my face. I
44:59 mean at this point
45:02 I I'm probably one of the more clonable
45:05 people out there. I've got a lot of
45:08 content online now. Thanks y'all
45:10 weirdos. Um you're not weird. We're
45:13 regular. Um whether you think humans are
45:16 good in the end or bad, we sure seem to
45:18 be getting there quicker. That's that's
45:21 actually a really good line. Like sands
45:24 through the hourglass. These were the
45:26 times of AI. Exactly. I already cloned
45:29 your voice. Damn it. Oh, that's right.
45:31 You did it for the other
45:35 movie. I call that a charge stang. That
45:37 is definitely a charge stang. Um, so let
45:40 me show you the the way this thing
45:43 works.
45:44 Um, I don't have a good easy way to make
45:48 this a bit or visible to to the Tik Tok
45:51 people, but whatever. Um, let's
45:58 see where
46:02 Let me hide that. I go timeline. No.
46:05 Reuse
46:07 prompt. So, okay. So, let's
46:11 see.
46:13 Uh, how do I get to scene
46:17 builder? Scene
46:22 builder. First frame. Last
46:26 frame. Where's my scene builder?
46:30 Oh, I see. Add to scene. Okay. So, let
46:33 me start something from scratch
46:38 here. So, we'll go to
46:42 flow. Um, you you do have to be to use
46:45 flow. You do have to be a paid member.
46:48 And on my individual account, I'm a paid
46:50 like 20 buck a month
46:54 person. All right. So there's my there's
46:57 my muscle car rolling down the Pacific
46:59 Coast Highway. It's very weird. It
47:01 starts out in a factory and then just
47:02 magically morphs into the Pacific Coast
47:05 Highway. Um the physics on the car
47:08 aren't bad. Like the wheels are at least
47:10 rolling. It does a little sliding of
47:12 that weird AI slide where it sort of,
47:15 you know, goes sideways on the Macadam.
47:18 Um but it's it's not horrible and they
47:21 put people in it and the waves look
47:23 pretty good. The Pacific Coast Highway
47:25 looks good. All right.
47:27 So, watch Flow TV. What's this?
48:24 Huh. So FL TV is kind of just like a I
48:29 want my
48:32 MTA, you know. Um,
48:37 this is basically just a player for ship
48:39 people have made.
48:42 All right.
48:48 All right. How do I do? Okay. New
48:51 project. And then we're going to go text
48:55 to
48:56 video. Now I'm going to go frames to
49:02 video. I'm gonna get rid of
49:05 everything. What's
49:07 this? Those are camera. Yeah. So,
49:10 they've got camera camera moves.
49:13 So frames to video is kind of like
49:17 the the fe the thing that we made fun of
49:21 an with the Ann Murphy with where we put
49:24 her in a New
49:26 York Central Park carriage ride with
49:29 William Defoe in a purple dress and he
49:31 was in a mobile gas station shirt. Um
49:36 that was done in Runway ML. That's now
49:38 that that basic feature is available
49:40 here as well. So the way the way it
49:42 works is you start with a you start with
49:45 making an image or uploading an image.
49:48 You then turn that image into a movie
49:51 clip, a film clip. You then add that
49:55 film clip to a timeline to I think they
49:58 call it a
49:59 sequence. So let me I'm going to click
50:02 the plus button. See, I've still got
50:05 I've got no upload here, but I can
50:09 generate. Um, so let's
50:14 say, all right, we're going to use the
50:17 muscle car image that I did, and then
50:20 we're going to add a new person. So,
50:21 we're going to say
50:22 generate.
50:25 Um, let's go to chat
50:29 GPT and let's say
50:37 um, based on what you know about
50:42 me,
50:44 describe me as an image
50:51 prompt. a charismatic, silverhaired,
50:57 creative. Um, I
51:05 just need the image prompt
51:10 of me, the
51:13 person,
51:14 not the
51:20 setting. Do you use chat GPT for any
51:23 legal stuff? Um, all of our legal stuff,
51:26 I mean, we do a lot of work with pharma,
51:27 so we can't we can't be chat GPT and
51:31 we'll we'll use chat GPT to summarize
51:35 contracts. Um, we'll use I've had it
51:38 hallucinate doing that though. You got
51:40 to be really careful with it.
51:46 um
51:48 we'll we'll use it to draft like initial
51:52 statements that we then give to a lawyer
51:54 to
51:55 vet.
51:57 So, all right. So, here's
52:01 my
52:04 prompt. So, let's go back to
52:07 flow. Where's flow?
52:10 flow text to image generate.
52:13 Okay, so now I'm going to make a picture
52:16 of me as as as chat GPT sees
52:20 me. This should be
52:28 fascinating. All right, that one's good.
52:31 Okay, so so reuse pro. Okay, so now you
52:34 say see where it says use this image. So
52:36 if I say use this image. So there's four
52:38 images over here to the
52:40 right. Right. This is where I can switch
52:43 between them. So this is the new image
52:46 gen thing. Oh, and in effect, let's say
52:48 let's put a
52:50 um
52:54 let's give him a
52:57 70s style
53:00 t-shirt under the
53:03 leather jacket.
53:06 that
53:07 has a
53:11 stylized
53:14 word challenge
53:20 stang in
53:23 bold
53:25 letters. All right. And then so now it's
53:28 over here generating images.
53:37 this this Google IO from from where I
53:40 sit
53:42 [Music]
53:46 um I think it puts Google squarely
53:49 squarely in the game. I think they've
53:51 been playing catchup for two years, but
53:54 I think they're squarely in the game
53:56 now. And I think listen, I think they've
53:58 been in the game with their large
53:59 language model and things like that, but
54:02 um like I they're they're showing up in
54:05 in all the arenas. So, you know, they're
54:08 certainly showing up in research and
54:10 science, but the fact that they're so
54:12 good on creative. Yeah. So, see how it
54:15 says challenge stang and it's spelled
54:17 right. And did I I probably spelled it
54:19 wrong, didn't I? Let's see. Reuse the
54:22 prompt. Yeah, I spelled it wrong. So
54:26 let's fix it with the
54:32 correct the image site here. Well, so
54:36 where I am right now is within their
54:39 their video editing tool called flow. So
54:42 it's at
54:46 flow.google, but you can also use image
54:49 im imagine for the image generation tool
54:53 at gemini.google.com. google.com. So,
54:55 you can do
54:57 both. All right, let's get Challenge
55:00 Stang with the correct spelling. There
55:01 it is. Wait, why did he Why does he have
55:04 [ __ ] brown hair now? Wait, why is
55:10 he 20 years younger with brown hair?
55:19 Read the
55:22 original
55:25 prompt, please,
55:27 and add the
55:31 t-shirt
55:32 like you have. All right. Damn
55:36 it. Yell at your AI.
55:39 Tell I hear this again and again and
55:43 again from people. Probably not from
55:45 many people in this room.
55:48 But one of the things I hear a lot is,
55:50 "Well, I tried it and it it wasn't
55:53 good." Well, did you tell it it wasn't
55:55 good? No. Look what it just
56:01 did. It's
56:05 baffled. So So it went from It went from
56:10 hip mid50s guy to hip mid30s guy.
56:18 the
56:19 hip
56:22 supermodel mid30s guy
56:27 to this is [ __ ] dumb. All right, hang
56:32 on. Let me
56:35 grab one of these
56:38 prompts. Reuse
56:40 prompt. No, that wasn't it.
56:44 That wasn't it. Just hang on. People,
56:47 calm
56:49 down.
56:50 Okay. So, I'm going to copy this
56:54 prompt. Oh, no. Here's what I'm going to
56:57 do. I go to the silverhaired
57:01 box and say reuse prompt. Oh, it ate my
57:05 other prompt. But
57:09 maybe a damn it.
57:14 Uh, his
57:17 t-shirt
57:19 should
57:21 say challenge
57:25 stang in
57:27 cool
57:29 70s t-shirt style.
57:34 All
57:39 right. The AI I'm using is
57:46 flow.google. Take a look at what I
57:50 made. Check the look what I made board I
57:54 just posted. All right. So, those things
57:56 are coming. Let's see. Wait. Let's see
57:58 if we got someone. Challenge. Stang.
58:09 All right. Which one's got the attitude?
58:11 That one's got the attitude. Yeah.
58:13 Okay. So, he's got he's got his
58:15 challenge stang shirt. So, we're going
58:17 to say use this
58:18 image. And so, now we should
58:22 have Oh, wait. We don't. We lost the
58:25 car. Okay. So, now we've got the car
58:28 plus the dude. So, we can
58:36 say the man in the leather
58:39 jacket walks toward the
58:41 car, gets in the driver's
58:45 side, and drives the
58:48 car out of frame. Period. And then we'll
58:52 we'll Oops.
58:58 We'll put a camera shot in
59:00 here of Do they have static?
59:04 Static. So, we just want him
59:07 moving. So, it's going to be a static
59:10 camera and then we put the first frame
59:13 is the car, the muscle car in the
59:15 factory. The second frame is him as a
59:17 dude. The man in the leather jacket
59:19 walks towards the car, gets in the
59:21 driver's side, and drives the car out of
59:24 frame. Now, I've tried to do a similar
59:27 prompt like this in all of the video
59:29 tools. None of them can do this. They're
59:31 always weird. So, let's see if if this
59:35 can do
59:36 it. All right, questions. You are
59:38 correct, Joy. I need to to fullon
59:41 confront my left
59:48 brain. Oh, check the
59:52 the thing. Look what I made. Look what I
59:57 made.
59:58 Steo. Wait, that was from yesterday,
1:00:01 wasn't
1:00:04 it? Hang on. I think I got to reload the
1:00:11 site. Khaki
1:00:15 fits Kyle broadcasting in New York City.
1:00:18 Hey, I don't have a five o'clock shadow,
1:00:20 do I?
1:00:25 with the kitchen. That's
1:00:30 cute. CJ Fletches. Some nice work from
1:00:32 CJ there.
1:00:35 Steo, where' you Where did Whoever told
1:00:38 me to go look here. I don't see your
1:00:40 post. What am I missing? Did you put it
1:00:42 in your regulars?
1:00:44 Maybe. AI learning lab. AI learning lab.
1:00:48 Dennis, this is what happens when you
1:00:50 had coffee. Since being up at 4:30 a.m.
1:00:53 AMI, a brand bible of grumpy genex
1:00:56 marketer plus suno. Oh
1:01:01 man. Yeah, but I'm not there. That's
1:01:04 Brandon, actually. Brandon's at Google
1:01:05 IO this
1:01:08 week. There's no likes tonight. I'm
1:01:10 working overnight. Oh, yeah. We're only
1:01:13 up to 4,000 likes, Mr. We. We miss your
1:01:16 We miss your clicks. Um, I haven't read
1:01:18 a regular novel for about 40 years. I do
1:01:21 all tech programming, math related
1:01:23 stuff. I find Sora and Pixverse and even
1:01:26 Meta to give me the best cinema. Um,
1:01:30 check out I mean, honest to God, like
1:01:33 all of them are getting so good now. But
1:01:35 check out um Hedra, check out Luma Labs,
1:01:39 check out Runway again if you haven't
1:01:41 checked that out in a while.
1:01:44 Um, there's there are so many video
1:01:47 tools now. There's so [ __ ] many. It's
1:01:49 just it's just
1:01:51 maddening. Um, all right. We can listen
1:01:54 to this song, but I've got to
1:01:57 change the way I share my tabs. I got to
1:02:01 change the way I share my
1:02:05 tabs. Oh, and this is going to spawn it
1:02:07 in a new window. God damn it.
1:02:12 Uh, hang
1:02:21 [Music]
1:02:24 on. Uh, wait, where did that
1:02:30 go? Now I'm
1:02:32 confused. AI learning labo
1:02:45 [Music]
1:02:57 Okay. I walk through the fires and lies.
1:03:00 Watch the boo dance in their digital
1:03:03 disguise. Manifest war while your reds
1:03:05 pass through. Tell me again how that
1:03:07 scam helped you. I ate bitter cheese and
1:03:09 call it what you will. Got to watch the
1:03:12 wisdom in a bone deep chill. So keep
1:03:14 your hats and your dopamine hits. I'll
1:03:16 be over here where the real [ __ ] sits.
1:03:18 No sugar, no sage, just blood truth.
1:03:21 This brand ain't a trend. It's a
1:03:23 marketing news for the mid crew done.
1:03:26 We're sharpening minds while you sharpen
1:03:28 your spikes.
1:03:32 I [ __ ] love Sununo.
1:03:35 Fantastic. Fantastic, Bob. That is just
1:03:39 awesome. It's really good. It's really
1:03:43 fantastic. All right. How we doing?
1:03:46 We're doing okay. Let's go back to
1:03:49 flow. That's not flow. Uh the tool that
1:03:52 I'm on is
1:03:54 flow. All right. So now we have I seen
1:03:58 flow width. Yes. So, I've seen it and I
1:04:03 kept missing getting a code and so I
1:04:05 whed about that today and Robert Scoble
1:04:08 DM'd me a code and so I got in today but
1:04:11 I haven't played with it but maybe we
1:04:12 can go we'll go turn it on here. So if
1:04:15 you haven't seen flowith flowith is
1:04:18 different than flow hang on physical
1:04:23 dexterity. So flow.google Google. This
1:04:26 tool is a video image and video
1:04:30 generation tool with an editor built
1:04:34 in.
1:04:36 Flowith.ai is a Manis and Genspark
1:04:40 competitor. It's an agent. Uh, and it
1:04:43 looks pretty incredible. Um, so but I
1:04:46 haven't played with it yet. So, all
1:04:49 right. I've also figured out for the
1:04:51 first time how to fit more hexagons into
1:04:54 a regular sidelength
1:04:56 hexagon.
1:04:58 Jesus, everyone's going to be solving
1:05:01 all the problems. It's going to be
1:05:02 [ __ ] crazy. Google Alpha Evolve blew
1:05:06 my mind with recent 4x4 Matt Mull
1:05:09 improvement. Wow. I don't know what that
1:05:12 is, but I do know it's a term.
1:05:19 All right, let's watch our videos here,
1:05:21 people. All right, so this looks like
1:05:24 the dude is carjacking the car. And
1:05:26 that's not the driver's side. I hate to
1:05:28 break it to
1:05:38 you. All right, that's a fail. Let's
1:05:41 call that a fail. Did I tell you it was
1:05:43 going to get weird? It got weird.
1:05:49 Wait, hang on.
1:06:13 I guess you'd have to
1:06:15 just do an edit. So, so, okay. So, let
1:06:18 me show you the way this
1:06:20 works. So, now that we've got a scene,
1:06:23 let's take this one that's not horrible.
1:06:25 And we'll say, so now, so what we did
1:06:27 before was we we generated images or you
1:06:30 can upload images or you can use text to
1:06:32 image. So, you create an image, then you
1:06:35 turn that image into a video, then you
1:06:37 drop that video into a scene. And so, so
1:06:40 now I've got this little timeline here.
1:06:42 What's cool
1:06:44 is I
1:06:49 can change the in and the out point. So,
1:06:52 I can trim it. So, like, you know, a lot
1:06:54 of times like like this one did, there's
1:06:56 this weird
1:06:58 ass like that thing. I can just trim
1:07:00 that out. And so now this is just
1:07:05 this. He just walks
1:07:08 there and bang, done. And so now I can
1:07:12 say add another scene. Now what I don't
1:07:16 know is I don't know how to add audio to
1:07:19 this. They demoed adding audio. Wait,
1:07:21 what's this audio thing here? No, that's
1:07:23 just
1:07:24 that. I don't I don't know how to do
1:07:28 that. But whatever. But so now I just
1:07:30 say add a new
1:07:32 one. And let's say let's extend
1:07:36 this. So this should he should walk a
1:07:39 little farther. Let's see if it attempts
1:07:41 him to climb in the
1:07:46 car. The plus mark will extend it.
1:07:50 Yeah. Are you able to put a file and
1:07:52 then bring it into your editor? Not
1:07:55 right now.
1:07:57 Audio is only on generations made with
1:08:00 VO3. I thought this was V3 in here, but
1:08:04 they obviously haven't added that. So,
1:08:08 where do you find V3? Is
1:08:11 that I posted some examples earlier.
1:08:15 What's the URL for V3? Does anyone have
1:08:18 that? That's still extending. While
1:08:21 that's extending, let me go look here.
1:08:23 Kyle, you're doing the some Benjamin
1:08:26 Button
1:08:27 [ __ ] I don't know what that means. My
1:08:30 wife loves the talking babies, but only
1:08:32 when the lips sync isn't off. Yeah, I
1:08:34 told her it will get
1:08:37 better. The audio is generated at the
1:08:40 same time. Oh, that's
1:08:42 interesting. But what's the
1:08:46 URL? Very cool. They have an audio
1:08:55 What's good is what good is if you can't
1:08:57 bring into a file. So, King Arthur, it
1:09:00 just it's the
1:09:02 upload functionality is there but grayed
1:09:05 out and it says it's coming soon. It
1:09:07 says dropping
1:09:08 soon. So, it's just temporary like like
1:09:11 that's this what I said earlier if you
1:09:13 didn't hear me earlier. Um, the good
1:09:16 news about the Google event today is
1:09:18 most of the stuff they announced was
1:09:21 available today. They they released they
1:09:24 launched a bunch of [ __ ] today. Most of
1:09:27 what they launched is is stubbed in some
1:09:31 way or whatever the term is not not
1:09:34 fully
1:09:35 complete. Um, V3 is the high quality. I
1:09:38 know, but I thought V3 was I thought
1:09:41 flow was using V3. Maybe it's only No,
1:09:43 it it has to be Well, I guess it's using
1:09:47 imagine 4 because it did text really
1:09:49 well and the the previous thing didn't
1:09:51 use text well, but whatever. If if VO3
1:09:54 is not there, I still need the read my
1:09:56 comment about
1:10:00 TGAT. Good hex and tool info. I love
1:10:03 this live. The B
1:10:05 button lives life from old to young, I
1:10:08 think. Oh, I don't know what that means.
1:10:11 Okay. I think V3 is only available on
1:10:15 the Google Ultra plan. Ah, okay. I do
1:10:19 not have the Google Ultra plan,
1:10:21 but let me jump over. Here's Gemini
1:10:26 Advance. Let me flip over to my
1:10:29 individual account. This is the paid
1:10:32 account. So, that says
1:10:34 pro and V2 and V3 is not in here. So,
1:10:39 yeah, I don't have it. All right.
1:10:41 Whatever. Whatever. I don't want your
1:10:43 stupid [ __ ] super
1:10:45 advanced video tool that makes audio.
1:10:48 Why would I want that? You stupid. Yeah.
1:10:53 $249.99. So, OpenAI has their
1:10:58 um $200 a month plan. Gemini just
1:11:02 launched a 249 plan. Let me tell you
1:11:05 this.
1:11:08 At least for the most part, OpenAI [ __ ]
1:11:13 is all in one
1:11:15 URL, right? You have external URLs for
1:11:18 Sora, Operator, and now Codeex, but
1:11:21 there are links to them directly within
1:11:23 the
1:11:25 interface. Google [ __ ] is all the [ __ ]
1:11:28 over the place. Some of it's in AI
1:11:31 Studio, some of it's in
1:11:33 labs.google.com, some of it's at
1:11:35 gemini.google.com, google.com. Some of
1:11:37 it's in workspaces within your Google
1:11:39 Teams [ __ ]
1:11:42 thing. It's they they are all over the
1:11:44 place. So, if they're going to charge
1:11:46 people $250 [ __ ] dollars a month,
1:11:49 they damn well better tighten this [ __ ]
1:11:51 up and and just put it all in one place
1:11:54 cuz it's it's a train wreck right now.
1:11:57 An absolute [ __ ] train wreck. I I
1:11:59 mean I think it's ballsy for them to to
1:12:01 you know to charge that much
1:12:06 money. Um
1:12:08 anyway, enough comedy
1:12:13 jokes. Did it extend
1:12:16 it? Oh, let's see what happens in the
1:12:19 extended thing. Let's see. Uh let's see.
1:12:26 man walks to the far side of the car and
1:12:33 gets
1:12:36 in. Let's see if it does
1:12:38 it. Let's see if it does it. Let's see
1:12:42 if it does it. I'm definitely not buying
1:12:44 that right now. That's for sure. Yeah,
1:12:46 let them get their [ __ ] together. I
1:12:48 mean, honest to God, Google,
1:12:52 like, it's a it's a little bit cart
1:12:54 before the horse, isn't it? Like,
1:12:57 they're charging as if they have their
1:12:58 [ __ ] together. And I got to say, the
1:13:01 Google IO event, there were not
1:13:04 uncomfortable moments that I saw. I
1:13:05 didn't see the beginning of it, so I
1:13:07 know I I I assume they didn't bring back
1:13:09 Mark, what's his name? The DJ that does
1:13:12 his DJ in pajamas.
1:13:17 Um, but
1:13:21 yeah, they they did much better, but you
1:13:23 know, still 250 bucks. Like, why don't
1:13:25 you get your [ __ ] together first and
1:13:27 then charge 250 bucks? That that would
1:13:30 be that would be my
1:13:34 thing. 47% done. Oh, thank you very
1:13:37 much. What the hell was that
1:13:39 thing? Winston sent the rising key. That
1:13:43 was very cool. What are you making? Is
1:13:46 this like a movie? Yes. So, Google today
1:13:49 announced this new tool called
1:13:53 Flow.google. Um, you have to be a paid
1:13:56 Gemini member to be able to use it. But
1:13:58 I I paid I paid for it on my individual
1:14:02 account, I don't know, six months ago,
1:14:04 and I forgot to unsubscribe. I don't
1:14:06 really Yeah, Mark Revolo, I think is his
1:14:08 name.
1:14:10 Um, was the DJ. Um, and so I still have
1:14:14 that paid account. I didn't think I did.
1:14:16 But what this allows you to do, it's
1:14:18 very similar to some of the other tools
1:14:20 like Lumalabs does this. Um, Runway ML
1:14:24 now does this where you can generate
1:14:26 images in the tool. You can then turn
1:14:28 those images into video. And then what
1:14:31 this tool does that which is really cool
1:14:32 is you can trim the in andout points of
1:14:35 any generation and then you can extend
1:14:38 it and then you can add other scenes. So
1:14:40 you can do multiple AI generations in a
1:14:43 single timeline without having to go to
1:14:46 other software. That's pretty [ __ ]
1:14:48 slick. It's very very rudimentary, very
1:14:50 simple. But like this is all you need if
1:14:53 you just want to cut a couple of scenes
1:14:55 together really really quickly. Like I
1:14:58 will likely be using this tool to cut
1:15:00 together some of my Sydney musical
1:15:03 stuff. Um cool. Yeah, it's very cool.
1:15:05 It's very cool. Does it keep style
1:15:07 consistency from shot to shot? Um, I I
1:15:11 would I would think that it's probably
1:15:14 going to be very very
1:15:16 similar
1:15:17 to Runway ML
1:15:20 um midjourney
1:15:24 um what's it called? Omni reference now
1:15:28 um chat
1:15:30 GPT all of them can do charact
1:15:33 consistent characters and and by
1:15:36 consistent I
1:15:39 mean every generation of me looks like a
1:15:44 first third
1:15:47 or third cousin
1:15:51 incest incest
1:15:55 product. Yeah, Pika does it as well.
1:15:58 Um, so, so the character consistency,
1:16:01 what I would say is if you have a start
1:16:04 image that is is your, you know, you
1:16:06 always use that, you always upload, you
1:16:10 know, the the starter image of that
1:16:12 character and it'll get it pretty close
1:16:14 to that. But if you do iterations of the
1:16:16 iterations, the characters start to
1:16:18 really look different. Wait, flow is now
1:16:20 available. Yes, flow is available now.
1:16:23 Fade to black. Let's see. That's at
1:16:27 99%. Consistency is drawn from memory.
1:16:30 Consistent equals drawn from memory.
1:16:32 Yeah. So, you basically just keep giving
1:16:35 it [ __ ] Was flow. Is that the pro? Is
1:16:38 that in probasic
1:16:41 plan? Yeah, it I do. I do not have
1:16:45 ultra. Ultra was the $250 one. And
1:16:47 that's people have access to V3 are
1:16:51 paying 250 a month. I don't have that.
1:16:54 But Flow is available for the I think
1:16:57 it's now 25 bucks a month for their pro
1:16:59 plan, the Pro Basic plan. Once I build
1:17:03 my video, I definitely need to bring it
1:17:04 into a file then go to my editor. Yeah,
1:17:07 exactly. That that's the thing. I think
1:17:09 this is I think this is um this is in in
1:17:12 the neighborhood of
1:17:13 usable. I I hate to I I hate to make
1:17:17 predictions like this doesn't suck, but
1:17:19 I have a feeling this doesn't
1:17:21 suck. All right. Add to
1:17:25 scene. Man walks to the far side of the
1:17:27 car and gets in. No, that's not the far
1:17:29 side of the car. So, it's not
1:17:33 understanding now. He just looks
1:17:35 confused.
1:17:37 [Laughter]
1:17:45 Uh, where am I going to go? Where am I
1:17:47 going? Oh, let's go to this side. I want
1:17:50 to drive it away. Let
1:17:56 me All right.
1:17:58 Whatever. Whatever. But, you know, but
1:18:02 now I could take that and I This is the
1:18:05 extended one. I could say add that to
1:18:07 the scene. And so, we now have
1:18:16 Wait, did I lose my previous
1:18:19 one? All right. It's buggy. It's buggy.
1:18:22 It's buggy. Can you just flip the in the
1:18:26 image 180 degrees and he's getting the
1:18:28 correct door? Yes, you could. You could
1:18:30 absolutely do that. Can you drag and
1:18:32 drop your script into it yet? No. This
1:18:35 is a very very very So the way Google
1:18:39 described this tool is they were Darren
1:18:42 Areronowski is the name of the director.
1:18:44 They were working with Darren
1:18:45 Areronowski and the indie filmmakers
1:18:48 that were using
1:18:50 VO and those filmmakers said, "Hey, we
1:18:54 need things like to we need to be able
1:18:56 to edit the clips that this thing
1:18:58 generates and we need to be able to
1:19:00 stitch them together and we need to be
1:19:01 able to, you know, generate images in
1:19:04 this thing." So, this very much has the
1:19:07 feel of an engineer worked three
1:19:10 weekends in a row with Darren.
1:19:13 Areronowski to give him what he wanted
1:19:16 to edit
1:19:18 movies. Like this is absolutely a
1:19:21 minimum viable product for generate
1:19:24 images, turn those images into movies,
1:19:27 movie clips, stitch those clips
1:19:29 together. This is the minimum amount of
1:19:32 technology you could use to do those
1:19:34 three things in a single
1:19:37 platform. So very, very rudimentary.
1:19:40 Anyway, all right. It's getting late.
1:19:42 I'm getting tired.
1:19:43 Um, is he getting in the passenger seat?
1:19:47 Yes, he is. Yes. But as someone pointed
1:19:49 out, we could just flip this video and
1:19:51 then he would be getting in the driver's
1:19:52 seat. There's There's always a solution,
1:19:55 people. When you're dealing with janky
1:19:57 technology, there is always a [ __ ]
1:19:59 solution. Always. There's always a
1:20:02 solution, right?
1:20:14 [Music]
1:20:26 All right. On the
1:20:29 YouTube, Kyle, you're doing some
1:20:31 Benjamin Button [ __ ] Uh, means getting
1:20:34 younger. That I know. But you mean I'm
1:20:36 looking younger? because I'm feeling
1:20:38 older than
1:20:39 [ __ ] Um I think VO3 Oh, it might be
1:20:43 because I'm I'm like literally sitting
1:20:46 underneath a light that's sort of given
1:20:48 me a I am I am basking in a glow um of a
1:20:52 light. Um I think V3 is Okay, that's
1:20:55 that. If only Google had AI that could
1:20:58 handle the organization for them. I love
1:21:00 that iMovie. Yep. Get rest, Kyle. Thank
1:21:04 you,
1:21:05 Tom. Thank you. Thank you. Thank you.
1:21:08 He's in Europe. No, I'm in uh I am in in
1:21:12 an exotic
1:21:16 location. It's called Paripony, New
1:21:23 Jersey. Here's a fun fact that you might
1:21:25 not have known if you're not from the
1:21:27 East Coast.
1:21:30 uh New Jersey. At some
1:21:32 point there was some some guy in the in
1:21:36 the curb
1:21:37 business. I call him the curb
1:21:39 empressario of New Jersey. In New Jersey
1:21:42 on main major highways, not major
1:21:46 highways, on major like streets that go
1:21:48 from town to
1:21:50 town, there are no left
1:21:52 turns. There's, you know, those concrete
1:21:55 things that look like a little pyramid.
1:21:57 They they spllay out at the bottom and
1:21:58 it's like a concrete wall and they stick
1:22:00 them together in between lanes of
1:22:02 traffic. Those are called Jersey
1:22:05 barriers because New
1:22:08 Jersey is all Jersey barriers. Every
1:22:11 road is divided with these concrete
1:22:13 barriers. You cannot turn
1:22:16 left. You can't turn
1:22:18 left. Can't turn left. You're like, "But
1:22:21 Kyle, what if you need to turn left?"
1:22:24 Well, you turn right to turn left. How
1:22:26 do you do that? They invented this
1:22:29 concept called a jug handle. So you you
1:22:33 take a right-hand exit and you circle
1:22:36 around and then you and then you turn
1:22:39 left. And if you've never driven here
1:22:42 before, it's [ __ ] insane because you
1:22:44 know you have to turn left up ahead, but
1:22:46 you have to be in the right lane and
1:22:48 then you have to not turn onto a road or
1:22:50 into a parking lot. You actually have to
1:22:52 take the jug handle. And my joke is that
1:22:55 the curb empressario came up with this
1:22:57 concept. He's like, "Here, here's my
1:22:59 vision. All right, stick with me
1:23:01 here. No left turns." No left turns. You
1:23:04 see what I'm saying? Every road gets
1:23:07 gets concrete dividers. Yeah, I know.
1:23:09 I'm in the concrete business. I know. I
1:23:11 know. It's going to benefit my business.
1:23:12 Whatever.
1:23:14 Okay, just hear me
1:23:16 out. No left turns. Why is that? Because
1:23:20 if someone's going to turn left, they
1:23:21 got to stop. And what's that do? that
1:23:23 stops traffic. So, here's what we do. We
1:23:26 put concrete barriers down the middle of
1:23:27 the streets and we build I'm going to
1:23:30 call them jug handles. These little jug
1:23:32 handles. You you just it's a simple
1:23:34 little loop. You do a simple little loop
1:23:36 and yeah, sure there's going to be curbs
1:23:38 all around the jug handles that, you
1:23:40 know, my company will install, of
1:23:41 course, but you know,
1:23:43 hey, it is it's the dumbest [ __ ]
1:23:46 state in all of the
1:23:48 states. [ __ ] ridiculous.
1:23:50 [Laughter]
1:23:54 think all the money we'll save on
1:23:55 traffic
1:23:56 [Laughter]
1:24:02 lights. I just got here in time to
1:24:04 gather. We we are gathered here today.
1:24:08 Yes. Carrying
1:24:09 on. Um so here's what you missed and and
1:24:13 thank you by the way to you and source
1:24:15 camp for co-hosting the salon tonight.
1:24:18 Thank you very much. I was at a business
1:24:20 dinner and could not be there. So, thank
1:24:22 you very much. But they pump your gas
1:24:25 for you in Jersey. That they still do.
1:24:28 Um, even though everywhere else in the
1:24:30 country you have to pump your own gas.
1:24:31 You don't have to do that in Jersey. But
1:24:33 let me tell you, it's
1:24:34 weird. It's weird.
1:24:37 I grew up, you'd pull into a gas station
1:24:39 and someone would come up to your
1:24:42 car, regular or premium, and then they'd
1:24:46 go do it and you give them a little
1:24:49 tip. Nowhere else but Jersey still does
1:24:52 it. It's [ __ ]
1:24:54 weird. All roads lead the neck. Texico.
1:24:58 Yeah,
1:25:00 exactly. Oh, man. Same in Oregon. Well,
1:25:04 it it would make sense in Oregon because
1:25:06 they're just polite, right? They're
1:25:08 like, "Could I get you some organic eggs
1:25:09 with that gallon of gas there, Miss
1:25:12 Murphy?" Miss Murphy, so good to see
1:25:15 you. Fant. I don't know why Oregon just
1:25:17 became or yeah, just became Minnesota.
1:25:20 Minnesota. Because they're so nice. Miss
1:25:23 Murphy, can I get you some premium
1:25:27 today? Fill her up.
1:25:34 Miss Murphy, you must have been on a
1:25:36 road trip. Weren't you just in here on
1:25:38 [Laughter]
1:25:50 Tuesday? Nice people are exhausting,
1:25:52 aren't they? It It's rough.
1:25:56 It I I feel like Okay, this is a theory.
1:26:01 Nice people are passive
1:26:05 aggressive angry people. Right. Right.
1:26:10 Because their overt
1:26:14 niceness basically puts you on the
1:26:17 spot. Like like Okay, if I were to
1:26:20 translate, "Hey, Miss Murphy, how you
1:26:23 doing today?"
1:26:26 I could easily translate that into, "Are
1:26:29 you going to treat me like an asshole?"
1:26:34 A, they put pressure on you. They're
1:26:37 like so nice that you're now like,
1:26:39 "Well, what if I'm having a shitty day?
1:26:42 I don't want to destroy their world.
1:26:44 They're in a happy place, so I have to
1:26:47 play along." And it it it puts a lot of
1:26:49 pressure on me. It's a lot of
1:26:51 anxiety, right? It's a lot of like how
1:26:53 are you doing today, Mr. Shannon? Well,
1:26:56 [ __ ] awful, but I can't say
1:27:00 that because that would make you sad and
1:27:03 clearly your universe would shatter if I
1:27:07 answered honestly. So, here's what I'm
1:27:09 going to say. Just fine,
1:27:12 thanks. See, it's passive
1:27:15 aggressive. Don't be nice to people.
1:27:18 Just be neutral. What's up? How's it
1:27:21 going? Everything okay? Not even that.
1:27:24 Even everything
1:27:25 okay. There's kind of an implied it
1:27:28 better be because I don't really want to
1:27:30 hear your
1:27:38 problems. I never figured it out that
1:27:41 way, but it's
1:27:44 funny. How you doing, Mr. Shannon? I'm
1:27:46 doing [ __ ] shitty.
1:27:49 Oh, I've I've done that before. I went
1:27:52 through a phase where I just I could not
1:27:54 tolerate
1:27:57 niceness. And I would just answer people
1:28:00 honestly. Hey, how's it going today?
1:28:03 Yeah, it's going pretty bad. My kids are
1:28:05 sick and I feel like crap and you know,
1:28:07 things are
1:28:09 challenging. They they don't know what
1:28:11 to do. They literally lock up. They're
1:28:14 like, "Well, at least it
1:28:17 it's gonna be sunny
1:28:23 tomorrow." I feel the
1:28:28 opposite. I have RBF, but I'm the nice
1:28:31 person. The invention of lying. Yeah,
1:28:34 exactly. They never listen to the
1:28:36 answer. Exactly. Hey, how you doing
1:28:38 today? I'm not doing so good. Oh, that's
1:28:40 awesome.
1:28:46 It's so [ __ ]
1:28:48 true. Oh man, I got to do that today.
1:28:52 Hit a deer on the way to work. There you
1:28:54 go. Hey, how's it going, Bob? Yeah, I
1:28:56 killed a deer on the way to work.
1:29:03 Oh, enjoy your meal. You too. Oh my god.
1:29:07 Nice waitresses there. You know what's
1:29:10 funny?
1:29:11 There are two kinds of nice waitresses,
1:29:13 and I suppose this is just people in
1:29:15 general.
1:29:17 Um, there are syrupy nice
1:29:22 waitresses that you can tell her or you
1:29:25 know, weight staff, doesn't have to be a
1:29:26 woman. There are syrupy nice weight
1:29:30 staff that you can tell are on the verge
1:29:33 of tears, right? And it's like their
1:29:36 nicenness is just like, "Please tip me.
1:29:38 Please don't hate me." And it's like
1:29:40 everything is sort of through this
1:29:42 gritted teeth. And then there are people
1:29:44 that are actually like human and they
1:29:48 just connect and they read the room.
1:29:52 Like a really good weight person will
1:29:55 come to your table and go, "Hey, here's
1:29:57 who I am. This is what I do." And if
1:30:00 you're [ __ ] tweaked out or you're in
1:30:03 an intense conversation, they will just
1:30:05 shut that [ __ ] down and get really
1:30:07 efficient. And if you're open, they'll
1:30:10 be open. Those are the best people. And
1:30:12 that's just people. That's just people
1:30:14 101. Just don't be a douche. It's not
1:30:16 that
1:30:19 hard. All right, I'm getting out of
1:30:21 here. Kyle, Larry, David, Shannon. I
1:30:25 know. It's a bit It's a bit of a
1:30:27 Seinfeld episode
1:30:30 tonight. Kyle Kyle on the road is
1:30:32 discombobulated.
1:30:36 So, okay man, I am dehydrated. [ __ ]
1:30:40 traveling just dehydrates me. ABC always
1:30:44 be
1:30:45 cool. Have a have great nice friends.
1:30:49 Yeah, like genuine genuine niceness is
1:30:53 [ __ ] amazing. It's that syrupy
1:30:55 niceness that Oh my god, the [ __ ]
1:30:58 pressure. Just leave me alone.
1:31:01 Now I'm all excited to get get into a
1:31:03 confrontation with a with an overly nice
1:31:06 person. Hey, how you doing today,
1:31:09 Shannon? I'm not doing great.
1:31:12 Okay. Well, let's let's turn that frown
1:31:16 upside
1:31:20 [Laughter]
1:31:25 down. All right, everybody. There's your
1:31:27 homework for tonight. Turn that frown
1:31:29 upside down and don't tolerate fake nice
1:31:33 people. They're just passive aggressive.
1:31:35 They're just angry and they're just
1:31:37 they're pulling you down into their
1:31:39 game. They're not happy. They're trying
1:31:41 to force you to be happy so that they
1:31:44 don't have to confront their own misery.
1:31:46 Don't put up with that
1:31:48 [Laughter]
1:31:52 [ __ ] Midwestern goodbye here. It's a
1:31:55 total This is as Danielle called it the
1:31:57 30 minute goodbye. It's true. Toxic
1:32:00 posit
1:32:01 positivity. Yes.
1:32:04 Affirmations. I'm not I'm not happy
1:32:07 until you're unhappy.
1:32:12 Exactly. All right. I'm leaving. Peace
1:32:15 out everybody. Uh I will see you
1:32:17 tomorrow. I know there's probably [ __ ] I
1:32:19 should talk about, but what's tomorrow?
1:32:21 Tomorrow's Wednesday.
1:32:23 So, tomorrow um the AI readiness podcast
1:32:26 is happening.
1:32:28 Um and it's
1:32:31 uh I forget his
1:32:34 name. Anyway, it's going to be awesome.
1:32:36 It's at 400 p.m. Mountain time, so 6 pm
1:32:40 where I am in Jersey. So, go to a
1:32:44 readiness
1:32:47 project.com and you can see past
1:32:49 episodes of that. This is Ann Murphy and
1:32:51 my um podcast about AI readiness and we
1:32:57 are annoyingly nice on that channel.
1:33:00 It's it's a pretty aggressive channel.
1:33:03 It's it's very
1:33:05 confronting because we're we're super
1:33:07 sweet. We're super sweet. We're just
1:33:09 like, "Hey, everything's awesome." And
1:33:12 then you're sitting there like, "But
1:33:13 what if the robots kill us?" We don't
1:33:15 address that. We don't. We just gloss
1:33:18 right over it. We're like, everything's
1:33:19 going to be
1:33:24 [Laughter]
1:33:29 fine. All right, let the healing
1:33:33 begin. That's really good. Oh my god,
1:33:37 Kyle, I feel like you got some some
1:33:40 wrong life
1:33:43 lessons or I got them right. Bye,
1:33:47 everybody. Don't be too nice. Don't do
1:33:50 it. Puts pressure on other people. They
1:33:53 can't take it. No one can handle it
1:33:54 right now. They're not okay. Don't
1:33:58 pretend like you are. Okay. Bye.