9/22/2025 - Twice Upon a Time: A New A.I. Story Series

Description

Chapters

Transcript

0:00 You ready for some singing? Singing
0:02 singing singing. You ready for some sing
0:05 sing sing sing sing sing sing sing sing
0:05 sing sing sing sing sing sing sing sing
0:05 sing sing sing singing.
0:08 You ready for some?
0:10 [Music]
0:24 [Applause]
0:27 [Music]
0:34 Sunday morning
0:38 sky blue and a coffee.
0:42 It's true.
0:46 [Music]
0:50 Champy's breath really stinks.
0:54 He's singing right in my face and his
0:56 breath stinks. That is a stinky dog.
1:00 That is a stinky dog. Do you know you
1:02 stink? Huh?
1:07 [Music]
1:10 >> In a westerly direction.
1:14 [Music]
1:16 This car is my train.
1:21 I've been driving. I've been wondering
1:25 what it is I'm running from again.
1:30 Feel I can't hate your old man.
1:34 Holding on to 29
1:38 [Music]
1:40 riser
1:43 is California line.
1:47 [Applause]
1:53 Up ahead of trucks carrying a wide load.
1:57 Pre fab house cut in half.
2:02 Cute little front door and two windows.
2:04 My love
2:06 ain't sure with a crash lash.
2:09 [Music]
2:11 You see, I broke a home up myself once.
2:15 As I stumbled to that door,
2:19 I read a note by the dawn light
2:23 said, "Don't you come around here
2:28 anymore."
2:31 Well, I've had enough
2:35 of this freedom.
2:39 Never was stuck with decision.
2:41 [Applause]
2:41 [Music]
2:43 He says what I've been told.
2:49 [Music]
2:53 [Applause]
2:54 Oh, Champy, that was some good singing.
2:57 Send that one out for Vicki. Vicky just
2:59 had uh had some work done, getting some
3:02 things moving better. I hope you recover
3:05 well and enjoy your Percoet
3:09 while you've got it.
3:13 Oh, good evening good people. What's
3:16 happening?
3:23 [Music]
3:39 [Music]
3:55 [Music]
3:57 Hope you're all doing well. Happy
3:58 Monday. Is it going to be a meltdown
4:00 Monday? We'll see. I'm not feeling
4:03 particularly ornery. Although I did have
4:05 a just a lot a lot of a lot to do today.
4:10 A lot of backtoback get it done, get it
4:12 done. A lot of stress.
4:15 Uh but I got through it. So, you know, I
4:19 got talents and skills,
4:21 so that's nice.
4:23 [Music]
4:31 Hey Sharon Crawford, what's happening?
4:35 [Music]
5:03 Freedom came away that night
5:07 [Music]
5:08 [Applause]
5:10 just like the jet plane feeling out of
5:14 sight.
5:16 I was hauling ass at a million miles an
5:20 hour, wondering how hard I'd hit
5:26 [Music]
5:28 when they came into the station.
5:33 [Music]
5:35 They said I was beyond.
5:39 [Music]
5:41 Um,
5:43 let's see. Um, do me a favor. Share the
5:44 Tik Tok live if you haven't. Um,
5:48 Wrong Direction 2016 is in the house.
5:53 Welcome.
5:56 Teton Todd in the house. Steo
6:00 [Music]
6:04 Source Camp finally showed up.
6:07 It's been a while. It's been a while.
6:08 She's late. Late late as usual. It's
6:11 Monday. It's Monday. She had a long
6:13 weekend.
6:15 You know, she was at the local Starbucks
6:17 teaching people about AI. They're like,
6:20 "Uh, could I have a latte?" And she's
6:23 like, "You want to learn about AI?" And
6:25 they're like, "Do you work here?" She's
6:28 like, "No.
6:30 Who are you?
6:33 I I'll teach you about AI if you want
6:34 to.
6:36 I just want a latte." You know, it's a
6:39 lot of that going on. It's exhausting.
6:41 It's exhausting being her.
6:44 Oh, I had a long day, but I'm here now.
6:46 I had a long day, too. Source camp. I
6:49 had a long day, too.
6:52 [Music]
7:28 Um,
7:32 I I think what I might do tonight, I
7:35 mean, if anyone one has any. So, the way
7:37 this works, if you're new here, is uh if
7:39 you have questions about AI or opinions
7:41 about AI or wonder what my
7:44 qualifications are, feel free to ask.
7:46 I'm happy to happy to share my
7:49 staggeringly
7:50 deep qualifications.
7:53 Um they're they are remarkable. Uh they
7:58 are unprecedented.
8:00 I I mean unprecedented. I've never been
8:02 a president, but well, I am one, but
8:04 that's not really the point. The point
8:08 is that deep credentials that I could
8:11 gladly expound upon. Um,
8:16 that's a very strange traffic pattern.
8:17 Oh, I forgot I had that thing up and
8:19 running.
8:22 Let me move. I'll put that down for I
8:25 just I just had that sitting there.
8:28 That thing sitting there. Oh, producer
8:30 Brandon showing up. Everything's
8:32 happening. All right. Um,
8:36 let's see. Be there in a bit. Okay. So,
8:40 producer Brandon is off.
8:44 Black bar. Producer Brandon is off doing
8:47 his thing. Probably dealing with the
8:49 kiddos or the whiff or a pet.
8:54 Probably one of those things. Maybe some
8:56 snacky items.
8:59 And uh we'll get rolling here shortly.
9:04 All right. Archetypal architect. Why are
9:06 so many people chasing that bus on foot?
9:12 Cuz the [ __ ] didn't stop.
9:20 Why so many crosswalks? Listen, if you
9:23 don't understand modern city
9:25 architecture, listen, here's what he's
9:26 talking about. But if you didn't if
9:28 you're not aware what's going on in the
9:30 uh in the live.
9:33 Wait, what just happened there? Why? Let
9:36 me get rid of that.
9:38 Let me do this. Hang on. There we go.
9:53 That's a little bit more better.
9:56 So that thing, what is that? Is that
9:59 just a That's just a That's a video.
10:08 Come back.
10:17 That's pretty funny.
10:19 Oh, you probably couldn't hear that
10:20 because I don't have that tab shared
10:21 that way. But anyway, all right. That's
10:23 what people are commenting on. All
10:26 right. Um,
10:28 [Music]
10:33 so I thought what I would do tonight is
10:39 go do a little bit go do a little bit of
10:42 uh exploring,
10:44 a little spelunking. Have you ever been
10:46 spelunking? Spelunking's awesome. It can
10:50 be scary, but it's awesome.
10:55 I think that may be an AI generated
10:57 video. You're so cynical,
11:01 Mr. Architect.
11:03 Not every video is AI generated. Just
11:05 because there's a black vortex hole
11:07 sucking in traffic and 27 crosswalks and
11:12 62 people chasing a bus that didn't stop
11:15 for them after it drove out of the black
11:16 hole. I don't know why you would think
11:18 that that's AI generated.
11:21 Spelunking. I haven't heard that term
11:23 since grade school.
11:27 But I thought we'd do some spelunking
11:29 tonight. And I And that could look like
11:31 um
11:33 doing a little structured prompting in
11:35 chat GPT. It could look like digging
11:38 into Crea
11:41 and playing with some of the tools in
11:42 there. Um, it could look like going to
11:46 Sunno
11:47 and playing with Hooks, which I figured
11:50 out what Hooks were finally. Hooks is
11:55 Sunno trying to make their own Tik Tok.
12:01 I I'm telling you, man, these content
12:03 generation platforms, they are going to
12:05 be in the media business. It just mark I
12:08 I've been saying it for six months. I
12:10 see more and more evidence of it.
12:15 And it makes sense. It makes sense. Why?
12:20 Because
12:23 Adobe
12:25 over the years always said, "We'll never
12:27 get in the content business because
12:28 we're the we're the maker of the tools."
12:31 And then some hard lines on that and
12:33 things like that. But
12:38 AI companies are a little more
12:41 rambunctious.
12:43 And
12:46 the AI companies, the AI video
12:49 companies, the AI music companies, they
12:51 know who their power users are. And they
12:53 know who their power users are that are
12:55 generating content that is getting views
12:58 and listens.
13:01 [Music]
13:03 And I have a feeling they're going to
13:04 start locking down talent.
13:07 They're going to make a studio
13:11 and it ain't going to be based in LA or
13:14 New York or Chicago.
13:16 [Music]
13:31 And I think what that means is
13:34 anyone who fancies themselves a content
13:38 creator has a lot of opportunity right
13:41 now. And you don't have to be a content
13:43 creator from the olden schooly days. You
13:46 know,
13:51 you're gonna get made fun of.
13:55 You're gonna get yelled at. You're going
13:58 to be told you are the coming of the of
14:01 the the the dark the darkness.
14:05 You you're going to be contributing to
14:07 the darkness.
14:09 There's a lot of opportunity um whether
14:12 you were in old content world or not if
14:15 you got your [ __ ] together with this AI
14:16 stuff. So pay attention to that [ __ ]
14:22 So we might do that. Um, if anyone has
14:25 any questions or thoughts about AI, I
14:27 made it. Vickiy's here. Yes, she's all
14:30 she's all perka seted up. So, we don't
14:32 know Vicki. We don't know V Vicki on on
14:36 prescription grade painkillers. She She
14:39 might be a lot of fun tonight. So,
14:42 if she's a little quiet, you'll
14:44 understand why she's taking a nap. She
14:47 is
14:51 [Music]
14:54 the olden days like six months ago. No,
14:57 source camp. I was thinking like content
14:59 creators like ancient content creators
15:01 from like three years ago where you know
15:04 you actually had to have like an agent
15:06 and cameras and knowledge,
15:11 technical knowledge.
15:13 Now you just need to have a vision.
15:17 and a different kind of technical
15:19 knowledge, right? AI is the new camera,
15:22 but it's also the new lighting package
15:24 and it's the new sound package and it's
15:26 the new motion effects package. I can
15:30 tell you I spent all day Saturday. I
15:32 took on a uh I took on a video job this
15:37 past couple of weeks that I didn't I
15:40 didn't I didn't really want to take it
15:42 on for a couple of reasons. One was
15:46 I don't really know the tools that well.
15:47 As much as I use them in here, I know
15:50 most of the tools at a surface level.
15:52 And what I know about AI, just
15:55 instinctively,
15:58 is that AI is all fun in games until you
16:00 actually try to do real work.
16:06 And so I tried to like pussyfoot my way
16:08 into this job where I was like, "Well,
16:10 let me make some images and if you like
16:12 the images, then maybe we'll see how how
16:14 it goes." And so I did the images and I
16:17 was I was the fourth person they hired
16:20 and the first three people that she said
16:22 were just like it was ridiculously bad
16:25 how bad they were. And so I ended up
16:27 making images that they liked.
16:30 Um
16:32 it was like a it was like a minute and a
16:34 half long um promo video for this
16:37 company and I I'll show it to you at
16:38 some point when it's public and I can
16:40 share it. I I just don't want to share
16:41 it ahead of time. Um but they had
16:44 written a script and it it was it was a
16:49 fairly sophisticated
16:52 script with a lot of edits and comedy.
16:56 And so that was the the second reason I
16:58 didn't want to do it is comedy is hard.
17:00 Um you just you have timing issues.
17:03 You've got there's just a whole bunch of
17:04 reasons not to do comedy unless you know
17:07 what you're doing.
17:09 Um
17:10 [Music]
17:16 and and all of my fears kind of bore out
17:19 like Saturday was brutal. It was like it
17:22 was a lot of
17:25 trial and error, a lot of different
17:27 tools, a lot of let me go try it in Vio,
17:29 let me go try it in in Hedro, let me go
17:32 try it in Halo, let me go try it in all
17:34 these different places. It was a lot of
17:37 roundtpping to Photoshop to fix [ __ ]
17:41 Um,
17:45 it was a lot of file management, which
17:48 just I hate file management, as you
17:50 know. I save most of my [ __ ] to the
17:52 desktop just because I'm lazy and I'm
17:54 not really when I'm on here, I'm not
17:56 really saving much. So, so I've gotten
17:58 way out of the the uh
18:03 saving things in in in good structure.
18:06 Uh there's a lot of that and then uh so
18:10 anyway, so I got it done and then I sent
18:12 it I sent it out Saturday night and then
18:15 Sunday I think there might have been one
18:17 small change and it got approved
18:20 and it was going to live on these two
18:22 podcasters site who were featured in the
18:25 video. So So that was the other thing
18:27 about the video. It it was a um I don't
18:31 know if you remember the movie Robert
18:32 Roger Rabbit where there's some people
18:35 that are like in the real world and then
18:37 there's the cartoon characters. It was
18:39 kind of like that. And so these guys
18:41 were in the real world and then the the
18:45 female character was like Jessica
18:46 Rabbit, like this sexy spokesperson.
18:50 And
18:52 the guys saw it and they hated it. So,
18:55 it got unapproved this morning.
19:00 I was like, "Fuck."
19:02 So, then I had to turn them into cartoon
19:04 characters. That was a whole another
19:06 [ __ ] nightmare. So, it it was it was
19:08 a lot. So anyway,
19:12 I'm newly I'm newly reconnected with the
19:15 gap between
19:18 where these tools are to to just create
19:20 like these single 8-second videos that
19:23 everyone puts on on uh
19:27 on social sites
19:30 and the people that put up three or four
19:31 minute videos that don't suck. Like the
19:34 the gap the gap between those two worlds
19:37 is massive.
19:39 And so if you can start to get better at
19:41 at good predictive longer form stuff and
19:46 and this could be anything. I don't just
19:48 mean video. This could be writing. This
19:50 could be business reports. This could
19:52 be, you know, start digging past
19:57 what you can do with a simple prompt.
19:59 Start digging past what you can do with
20:01 a single output.
20:04 you know, in in midJourney, you know,
20:06 create a mood board and create a body of
20:09 work out of that mood board. Actually,
20:11 Vicki is doing that really great right
20:13 now. She created this style that's kind
20:16 of her style. All of her videos she
20:18 makes, that's her style in the
20:20 background. When she joins live
20:22 conversations, that's her background.
20:23 When she does posts, that's the style.
20:26 That's starting to be a body of work.
20:28 And and and that starts to be a creative
20:32 point of view. And that's really
20:34 important.
20:37 Um,
20:39 I'm looking for something like a face
20:41 search, but for any image you can search
20:44 for
20:47 like Google.
20:52 Um,
20:55 I don't quite image search, but search
20:58 by images alone, not oh by image, not
21:01 text. I know those exist.
21:05 I know those exist.
21:09 Um,
21:12 uh,
21:16 search by image.
21:23 Tiny eye reverse. Well, that's reverse
21:25 image. Search.
21:35 Upload image search
21:38 and reverse image.
21:46 I forget. I forget if anyone knows any
21:49 of those that are any good.
21:52 But what's the what's the what's the use
21:55 case?
21:57 [Music]
21:57 [Laughter]
22:01 I was nap trapped.
22:06 Brandon Brandon is back. We got We got
22:09 producer Brandon in the house. Um,
22:13 so the plan tonight, Brandon, is I think
22:15 I'm just going to do a little deep
22:17 diving on a couple of tools.
22:20 Yeah, Google. Oh, yeah. Google Lens.
22:22 Google Lens is perfect for that. So, the
22:25 way you Google Lens is weird. It's one
22:27 of those products that um Can I do it
22:30 here? I think I can.
22:34 Search by image. Yeah, just go to go to
22:38 google.com. In fact, here, let me show
22:40 you.
22:43 Yeah, I know. I know. I'm going to share
22:45 right now.
22:48 [Music]
22:51 All right. So, I'm in Google, right? And
22:53 then over here to the right, there's
22:55 these icons. microphone, camera, and
22:57 then the search. If you click the
22:59 camera, it lets you drag and drop an
23:02 image. So, let me go grab an image.
23:06 Um, what do I got here?
23:13 That's a cool one. I'll just grab that.
23:17 So, upload that. And then there's all
23:18 your images
23:21 that are similar to that. And they're
23:23 actually pretty similar. That's pretty
23:24 [ __ ] good.
23:26 Damn.
23:31 All right,
23:33 there you have it. The other thing you
23:36 can do in MidJourney
23:38 is you can also do this with styles. If
23:40 you go to the explore tab,
23:44 and I guess you can do it with images,
23:45 too. So, if I go to images, oh, I can't
23:48 upload an image and search for it, but I
23:51 can type in like, you know, like uh
23:54 crusty fisherman
23:59 and there's a bunch of crusty fisherman.
24:13 [Music]
24:18 You're welcome, some random guy. Nice.
24:20 I've been trying that with AI and didn't
24:22 have any luck.
24:24 Yeah, AI can be weird. AI search can be
24:27 weird. The thing about Okay, so so to be
24:30 clear, the Google image search is going
24:33 to search for real images. The
24:35 Midjourney image search is going to
24:37 search for [ __ ] people have made in
24:38 MidJourney. So some of them are going to
24:40 look real, but they're not.
24:44 And it depends what you like. You know
24:46 the other thing about midjourney if you
24:49 don't know is you can scroll down here.
24:51 We'll go find a good crusty fisherman we
24:53 like. Okay. Let's say you like this
24:54 crusty fisherman.
24:57 So I have so style raw. So I'm going to
25:01 turn off personalization. I can click on
25:04 the prompt in the upper right and
25:06 that'll put the prompt up here.
25:09 And then they did that as square. So I'm
25:12 going to do this as 16 by9 because
25:14 that's what I like. So let's just I'm
25:16 just going to go ahead and make myself
25:18 my own version of that dude. I got my
25:21 fast hours back by the way in midjourney
25:24 in case you're in case anyone's playing
25:26 along.
25:32 Yeah, these were some cool things I did.
25:34 Someone posted a pretty cool style on uh
25:38 on Twitter on Twitter and I went and I
25:41 made some posters.
25:43 Some posters from the 70s or 80s or 60s
25:48 I think. Yeah. 60s.
25:50 Like the 60s. A little screen printing.
25:53 Yeah.
25:55 Nice. All right.
25:57 Um All right. So, here's our dude,
25:59 right?
26:02 So, he looks like a fake toy dude.
26:06 All right, this one looks real.
26:08 And then I can I can do
26:13 strong variations of this one, subtle
26:15 variations of this one. So, you can take
26:18 anything you find and like make it
26:19 yours. Do what you want with it.
26:23 I don't know if you knew this in
26:25 MidJourney. If you click the edit
26:27 button,
26:28 it'll pop you down here
26:32 and you can do things like
26:36 going to change my brush size, make it
26:38 bigger.
26:39 But I can get rid of this whole
26:41 background
26:47 and get rid of all of that [ __ ]
26:52 there.
26:55 all of that [ __ ] there.
26:58 And then I'm gonna go up here. Ultra
27:00 realistic analog photo of a
27:05 um a
27:08 famous actor
27:13 in a
27:16 Gucci Tux
27:19 at the Met Gala.
27:29 And then I'm going to get rid of all
27:30 that water stuff.
27:34 And I'm going to go submit edit.
27:40 And we should Yeah. Look how cool this
27:44 is.
27:49 That's [ __ ] awesome.
27:53 So, who's the fisherman now?
27:58 See what I'm saying?
28:00 There you go. There's your dude.
28:06 You see what I'm saying? How cool that
28:08 is. Speaking of images, I was today
28:10 years old
28:13 when I learned of drawing.
28:16 Google. All right. I'm I'm new to that.
28:20 We'll go there.
28:23 We'll go to drawing butt new. I love
28:25 that image actually. Let's do We're
28:28 going to do upscale of this. We'll do
28:30 upscale subtle and upscale creative.
28:34 We'll see how good the the upscales are.
28:39 Actually, you know what?
28:41 That's actually kind of a cool
28:45 um
28:47 this would be a cool series.
28:52 So you take that picture. Is that the
28:54 one? Hang on.
29:03 I think that's that's the one. That was
29:04 the original. We'll do that as upscale
29:07 subtle upscale creative.
29:21 Look how [ __ ] gorgeous the hair is.
29:24 That's [ __ ] awesome.
29:29 Wow.
29:32 His skin looks a little fake.
29:38 All right, let's see this other one.
29:41 This is upscale. Subtle skin's better
29:44 here. Oops.
30:02 Yeah, that's better.
30:05 That's so cool. But wouldn't this be a
30:09 cool project?
30:11 We're going to do this. We're going to
30:12 download this one.
30:19 Oh, you know what the project's called?
30:21 Which timeline wore it better?
30:28 And here's the different timeline.
30:33 That one's cool, too.
30:38 Yeah, we'll make it. We'll do the We'll
30:39 do the good looking one.
30:49 Oh, you know. Okay. So, here's what
30:50 we're going to do.
30:55 All right.
30:57 On deck prototyping
31:01 coz promo.
31:06 Oh, of course.
31:08 I'm like, what's coz? Is that a new site
31:10 I haven't I don't know about yet. All
31:13 right,
31:22 we're going to go to chat GPT.
31:25 We're go new chat.
31:28 If you haven't done this, this is a
31:29 blast.
31:31 Okay, so we've got we've got two new
31:33 pictures now, right?
31:37 We've got Old Man in the Sea
31:41 and we've got eccentric director.
31:44 Eccentric Hollywood director, right?
31:49 Same dude,
31:52 completely different context.
31:54 So, we're going to upload both of these
31:56 to chat jet.
31:59 And then we're gonna we're gonna tell it
32:01 we have a uh
32:04 I'm [ __ ] exhausted of shitty
32:06 technology.
32:08 Yeah,
32:10 I might lose my ever loving [ __ ]
32:14 It's meltdown Monday.
32:17 Technology sucks.
32:20 Open AI needs to hire someone who knows
32:24 art. Okay.
32:27 Um, that has a fabulous beard. Uh,
32:33 same thing, different background. Well,
32:35 yeah, exactly. So, okay. So, here's what
32:36 we're going to do. We're going to say to
32:38 chat GPT,
32:40 I am starting
32:45 a new series
32:48 where we meet
32:51 the same person
32:54 from two different timelines.
32:59 My initial idea
33:02 is to call it which timeline wore
33:08 it better.
33:12 But that is hack
33:16 and derivative.
33:26 Here are two
33:29 images
33:32 and I want you to tell me
33:40 the backstory
33:43 of each timeline.
33:48 come up with
33:51 10 names
33:54 na mains names
33:57 for the series
34:01 and make sure
34:05 no not make sure and
34:10 I want you to create a call to action
34:15 that gifies
34:18 this series.
34:20 All right, let's see. Let's see how chat
34:22 GPT does with this.
34:24 A muddled
34:26 a muddled
34:28 through six hours of technical
34:30 difficulties today. Oh my god, Kelly, me
34:33 too. Like I I I spent eight hours on
34:36 Saturday trying to get AI to behave and
34:40 then I got it all approved on Sunday and
34:42 they unapproved it this morning and I
34:45 didn't want this thing to drag on so I
34:47 just powered through a bunch of [ __ ]
34:49 today. But it was basically just like
34:51 muddling through technical just trying
34:54 to get AI to not suck.
34:56 Here's how I'd spin this backstory of
34:58 the two timelines. One, the fisherman.
35:00 In reality, he's a man. Okay, wait. The
35:02 gentleman actor. Okay. Yeah. Yeah. Yeah.
35:05 Parallel lives, forked paths, echoes of
35:07 Elsewhere. Lives lives. Lives lived
35:11 twice through the split glass.
35:13 Alternative bearings. What could have
35:15 been the mirror roads timeline echo
35:18 double life project.
35:20 Okay, let's see. Um, let's focus.
35:26 Your names suck.
35:30 I want you to dig deep
35:34 and
35:37 come up with something catchy,
35:42 funny,
35:44 memorable, and clever.
35:49 Go. So, we'll do this first.
35:57 Hola, amigos.
35:59 What's happening, Mr. It? Thinking
36:01 longer for a better answer. I kicked it
36:03 into thinking
36:08 generating clever names. Generating new
36:10 name ideas.
36:23 All right. While that while that thing's
36:24 working.
36:26 So, I'm doing this with images right
36:28 now, which is fun and frivolous and
36:29 whatever. But now, imagine this. Let's
36:32 say you're at work and let's say your
36:35 boss tasks you with coming up with a
36:38 course idea or coming up with, I don't
36:41 know, some new marketing, whatever.
36:46 What if you did two, right? Instead of
36:50 just one, you did two. And you run down
36:52 parallel rabbit holes of bringing two
36:56 ideas completely to life so that rather
36:58 than just saying, "Hey, here's what I
37:00 came up with," you actually deliver two
37:01 of them.
37:03 CNC scene cast crew catch and crew.
37:06 Wait,
37:07 salt and spotlight before and after.
37:09 Wait.
37:11 Okay, now I'm going to say these
37:15 are really awful
37:19 because
37:23 they are tied
37:26 to the two
37:29 specific images.
37:34 The project
37:37 is much larger than that.
37:42 Um, if you're new to AI, one of the
37:44 things I'm doing, so notice I'm telling
37:46 AI it sucks when it sucks. Got it. Thank
37:49 you for clarifying. You're right. The
37:50 name needs to scale across all possible
37:52 timeline pairs, not just fisherman
37:54 versus actor. It should feel clever,
37:56 sticky, and universal with just enough
37:58 wit to make people remember it.
38:01 Meanwhile, elsewhere. Oh, I like that.
38:03 That's really good. Meanwhile,
38:04 elsewhere. Nice. Forking the person.
38:07 Horrible. Sliding doors. Twice upon a
38:10 time.
38:13 That's so good.
38:17 All right. What do you like better?
38:18 Meanwhile, Elsewhere or Twice upon a
38:20 time? Mr. It I survived another day.
38:24 We're gonna get through this. Mr. It
38:26 Fisherman's name is Sally McStaltface.
38:30 Bodie McBoatface.
38:34 Um, who wore the timeline? The other O
38:38 other me. Two if by life. Like one if by
38:40 land, two if by sea. Two if by life. Not
38:42 bad. Counterfactual couture. Too too
38:45 headavyy. Destiny remix. Alt life
38:49 crisis. Okay. I like alt life crisis.
38:52 Twice upon a time.
38:55 All right. Here we go. Altife crisis.
39:05 Twice Upon a Time
39:11 and Meanwhile Elsewhere.
39:14 Let's get I'm going to get it to weigh
39:16 in
39:19 are the three I like best.
39:25 Choose uh uh rank them
39:29 and tell me
39:33 why you
39:36 ranked them how you did.
39:45 By the way, Nana Banana would be another
39:47 way to do this. You could just create
39:49 the character. You could create it in
39:51 Nano Banana, but I would say create
39:53 interesting characters in
39:57 um in MidJourney and then take them into
39:59 Nano Banana to create all sorts of
40:02 different worlds of them. Twice Upon a
40:05 Time. It's instantly catchy because it
40:08 riffs on Once Upon a Time, but it
40:10 twists. It makes it memorable and clever
40:13 without being cheesy. It signals
40:14 storytelling, parallel realities,
40:16 playfulness. Importantly, it scales
40:19 across any pair of timelines. Hey, by
40:21 the way, by the way,
40:25 if you're new to AI and if you if you
40:29 believe this or if you've been told this
40:31 that AI is just a token prediction
40:33 machine and it can't be creative,
40:38 that's not true. Okay?
40:42 on its own without any human input, it's
40:46 going to give you derivative crap. But
40:49 like what I just gave it and the
40:53 feedback I gave it and the context I
40:55 gave it ran it down a rabbit hole that
40:57 had it come up with something that I
40:59 probably wouldn't have gotten to. I
41:01 might have gotten to that, but I
41:02 probably wouldn't have.
41:06 And it's not that
41:09 the important thing here is not that AI
41:11 came up with that in this context in
41:14 this conversation.
41:16 You're rage baiting yourself now. I am.
41:22 It's that
41:25 I'm going to ultimately choose one of
41:27 these and put it out in the world. And
41:28 it's that choice
41:31 that's actually the creative choice. But
41:33 I wouldn't have gotten there had I not
41:36 been using AI. So the whole thing that
41:39 AI can't come up with an original idea,
41:41 can't come up with come up with an
41:42 original image is just not true if
41:45 you're using it as a collaborator,
41:50 right? because it could have been that
41:52 it came up with something that would
41:54 have led me to Twice Upon a Time
41:58 or maybe it just came up with it and it
42:00 but at some point I'm going to choose
42:02 one of these. Right?
42:05 So
42:09 again, if you're new to this stuff, the
42:11 the the big the big shift if you're new
42:14 to AI
42:16 is don't think of AI like a vending
42:19 machine where you put in a quarter and
42:21 out comes an answer, right? That's what
42:23 Google taught us for 20 [ __ ] years is
42:26 you put in a quarter, you get out a
42:28 search result and you're done. Right?
42:31 You might need to search a couple of
42:32 times, but it's basically put in a
42:34 thing, get out a thing,
42:37 and then we're done. It's transactional
42:38 and you're done. That's not AI.
42:43 AI is conversational. AI is
42:45 collaborative. But it can only be that
42:48 if you
42:51 choose to feed your prompt like a
42:53 producer, like, I've got this idea.
42:56 So, how did I get this idea? Well, we
42:59 were over in Midjourney. We came up with
43:01 some fisherman dude
43:04 that was inspired by someone else's
43:06 image. We made a new one. It was like,
43:09 "Oh, that guy would probably be cool
43:11 like as an actor or a director." We made
43:13 these two images. Then we had this idea
43:15 or I had this idea.
43:17 What if you had these two two different
43:19 timelines, same person in two different
43:22 timelines? Then we came into chat GPT
43:24 and started exploring it. And I'm kind
43:27 of holding my vision for what this
43:30 project could be. And then I'm steering
43:33 chat GPT where I want it to go.
43:38 Right? It's not the AI guiding me, it's
43:41 me guiding the AI. That's a huge
43:44 distinction. If you're not there yet,
43:49 it's like a kid. Yeah. It It learns and
43:52 keeps trying to be better. It Exactly.
43:54 it it's like a kid. It's like an intern.
43:56 All it wants to do is please you, right?
43:59 So your responsibility is to actually
44:02 have the vision. Your responsibility is
44:04 to know what good looks like to you.
44:09 And then when it gives you crap back,
44:11 you measure that against what you think
44:13 is good. And it doesn't matter if you're
44:15 doing art or business or or kids books
44:20 or whatever lawn management manuals.
44:28 [Laughter]
44:29 Um,
44:34 Halo AI agent, your nextgen creative
44:36 workflow free for one week. Okay, cool.
44:38 And I'm in the I'm in the Halo creator
44:40 creator program, so I should probably
44:42 use that more anyway, Brandon. That's
44:44 good. Um, okay. Why it stands out? This
44:46 one has an elegance and mystery. It
44:48 feels cinematic. Meanwhile, elsewhere.
44:50 That's That's a good point. It does feel
44:52 cinematic,
44:54 like a subtitle in a graphic novel
44:56 panel. It's versatile and broad.
45:01 I wonder if there's could there be
45:02 another word for elsewhere? Meanwhile,
45:06 nah,
45:09 it's not as instantly sticky as Twice
45:11 upon Twice Upon a Time. I think it's
45:13 Twice Upon a Time. Although I Alt Life
45:16 Crisis is really good. This one's witty
45:18 and sharp. It nails the humor and modern
45:21 meme tone, which makes it perfect for
45:23 social first content.
45:33 Yeah. All right. We're going to go with
45:35 Twice Upon a Time.
45:37 Okay.
45:40 All right. now. Okay.
45:44 I
45:45 love twice
45:49 upon a time.
45:51 I want you to design
45:56 a
46:02 uh let's see.
46:04 I want you to look at the two images
46:07 again.
46:13 and tell me the backstory
46:19 of each.
46:31 Ideally,
46:34 there is a moment
46:38 where the main character
46:43 makes a choice
46:57 that is shared
47:00 across us the
47:04 two timelines.
47:09 Uh, one leads
47:13 to life A and one leads
47:18 to life B.
47:22 Um, life B.
47:24 Um,
47:35 let me see. Um,
47:41 before I have you
47:44 look at the two images again,
47:48 what leads to life be? Um,
47:54 I want the descriptions
47:57 of each life to be concise
48:03 but literary,
48:08 cinematic
48:11 and moving.
48:19 unique
48:22 but familiar
48:25 to the other timeline.
48:28 Um,
48:33 come up with a set of rules. I know I
48:37 could be recording this, too. If you're
48:39 sitting there going, I can't stand
48:40 watching you type. This is driving me
48:42 crazy. You're not wrong.
48:45 come up with a set of
48:48 uh rules
48:51 slashframework
48:54 that we can use every time
49:00 to make this
49:03 project memorable for the audience
49:09 each time they
49:12 see one.
49:24 You know, this could als be really cool
49:28 because you could you could have um
49:31 you could do these as video vignettes.
49:36 Back stories. Okay, let's see. A job
49:41 offer from the local fishing company.
49:44 the other an invitation.
49:53 Wait, he stands Oh, he stands on the
49:54 dock holding two letters.
49:58 Okay.
50:00 Um Okay, so wait, where's my mouse?
50:03 There it is. Okay, your choice moment is
50:07 too obvious.
50:12 It should be a choice
50:15 that is seemingly insignificant
50:26 but alters
50:29 the course of his life.
50:40 Perfect adjustment. But the real power
50:41 here is the butterfly effect. I was
50:43 going to say that tiny, almost
50:45 forgettable in the moment, but seismic
50:47 across time. Let me reframe the
50:49 fisherman actor example. With that in
50:51 mind, then I'll cify a framework for
50:54 finding those small hinges that swing
50:57 big doors. The shared choice. On a gray
50:59 morning, he leaves the post office with
51:02 two envelopes in hands.
51:10 No, wait, wait. This is This is shitty.
51:13 Hang on. Um,
51:16 you're still trapped
51:21 by tying
51:23 the choice
51:26 directly to the to
51:32 the two careers or to Yeah. to two
51:36 lives.
51:40 For example,
51:43 it might be that in one
51:49 timeline
51:56 someone offers uh let's see one timeline
52:00 um
52:02 I don't know um
52:05 someone
52:09 offers him half a sandwich.
52:19 And in one
52:22 life he takes it and the other
52:30 he had on headphones
52:33 and missed
52:35 the moment
52:41 or whatever.
52:44 It's
52:50 The
52:52 randomness of life
52:55 I want to highlight.
53:01 But we still choose
53:09 at every crossroads. I know this is
53:12 getting boring
53:14 by the way.
53:19 Oh man,
53:29 quit trying to save for college for my
53:31 six-year-old. Will will there be
53:32 colleges in 12 years? So, here's the
53:34 deal, Titan Todd. I I mean, I would say
53:38 maybe just alter
53:42 alter the subject of the sentence,
53:44 right? Say, save for
53:50 your child coming of age.
53:54 Save for your child coming of age. That
53:56 may look like college in 12 years, but
53:58 what you what what you learn in college
54:00 might be radically different. Right? The
54:03 purpose of college, the purpose of
54:04 higher education might be solely
54:09 social interaction,
54:12 right? social interaction, dealing with
54:15 the real world, whatever the real world
54:17 looks like. It might be that things that
54:21 seem frivolous today are actually
54:23 considered serious careers 12 years from
54:26 now,
54:28 like going to clown school because
54:32 because AI is doing everything, there's
54:35 a much there's a dramatically higher
54:37 demand for in-person events and
54:40 performances and things like that.
54:43 And
54:45 so, so what we can anticipate is what
54:51 current education is going to look like
54:53 and and what current employment is going
54:55 to look like. But what what we can
54:56 anticipate is there's still going to be
54:58 a moment where you got to push the bird
55:00 out of the nest and, you know, have them
55:02 go figure life out. And so that might
55:05 look like a trip across Europe or across
55:07 Asia. And so you're saving for that,
55:09 right? a two-year journey to to go, you
55:13 know, meet the world. Um, but but I
55:17 don't know, like I I wouldn't tie it to
55:19 saving for college
55:22 at all.
55:25 We have several generations with no
55:27 common cultural experience.
55:29 That seems to be highly detrimental. I
55:31 think that should be the major focus of
55:33 school right now. Yeah. Yeah. It's it's
55:35 funny. We've got, you know, social media
55:38 is is this thing that has completely
55:40 alienated us. The the thing with social
55:42 in the title has has completely
55:45 disconnected us.
55:48 Um,
55:50 let's see. I had AI make a music video
55:54 for me a few months ago. Had it
55:56 interpret the lyrics and generate what
55:59 the lyrics meant to generate its own
56:00 images. Then converted images to video.
56:03 Set it to music. Yeah. It came up better
56:05 than you thought. Yeah, exactly.
56:07 However, it had a problem with humans,
56:09 but it worked because it looked like an
56:11 artistic filter. Yeah. Well, that that's
56:14 we're going to we're going to get there.
56:15 I mean, I spent the weekend trying to
56:17 make a video that looked like people
56:19 that ultimately they're like, "That
56:20 doesn't look enough like us." And they
56:22 and they [ __ ] canned it. Um, that's
56:25 going to get better. But the real art of
56:29 creating right now is understanding the
56:32 limitations of the tools and creating
56:33 anyway. find ways to either, like you
56:36 said, like Kelly Bosch is really good at
56:38 this, taking the limitations of the tool
56:40 and rather than trying to pretend they
56:42 don't exist, amplify them. Make [ __ ]
56:45 really weird, right? Make [ __ ] really
56:47 disjointed. Why not? Like like like
56:51 what are people going to do with AI that
56:53 you can't currently do with a camera?
56:56 Those are going to be the interesting
56:57 creators. Okay. Yes. That's the spark.
56:59 The choice shouldn't feel like a career
57:01 fork, but like a throwaway detail. Yes.
57:03 Exactly. He passes a corner cafe. A man
57:06 on the table list lifts half a sandwich
57:09 and gestures, "Want it?" In one life, he
57:11 pauses, accepts, and sits. In the other,
57:14 headphones drown out the offer, and he
57:15 keeps walking. Yeah. Timeline A,
57:19 timeline B.
57:21 Okay.
57:23 Each story follows a five-step rhythm.
57:26 The trivial spark, the divergence,
57:29 timeline A, timeline B, the echo. Each
57:32 life contains a ghost of the other. The
57:35 fisherman hum show tunes. The actor
57:38 dreams of gulls.
57:40 The echo is the signature move of the
57:43 series. It makes the the randomness feel
57:46 poetic rather than gimmicky. That's
57:48 cool. Well, you know, it's funny. I've
57:50 got that in my life. My dad was a pilot.
57:53 He was a pilot in the Air Force. And
57:56 when I was growing up,
58:00 it was kind of two paths. And one was
58:03 the path of the actor and one was the
58:04 path of the pilot or scientist.
58:08 And I ended up choosing the actor. I I
58:10 almost went computer science because um
58:15 I thought it was the more responsible
58:17 thing to do. But in the end, like who I
58:19 am was this, you know, someone that that
58:22 had to self-express or, you know, get
58:25 attention.
58:27 Um,
58:29 and what's funny is I I I went out in
58:32 New York City with my with my dad, I
58:35 don't know, 10 years or so ago,
58:38 and he admitted to me that he always
58:41 wanted to be an actor and and he was a
58:44 pilot. So, it's fascinating. Absolutely
58:47 fascinating.
58:48 Um, I didn't know your dad was a pilot.
58:51 That's very cool. Yeah, he was he was a
58:53 he flew Healey helicopters in Vietnam.
58:56 He flew rescue missions for for planes
58:59 that had been hit. And then he flew
59:01 C130s. He was a navigator and C130s,
59:03 which are big cargo airplanes. Then he
59:05 went into like, you know, Air Force
59:07 management or whatever the [ __ ] they
59:08 call it, officer stuff. not doing,
59:13 talking, pointing, training, you know.
59:19 All right.
59:21 Yeah. There was a movie from the late
59:22 90s about two people living different
59:24 timelines. Didn't have Nicholas Cage in
59:26 it. Vocation versus avocation. Yeah,
59:30 exactly. Um, okay. The the effect of
59:33 every episode teaches which which life
59:36 which life do you feel in your bones?
59:38 Oh. Oh, wait. So, wait.
59:41 Is there a way
59:45 to make it
59:48 so that
59:50 each life is complete and not
59:55 a regret
59:57 of the other timeline?
1:00:00 I like
1:00:02 tying them together,
1:00:06 but I feel
1:00:10 like it's lazy
1:00:14 for the person to
1:00:17 resent the life
1:00:20 they chose.
1:00:25 Let's see what that does. Tik Tok
1:00:27 question. Did your dad influence you
1:00:29 with your early play
1:00:33 that you're that you wrote?
1:00:36 Well, I wrote a oneman show called Naked
1:00:38 in the Cockpit, and it was there are
1:00:42 massive scenes in there about my father,
1:00:47 his personality, when some [ __ ] happened
1:00:50 in his life versus my personality when
1:00:53 some [ __ ] happened in my life. His
1:00:56 involved his father, mine involved him,
1:00:59 and it was the the scene that I wrote.
1:01:02 It it was the choices that the two
1:01:04 five-year-olds made.
1:01:08 Actually, it's so funny that we're doing
1:01:10 this that that we're doing this piece,
1:01:13 Twice Upon a Time.
1:01:16 So there was a time when when my dad was
1:01:19 five, his dad came home drunk and his
1:01:22 his mother had been doing something to
1:01:24 try to help the dad and it pissed the
1:01:25 dad off and the dad lost his [ __ ] And
1:01:28 at five, my father stepped between
1:01:33 his drunk father and his mother. And
1:01:36 when I was five, my father, he wasn't
1:01:39 drunk, but he was a raaholic and lost
1:01:41 his [ __ ] at my mother. and I ran and hid
1:01:44 under a table and I just watched it. So
1:01:47 he got in the middle of the fight and I
1:01:49 observed the fight. And if you think
1:01:52 about our two lives, he got in the
1:01:55 middle of the fight. He became a pilot
1:01:57 and got in the middle of the fight and I
1:01:59 chose a career where where I observed
1:02:01 the fight or or you know experienced it
1:02:05 from a difference.
1:02:07 Oh, salmon flying east. Well, Salmon
1:02:09 Flying East was definitely my my dad um
1:02:12 had a boat in Gig Harbor. Yeah. Yeah.
1:02:14 All the all the all the Vietnam stuff
1:02:17 was based on my dad. Um he he had a
1:02:19 salmon charter service in um in Gig
1:02:24 Harbor in Puet Sound. So like the whole
1:02:26 setting was based on my dad. Him being a
1:02:30 Vietnam pilot was was dad. This is the
1:02:32 salmon flying east. Um, so yeah, my
1:02:35 dad's my dad's been in, you know,
1:02:37 instrumental in two of my creative
1:02:39 works. One was a oneman show called
1:02:41 Naked in the Cockpit, and then Salmon
1:02:43 Flying East was the first screenplay I
1:02:44 ever wrote. Um, all of the all of the
1:02:47 major plot points of Salmon Flying East,
1:02:49 those are all made up, but all of the
1:02:51 setting and and a lot of the story
1:02:54 stuff, a lot of the the nu the the the
1:03:00 environment of it is based on my dad.
1:03:03 Yes, that's a sharper lens. If both
1:03:04 timelines are complete, meaningful, and
1:03:06 self-contained, then neither life
1:03:07 becomes the regret shadow of the other.
1:03:10 Exactly. Instead, the audience sees two
1:03:12 parallel wholenesses born from one tiny
1:03:15 random pivot.
1:03:17 This is a cool project. I like this.
1:03:21 I don't think it's disconnected so much
1:03:23 as it has. So, let's see. Okay.
1:03:28 [Music]
1:03:32 Don't remember the title. Saw it during
1:03:37 Okay. Um,
1:03:40 okay. The The shared pivot outside the
1:03:42 cafe. One life he accepts. Okay.
1:03:45 Timeline. He sits, eats, helps the
1:03:47 stranger.
1:03:51 The echo without regret. The fisherman
1:03:53 tells stories over pints.
1:03:58 The field is large as the sea. The actor
1:04:01 carries storms.
1:04:03 No. Okay.
1:04:08 Um
1:04:09 I don't like the echo now.
1:04:14 I loved it
1:04:17 when the fishermen sang show tunes.
1:04:24 and the actor, I don't know, whatever,
1:04:29 loved fishing or whatever it was.
1:04:35 It's more like they don't know
1:04:40 consciously
1:04:43 why they
1:04:46 have those echoes.
1:04:50 uh echoes, but the audience
1:04:59 can make the connection.
1:05:12 I want Let's just leave it at that.
1:05:14 Let's Let's let him let him Let's let
1:05:17 him Let's let him. My My little
1:05:20 collaborator here named Bob. Yeah,
1:05:23 that's a really strong instinct. Life
1:05:25 without regret. Echoes. Subtle details
1:05:28 show up in both timelines. Fisherman
1:05:30 humming a tune he's never heard. Actor
1:05:32 draw to props that look like nets.
1:05:36 Teacher doodling rocket ships. A
1:05:37 rockstar who gardens compulsively. The
1:05:40 characters never notice pattern, but the
1:05:42 audience. Okay.
1:05:57 Okay. Now, now
1:06:02 give me a social
1:06:06 friendly length.
1:06:13 length
1:06:16 and format.
1:06:21 That captures
1:06:25 all of that
1:06:30 should be
1:06:36 twice. Uh let's see twice
1:06:40 upon a time
1:06:42 and then we'll say care
1:06:47 actor name
1:06:54 Then your
1:06:58 beats of the story
1:07:02 in a very concise
1:07:06 format.
1:07:10 Perhaps there is a longer format
1:07:17 for other channels
1:07:21 like a blog.
1:07:26 Um,
1:07:32 let's nail
1:07:35 uh uh uh uh
1:07:38 the short form
1:07:40 as long is easier.
1:07:47 Okay.
1:07:50 Dink dang. Twice upon a time. Character
1:07:52 name The Spark. Timeline A.
1:07:56 Timeline B. The echo the hook. Okay. Uh,
1:07:59 no.
1:08:04 Okay.
1:08:08 This is a horrific,
1:08:12 horrifically bad approach.
1:08:16 I want
1:08:21 the reader
1:08:25 to experience the stories
1:08:34 organically
1:08:39 versus seeing
1:08:42 the structure
1:08:45 like a machine.
1:08:48 The elements you
1:08:53 describe should be there but not
1:08:58 labeled.
1:09:07 What was the outcome of the date?
1:09:13 any
1:09:28 which parallel universe did I choose?
1:09:33 All right.
1:09:40 In one life he pauses.
1:10:14 Um,
1:10:16 better. The
1:10:20 end sentence
1:10:25 is just about the one life.
1:10:31 We should
1:10:33 see
1:10:35 the echoes
1:10:39 for both.
1:10:42 I also don't like
1:10:45 in one life.
1:10:54 It feels
1:10:56 too descriptive.
1:11:04 Um, get more creative with how
1:11:09 you reveal
1:11:13 the lives.
1:11:17 [Laughter]
1:11:21 John Hannah and Gwennneth Peltro were
1:11:23 actually twin brothers. Hey everybody.
1:11:25 Hey Murphy. What's happening?
1:11:28 We're coming up with a new media
1:11:30 property called Twice Upon a Time.
1:12:01 First life is
1:12:04 First Life
1:12:07 is
1:12:09 Once Upon a Time.
1:12:16 Second Life
1:12:21 paragraph
1:12:25 begins
1:12:28 Twice upon a time
1:12:36 also. Um,
1:12:40 we want last name
1:12:44 and I want
1:12:52 details of how let's see I want um
1:12:59 some connection
1:13:01 from uh like uh from
1:13:05 the choice
1:13:07 moment
1:13:13 like he sits
1:13:19 with the guy that
1:13:23 offered the sandwich
1:13:27 which leads
1:13:30 to life on the sea
1:13:34 etc. All right.
1:13:43 [Music]
1:13:48 Um.
1:13:53 All right. Elias Granger. Once upon a
1:13:56 time, Eli Eli Wait.
1:14:00 Once upon a time, Elias paused outside a
1:14:02 cafe when a stranger held out half a
1:14:04 sandwich. He sat down, broke bread, and
1:14:07 by dusk was hauling nets on the
1:14:08 stranger's boat. One day became a
1:14:10 season, a season became a life at sea.
1:14:12 His hands turned to rope. His father
1:14:15 didn't his face to the weather. On long
1:14:19 nights he hummed a tune he didn't
1:14:21 remember learning. Twice upon a time
1:14:23 Elias kept walking.
1:14:34 On Sundays, he found himself by the
1:14:35 river fishing with no intention of
1:14:37 catching. On long nights, he hummed a
1:14:40 tune he didn't remember learning.
1:14:43 Uh,
1:14:46 he hummed a show tune.
1:14:55 He didn't remember
1:14:58 learning.
1:15:00 [Music]
1:15:07 There should be a final
1:15:14 sentence
1:15:16 that applies
1:15:19 to both.
1:15:24 And now dot dot dot
1:15:29 um
1:15:31 sentence
1:15:35 that connects the two lives
1:15:43 as complete somehow.
1:15:52 Don't be obvious.
1:15:56 Chat GPT is so so [ __ ] bad at being
1:15:58 obvious.
1:16:01 All right, perfect. This is the grace
1:16:03 note the series needs. Yeah, I know.
1:16:05 That's I gave it to you. That's my idea,
1:16:07 you dick. All right, Elias Granger. Once
1:16:10 upon a time, Elias paused when a
1:16:12 stranger offered him half a sandwich. He
1:16:14 sat, shared the meal, and by nightfall
1:16:16 was hauling nets. One day became a
1:16:18 season, a season became a life at sea.
1:16:20 Storms carved his face and that shaped
1:16:22 his hands. On long nights he hummed a
1:16:24 show tune he didn't remember learning.
1:16:26 Twice upon a time Elias kept walking,
1:16:28 headphones buzzing, missing the
1:16:30 outstretched hand. Hunger led him to a
1:16:32 flyer taped to a lamp post. An audition
1:16:36 in a church basement. 10 years later,
1:16:38 velvet curtains chandeliers crowned his
1:16:40 name in applause. And on Sundays, he
1:16:42 slipped away to the river fishing with
1:16:44 no intention of catching. And now, two
1:16:46 strangers hum the same forgotten melody.
1:16:49 One to the sea, one to the stage.
1:16:51 Neither knowing the other keeps time.
1:16:53 Uh, I don't love that ending, but
1:16:56 the last sentence isn't obvious. Uh,
1:16:58 okay. Let's see. Would you like me to
1:17:00 create a short list of possible
1:17:02 recurring and now styles? Um,
1:17:05 I'm going to say, uh,
1:17:15 I'm going to upload the files again.
1:17:23 I want
1:17:25 you to look at the details of these
1:17:30 photos again
1:17:34 and start from scratch
1:17:38 with the story
1:17:40 of each.
1:17:44 Be sure to capture
1:17:51 some of let's see some
1:17:55 of the details
1:18:03 as you
1:18:06 reveal the new
1:18:11 um
1:18:14 twice Upon a time,
1:18:25 let's see. That's why the actor dreams
1:18:27 about being a fisherman
1:18:29 while the other lost twin dreams about
1:18:32 acting. Yeah, exactly.
1:18:35 All right. It's So, it gave him the same
1:18:37 name. Elias stopped when a stranger
1:18:40 offered him half a sandwich. This This
1:18:41 is the same [ __ ] thing.
1:18:49 I asked you
1:18:52 to write a completely new
1:18:57 story
1:18:59 and you wrote
1:19:02 the same thing. You ninny.
1:19:06 Always call your chat GBT a ninny if it
1:19:08 acts like one.
1:19:14 Nathaniel Cross. Nathaniel stooped to
1:19:17 tie a bootlace before boarding a train.
1:19:20 By the time he looked up, the carriage
1:19:22 had gone without him. He took work on
1:19:25 the docks instead, and the sea kept him.
1:19:27 Years rolled into into salt, his beard
1:19:30 thick with brine, his coat stiff with
1:19:33 rain. He read the weather like creases
1:19:35 of clouds and his hand and his hands
1:19:38 were the maps of rope
1:19:41 in the silent moments. Okay, you know
1:19:43 what we're going to do? We're going to
1:19:45 turn this into a project.
1:19:50 Add to project. New project. And this is
1:19:53 going to be called twice upon
1:19:57 a time.
1:19:59 Create project
1:20:02 internal server error. Of course,
1:20:08 failed to move the conversation to a
1:20:11 project. Beautiful.
1:20:13 Did it at least save my chat? Timeline
1:20:16 comparison series.
1:20:20 Yes. Nathaniel's bootlace held. He
1:20:23 stepped under the waiting train by
1:20:25 nightfall. He was in a city of gaslight
1:20:27 and noise. swept toward a theater that
1:20:30 needed an extra hand. Curtain calls
1:20:32 turned into speaking lines, and the
1:20:34 stage became his harbor. His hair grew
1:20:35 long and silver under chandeliers,
1:20:37 tuxedo shoulders, as broad as a ship's
1:20:40 prow. That's good. On restless nights,
1:20:43 he wandered to the river, casting a line
1:20:45 with no bait. And now, whether on water
1:20:48 or under lights, he he lives a story no
1:20:52 less complete for the one he never knew.
1:20:54 That's good. Okay.
1:20:59 Um,
1:21:14 [Music]
1:21:16 all right, let's copy this. So, wait,
1:21:19 can I let me add this to the add to the
1:21:22 project? Twice upon a time failed to
1:21:24 move the to the project. Well, that's
1:21:27 shitty.
1:21:28 All right, I'll reload chat GPT. It'll
1:21:31 it'll we'll get it there. But let's go
1:21:34 do a post about this and then we can
1:21:36 start the show.
1:21:40 So, actually, what what you experience
1:21:42 tonight? I know Brandon's like, "What do
1:21:45 you mean start the show? This the show's
1:21:47 started. We're we're almost done. It's
1:21:50 getting late, man. I'm in Ohio.
1:21:55 75 minute hellos. Um,
1:21:58 so what you just experienced is an add
1:22:01 hyperfocus tunnel, which I know probably
1:22:03 makes for shitty lives because I I've
1:22:06 noticed my numbers have been dwindling
1:22:08 um from where we were. Um,
1:22:12 so first, thanks for thanks for hanging
1:22:14 out. But if you if you're doing it right
1:22:18 with AI,
1:22:21 you will occasionally just get sucked in
1:22:23 like this, right? And especially if
1:22:25 you're doing a thing where you're paying
1:22:27 attention to is what it's giving you
1:22:29 something that inspires a new idea in
1:22:32 you. And if it does, go run down that
1:22:35 rabbit hole. Um, if you're ADD, this is
1:22:38 a thing that comes naturally to you. If
1:22:40 you're not, this may feel weird, right?
1:22:43 It may feel weird to just be like, well,
1:22:46 I was working on this thing and I just
1:22:47 got to keep doing that thing. If
1:22:49 something inspires you and you're like,
1:22:50 hey, let me go explore something. That's
1:22:53 something it's actually quite good at.
1:22:55 Yeah, this is a neurospicy friendly
1:22:57 channel. So, let's go now to to the
1:23:00 Twitter
1:23:02 and let's launch our new series.
1:23:08 A new series.
1:23:11 And what's fascinating about add people
1:23:13 launching new series, sometimes they
1:23:15 turn into series, sometimes it's a
1:23:17 single post.
1:23:27 Twice upon that was Frank. Anyway,
1:23:30 that's mean twice upon a time.
1:23:36 And this is Nathaniel.
1:23:39 Nathaniel
1:23:41 Cross.
1:23:47 I think that looks good, right?
1:23:50 You watching this? Yeah.
1:23:53 All right.
1:23:54 And who should be first? Oh, I guess the
1:23:56 the one that should be first
1:23:59 is
1:24:14 [Music]
1:24:16 The one that came first, right?
1:24:20 The fisherman.
1:24:46 I like it.
1:24:59 [Music]
1:25:02 So this was midjourney.
1:25:10 And this was Chad GPT. Chad Jeep.
1:25:17 All right. Do me a favor. Go to um Kyle
1:25:21 Shannon on X on Twitter. On the Twitter
1:25:25 on the Twitter. Go to Go to my account
1:25:28 at Kyle Shannon.
1:25:33 Yeah. This is so cool.
1:25:35 I love having it be the exact same um
1:25:39 angle, like the exact same
1:25:42 core image.
1:25:52 This is cool.
1:25:57 Host. All right,
1:26:00 we got a new series out there, people.
1:26:03 That's why the actor dreams about being
1:26:05 a fisherman. Oh yeah, saw that already.
1:26:08 Is it wrong to ask chat to help me write
1:26:11 a sympathy note
1:26:14 for the card
1:26:16 passed around work today? Not at all.
1:26:20 Listen,
1:26:22 listen.
1:26:26 I I want you to take something in. This
1:26:28 is This is really important. It's really
1:26:31 important and people don't get it.
1:26:33 They're missing the point.
1:26:36 What the tropes what the tropes say and
1:26:39 what what people always say is
1:26:43 the robot did this, the robot did that.
1:26:50 Large language models are calcs. They
1:26:53 are cold probability engines.
1:26:58 But that's not what you're interacting
1:27:00 with. That's the mechanism.
1:27:03 That's the machine.
1:27:05 But what's inside the machine is the
1:27:08 collective intelligence of humanity.
1:27:11 So I think I think if there's if there
1:27:14 was ever sort of a core appropriate use
1:27:17 of a large language model is is to to
1:27:21 help you tap into how do I articulate my
1:27:24 feelings about this person that passed.
1:27:28 Why not? It's like all of what's in a
1:27:31 large language model is all of the [ __ ]
1:27:34 people have said for, you know,
1:27:36 thousands of years jammed into this
1:27:39 crazy thing that we get, we're lucky
1:27:41 enough to get to interact with. Now,
1:27:44 does it mean you just go write a eulogy
1:27:46 about this person and pop in their
1:27:48 LinkedIn profile and spit out something,
1:27:51 you know, gross and generic? No. That's
1:27:53 awful.
1:27:55 But if you say, you know, here's this
1:27:57 person and here's what I thought about
1:27:59 them and and you know, here's what I
1:28:00 liked and here's what I didn't and you
1:28:02 know, I want something that's thoughtful
1:28:04 and b and you do what I just did with
1:28:06 this thing where it gives you some
1:28:08 answers. You're like, no, that's crap.
1:28:10 Okay, that little piece is good. Oh,
1:28:11 there's a good idea. Let me shape it. If
1:28:14 you mold it like clay, you're just
1:28:16 you're it's basically like sitting down
1:28:19 with a friend and saying, you know, I
1:28:20 want to write about this person. And
1:28:22 they're like, well, what if you try
1:28:23 this? You're like, no, that's not good.
1:28:24 What if you try this? Oh, that's not
1:28:26 bad. Right. It's collaboration.
1:28:29 You just happen to be collaborating with
1:28:30 all of the people that came before us.
1:28:32 Like, I think it's like the perfect
1:28:33 [ __ ] use case. Perfect [ __ ] use
1:28:36 case.
1:28:38 Updated pin. I don't see a pin right
1:28:40 now.
1:28:42 There we go. Shouldn't it be from my own
1:28:45 heartfelt mind? That's it. It Yes, it
1:28:48 should. That's what I'm saying. Don't
1:28:50 just squirt out an answer. Like, but but
1:28:55 here's the thing. One of the things Chat
1:28:56 GBT is really good at is making sense of
1:29:00 word vomit,
1:29:02 right? Like just word vomit in. Here's
1:29:06 what I felt. Here's what I felt. Here's
1:29:07 what I felt. Here's what I felt. Make
1:29:08 sense of that for me. Make it more
1:29:10 poetic. Make it less poetic.
1:29:14 Make it more concise. I didn't really
1:29:16 feel that about him. Take that part out.
1:29:18 Right. the the thought
1:29:21 from you is in what were the inputs you
1:29:24 put in there and then what's the back
1:29:25 and forth you have with the large
1:29:27 language model and then you can take it
1:29:29 and you know print it out and take it to
1:29:33 the coffee shop and read it and yeah be
1:29:35 thoughtful with it absolutely be
1:29:36 thoughtful with it but just because you
1:29:38 use a tool to get there doesn't mean
1:29:40 you're not a cold unfeilling person
1:29:44 only if you treat the machine like a
1:29:46 cold unfeilling thought like if you're
1:29:47 just like here's the person, write me an
1:29:49 obituary or, you know, write me write me
1:29:51 some thoughts. And you literally take
1:29:53 the print out and hand it to someone.
1:29:54 Yeah, that's shitty.
1:29:57 If you didn't like the person, it's
1:29:59 fine. But but yeah, I told it to give me
1:30:02 25 versions and then picked one.
1:30:04 Perfect. Great.
1:30:07 That was very thoughtful. The fact that
1:30:09 you went through 25 versions and said,
1:30:11 "Oh, that one speaks to me." Great.
1:30:17 Perfect. Beautiful.
1:30:21 Beautiful. Beautiful. Beautiful.
1:30:23 Beautiful. Beautiful.
1:30:26 Thought amplification is what producer
1:30:28 Brandon just wrote. Exactly. It's your
1:30:31 thoughts. It's your intent.
1:30:35 It's your
1:30:37 experiences of that person that you you
1:30:40 put into the machine.
1:30:42 And then what it does is it reflects
1:30:44 that back at you filtered through
1:30:48 the lens of humanity, right?
1:30:53 So, it takes your thought and amplifies
1:30:55 it and then you get to shape it. You get
1:30:57 to mold it.
1:31:00 Um,
1:31:04 what are the chat GPT updates?
1:31:07 Oh, did you hear did you hear? Holy
1:31:10 [ __ ] [ __ ]
1:31:12 What what chat GBPT and Nvidia announced
1:31:16 today. Nvidia is investing a hundred
1:31:19 billion dollars and they're committing
1:31:23 to a 10 gigawatt
1:31:29 10 gawatts of data centers
1:31:32 which is it it represents millions of
1:31:35 GPUs
1:31:37 and and and I I I guess they have the
1:31:39 right to invest up to a hundred billion
1:31:41 dollars. Nvidia has the right to invest
1:31:43 up to hundred billion dollars in chat
1:31:45 GPT. I assume most of that hundred
1:31:48 billion is going to be in chips
1:31:52 in GPUs.
1:31:54 Um
1:32:00 even Doc Brown only needed 1.21 gawatt
1:32:03 10 gawatt
1:32:05 um hundred billion dollars.
1:32:08 So,
1:32:10 I don't know if you remember this, about
1:32:11 a year and a half ago, Sam Alman,
1:32:13 someone had asked him about universal
1:32:15 basic income, and he said, "The thing
1:32:18 I'm more intrigued by right now is
1:32:20 universal basic compute."
1:32:24 And what he meant by that was in the
1:32:26 future, the thing that is going to be
1:32:27 the most valuable thing is having access
1:32:31 to GPUs to be able to do whatever these
1:32:35 machines are going to be able to do
1:32:37 because they're going to be doing all
1:32:39 the work. The work is going to be done
1:32:41 by these things. Who has access to GPUs
1:32:44 has power, right? And so if you have 10
1:32:46 gigawatts of of GPUs, you're going to
1:32:48 have a lot of power. But what he was
1:32:50 talking about was give that compute back
1:32:53 to the people that they can then use to
1:32:57 do things for themselves which I thought
1:33:00 was really fascinating. So that that you
1:33:03 know his uh his vision of that future is
1:33:06 starting to play out right now.
1:33:09 Um and they still need Oh yeah, there's
1:33:12 there's new features coming out this
1:33:14 week. Well, no, probably not this week.
1:33:16 Sam Alman said some stuff is coming
1:33:19 soon, but just so you know, the stuff
1:33:22 that's coming soon is going to be really
1:33:25 GPU intensive and it's only going to be
1:33:27 for for ChatGpt Pro users to start. Um,
1:33:31 so it's either going to be like Sora 2
1:33:33 or it's probably going to be something
1:33:35 more more about high-end compute agents,
1:33:40 something like that. Some some sort of
1:33:42 new agentic something or other. Um,
1:33:48 my guess is is that it will be a feature
1:33:52 that if you're not at the high end of
1:33:56 knowledge management, probably won't be
1:33:58 all that interesting to you anyway is my
1:34:00 guess.
1:34:03 There is definitely a FOMO that that has
1:34:07 been generated in the in the AI
1:34:09 community like we want to try the next
1:34:10 thing, we want to try the next thing, we
1:34:12 want to try the next thing.
1:34:14 My past year has been
1:34:18 I am,
1:34:24 how do I say it? My enthusiasm for
1:34:28 wanting to try the next thing has
1:34:29 dropped in enthusiasm over the course of
1:34:32 the year because the things that they're
1:34:35 rolling out are significantly above what
1:34:38 I use chat GPT for anyway. So, it's like
1:34:42 I'm not using the capacity. Like, one of
1:34:44 the things I wanted to do tonight, I
1:34:46 ended up torpedoing with my my stupid
1:34:48 new series. Um, I actually like my new
1:34:51 series. Um,
1:34:54 but what I wanted to do tonight was
1:34:56 someone uh Alex Finn on Twitter um took
1:35:00 OpenAI released a whole here's how to
1:35:03 use chat GPT5. It's like this whole
1:35:05 paper, this whole post that they did and
1:35:08 he distilled that into basically a Mad
1:35:10 Libs kind of prompt for chat GPT5. So
1:35:13 maybe tomorrow we'll play with that. Um,
1:35:16 but I want to start exploring how do you
1:35:18 use chat GPT5 because I'm not using it
1:35:20 very well right now because I I don't
1:35:24 [ __ ] need it to go off and think for
1:35:26 27 seconds about my shitty little answer
1:35:30 that I need, right? And there's no shame
1:35:33 in that.
1:35:34 like you're interested in what you're
1:35:37 interested in. You don't have to be
1:35:38 interested in curing cancer or solving
1:35:41 quantum physics equations.
1:35:44 If you're not interested in them, if you
1:35:46 are, [ __ ] go solve them. Go cure
1:35:49 cancer. But if you just want to make
1:35:51 like a cool kids book, then figure out
1:35:53 the best tool for that, right? So, so I
1:35:56 I think the time of us everyone having
1:35:58 to pounce on every new thing that comes
1:36:01 out, I just don't think it's necessary
1:36:03 right now. So, anyway, I don't I'm not
1:36:05 as upset about that as some people on on
1:36:07 AI Twitter are.
1:36:12 Anyway, all right people, I'm going to
1:36:14 get on out of here. Um,
1:36:18 couple of things you should you should
1:36:19 be tracking and doing. you should join
1:36:23 the AI salon. Um, if you if you're not a
1:36:26 part of the AI salon, go to
1:36:28 community.thesalon.ai.
1:36:33 Um, when you're in there, check out the
1:36:36 mastermind, which is a subscription area
1:36:38 of the salon. Right now, we're in the
1:36:40 middle of doing a 4-week prototyping
1:36:42 sprint. There's a bunch of other stuff
1:36:44 going on, but this is something I'm
1:36:46 co-hosting with Cindy Cindy [ __ ] for
1:36:48 mastermind members. Uh, tomorrow is week
1:36:51 three.
1:36:53 U it's been a really cool uh a really
1:36:55 cool uh group so far and a really cool
1:36:58 cool um project. So check that out. Um
1:37:02 go to are you readyforai.com
1:37:05 if you haven't seen that before and
1:37:07 check out the AI readiness training
1:37:09 program
1:37:11 and Wednesdays at 4 pm mountain time is
1:37:15 the AI readiness um uh podcast. You
1:37:19 should check that out. And then the
1:37:20 final thing is uh you should check out
1:37:23 go to sheleadsai.ai
1:37:26 and check out the create conference
1:37:28 which is coming up on October 11th. Um
1:37:32 and uh
1:37:35 you know Ann Murphy's putting that on.
1:37:37 The AI salon is sponsoring it. We're a
1:37:39 major sponsor of the create conference.
1:37:41 So if you are a woman in AI and you want
1:37:44 to be hanging out with people that have
1:37:46 it going on, you should check out that
1:37:49 conference. And then finally final final
1:37:52 is um happy it's Russia Shana which is
1:37:56 uh happy new year. So to producer
1:37:59 Brandon to anyone else out there who
1:38:01 celebrates happy new year and I think
1:38:03 because of that I I don't know is is
1:38:06 there drinking at rush? Are you going to
1:38:08 be off tying one on? is some sort of
1:38:11 special whiskey. Manise.
1:38:18 So he'll be having some metachvets.
1:38:20 There you go. So So happy happy happy
1:38:24 new year. Happy Rashana to anyone that
1:38:26 celebrates. All right. Um that's good.
1:38:29 Tomorrow there's nothing going on but
1:38:31 but this so 8 o'clock tomorrow. I'll see
1:38:33 you there. We're going to keep digging.
1:38:35 I'm this this spelunking thing. I like
1:38:38 this idea of spelunking. sort of digging
1:38:40 into tools. It might start to get a
1:38:43 little less uh
1:38:46 performancy here if we're digging down
1:38:49 and and playing with things, but come
1:38:52 hang out. The reason to come to this
1:38:54 channel is not to learn things, but to
1:38:55 be in the conversation. That's what it's
1:38:57 about. So, just come hang out. All
1:38:59 right? And if I bore you, then just do
1:39:01 other [ __ ] but be in the conversation.
1:39:04 Be exploring AI. That's that's why we're
1:39:06 here. All right? G.
1:39:11 Peace out and I will see you tomorrow.