6/4/2025 - The Secret of AI: Amplifying Yourself

Description

Chapters

Transcript

0:01 [Music]
0:14 Load the car and ride the
0:19 note. Grab your bags and grab your
0:25 coat. Tell the ones that need to know.
0:31 We are headed north.
0:35 [Music]
0:40 One foot in and one foot
0:44 back. Well, it don't pay to live like
0:49 that. So, I cut the ties and jump the
0:54 tracks for never to return.
0:59 [Music]
1:05 take me
1:06 [Music]
1:08 in. You wear the shape I'm
1:13 in. My hands and shake my head. It
1:18 spins. Oh, Brooklyn, Brooklyn, take me
1:22 [Music]
1:25 in. When at first I learned to
1:28 [Music]
1:31 speak, I used all my words
1:34 for something. I don't know the words.
1:37 What's happening, people? What's
1:39 shaking? What's going down? Bring your
1:42 friends. Tell your friends. Tell your
1:43 neighbors. We got We got We got stuff to
1:45 do. We got people to see. We're going to
1:48 We're going to make a difference
1:49 tonight. We're going to make a
1:50 difference tonight.
1:53 [Music]
1:58 If you pay attention to this
2:01 channel and my activities, you probably
2:03 know that I I discovered I discovered
2:06 something. I discovered something. I
2:08 discovered the secret to
2:11 AI. And I'm not [ __ ] around. Not
2:13 joking.
2:16 [Music]
2:23 I'm trying to I'm trying to figure out
2:24 if the word discover is the right word.
2:29 [Music]
2:38 But like there's all these there's all
2:41 these educational
2:44 courses. How to prompt, how to do
2:47 agents, how to do this, how to do that,
2:49 how to do
2:51 that. None of that
2:54 [Music]
2:57 matters. I mean, sure, you may need some
2:59 skills at some point.
3:02 [Music]
3:33 Mr. He is cursing. I did curse. I did
3:36 curse. Yes, I am
3:38 cursing. Hey, Brian Whitney. Brian
3:41 Whitney knows the
3:43 secret. So, I'm going to reveal the
3:45 secret to AI tonight. Generative
3:51 AI. And honest to God, I this is it is
3:55 such a
3:57 subtle it's a it's such a subtle
4:01 shift, but what if I look back at my two
4:04 years, two and a half, three years of
4:07 doing
4:09 this, all the stuff that I've done that
4:11 I'm like,
4:13 yes, lives in this secret. and all the
4:17 stuff that I did that was like, "Yeah,
4:19 whatever." Lives outside this
4:23 [Music]
4:28 secret. We also don't have uh producer
4:31 Brandon tonight had a a pet emergency.
4:34 There's there's some there's a pet
4:36 situation happening. So, uh producer
4:39 Brandon is is dealing with that. I don't
4:41 know if you recall here, we had a
4:43 champion problem. We fed him grapes one
4:44 night. Apparently grapes, they'll take
4:47 they they'll take a dog down. They'll
4:49 take a dog down. So, one night I didn't
4:52 do this because we were at the emergency
4:54 room getting a grapes extracted from the
4:57 dog's
4:58 belly. So, I don't know if it's a grape
5:01 situation, but Brandon is taking care of
5:04 the care of the kitty cat. So, he won't
5:06 be here tonight. So, what that means on
5:08 TikTok, I will likely be screwing things
5:11 up. So just yell at me, you know, do
5:16 what you do.
5:18 Troll, call me a
5:20 [Laughter]
5:24 savant. Not not The
5:27 secret. It It's just like the movie The
5:30 Secret. It's just like the movie The
5:32 Secret. It, you know, in a lot of ways
5:35 it kind of is, right? It's kind
5:37 [Music]
5:40 of It's kind of that
5:42 [Music]
5:44 simple. But this is the AI secret. It's
5:47 [Music]
5:53 different. And I think it's actually a
5:55 book. I think there's a whole book here.
5:57 It's a very simple idea, but there's a
5:58 whole book here.
6:04 [Music]
6:07 through blue
6:12 telescope. Looking at the world
6:14 tonight through blue
6:18 telescope. Wish I might wish I might not
6:22 see what I
6:24 see. Sheet metal on sheets of ice.
6:30 [Music]
6:34 Looking through this blue
6:38 telescope down a moon struck road
6:41 [Music]
6:52 tonight dee.
6:55 [Music]
7:01 All right.
7:03 Um, hello Sharon Crawford, Steo, Brian
7:06 Whitney, Jeff Flanigan, Vicky Baptiste.
7:09 We've got most of the crew here. Alio's
7:11 wife on time once again. Very nice.
7:14 Solid performance. Source Camp. Nowhere
7:16 to be found. Nowhere. Nowhere. Nowhere
7:20 to be found. Or is she
7:22 here? Side Hustle Mimi's in the house.
7:25 We got Silver. Oh, Source Camp's here.
7:28 Damn, it's What is it? 10 after. She's
7:31 She's effectively early. All
7:34 right. I almost have my book finished
7:36 for the GPT
7:38 competition. Oh, yeah. There's a chat
7:40 GPT
7:42 competition. If you go to the uh in the
7:45 AI salon, go to what's it called? Uh
7:48 water cooler. There's a discussion about
7:50 a new chat GPT competition for a
7:53 children's book. All right. I heard
7:56 that.
7:57 [Music]
8:03 Mr. It where? AI salon. AI
8:07 salon. Uh AI salon
8:10 is if you go
8:14 to go to that URL and then click on join
8:18 our community. That'll take you over to
8:19 the community website. And then there's
8:22 sort of an there's sort of a sevenstep
8:24 welcome. And I think step number six is
8:27 water cooler. Might be five or six, but
8:30 water cooler is just an area to chat.
8:32 And I would imagine if you also just
8:34 went to uh to chat JP or to the Google
8:39 and search in chat GPT children's book
8:42 contest, there will likely be a link
8:44 there as well. All right. All right.
8:47 Good
8:49 people. Beautiful. Beautiful. Fantastic.
8:53 Fantastic. Bob, tell him what he's won.
8:56 I wasn't wanting anything because it's
8:58 not really a
8:59 contest. Yeah. This is This is my
9:02 impersonation of Dana Carvey doing Jimmy
9:06 Stewart. Yeah. See? Yeah. It's It's not
9:09 quite Jimmy Stewart. It's more like a a
9:12 Dana Carvey. A young Dana Carvey
9:14 impersonation of a Jimmy Stewart. Yeah.
9:23 Yeah, this is what this channel is. This
9:26 is what this channel
9:28 is. I thought we were going to be
9:30 learning something. It's called the AI
9:32 learning lab. Are you giving lessons? I
9:34 need lessons. How do you do AI? Can you
9:36 make money with AI? How does chat GBT
9:38 work? Is it evil? Does it steal? Isn't
9:40 it the world's greatest plagiarism
9:42 machine? It's stealing from artists,
9:44 right? You can't really write with it
9:46 because it's just stealing. That's
9:48 cheating, right? If it does it fast, it
9:50 means it's cheating. It means it's bad.
9:52 Is that what this is? This is a cheaters
9:55 channel. This is where the cheaters come
9:57 to cheat, to learn how to
9:59 cheat. Is this That's what this is. Yep.
10:03 That is what this is. Welcome. Welcome.
10:06 Welcome. Welcome
10:08 cheaters. You You heathens. You thieves.
10:11 Thieves of the night. Your large
10:14 language model
10:15 embedding frequencies. your
10:18 probabilistic mathematical latent space,
10:21 thousand-dimensional mathematical space
10:26 cheating. How dare
10:28 you? How dare you use the world's most
10:31 profound
10:33 knowledge
10:35 system to
10:37 cheat? Life is supposed to be hard,
10:42 people. Don't you understand?
10:45 So you don't use AI, you
10:55 suffer. AI learning lab. Want proof?
10:58 Early images. I need a golden golden
11:01 button that explains privacy and
11:03 security with chat GPT. That
11:04 conversation's getting old. I know.
11:08 Can I just post that I have 10 pages
11:10 from ChatGpt with perfectly consistent
11:12 images. That's crazy. which if you don't
11:15 know, if you've ever tried to make
11:17 consistent images for a children's book
11:18 before, it's almost impossible, but it's
11:20 not
11:22 now. Sharon Crawford, I was invited to
11:25 be part of our company's AI council this
11:28 week.
11:31 That is not the first time that members
11:33 of this
11:35 channel have either survived layoffs
11:38 because they knew AI or have been
11:41 invited to join AI councils or
11:45 both. Congratulations. That's
11:48 awesome. Fantastic. Fantastic. Sharon
11:51 Crawford, I like it. I like it a lot.
11:54 The leading lab. Yeah, exactly.
12:00 [Music]
12:14 Now, if you're here and you're
12:16 wondering, did didn't he promise us he
12:19 was going to tell us the secret of AI?
12:21 He seems to be doing some bad Saturday
12:23 Night Live skit with Adam Sandler level
12:26 guitar skills.
12:29 We are we gonna is when is the learning
12:32 start? We I I came for the learning.
12:37 It's called the learning lab. And
12:38 where's the Is he a doctor? Where's the
12:41 lab coat? Why aren't there
12:44 chemicals? I'm very
12:47 confused. I'm I'm on TikTok. I spent
12:52 Well, I didn't spend any money, but I
12:53 bought an iPhone and that costs a lot of
12:56 money. And so I'm here. I'm expecting my
13:00 learn. Where is my
13:02 learning? Where is my
13:05 learning? This is a very distressing
13:07 channel. I feel like he's mocking me. Is
13:10 he talking directly to me? I don't talk
13:13 like that. But you do.
13:18 [Music]
13:29 [Music]
13:32 Wait. Chat GBT40 helped me understand an
13:35 equation parameter capital PSI, a list
13:39 of
13:40 ordered egen
13:42 values.
13:43 Tom, that's freaking awesome.
13:47 One of the things I'm most excited about
13:49 is I can use chat GPT to learn what an
13:51 Egan value
13:55 is. Anytime anyone comes in here and and
13:59 has has has math skills, I am
14:01 intimidated. I I I shut down quickly. I
14:04 shut down
14:07 quickly. But yeah, I you know, it's
14:10 absolutely crazy. And if you if you
14:12 haven't tried it, Tom, um if you're kind
14:15 of new to the chat GPT thing, flip over
14:17 to 03,
14:19 um the 03 model here, let me show you.
14:21 Let me show you on the on big screen
14:24 here. Let me show you on big screen. Oh,
14:27 this is this is my uh my part of my new
14:30 concept that part of the secret is this.
14:33 It's a bad graphic. I'm trying to get a
14:35 not shitty graphic.
14:38 It's the worst graphic ever. But uh if
14:41 we go to chat jp tai here and we do new
14:44 chat. If you go to 03 wait what did you
14:48 say about egan values? What
14:52 um hang on you said help me understand
14:57 an equation
14:58 parameter capital psi a list of ordered
15:02 een values eigen values. Okay. So, I'm
15:06 going to
15:07 say let's see. Uh, I need you
15:13 to help me
15:17 understand. I'm just going to put I'm
15:20 just going to put [ __ ] in here.
15:22 equation
15:23 [Music]
15:25 equation
15:28 parameters and
15:33 PSI ei values
15:37 um show your
15:40 work. So if you do it in 03 rather than
15:43 40 or did you do it in 04? No you did it
15:45 in 40. If you do it in 03, what 03 does
15:49 is it does reasoning and then within
15:52 that reasoning, it can actually write
15:54 and execute Python code, which I don't
15:56 think 04 can do that yet. It might be
15:58 able to um so if it wrote Python code,
16:01 then it was already doing it. But um but
16:05 let's see. The user's question covers
16:06 topics like equation parameter. Oh boy.
16:09 See, look Tom Python Mattplot
16:14 live.pipplot. Pi plot pi plot as plt
16:18 with h barme and l l equals le e-9 which
16:23 everybody knows that right n= a,000
16:27 which why wouldn't it? dx equals x^ the
16:30 1 - x
16:31 0 because if you did x^ the 1 - x to the
16:35 1 that would be zero and it wouldn't
16:37 even make sense to define it as a as a
16:39 parameter right you understand what I'm
16:42 saying probably more than I do um but
16:45 then t equals minus hbr star* star 2 to
16:49 the double exponential
16:51 [Laughter]
16:58 Nice. Yeah, I don't listen Tom. I don't
17:01 mean to brag, but uh but you know, I
17:04 kind of I dabble I dabble in Python with
17:07 Mattplot Live SP plt because it's you
17:12 know what it is for me. It's
17:14 enjoyable. It's enjoy. You know, some
17:16 people like making bagels. I I like
17:19 writing Python plots. The first three
17:22 egen functions of a one nanometer
17:24 infinite square. I I sort of plot them
17:26 out um with uh with mapplot lib lib lib
17:31 li the lib tards. I I don't know how you
17:35 pronounce it because I just because I
17:36 just dabble and I and I play in Python.
17:40 Um, but the physical constants, you fix
17:43 the scale of the problem with MHQ and
17:46 then K to the underscoreB, capital B,
17:49 not not the small B. That's a rookie
17:51 mistake. You got to use the capital B.
17:54 Geez, I thought everybody knew that.
17:57 Yeah. Yeah. Um, but the potential
18:00 boundary parameters where you carve out
18:01 where the solution is allowed to be
18:03 where you got V to the X and then L and
18:05 then R to the zero and E to the F, you
18:08 know, E like to the F. It's just like a,
18:12 right? You know, it's got and it's like
18:15 four or five, right?
18:21 No, I don't do maths. Well, clearly
18:24 source camp, you're not as sophisticated
18:25 as I am with the mathematics because
18:28 egen plots are are
18:31 uh they're relatively trivial. I mean,
18:35 you know, everyone can do them
18:38 now. So, anyway, there you go. We got
18:41 our plots. Oh,
18:44 man. Actually,
18:48 Tom, you just proved
18:53 We we've got a
18:54 theorem and and you are proof of the
18:58 theorem. You're proof of the theorem. I
19:01 like this a lot. Now, here's the
19:05 theorem. I've discovered the secret to
19:08 AI. And the secret to AI is this. We're
19:12 asking the wrong questions as purveyors
19:16 of generative AI.
19:20 The question that we ask and the
19:21 question that I base this channel on
19:25 is how do I how do I get the most out of
19:30 AI, right? How do I get the most out of
19:33 AI? I'm going to go to chat GPT. I'm
19:35 going to put in a thing and it's going
19:36 to come a thing. How do I get the most
19:37 of that? That's the wrong question.
19:41 The right question is how do I get the
19:44 most out of
19:45 myself using
19:48 AI? And so what you just illustrated
19:52 really well is I don't I so don't know
19:55 anything about math that I don't even
19:57 know what the [ __ ] you're talking about.
19:59 Now I can go into chat GPT and I can put
20:01 in words that you
20:04 said and out will come a thing that
20:06 probably makes sense to someone. And if
20:10 I wanted to take the time, I could learn
20:11 about all this.
20:13 But that's something that you were
20:16 interested in, right? And so you're
20:18 using Chat GPT in this very particular
20:20 way that's personal to you. I use it in
20:23 a way that's personal to me. When Chat
20:25 GPT is at its best or when any AI is at
20:28 its best is when we're using it to
20:32 amplify
20:33 us, right? So you've got the the prompt
20:36 box and it's hungry. That prompt box
20:39 will eat anything you throw at it. And
20:41 what most people do is they throw shitty
20:44 prompts at it because they're thinking
20:46 about the tool. What am I supposed to
20:48 put in the prompt box to get out a good
20:50 answer from chat jpe? I need to make
20:53 money with chat jpate. You can make
20:55 money with jpe.
20:58 And they forget. They get seduced. They
21:00 get seduced to think that the tool is
21:04 important. It isn't.
21:08 The tool is
21:12 irrelevant. The only thing that the tool
21:15 serves is what do you
21:19 want when done
21:21 right? What are you trying to
21:23 accomplish? Who are you speaking to? Are
21:26 you trying to brainstorm? Are you trying
21:27 to write a book? Are you trying to write
21:29 a
21:30 musical? Do you have a Do you have a
21:33 question? Is your eyeball doing weird
21:35 things and you want to know what the
21:37 [ __ ] is going
21:39 on? That's when it comes to life. When
21:42 it takes who you are, what you want,
21:45 your intentions, your ideas, your
21:49 frustrations, and it amplifies them and
21:52 brings them to
21:53 life. And that's when it's remarkable.
21:56 That's when people have their Kevin
21:57 Mallister moment.
22:00 If you do things that are not personally
22:03 relevant to
22:04 you, it's just [ __ ] cycles. It's just
22:07 activity. It doesn't [ __ ] matter. A
22:09 lot of what I do on this channel night
22:11 after night is just throw gobbledygook
22:14 at the machine. And then
22:17 occasionally it will generate something
22:19 that I'm like, "Oh, I like that." And
22:21 then I'll sort of run down that rabbit
22:23 hole. That's when things get
22:24 interesting. Like the night we made the
22:26 Weird Mary
22:27 song, we were just cranking out songs.
22:30 cranking out songs and then this one
22:31 song came and it was like, "Oh, that's
22:33 really good. There's something really
22:34 interesting there." It's like chat GPT
22:36 wrote the lyrics or forget how it went
22:38 and then we went to Suno and it did some
22:40 [ __ ] and it was kind of cool and we just
22:41 kept digging and digging and digging and
22:43 digging but that was all filtered
22:46 through that sounded good to
22:49 me. And
22:51 so, so one of the things that I want to
22:53 start to articulate
22:56 is, okay, if the prompt box is hungry,
23:00 you've got to feed the
23:02 prompt. Well, what do you feed the
23:04 prompt? Well, if you take that that new
23:08 question, how do I make the most of
23:10 myself using AI? What you feed the
23:13 prompt is
23:15 you. Well, what's that mean? That sounds
23:17 kind of
23:19 esoteric. Yeah,
23:22 except
23:23 you are whoever you are. You've got
23:26 ideas. You're interested in egen
23:28 functions and
23:31 mathematics. I'm interested in, I don't
23:34 know, storytelling and and marketing and
23:37 and entrepreneurship and business stuff.
23:40 And sometimes I have ideas and sometimes
23:42 I want ideas. And sometimes I have a
23:45 problem that I want to solve. And
23:46 sometimes I want to expand on something
23:49 that I thought about in the
23:52 car. If all your focus
23:55 is take a moment, we had we had CJ
23:58 Fletcher on the uh podcast today with
24:00 the AI readiness project podcast with
24:03 Ann Murphy and myself. And one of the
24:05 things that CJ talks about in in
24:07 prompting images is he says when you sit
24:10 down to to prompt images to to generate
24:15 art. He said take a
24:18 moment. He said close your
24:22 eyes. Ask yourself where are you? Where
24:24 did you
24:27 go? And then he said, "Where wherever
24:30 you end
24:31 up, there were things that happened
24:33 before that and there were things that
24:35 happened after that." He goes, "That's
24:37 that's a
24:39 story." He goes, "Find that find that
24:41 moment. Find find that instance. Think
24:44 about the before. Think about the after.
24:47 And then start prompting." What's he
24:49 doing
24:50 there? He's separating the tool from
24:54 you. And he's saying, "Go to you. check
24:57 in so that when you come back to tool,
24:59 what you're feeding the prompt is
25:03 you. So that's the
25:06 secret. And if you've been in here for a
25:09 while, if you're in a
25:10 regular and and Murphy said this today
25:13 on the podcast, she goes she goes,
25:15 "Yeah, I get that." Like I get it, Kyle.
25:17 She goes, "But isn't that essentially
25:19 what the salon and what the irregulars
25:21 have all organically come to?" And I
25:23 think I think that's right.
25:26 I think that's
25:27 right. But the AI is a sneaky little
25:30 devil. It's very
25:33 seductive. Right. It's very seductive to
25:36 go, you know, come here. Just, you know,
25:38 if you pay my $150 a month, I'll make
25:42 you great
25:43 videos. Well, who gives a [ __ ] about
25:46 great videos? What's the story you want
25:48 to
25:50 tell? And then if you got that, then
25:52 maybe you want great videos, right?
25:55 So, it's just shifting. It's such a It's
25:57 such a minor shift, but huge. All
26:02 right. What is happening, Rachel Kimber?
26:05 Uh, I'm We're talking
26:07 philosophy. We're talking philosophy and
26:10 and the secret to generative AI, which
26:15 is
26:17 you are what the prompt box is hungry
26:21 for. When you feed your
26:23 prompt, feed it
26:26 you, your dreams, your ideas, your
26:31 frustrations. And then what the AI will
26:35 do is
26:38 take that idea, that spark, and it'll
26:42 amplify it. And it allows you to get you
26:45 out of your head and in front of you in
26:48 a really powerful way. And that's where
26:51 the [ __ ] magic happens. So, all
26:55 right. I've got lots and lots of stories
26:57 to tell. I wish I had any skill seeing
27:00 them as
27:01 images.
27:03 Dude, dude.
27:06 Dude.
27:11 Dude, you do have
27:14 skill. You do have skill.
27:18 If you can tell the story. So, so give
27:21 me give me one of your
27:26 stories. Give me one of your stories.
27:29 Just a couple of words of
27:32 it. Eat me. Exactly,
27:35 Joker. Eat
27:38 me. That's what the prompt box is
27:40 saying. It's saying, "I'm hungry. What
27:41 are you going to give me?"
27:43 And if you give the prompt box crap,
27:46 you're going to get back crap.
27:50 Crap. Guest appearance time. Uh, we
27:54 could bring a guest up, but we don't
27:56 have producer Brandon.
28:03 [Music]
28:22 [Music]
28:29 You know,
28:34 this this idea of
28:43 making the most of yourself using
28:47 AI. Someone was asking me today about,
28:49 well, how does that relate to
28:52 business? Well, what are businesses?
28:57 Businesses are essentially just groups
28:59 of people that agree to do the same
29:04 thing, you know, to to, you
29:08 know, participate together in moving
29:11 some idea forward, right? It's just
29:16 [Music]
29:17 people. But because AI is so
29:22 fickle where it can seduce you into
29:25 thinking that crap is
29:33 good. Most people in business don't even
29:36 know what it makes possible. Don't even
29:38 know what it can do.
29:41 And so I think everyone from the CEO
29:44 down needs to have this kind of personal
29:48 experience with chat GPT. The maybe this
29:51 is source camp. Maybe this is one of the
29:53 reasons that explaining this [ __ ] so
29:55 hard. All right. Archetypal architect
29:58 has a story for us. That's great. Okay.
30:04 Um, one of the reasons that talking
30:08 about AI to people who've not
30:10 experienced it is so hard is because it
30:12 is so personal. One of my favorite
30:14 things, one of my favorite um, parlor
30:18 tricks that Source Camp ever did. She
30:20 was with a bar owner and he said
30:22 something to the effect of, "I AI is not
30:25 good for me. I just I just own a bar.
30:27 It's, you know, I I can't use it.
30:31 And she said, "Well, hang on." And she
30:33 took a picture of the bottles behind the
30:35 bar. And then she told Chat GBT, "Make
30:38 me an inventory." And any bottle that's
30:40 less than a third full, marked that as
30:43 reorder. And as she was saying what she
30:46 was doing, he was like, "No, it can't do
30:49 that." And then it [ __ ] did it.
30:56 And so what you did source camp in that
30:59 moment is you put him in the chat box,
31:05 right? And so only when someone
31:08 experiences that thing that's relevant
31:10 to them do they actually get
31:13 it. And so you're a CEO of a big Fortune
31:17 500 company, you're like, I got the IT
31:18 group. We've had AI for decades. I got
31:20 the IT group on
31:24 that. That CEO is going to get
31:28 blindsided by a bunch of punky Gen Xers
31:32 that are all AI native. All they think
31:35 about is
31:37 AI. And three of them know his industry
31:40 really
31:41 well. And they're going to get together
31:44 and they're going to with their little
31:45 10person company come in and [ __ ]
31:47 decimate that big company.
31:50 Yeah, it's not possible. A small team
31:53 can't decimate a big company. You
31:56 watch. Mark my
31:59 words. Small companies are going to you
32:03 know you know the the big the big ships,
32:05 the big cargo ships and the big oil
32:07 tankers and stuff and they go through
32:09 the wherever the hell they go where
32:11 where the little pirates and their
32:12 little zippy boats run up and you know
32:15 hijack those things.
32:18 That's what corporate America is going
32:19 to look like in the next five years.
32:22 Lots of zippy little boats going, "We're
32:25 [ __ ] with
32:27 you. We're a 10 person company. We just
32:29 two did $200 million. Woohoo.
32:35 [Music]
32:42 Sharon Crawford watching our she seuite
32:45 shaky shaking in their
32:48 boots and and Sharon here's here's the
32:51 thing like here's an opportunity for you
32:54 if you if they got if they put you on
32:55 the AI council a big opportunity for you
32:58 and this is anyone that's working in a
33:00 big company right
33:02 now if you're like the AI pi piper or
33:05 you know AI I
33:07 weirdo. If you're the one that just
33:09 won't shut up about
33:16 AI, I would do everything I could to to
33:27 to
33:31 invite anyone from the executive team
33:34 all the way up to the seauite.
33:38 like on a weekly basis, invite them to a
33:41 I'll do I'll do
33:44 one-on-one, you know, tutorials for you.
33:46 I'll show you what's going on with this
33:47 stuff. Just just keep putting in front
33:51 of
33:52 them. You can't
33:59 possibly you can't possibly
34:03 lead the company in the face of
34:06 generative AI without using it yourself
34:09 and without having and without having
34:11 some
34:13 personal experience with
34:15 AI where it finally sets in. Oh, this is
34:20 different. And I think the only way you
34:22 make it different is you make it
34:23 personal to
34:26 them. The more I learn about this stuff,
34:29 the more I dig into this, the more I am
34:31 convinced you cannot
34:35 possibly run a
34:43 company without understanding what what
34:45 what's happening right now. what these
34:47 tools make possible
34:50 today. It's like I I feel like it's
34:54 irresponsible. So, I think there's a big
34:56 opportunity for for you there. Ootr
35:00 trucking. It amazes me how many still
35:02 haven't used chat GPT. Yeah. Yeah. and
35:04 like like like
35:07 like I don't know maybe it's like you
35:10 know Friday night get get you know
35:13 someone from the HR department to
35:15 fund Friday night pizza night and you
35:18 know 400 p.m. every Friday you're going
35:19 to have two pizzas and you're going to
35:21 teach people how to use chat GPT or
35:23 something like just and and [ __ ]
35:26 invite the CEO every
35:28 time till they show up. Okay.
35:33 My go-to move is ask them a few qu
35:35 questions and spit out a badass song
35:36 about it while we're having the
35:37 conversation. Thanks for the ideas.
35:39 Yeah, listen. I mean, here's the good
35:42 news. You got asked to be on the
35:43 council. Now, you were squeaky enough of
35:45 a wheel to get asked to be on the
35:47 council. Now that you're on the council,
35:48 keep being that squeaky
35:52 wheel. You know,
35:55 like as silly as this channel is, right?
35:58 What what this channel is is effectively
36:01 like the the silly version
36:06 of what needs to happen in businesses,
36:10 right? Like when I do this for
36:12 businesses, it's a very it's not this
36:14 zany
36:16 experience, but it's it's it's no less
36:19 important, right? It's like it's like
36:21 this.
36:24 The only way you can learn this stuff is
36:26 to be inside it, to be in the
36:28 conversation. And until someone I don't
36:31 give a [ __ ] if they've launched Chat GPT
36:34 50 times and they've written emails or
36:36 whatever they've done. If they haven't
36:38 had that moment where they're like,
36:39 "Holy
36:42 shit." That then they don't understand
36:45 what's
36:46 happening. Oh, I have a page and a half
36:48 of ideas. That's great. Something about
36:50 music made from their own life really
36:53 gets people. Yeah, that that one's
36:55 pretty amazing. Uh, Architect, the uh
36:57 the Jim Ross thing where he landed a
36:59 client because he did that. All right,
37:00 let's let's let's do Archetypal's
37:03 story. Grayson's parents sent him back
37:06 to Earth on the day of his 18th
37:08 birthday. One mission, saved the world,
37:10 not so hard. So, we're gonna do a couple
37:11 of things. One is, can I copy
37:15 that? No. But you know what I can do?
37:18 Watch this. This is going to be
37:20 fun. So, I can't copy that.
37:24 Oops. So, I am going to screenshot what
37:27 he
37:34 wrote. All
37:37 right.
37:40 So, there's what he wrote. I'm going to
37:43 paste this in here. And then I'm going
37:46 to say uh OCR
37:56 this. All right. So here's your story.
37:59 So So you said you don't have the
38:02 skills to make an image of the stories
38:05 that you have. Rick Olsson, I just
38:07 accepted a position as data and AI
38:09 consulting
38:10 partner. Cover the keyboard, Kyle. Huh?
38:15 Oh, you mean this
38:17 thing?
38:20 Yeah. Wait, there we go. I'm writing a
38:24 screenplay based on a true story.
38:25 Cornbread mafia. Love it. That's
38:29 great. Um, okay.
38:36 So, so now we've got that. So now I'm
38:39 just going to do um make this a
38:46 cinematic
38:49 image that
38:52 brings this idea to life.
39:07 Rick Olsson, great job on the AI
39:09 readiness podcast today. Thank you very
39:10 much. I know Ann's amazing, right? And
39:13 like why she puts up with me is beyond
39:16 [Laughter]
39:23 meh. Never a dull moment. Yeah. Well,
39:27 specifically, I don't create visually in
39:29 my head. I get no coherent plans, but
39:32 that's okay. So, architect, here's the
39:34 here's the
39:39 trick. You know what the story is? It's
39:41 funny. You should go watch the AI
39:43 readiness uh project today because I
39:46 asked CJ CJ Fletcher about this. Oh,
39:49 that's funny. It wrote the words like
39:50 like a Star Wars
39:53 movie. This is pretty funny.
39:58 Have you published project cardboard? I
40:00 have. If you go to
40:03 project-cardboard Oh, but published it
40:05 like you can edit it. No, I haven't.
40:07 I'll go do that right now. I think
40:09 weren't you the one that asked, "Have I
40:10 published it so it's editable or
40:14 remixable?" But if you just if you just
40:17 want to Yes. Okay. You were the one.
40:18 Okay. So, I'll go do that. But anyway,
40:20 this is funny. Grayson's parents sent
40:22 him back to Earth on the day of his 18th
40:23 birthday. One mission. Save the world.
40:26 So, not so
40:28 hard. Okay, we got that. That's great.
40:30 We'll say, okay, let's see. Lovely. Um,
40:34 now no
40:38 words.
40:40 Um, archetype, do you have access to
40:43 midjourney? I I always I kind one of the
40:45 mistakes I make in here is I kind of
40:47 assume everyone's got all sorts of
40:49 subscriptions.
40:52 [Music]
40:59 We'll go to lovable in a second. Yeah,
41:01 of course.
41:02 Okay,
41:07 so I'm going to go grab this
41:11 thing. Just that
41:15 sentence. So, it's kind of sci-fi. We're
41:18 going to go to
41:22 midjourney. That was that cool image we
41:24 did last night. Yeah, this was this is
41:26 some cool [ __ ] last night. Okay, so let
41:29 me show you something.
41:31 Archetype. So, one of the things that I
41:33 was doing last night was I was just
41:36 taking
41:39 um negative what what was it? negative
41:42 self-belief, ne negative self-t talk or
41:45 whatever. I whatever the [ __ ] it was we
41:47 were talking about, but I was just
41:48 putting them into different styles,
41:51 different midjourney styles, right? And
41:54 then these are all cool. They're all
41:55 cool. They're all cool. They're all
41:56 cool. Yeah, that's cool. That's cool.
41:58 Looking at that. Oh, that's cool. That's
41:59 interesting. That's cool. That's cool.
42:01 That's cool. That's cool. Yeah. Nice.
42:02 Nice. And then this image came
42:04 up, right? Do you remember this? You
42:07 were here last night.
42:10 this image came
42:12 up and this is part of my epiphany today
42:17 about the question is not how to make
42:20 the most of AI but how to make the most
42:21 of yourself using
42:24 AI. There's something about this image
42:28 that amongst all these other
42:31 images felt like
42:33 me. It was personal to
42:37 me when CJ Fletcher in the in the AI
42:42 readiness podcast today, I asked him
42:44 about his prompting because when he does
42:45 prompting, he'll prompt an image and
42:48 then he will
42:50 do iterations of the prompt. He just
42:53 keeps going and going and going and
42:55 going and going like like far beyond
42:58 where I would have stopped. Far beyond.
43:03 And I said,"Can you explain that
43:04 process? Is it are you in like a really
43:07 logical kind of process where you're
43:08 sort of in prompt engineering mode and
43:10 then how do you switch back and forth
43:12 between, you know, that mode and
43:14 creative mode?" And it turns out I I had
43:18 interpreted it all wrong. What CJ said
43:21 he had is he said he's got stories in
43:23 his
43:24 head and
43:26 then he just takes the story out of his
43:30 head, puts that in as the prompt, and
43:33 then when he looks at something, it
43:35 either looks like that story or it
43:37 doesn't. So, what he was talking about
43:39 today is what AI allows him to do is get
43:43 the stuff out of his head and in front
43:45 of him. And then he basically just keeps
43:49 going until he sees something in front
43:51 of him that matches the story. He
43:53 basically described exactly what you're
43:56 saying is a big
43:58 limitation. like yeah it's a big
44:00 limitation for you as a human if you
44:03 wanted to visualize images right in your
44:06 head and you don't have that capacity
44:08 okay but now you do right and so I think
44:12 I think what could be fun is just go
44:15 into midjourney and I'm just going to
44:17 pop in your story as it was with no let
44:20 me see I've got stylization atund at 100
44:23 or was it 200 no
44:26 100 Grayson's parents sent him back to
44:28 Earth on the day of his 18th birthday.
44:31 One mission uh to save the world. So,
44:34 not so hard.
44:37 Boom. And then, you know what we were
44:39 doing last night is we were going into
44:43 um into my little collection of style
44:47 references and you know, maybe there's
44:52 kind of a style you want it in. And that
44:55 doesn't have to be in your head. That
44:57 can be outside of your head. But like,
44:59 what made me choose this style? I don't
45:01 know. It's a thing that spoke to me.
45:03 There's no right answer, right? So,
45:05 there's that prompt. There's that style.
45:11 Boom. And it'll go to that. Here's that
45:15 prompt. Um, I'm gonna hit this with my
45:18 personalization now. Boom. Here's that
45:20 prompt. I'm gonna turn stylization up
45:23 to 900.
45:28 Boom. Just vomit out
45:31 work until you see something that starts
45:34 to look like the story and then run down
45:36 the rabbit hole,
45:41 right? Yeah. Hard
45:47 sci-fi. You know, here's not so hard
45:49 sci-fi. There's cartoon. Kyle, can you
45:52 tell the nice
45:53 people that Kim Kim Offford's film
45:57 content screening is tomorrow? Yeah.
45:58 Where where is that happening?
46:03 Ann, that is Midjourney. Yeah, we're in
46:05 Midjourney right now. Noon to 8 central.
46:07 So Kim offered um did a did a AI film
46:11 contest. She was making AI films and she
46:14 did some music videos and stuff like
46:16 that and then she's like, I want to do
46:17 this for other people. and she put
46:19 together a a contest. And so tomorrow
46:22 she's screening all the winners, I
46:24 guess. Go to her Instagram. Okay. Go to
46:27 Kim Offford's Instagram. Let me jump
46:29 over to Instagram quick and see if I can
46:31 find her. I'll give you the
46:34 um
46:43 Instagram. So, we're going to do a
46:45 search. I'm going to search for
46:48 Kimberly
46:52 offered offered oF right not of
46:58 Fred Kimberly offered. There it
47:02 is. All
47:07 right. Viewing party June 5th. There you
47:11 go. There's your info. You get swag and
47:15 it's all
47:17 free. Oh, off ford. Not off offer. Off
47:22 ford. Yes. What you said,
47:28 Ann. All right, back to archetypal
47:31 architect and
47:33 midjourney. So, those aren't hardcore
47:36 sci-fi. This is kind of
47:39 stylized sci-fi. Could be in the
47:41 neighborhood.
47:42 There's your beautiful 18-year-old boy
47:44 with his little space
47:46 helmet,
47:49 right? There's There's sexy hot sexy hot
47:53 space
47:55 cowboy. There's There's brooding
47:58 brooding. Daddy didn't love me. Space
48:01 cowboy. There you go. There's some
48:03 gritty sci-fi. I think this is the one
48:05 that uses my personalization. Yeah.
48:12 There you go. And and like at some point
48:15 something's going to speak to
48:17 you and then you just run down that
48:19 rabbit hole and it's like so it's like
48:22 the way CJ was describing it. It's like
48:25 you've got a story in your head.
48:27 Articulate that story. Put it in these
48:30 tools. Just have it vomit. Vomit. Vomit.
48:34 Vomit. And then if it's not right, just
48:35 keep just keep pushing. Keep pushing.
48:37 No, that's not it. No, that's not it.
48:39 And at some point it's going to be it.
48:41 It's going to be close enough
48:46 anyway. Cool. Makes more sense to
48:49 me because it's the I I've I I've talked
48:53 about this before here. I'll tell you
48:55 when I had my first epiphany with this
48:58 approach.
49:06 Um, so I
49:09 did I did this project called Kyle
49:11 Shannon Dreams. This was the very first
49:13 image I did. This was I taught myself
49:15 stable diffusion. This is like late
49:18 summer, early
49:19 fall of uh 2022, right before Chat GBT
49:24 came out.
49:25 And you know, I did I taught I I used
49:28 Dream Booth and and had it um you know,
49:31 I uploaded a bunch of pictures of myself
49:33 and then I started making these crazy
49:35 images and they're all kind of
49:37 interesting,
49:39 right? I keep thinking I should be able
49:42 to describe the scene in a way other
49:44 than what's on the
49:45 page. I you
49:47 know architect
49:51 I I would forgive
49:55 yourself
49:59 like maybe that's something you
50:04 develop maybe you take the
50:08 stories let the machine generate the
50:11 images find your
50:13 style and then have chat GPT describe
50:16 the image back to
50:19 you. Like maybe you can actually teach
50:23 yourself the language of imagery, but do
50:26 it in reverse. Start with the story. Let
50:29 the machines do the heavy lifting on the
50:31 image generation stuff. And if you
50:33 really are still passionate about
50:35 learning how to describe images, have
50:37 chat GPT. You can upload any image to
50:39 chat GPT. It'll describe it. Right. I
50:41 like that idea. Yeah. Because if this is
50:44 if it's it's like what we were talking
50:46 about yesterday was limiting
50:47 self-beliefs. Like that's absolutely a
50:49 limiting self-belief that you're like I
50:51 can't move forward until I can actually
50:53 describe the image that I want.
50:55 Well, that's not a skill in your arsenal
50:58 right now, but it doesn't need to be.
51:01 So, so let it let it be this other way.
51:04 It's this mort sort of, you know,
51:06 outside in approach and then yeah, use
51:08 chat GPT to train you how to look at
51:12 that image. Like what do you call it
51:13 when there's misty fog and you just can
51:15 barely see things in the distance? I
51:17 don't know. It'll tell you
51:19 that, right?
51:22 Anyway, so I'm making all these
51:26 images like that was one that this was
51:29 the first one that I went like, "Oh,
51:32 like I know I know that I know that
51:35 guy's story. I can tell that
51:37 story." Side hustle Mimi. The way I do
51:39 it is I tell GPT all of my favorite
51:41 artist, paintings, techniques, as well
51:44 as some other
51:47 stuff. Side hustle me. It seems like I
51:50 missed a lot. Ah, it's just me rambling.
51:52 You know, you know how it is in
51:54 here. Most of the time it's just I'm
51:57 just background noise. All right. So,
51:59 I'm gonna I'm gonna flip through these
52:01 images.
52:03 So, all of these
52:05 images were giving me some level of
52:08 satisfaction,
52:11 right? And and like they gave me an idea
52:13 for a project and the project was Kyle
52:15 Shannon Dreams. And the basic idea was I
52:18 was going to take these weirdass
52:20 self-portraits and I was going to start
52:22 telling the stories of these
52:24 people that were kind of like me in
52:27 another
52:28 universe. And so I was doing this, you
52:31 know, for
52:32 weeks. Like these are all fine. They're
52:34 all fine. They're all
52:40 fine. Some make me happy. Some are like,
52:43 why'd you make me so fat? I'm not really
52:45 that fat. Am I really that fat? I have
52:47 low self-esteem. Oh, I'm a superhero.
52:50 Yeah, cool. Maybe I'm a superhero,
52:52 right? There's, you
52:54 know, what's his name? Ralph Steedman
52:57 style.
53:00 Me. So, I'm making all these things. I'm
53:02 making all these things. I'm writing
53:04 stories about all these
53:05 things. Annie
53:07 Liuititz made a photo of me in the 70s.
53:17 You're like, "What's the point, Kyle?
53:18 Where's this going? It's going
53:24 [Music]
53:29 somewhere. I think we're in the
53:30 neighborhood.
53:39 [Laughter]
53:43 You imagine someone flipping through
53:44 TikTok right now? What the hell is going
53:46 on
53:48 here? Those were
53:52 funny. Uhoh. Hang on. I got to
53:55 do physical dexterity on Tik Tok. All
53:58 right. I I passed. Wait, was this Was
54:01 this it? No, this is in the
54:07 neighborhood. Uh uh uh uh uh. All these
54:10 images, all these images were just like
54:13 I was just putting them out in the
54:14 world. Putting them out in the world,
54:15 putting them out in the world, making
54:16 them like I didn't share most of these,
54:18 right?
54:37 Sunflower.
54:45 Dude, there's a point to this. I didn't
54:48 think it would be this long. I didn't
54:49 realize how many of these [ __ ] images
54:51 I made.
54:57 There's military me, Space Force
55:02 me, kid
55:05 me, fat rocker me, cabbage me, cabbage
55:09 woman me. Look how sexy I am. I'm a
55:12 handsome
55:13 [Laughter]
55:17 woman. Oh my god.
55:23 [Music]
55:30 That
55:31 image. This
55:33 image. These were cool. It was fine.
55:37 That
55:40 image. So archetype. This is for you.
55:42 See? See that's interesting. You see an
55:44 image and it tells you a story. Just
55:47 have a story. Looking for an image.
55:49 Yeah. Yeah. Exactly.
55:52 So all of those images I just went
55:54 through, there were probably 1500 images
55:56 that I just blasted through, right? So
55:59 this was I don't know two, three months
56:02 in of this project. I'm making just
56:05 every night I was making 20 30 50
56:09 images and then one night this image
56:12 came
56:13 up and I actually got
56:17 emotional like I actually like this it
56:20 took my breath away. It was like like oh
56:27 [ __ ] And I couldn't figure out what
56:29 that what that feeling
56:32 was. And what that feeling was was of
56:36 all those other images. This
56:41 image was was an image that is who I
56:46 am. that if I had the talent and the
56:49 skill and the Photoshop or the
56:51 illustration skill or the compositing
56:54 skill to craft something from scratch,
56:57 this is the kind of image I would
57:02 make. This was this was this image was
57:06 probably my first
57:08 epiphany
57:12 around the idea of
57:14 curation, right? what Rick Rubin talks
57:17 about where he doesn't give a [ __ ] about
57:19 the audience. What Rick Rubin cares
57:21 about is, is this a song I would want to
57:24 listen
57:27 to? This was an image I would want to
57:30 make. And I had this weird relationship
57:33 with this image where there was a part
57:35 of me that was like, "Oh my god, isn't
57:38 it amazing that I made that image?" And
57:39 then there was a part of me that was sad
57:42 that I couldn't make that image.
57:45 And yet I just had. But had I really?
57:48 And it it's kind of like what you're
57:50 talking about. Right. Gareth Hood. Yes,
57:53 Kyle. I absolutely know that
57:55 feeling and love that you got to
57:57 experience that. So cool. Right. Yeah.
58:00 And it's it's not that this image is is
58:02 in and of itself better or worse than
58:05 any of those other images. It's just
58:07 that this image is me.
58:10 This image is like a full
58:13 self-expression
58:15 moment that was generated like that. But
58:19 I couldn't have gotten to this unless
58:20 I'd done all those other 1,200
58:22 images. And that idea that people have,
58:25 that [ __ ] trope that people have
58:27 that, oh, AI is lazy because you just
58:30 push a button and out comes an image.
58:33 Yeah. And most of them are
58:37 crap. For every image that CJ Fletcher
58:40 shares with the world, he's made a
58:42 thousand images that didn't make the
58:44 cutting room, you know, that that are on
58:46 the cutting room
58:49 floor because this is about you and what
58:53 you want and what your idea is. So
58:57 anyway, great question. I don't even
59:00 know if I answered
59:03 it. I'm such a pain in the
59:06 ass. It moves you from the inside.
59:10 Exactly. And like that's the that's the
59:12 kind of miracle of
59:16 of AI in general. Sometimes it happens
59:19 with the music programs. Sometimes it
59:22 happens with chat GPT. Sometimes it
59:24 happens with Claude when it it makes you
59:26 those interactive
59:28 dashboards. You know, it did for me. In
59:30 fact, let's go to lovable right now.
59:37 Um, the project cardboard thing. This
59:40 one, the to-do list
59:45 maker. I wrote a bunch of to-do lists on
59:48 my whiteboard and I I thought I those
59:50 should be cards. I want that dynamic.
59:52 And then I thought, "Oh, I should just
59:53 go build that." And then I
59:55 did. And I built this thing that works
59:58 the way my brain works. I can sort by
1:00:02 different categories of projects and
1:00:03 then within them I've got little mini
1:00:05 to-do lists and there's like zero
1:00:07 structure about it. I can I can drag
1:00:10 to-do lists up and down. I can add
1:00:14 subtasks. All right. Now I need to how
1:00:17 do where do I
1:00:19 go to make this editable?
1:00:24 Uh, whoever was asking before Brand
1:00:26 Brandon
1:00:29 Jet, you can publish empty.
1:00:37 Well, yeah. Like it publishes with a
1:00:40 let's see where do
1:00:43 I
1:00:46 publish two superbase warnings review
1:00:50 security.
1:01:03 No, but you said there's there's some
1:01:06 play. Well, let me ask
1:01:08 Lovable. How do I publish this so people
1:01:16 can remix it?
1:01:27 [Music]
1:01:38 Come on, lovable. What's going
1:01:45 on? Stop.
1:01:57 Hang
1:02:05 on. What's the URL for that? The
1:02:10 URL. Click the publish button. This will
1:02:13 deploy your app and make it publicly
1:02:15 accessible. Once published, you'll get a
1:02:17 URL that anyone can visit, making it
1:02:20 mixable. After publishing, anyone can
1:02:24 remix by visiting the published app and
1:02:26 looking for the edit and lovable badge
1:02:30 on your app if
1:02:33 enabled. Project name, lovable badge.
1:02:36 Keep the edit. Okay. You can control the
1:02:40 remixing experience in your project
1:02:42 settings. Where are my project
1:02:46 settings?
1:02:49 Settings.
1:02:51 Okay. Keep wait. Private project. Oh,
1:02:55 keep your project hidden and prevent
1:02:58 others from remixing it. Well, that's
1:02:59 shitty. Let's turn that off. Hide the
1:03:03 lovable badge. Remove the edit with
1:03:04 lovable. No, we don't want to do that.
1:03:05 We don't want to hide
1:03:07 it. Okay, so I did that. So now, let me
1:03:11 publish it.
1:03:15 update. I've got two superbase warnings.
1:03:18 I'm sure I've got, you know, API keys in
1:03:21 there and [ __ ] like that. So, don't
1:03:23 steal my
1:03:27 [ __ ] Project updated. Okay. So, if you
1:03:31 go
1:03:36 to I'm putting this on the YouTube
1:03:38 channel.
1:03:40 It's
1:03:46 project-cardboard.lovable.app. And now
1:03:47 it says edit with lovable. And you can
1:03:49 now go edit it. There you go. Thanks for
1:03:52 coming back and asking about that
1:03:53 because I knew I knew I said I was going
1:03:54 to do it and I didn't do it because I
1:03:56 have
1:03:57 ADD. It's the universal excuse for not
1:04:00 doing
1:04:04 [ __ ] Oh, wait. Sorry. Here. Let me put
1:04:06 this up on screen.
1:04:11 There is the
1:04:14 URL. There you go people. You see what
1:04:18 I'm saying there? You see what I'm
1:04:20 saying? You understand? You understand
1:04:22 how this works,
1:04:26 people? Save it in your PC. Publish it
1:04:28 empty.
1:04:30 Why? It's it's already like it's
1:04:32 published as an app. You can go create
1:04:34 your own account on it. It's got a
1:04:36 superbase back end. You're not going to
1:04:38 see any of my
1:04:42 to-dos. You get to make your own to-dos.
1:04:44 This is a this is a cool
1:04:46 project. I mean, I can tell you about it
1:04:48 if you want me to walk you through the
1:04:50 application, but
1:04:52 whatever. All right.
1:04:56 Um, there's there's architect's movie
1:05:01 started, by the way. Good movie. Good
1:05:04 story. Go make some images.
1:05:08 Oh, you know the other thing we can do
1:05:10 there, archetype. Uh, I gota I got to go
1:05:13 get the prompt again. Hang
1:05:15 on. Copy. Let's go to
1:05:19 flow. Is this
1:05:22 flow? Yeah. Am I Is this
1:05:26 flow? Yes. Okay. So, where I am is I'm
1:05:34 in the
1:05:39 new
1:05:41 VO3 from
1:05:45 Google. And this is bonkers. So, we're
1:05:47 going to do text to
1:05:48 video. Oh, we could do frames to video.
1:05:55 Oh, but I don't think I don't think
1:05:56 frames to video you can do. You can't
1:05:59 have it do talking. So, we're going to
1:06:00 do text to
1:06:01 video. So, we're going to say in this
1:06:10 scene, I'm going to put it in quote
1:06:15 marks. Grayson's parents sent him back
1:06:17 to Earth on the day of his 18th birthday
1:06:20 with one mission. Save the world. So,
1:06:22 not so
1:06:30 hard. Then I'm going to
1:06:34 say
1:06:44 Grayson fogs up
1:06:47 the
1:06:49 inside of his
1:06:53 Visor as
1:07:03 he says out
1:07:06 loud how much he
1:07:10 hates his
1:07:13 parents. All right, as he says out
1:07:17 loud how much he hates his parents. All
1:07:20 right. So, we're going to make a movie
1:07:21 scene now. Architect Brian Whitney. I
1:07:24 got it. I just had to create an account
1:07:26 to get it. Ah, okay. Great.
1:07:30 Perfect. Yeah, you got to create an
1:07:32 account because it's it's uh it's got
1:07:35 authentic it's it's got it's got user
1:07:38 authentication. That app, it's that's
1:07:39 actually a real app. Like, you can use
1:07:41 that for your own uh for your own to-do
1:07:44 lists and [ __ ] like that. It's all
1:07:47 sitting in Superbase there. And I don't
1:07:49 have access to it. Just you have access
1:07:50 to your
1:07:57 data. Is it free? Of course it's
1:08:05 [Music]
1:08:11 free. Oh, that's not the right
1:08:20 Why did it not make
1:08:22 audio? Oh, because I did it. Oh, okay.
1:08:25 This is one of the This is one of the
1:08:26 shitty things about Flow is it defaults
1:08:28 to
1:08:30 V2.
1:08:36 Um, hang on. Let's see. V2 fast V3
1:08:41 highest quality. So, this will be
1:08:56 And I think reuse the prompt. Grayson
1:08:59 fogs up the inside of
1:09:01 his space suit
1:09:06 helmet
1:09:09 visor as he says out loud how much he
1:09:12 hates his parents. All right, so we'll
1:09:13 do two of these.
1:09:17 That one's pretty
1:09:18 bad. But he does look like he hates his
1:09:21 parents for putting him in welding
1:09:24 [Laughter]
1:09:32 class. All
1:09:34 right. All right. So So now we're going
1:09:37 now we're making Now we're making actual
1:09:39 video. All right. Beautiful. Beautiful.
1:09:42 Beautiful. Beautiful. Beautiful,
1:09:44 beautiful, beautiful people. You're all
1:09:46 beautiful people.
1:09:48 Okay. Does anybody know what we're
1:09:50 talking about? I didn't bring the
1:09:52 syllabus
1:09:54 today. He's weed eating the world.
1:10:01 Exactly. Visor is going to confuse it, I
1:10:04 think. Yeah, it
1:10:05 might. Here I am giving archetypal
1:10:08 architect lessons on how to prompt and
1:10:10 prompting shittily. Which again gets
1:10:13 back to my point. This is about each of
1:10:16 us
1:10:18 individually finding ways to put
1:10:20 ourselves in the prompt, our point of
1:10:23 view. You get your point of view in
1:10:26 there, it amplifies
1:10:28 you. That's the [ __ ] magic. That's
1:10:30 the magic. That is the
1:10:33 magic. Feed the prompt. Why do you feed
1:10:37 it? You. What's that mean? Do you have
1:10:39 ideas? Yeah. Do you have problems? Yeah.
1:10:42 Do you have health conditions? Yeah. Do
1:10:44 you want to create a pizza party for
1:10:45 your friends? Yeah. Feed it that it's
1:10:50 hungry. Feed the
1:10:53 prompt. Feed the
1:10:55 prompt. Man, that does not work for me.
1:10:58 It's not catching on at all that this is
1:11:01 anything sci-fi or serious. put put in
1:11:04 things like sci-fi, like put sci-fi in
1:11:08 the style of Bladeunner.
1:11:17 Oh, you know what to do, archetypal
1:11:19 architect, take your prompt that's not
1:11:21 working, go to chat GPT, put in put in
1:11:24 your thing that's not working and say to
1:11:27 chat GPT, describe this as a sci-fi
1:11:30 scene, and then take all of what it
1:11:33 generates and and put it out there. All
1:11:35 right, here's here's
1:11:37 Grayson. You ready,
1:11:39 people? Oh, I got to change my sharing.
1:11:41 Hang on a sec.
1:12:00 I hate my parents.
1:12:07 It's good. I hate my parents.
1:12:11 Good acting.
1:12:15 I mean, the fog comes through the visor,
1:12:17 so that's not great, but that's not bad.
1:12:21 Let's see. Do we have a second one?
1:12:23 Yeah. Here we
1:12:26 [Music]
1:12:33 go. I hate my parents so much.
1:12:42 I hate my parents so much. Huh?
1:12:51 I hate my parents so much.
1:12:55 Does my prompt have any personalized
1:12:57 numbers added to the end of it? No. What
1:12:59 do you mean?
1:13:02 I hate my parents so much.
1:13:06 Very futuristic visor.
1:13:10 You'd think if they were this far in the
1:13:12 future, they would have figured out how
1:13:13 to make it not fog
1:13:14 [Laughter]
1:13:22 up. Yeah. Architect, maybe what you do
1:13:25 is you just skip skip making images and
1:13:29 just go right to making a
1:13:33 film. Oh
1:13:35 man, that's pretty cool.
1:13:43 [Music]
1:13:46 All right, everyone. All right,
1:13:48 everyone.
1:13:49 Fantastic. Fantastic. Oh, talking to
1:13:52 talking to Archetypo. Got it. That's
1:13:55 what I like about this channel. You you
1:13:57 you all are in here solving each other's
1:13:59 problems. Well, I'm just waxing poetic
1:14:03 about the right questions to ask.
1:14:07 Can you use a picture in the prompt? You
1:14:11 can, but I don't think so. Let's try it.
1:14:14 Let's go grab this image. We're going to
1:14:17 download this image. I love that image,
1:14:19 by the way. Like that. That one. If if
1:14:22 we could make this one talk. That would
1:14:25 be great. So, let's
1:14:29 see. We'll go back to Flow. Is
1:14:33 this No, that's Flow TV.
1:14:36 Does this blow? Yeah.
1:14:38 Okay. So, let's go to image to
1:14:45 text upload.
1:15:18 Why did that not
1:15:22 upload? Did
1:15:30 it? There we go. Crop and
1:15:40 save might violate our
1:15:43 policies. Oh, you got to be [ __ ]
1:15:45 kidding
1:15:47 me, Google.
1:15:50 Seriously, what is it? Because he looks
1:15:53 sub
1:15:56 18. Unfucking believable.
1:16:01 All right, here. Watch. We're gonna fix
1:16:04 this, [ __ ] We're gonna go edit. I'm
1:16:08 in midjourney now. I'm gonna erase his
1:16:11 lip and I'm just going to put um thin
1:16:14 mustache.
1:16:19 I am so sick
1:16:28 of [ __ ] safety policies that are
1:16:40 horshit. Sick of it, people. Sick of it.
1:16:47 Why did I get no mustache? Look at this
1:16:49 one. He made a He made a mustache out of
1:16:53 [Laughter]
1:17:00 Twizzlers. Oh my god. We got nothing. We
1:17:03 got
1:17:03 nothing. Ah
1:17:06 [ __ ]
1:17:09 I I'm exhausted.
1:17:13 Let's
1:17:14 do we'll do edit. We'll erase his whole
1:17:26 face
1:17:29 18year-old
1:17:32 with facial
1:17:35 hair. See if we see if we get anything
1:17:38 by completely obliterating his face.
1:17:45 I got that yesterday, too. Oh, this
1:17:48 one's
1:17:49 good. I guarantee you it'll let us do
1:17:51 that one, but it won't let us do the one
1:17:53 where he looks slightly
1:18:02 innocent. Somehow made him look
1:18:07 younger. I I'm exhausted. I can't. Well,
1:18:10 we're gonna do we're gonna do a zombie
1:18:12 film
1:18:18 now. All right, back to flow. Now, we
1:18:22 just turned this into a zombie
1:18:24 film. Way to go, architect. You know, we
1:18:28 we wanted to sleep tonight. Now, we're
1:18:29 not going to be able to sleep. Thank
1:18:31 you. Thank you for Thank you for your
1:18:34 service. You can create anything as long
1:18:37 as it's what they want.
1:18:39 H rabbit hole. Yeah, welcome to the
1:18:41 rabbit hole. Okay, so we've got our dude
1:18:43 in there. Now, let's say generate video
1:18:45 with first and last frames. So let's
1:18:49 say
1:18:53 astronaut
1:19:04 Astro
1:19:05 astronaut
1:19:07 size and says how
1:19:13 much he hates
1:19:16 his parents. I don't think that this one
1:19:21 generates
1:19:23 voice, but we'll find out
1:19:27 together. Uploaded my kids book. Vicky.
1:19:31 Vicki, while we've all been um being
1:19:35 productive talking about the future of
1:19:39 mental health with AI, Vicki's been
1:19:42 producing a book
1:19:46 Gareth is working on his. That's
1:19:49 awesome.
1:19:51 Detour. Sure. 18-year-old zombie sent to
1:19:54 Earth to save the world. Exactly.
1:19:55 Wolfman
1:19:57 Clint. Oh, Champy. It is not cheese time
1:20:05 yet.
1:20:07 Mew. You want to go out? Do you want to
1:20:09 go visit
1:20:10 mama? Yeah. All right. Hold,
1:20:19 [Music]
1:20:27 please. All
1:20:56 right.
1:21:04 43%. Literally during the live. That's
1:21:07 great. Source Camp, did you all write a
1:21:10 children's book or something? Yeah,
1:21:11 Source Camp, there's a uh OpenAI um
1:21:15 contest for children's books.
1:21:30 Vickiy's book will win the
1:21:34 contest. I missed the last 30 minutes.
1:21:37 Did I miss anything mind-blowing? This
1:21:39 is the AI learning lab. Of course
1:21:45 not. No, nothing mind-blowing.
1:21:49 [Laughter]
1:21:55 You're good. You're good. You're caught
1:22:02 up. Oh
1:22:04 man.
1:22:06 76% till we get our zombie
1:22:12 thing. Zombies with popping out eyes are
1:22:15 fine, but young men are not.
1:22:17 Unbelievable.
1:22:33 It does do audio.
1:23:00 He come
1:23:03 on. Jesus Christ. Honest to God,
1:23:08 Apple. The the [ __ ] window reordering
1:23:11 that Apple does just [ __ ]
1:23:16 randomly. Could Could someone please use
1:23:19 AI to resurrect Steve Jobs so we can get
1:23:21 some quality control going at Apple? I
1:23:23 mean, honest to God, could someone like
1:23:26 someone any Does anyone even work there
1:23:31 anymore? The breath is impressive. Yeah,
1:23:33 it looks so alive. Yeah, that's that's
1:23:35 pretty
1:23:36 crazy. Um, let me try let me try another
1:23:41 one. Let's go grab a different image.
1:23:51 The good ones we can't use. But let's go
1:23:53 grab one of these arty
1:23:58 ones. That one's good. That one's good.
1:24:01 Download
1:24:04 that. Go back to Flow. Hello,
1:24:09 Flow.
1:24:12 Upload Grayson. Grayson is looking for
1:24:16 real, Kyle. I think Steve would be so
1:24:18 disappointed in Apple right now.
1:24:20 Seriously.
1:24:23 Right. Like Steve was the only [ __ ]
1:24:26 willing enough to say,
1:24:30 "No, I don't care if you have to work
1:24:32 all weekend. Go [ __ ] fix it."
1:24:39 Um,
1:24:41 Grayson
1:24:43 size
1:24:44 relief he is
1:24:48 alive and
1:24:53 says how much he hates his
1:24:59 parents under his breath.
1:25:05 All right, let's see if we can get
1:25:08 one. Uh uh uh. I forgot the tapper is
1:25:11 still running. Love it. Brandon Jet
1:25:15 Kyle, I'm in 100% geek out right now.
1:25:18 About about to be all night. Um team no
1:25:21 sleep project cardboard. So So Brandon,
1:25:24 um if you if you remix it and do a cool
1:25:28 upgrade of it, let me know. I'm
1:25:30 excited. I'm excited. You're going to
1:25:33 take my three hours worth of work and
1:25:35 put in
1:25:37 eight. Um, that's super cool. Awesome,
1:25:40 man. Have fun with it. Um, yeah, that's
1:25:44 really cool. Love it. Love it. Love it.
1:25:47 Love it. Love
1:25:49 [Music]
1:25:53 it. That hole in his cheek. I know,
1:25:55 right? Steve Bashi as a kid. Yeah,
1:25:59 totally.
1:26:05 That's that's that's actually I mean,
1:26:07 you know what you would actually do with
1:26:09 this? I what I would do with this is you
1:26:11 would actually lose that that part of
1:26:13 the thing and you'd start it right
1:26:15 there,
1:26:17 right? Like he just had a crash landing.
1:26:20 His visor got smashed. He's got a hole
1:26:23 in his
1:26:24 face. He's in
1:26:28 shock, right?
1:26:30 Like that's a good I'm in shock and it
1:26:33 sounds
1:26:35 good. Did I Where are we with sharing?
1:26:39 What am I sharing right now? Can you
1:26:41 hear audio of this if I play
1:26:45 it? Wait, why is it not
1:26:50 playing?
1:26:53 Reload. Reload.
1:27:02 No
1:27:05 audio. I guess I shared it wrong. Oh, I
1:27:08 think I went back to sharing the
1:27:09 oldfashioned way. Hang on.
1:27:31 Absolutely. My managers are to be using
1:27:33 this also. All right. Good. I'm glad
1:27:36 we're Well, listen. The the whole point
1:27:38 of that of that to-do list app is it's
1:27:42 for ADHD brain where it's like it's not
1:27:47 meant to be a replacement for Trello. a
1:27:50 bunch of
1:27:51 [ __ ] wherever I posted it. Maybe it
1:27:53 was
1:27:54 LinkedIn. Oh, no. It was on It was on
1:27:56 Tik
1:27:57 Tok. They came in. They're like, "Yeah,
1:28:00 uh, it already exists. It's called
1:28:01 Trello." I'm like, "Uh, yeah, Trello's
1:28:05 like, you know, a [ __ ] bazooka and I
1:28:07 needed a BB gun." You [ __ ] turd
1:28:13 knuckle. Those tick tockers.
1:28:18 Oh
1:28:20 [Music]
1:28:23 man. Yeah, on on Tik Tok you can hear
1:28:26 things, but I've got to share it
1:28:28 different if they want to hear it over
1:28:29 there. He reminds me of the movie Warm
1:28:31 Bodies. Oh, yeah. It does have that
1:28:33 aesthetic, doesn't it? Did I lose that
1:28:35 other
1:28:37 video? I think I
1:28:43 did. This is so [ __ ] buggy.
1:28:48 God damn
1:28:51 it.
1:29:01 Flow. Oh, here he is. Okay, we got we
1:29:03 got it. We got one. We got a live
1:29:06 one. All right, let's see what we got.
1:29:09 Is it any
1:29:11 good? No audio.
1:29:21 Looks pretty
1:29:23 good, but no
1:29:28 [Music]
1:29:35 audio. Thank you,
1:29:43 Vicki. Mm. All right. So, apparently I
1:29:47 don't I don't know what's going on. This
1:29:49 this this app is really buggy. I don't I
1:29:53 don't know what you want for 250 bucks a
1:29:55 month. You actually wanted it to work.
1:29:58 Oh, no. I'm sorry. That would be
1:30:02 more. You all are silly. We've done the
1:30:05 remarkable thing here. This is video
1:30:07 that almost works sometimes. You wanted
1:30:09 it to actually work. It's cute. It's
1:30:12 cute. Fantastic. Oh, dog's back. Champy.
1:30:16 Champy got his
1:30:18 cheese. All right, everybody. I think
1:30:20 I'm going to get on out of here.
1:30:22 Fantastic. Oh, I didn't plug in my
1:30:24 phone, so my phone's about to die. The
1:30:26 Tik Tockers are about to lose me alto
1:30:27 together. Where did Kyle go? Why did he
1:30:29 just leave us? Why didn't he say
1:30:32 goodbye? Because no one's here to remind
1:30:34 me to plug [ __ ] in. That's
1:30:38 why. All right.
1:30:41 Um, it's Wednesday night. Tomorrow,
1:30:45 Thursday. Ah, I've got a Thunderview CEO
1:30:49 dinner tomorrow night. So, I will we
1:30:51 will probably be late. What app is this?
1:30:53 The app. The janky. Yeah, the $25,000
1:30:56 level is the one that works. The $250
1:30:59 buck a month level. It's not so not not
1:31:01 really working. Um, the app that I'm
1:31:05 using is flow.google. So, if you go to
1:31:08 flow.goole Google right now. You can go
1:31:11 there, but if you if you don't pay
1:31:13 anything, I don't think you can use it
1:31:15 at all. If you pay 25 bucks a month, you
1:31:17 get like eight
1:31:19 videos. If you
1:31:21 pay you subscribe for 250 a month, but
1:31:24 the first three months are
1:31:27 125. So, for 125 bucks a month, you get
1:31:30 12,000 credits. And I think it's like a
1:31:33 100 credits a video or something like
1:31:35 that. So, you get a decent amount of
1:31:36 videos, not a ton.
1:31:39 Um, but it's it's really good, right? It
1:31:44 does like acting like this. This one in
1:31:47 terms of acting, this one was pretty
1:31:49 good.
1:31:54 I hate my
1:31:57 parents. Pretty
1:31:59 good. Pretty good. This one, not as
1:32:03 much, but you know, better physics.
1:32:09 I hate my parents so much. You know
1:32:12 what's amazing to me is it really got
1:32:15 like the age of his voice really matches
1:32:19 the age of him, man. Yeah, I know.
1:32:22 Exactly. This one's Matt
1:32:24 acting. I use VO3 through Po. Okay,
1:32:27 cool. So, yeah, if you're a subscriber
1:32:29 to Po, Po gives you access to so many
1:32:32 different models. If you're tired of all
1:32:34 the model [ __ ] and where to access
1:32:37 it, um, Po's really good. I don't like
1:32:39 the interface personally, but they give
1:32:42 you access to lots of models and and
1:32:44 they're they're really quick on it. All
1:32:46 right, cool. So, I will be late
1:32:48 tomorrow. Tomorrow be probably more like
1:32:50 nine o'clock. It costs about 50,000
1:32:53 credits a vid. Oh, it does? No, it can't
1:32:56 it can't be that many. Oh, it it costs
1:33:00 50,000
1:33:01 credits on PO per vid. Got it. Okay.
1:33:05 Yeah, V3 is really expensive. It's like
1:33:08 a buck 50 a video or something like
1:33:10 that. I have PO. It's nice. All right.
1:33:13 Beautiful. Okay. All right. Listen,
1:33:16 thanks for hanging out tonight. Hope you
1:33:17 had fun. Hope you learned something.
1:33:19 Don't
1:33:20 forget the question we don't ask anymore
1:33:23 is, "How can I get the most out of AI?"
1:33:25 The question we do ask, how can I get
1:33:27 the most out of myself using AI?
1:33:31 Okay, feed the prompt. You, the prompt
1:33:35 is hungry. And what it wants is you,
1:33:38 your problems, your
1:33:40 ideas. Feed it. And I'm telling you,
1:33:43 [ __ ] will
1:33:45 change. It'll make a difference. Man, I
1:33:48 missed this. Yeah, it was a fun night
1:33:49 tonight. We did some actual stuff. We
1:33:52 had some good
1:33:54 philosophy. All right. AI can give you
1:33:57 wings. All right. Peace out everybody.
1:34:00 I'll see you tomorrow night. Bye.