5/29/2025 - GPT5, Grok 3.5, Krea.ai and DeepSeek R2, also highlighting RunwayML's new features

Description

Chapters

Transcript

0:01 Data.
0:05 [Music]
0:27 No.
0:30 [Music]
1:04 [Music]
1:17 Freedom came my way that night.
1:20 [Music]
1:24 Just like a jet plane in and out of
1:28 sight. I was calling ass at a million
1:32 miles an hour wondering
1:34 how I'd
1:36 [Music]
1:38 hit when they came into the st.
1:46 They said I was bad beyond
1:49 repair, but I got no qualms with my
1:55 situation. Hey, here I am.
1:59 [Music]
2:05 So say Sheree Sheree Sheree won't you
2:09 dare to say
2:12 Sheree Sheree won't you dare to say
2:17 Sheree Sheree
2:20 [Music]
2:22 Shereeh yeah leave a message and your
2:26 number
2:27 [Music]
2:29 please take your time to want to satisfy
2:35 Take all these old fantasies and send
2:38 them care of
2:42 me. Oh, hello good people. Happy
2:44 Thursday. How's the hair? How's the
2:46 hair? Really important that we get that
2:48 right. As you all know, I'm very
2:52 vain as evidenced by all of
2:59 this. Oh, we got seven people in here.
3:02 Let's get this thing rolling. If you
3:04 would be so kind, share the
3:07 live. Tappy tap the
3:09 screen. Let Tik Tok know we're having
3:13 fun here.
3:16 [Music]
3:17 [Applause]
3:18 [Music]
3:31 There's been something, baby, I've been
3:33 trying to
3:36 say for an age, and it seems I don't
3:39 know
3:40 [Music]
3:42 how with past and a future now
3:45 surrounding me.
3:49 Surrender to whatever cheap thrill can
3:51 be
3:52 [Music]
3:55 found. There's been a little
3:58 trouble since you came to my
4:05 rescue. And if you're like all of the
4:08 rest, I would have quit you long ago,
4:12 but I couldn't do
4:14 [Music]
4:16 that. Oh man.
4:20 lot lot of stuff today. Lots of little
4:22 things sort of perking and popping. I
4:24 think there are folks waiting for GPT5
4:28 and Grock 3.5
4:31 and DeepSeek R2 and there there's a
4:34 bunch of big things in the
4:36 wings. Um, but stuff's still moving.
4:39 There's stuff still
4:43 happening. Steo's in the house. What's
4:45 happening? What's shaking? What's going
4:47 down there, Steo? Hope you're doing
4:50 well. Yeah, hope you're doing fantastic
4:52 there down there in uh in the in down
4:56 under where they have the shrimps on the
4:58 barbie. Yeah. Hope it's probably
5:00 fantastic down there, right? Isn't it?
5:02 Like you're the opposite. So like you're
5:04 going into winter, we're going into
5:06 summer. I think it's like that. Your
5:08 toilets, they go the other way. Yeah,
5:10 it's fantastic. I hope you're hope
5:12 you're mowing the yard and and and
5:14 getting your granddaughter to do the
5:15 tappy taps. Yeah, it'll be fantastic.
5:18 Fantastic there. There, Steo. Yeah, it's
5:21 it's really nice down here in Australia.
5:23 I was I was never down there, but I'll
5:25 tell you what I did do. I went on the my
5:28 wife my wife, she showed me how to use
5:30 the Google Maps. And what's what's
5:32 remarkable about the Google Maps is it
5:34 it's got the whole world. It's got the
5:37 whole world. So you go right down there
5:38 and I went on a little what do they call
5:40 those in the in the outback? A little
5:42 foot foot path journey, right? And I did
5:46 one on the Google Maps. It was
5:48 fantastic. Took took me about three
5:50 days. I I was exhausted after. I mean,
5:52 yeah, I had lemonade. I had lemonade,
5:54 you know, to to to get me through it,
5:56 but it was fantastic. I I did it in real
5:58 time. I zoomed into one one to one and
6:01 was just walking there in a Google map.
6:02 That was awesome. It was just awesome.
6:05 All right.
6:06 Fantastic. Fantastic. What's this? What
6:10 is going on here? I've got something is
6:12 bad with my screen redraw. Ah, there we
6:15 go. Perplexity. Wait, what's it
6:20 say?
6:22 Perplexity has what looks like an
6:25 awesome new labs feature. All right,
6:27 let's add Perplexity to the list.
6:29 Fantastic. The black bar. Let's put the
6:31 black bar
6:33 there. Kak Kyle, look at my way of code.
6:35 Where is it? Kak, where'd you put
6:39 [Music]
6:44 [Applause]
6:48 it? All right, look what I made over a
6:52 fellow. It's at the AI fellow. And look
6:54 what I
6:56 [Music]
7:01 made. What is is Oh, that's Yeah. Yeah,
7:05 that's those. That's
7:06 those. Earn your first badge. Look what
7:10 I made. There it is. There it is. That's
7:14 fantastic. Cactus Simmons, the way of
7:17 code. Just a moment. All right, let's go
7:19 see what we got here. This is from KKE.
7:23 Oh,
7:25 damn. Gorgeous.
7:28 And did you add
7:31 the Wait, where's the music coming from
7:34 when you
7:40 click? How did I do
7:48 music? Prompt is in the
7:51 salon. Tabs for audio. I can't get it to
7:54 consistently make audio.
8:01 I don't know what I'm
8:03 doing. Move your
8:11 [Music]
8:15 cursor. That's
8:18 beautiful. How did you do that?
8:21 Wait. Audio. Move mouse for interactive
8:24 sound. Click.
8:31 Click. That's very
8:34 beautiful. How did you get the the uh
8:37 the poem to show up on top of it? Did
8:39 you just tell it to do
8:40 that? It's an ambient meditative sound.
8:43 That's beautiful. Did you publish it to
8:45 the project? And the prompt is in the uh
8:49 the prompt is in the is here.
9:03 Where's the prompt? I don't see the
9:05 prompt. I don't see the prompt there.
9:08 Producer
9:09 Brianna. Anyway, that's
9:13 gorgeous. Gorgeous. Po just added
9:16 VO3 hash the way of code.
9:23 #thewayof code. By the way, if you
9:26 haven't seen No, I just see this one
9:28 this one
9:29 post and it does not have the prompt in
9:32 it that I can
9:40 [Music]
9:44 see. That's pretty gorgeous. It's pretty
9:47 gorgeous, I tell you.
9:55 [Music]
9:58 Um, all
10:01 right. What I had a busy day today, so
10:03 my head my head was not I did a fair I
10:06 did a lot of AI stuff today. So at work,
10:10 I've been doing a lot of video avatars.
10:12 So, I've been doing a lot of voice
10:14 synthesis work and voice designing and
10:18 people designing and then trying. So,
10:23 here's one of the challenging things
10:25 with AI when you do stuff for clients is
10:29 they're like, "Oh, could you just make
10:32 her shirt a little different or this or
10:33 that?" And it changes everything else.
10:35 And it's it's a it's a bit of a struggle
10:38 to get it to get subtle things like big
10:42 things are easy, subtle things are hard.
10:44 So, one of the things I'm going to show
10:46 tonight is Korea just added
10:48 um a new thing called context, which is
10:52 a new image editing tool that keeps the
10:54 base image intact while you change
10:56 things within it. The prompt was, "Oh,
10:59 what if it was weightless?" Oh, that's
11:01 fantastic. Oh, that's super cool. Nice.
11:06 [Music]
11:09 Yeah, that's one of the cool things
11:11 about things like vibe coding where
11:14 you're just modifying someone else's
11:16 thing
11:17 is the prompt doesn't have to be that
11:19 big. I think some of my favorite
11:21 midJourney images were single word
11:23 prompts.
11:26 [Music]
11:52 Granddaughter just arrived. Let the tip
11:54 tapping begin. But I don't want to,
11:56 Grandpa. My finger hurts. Keep tapping,
11:59 Sally. Exactly. You like it? That's
12:02 solid. That's
12:04 solid. Child labor down under. I like
12:07 it. I'm going to finish a
12:10 muffin.
12:11 Okay. What do we want to start with?
12:14 Let's
12:17 [Music]
12:25 start. Okay. If you
12:31 go, if you got access to
12:38 runway, where is
12:41 my that one? Yeah. If you've got access
12:45 to
12:46 runway runway ML, go to
12:53 academy.runwayml.com. And hang on, let
12:56 me straighten things out here for the
12:57 tick
12:58 tockers. You know those tick tockers, if
13:00 you don't treat them right, they get
13:03 sassy.
13:07 Um, hang on. Things are all screwed up
13:10 over here. I don't know what producer
13:13 Brandon did, but it's not
13:16 good. It's not his fault, actually.
13:19 Okay.
13:23 do a little new black bar action. Ann
13:25 says, "Hi, terrific." Hello, Ann Murphy.
13:28 What is shaking? Okay, so if you go to
13:31 academy.runwayml.com,
13:35 um it's it's a bunch of uh you know,
13:38 lessons, right, on how to use Runway.
13:40 But I was on a I'm in the creative
13:42 partner program for
13:43 Runway. And one of the things that we're
13:46 doing today in the upper right hand
13:47 corner if you go to this comm community
13:49 community
13:50 workflows, it takes you into a bunch of
13:53 basically Twitter posts
13:55 um of people from the community sharing
13:58 how they do what they do. And they were
14:00 they were going through through some of
14:02 these today. And it's, you know, it's
14:04 things
14:08 like being able to start with that
14:12 image, you know, and then say like a man
14:15 shooting an apple off a woman's head and
14:17 there's the composition and then runway
14:20 ML will generate that and then you can
14:22 turn that into a video. And so it was
14:26 just a bunch of those. And so they're
14:27 they're all sort of quick hit quick hit
14:30 little things.
14:33 of of just things you can do within
14:35 runway. This one was really
14:39 cool. So, this one
14:42 was someone uploaded a little flower
14:46 vase, you know, full of
14:48 sunflowers. And then it says, "Create a
14:51 board game based on the following visual
14:54 theme shown in image one." Right here's
14:56 image one.
14:58 uh put the board counters and any cards
15:00 into a nulling style layout. So laid out
15:03 in a nice organized fashion on a plain
15:06 background with studio lighting as if
15:08 for a product marketing campaign. And
15:10 then Runway ML does
15:13 this and just generates like an entire
15:16 [ __ ] board game called Bloom and
15:21 Harvest. The dice are a little weird.
15:23 The game board's janky. The pieces are
15:26 almost bigger than the board itself, but
15:29 still. I mean, holy crap, you
15:34 know. Oh, I did a Mad Lib like that for
15:37 everyone. Oh, that's very cool. That's
15:39 very cool. And and Runway ML's um
15:42 ingredients, this thing where you can
15:44 put three things together is is really
15:46 quite good. It's really quite
15:49 good. Um let me see if there was another
15:51 one in here that I thought was powerful.
15:55 that I super
15:58 liked. Um, you can do things like um put
16:02 raw 3D models in it, like position them
16:04 in a 3D program, screenshot it, and then
16:07 it'll follow that
16:08 layout, which is
16:13 cool. Tik Tok pin. Did someone say
16:17 sunflowers? Amelio's wife. What's
16:21 happening? Yeah, here's a cool one. This
16:23 was a cool
16:25 one where, you know, they put in the 3D
16:28 model with with no texturing. Then they
16:31 said, "Here's the texture of the plane.
16:32 Here's the sky it's flying in." And then
16:35 they basically just, you know, it it
16:37 generated that image and, you know, then
16:40 they put it that way and it generated
16:41 that image and then that one. And if you
16:44 look close, it doesn't follow the model
16:46 exactly. Everything's a little janky,
16:48 but it's still pretty
16:50 good for a lot of stuff. If you're going
16:52 to turn these into movies, you
16:54 know, people aren't going to notice
16:56 those sort of
16:58 details right now. And the stuff's going
17:00 to get better. Uh oh, this is one of my
17:05 favorites.
17:11 Um, so here's the plan. Here's the
17:15 furniture, right? So, wait, let me zoom
17:17 in so you can see
17:19 it.
17:24 Right, right. Here's all the furniture
17:27 all laid out on a single image. Because
17:30 what's what's fascinating about
17:31 references in Runway ML is that they
17:34 only give you three imi reference
17:36 images, but you can put, you know, a
17:39 dozen different pieces of furniture in
17:42 this image and then this is a plan. And
17:45 then the way they write the prompt is
17:47 you know generate an interior image
17:50 using the furniture style colors and
17:52 materials shown in style you know
17:56 underscore internal one or whatever it
17:58 is. The layout must strictly follow the
18:01 furniture positions and shapes from the
18:03 floor plan in plan two. And
18:07 then right so there's the floor plan,
18:10 there's the furniture and there's what
18:12 it did. Now, let's go look. Did it
18:15 follow the floor
18:17 plan? Single sofa, multi-person sofa,
18:21 wall cabinet, round table in the
18:25 middle. Round table, single sofa, multi-
18:28 sofa, cabinet over here. Yeah, it
18:30 followed
18:31 it. Amazing,
18:34 right? Groovy. It's groovy. I tell you
18:39 it's gravy. Get in my belly.
18:43 [Laughter]
18:45 You tasted,
18:51 baby. Wow, you're great at prompting.
18:53 Well, I didn't do
18:56 this. Wonder if AI got the idea for the
18:59 game board from one of the billions of
19:01 files on someone's computer. I think
19:04 that might be a troll. That's an anti-
19:06 AI troll, isn't it?
19:08 um just like humans got the idea from
19:11 board games from playing all the board
19:13 games they loved as a
19:16 child. It's it's really no different.
19:19 What what I would encourage you to do if
19:22 you're bitter about AI and you don't
19:23 actually use it, I would encourage you
19:25 to use it. And if you use it and you're
19:27 still bitter about it, I would encourage
19:29 you to learn how it works and learn what
19:32 it actually does because it's not
19:34 copying and pasting. Now, were these
19:37 models trained on stuff they weren't
19:39 supposed to be? Absolutely. Is it a is
19:42 it a crazy gray area? Absolutely. But
19:47 what I can tell you is this. If Open AI
19:51 and Enthropic
19:52 and X and all of
19:55 those frontier companies, if they had
19:58 gone to the creative community and said,
20:00 "Hey, we want to do this thing that's
20:02 going to be the equivalent of mashing
20:05 together all of the museums and all of
20:08 the libraries and all of the
20:09 universities and all of the collective
20:12 output of humanity. We're going to we're
20:14 going to put it into this magical little
20:17 ball that everyone can tap into and
20:20 everyone can benefit from all of what
20:22 humanity has contributed. Can we have
20:24 your permission to use this little
20:26 sliver of content that was done in like
20:29 the last, you know, 50 years or 30 years
20:33 where people are still
20:35 alive. Absolutely everyone would have
20:38 said no. None of this [ __ ] would exist
20:40 if they had asked per for permission.
20:45 The other thing I'll say is
20:47 this. Even if you're still resentful of
20:50 the fact that this was trained on stuff
20:53 it shouldn't have been trained
20:55 on, it's already been trained. Like the
20:58 water's under the bridge. AI is not
21:00 going away. and being a conscientious
21:03 objector and just not using it at all
21:06 and coming on channels like this and
21:07 just, you know, making smmy comments.
21:10 All it does is going to isolate you from
21:13 understanding what these things actually
21:15 make possible. And that makes me really
21:17 sad. Like the whole purpose of this
21:19 channel is not to say AI is good or AI
21:22 is bad. I agree that a bunch of the [ __ ]
21:24 that they do is bad. I agree that a lot
21:27 of what these things are capable of
21:29 doing is
21:31 scary and it's not going
21:35 away and it's going to transform every
21:39 [ __ ] job on the planet. Every one of
21:43 them. And so then you have a choice.
21:47 It's a binary choice. It's it's crystal
21:50 clear. If AI is not going away, if you
21:52 accept that
21:54 premise, take your opinion out of it.
21:56 Just if you accept the premise that it's
21:58 here, it's getting stronger, it's
22:00 accelerating, and it's not going away,
22:02 then you have two
22:03 choices. You deal with it or you don't.
22:07 Those are the only two choices.
22:10 And so if you're using the fact that
22:12 it's got some ethical shitness to it as
22:16 a reason not to participate and
22:19 understand what it makes possible,
22:20 you're actually just hurting yourself.
22:22 So this channel is just about not
22:25 judging it good or bad. It's just it's
22:27 here. Let's go look at what it makes
22:29 possible. And then let's talk about
22:32 how's that going to impact me as a
22:34 human and how can we create a community
22:37 where we can support one another and
22:39 learn as fast as possible what's coming,
22:42 what's
22:43 here. Like if you haven't seen V3, it's
22:46 [ __ ] insane. To which your response
22:48 would be, I guess that was trained on
22:50 every YouTube video. Yeah, it was made
22:53 by Google, so I'm pretty positive that
22:55 it
22:57 was. Did Google get everyone's
22:59 permission to do that?
23:01 No. Did Did Elon Musk get everyone's
23:05 permission to train Grock on every tweet
23:08 that's out there?
23:11 No. But it ain't going away. So anyway,
23:14 that's my two cents on it. All right. So
23:17 AI will benefit society and not the
23:20 billionaire billionaires that own it. It
23:23 already
23:24 is. It already is. Chat GPT has a user
23:28 base of like 600 million people. Like
23:31 every [ __ ] night on this channel,
23:34 every
23:35 night I get to witness people who are
23:38 coming out of retirement. I get to
23:40 witness people like Jim Ross who is he
23:44 takes every new AI innovation that comes
23:47 out. He runs um uh like self- storage
23:50 management
23:52 business, you know, like like you know,
23:55 lockers with, you know, like like self-
23:59 storage units. Those properties he
24:01 manages those. You're like, what's that
24:02 got to do with AI? He's [ __ ] figuring
24:04 out how to make it make AI relevant to
24:07 him. He says he's operating now like a
24:09 10-person company. He's a single person
24:12 company. He's making ads. He's making
24:15 advertising campaigns. He writes songs
24:17 for his clients. He creates images for
24:20 his clients. He creates marketing plans
24:22 for his clients. He'll get off
24:24 complicated pitch calls,
24:28 um, synthesize the conversation, turn
24:31 that into an interactive dashboard, and
24:33 send it to the person within 10 minutes
24:35 of them talking. And, and people are
24:37 like, "How did you do that? What? Who
24:39 are
24:40 you? He's occurring to people like a
24:44 [ __ ] wizard.
24:48 because he's curious and he's
24:51 adventurous and he's like, "Huh, if if I
24:54 can now just think of a song concept and
24:58 create a song about self storage for the
25:01 guy Bob I met with, let me go do
25:04 that." It's absolutely
25:07 remarkable. But you have to be willing
25:10 to step into it and try.
25:15 So
25:18 anyway, that's my two
25:20 cents. That is my two cents. Okay, I
25:24 stopped ranting. You can pin something
25:25 [Laughter]
25:31 again. Never a dull moment here at the
25:33 AI learning lab. My name is Kyle
25:36 Shannon. I don't know nothing about
25:39 nothing. But I'll tell you what what I
25:41 do know is AI is gonna [ __ ] everything
25:46 up. And you know, I personally
25:50 think in the long run it is going to
25:54 make things dramatically better. Like
25:58 dramatically.
26:00 Um might it [ __ ] everything up? Might.
26:05 Is it going to [ __ ] everything up in the
26:06 next three years? Oh, yeah. Next three
26:09 to five years are going to be brutal.
26:12 It's going to be
26:14 insane. So, pin is back. All right. I'm
26:18 a maintenance engineer. How do you see
26:20 AI changing what I
26:23 do when I have to be present to fix?
26:27 Well, right now, do I have is my
26:31 iPad charged at all? No. Hang on. Let me
26:34 charge my iPad and I'll I'll I'll
26:36 demonstrate it for
26:37 you. Um right now in Gemini and in chat
26:42 GPT both of them have
26:45 um live. So So basically you go into
26:50 your um you go into chat GPT, you put it
26:54 into voice mode and then you can turn on
26:57 the camera and you can do the same thing
26:59 in Gemini. If you go to the Gemini app
27:04 um and you turn on the camera and then
27:06 it can now see and it's connected with
27:09 chat GPT which knows everything and is
27:12 connected to the
27:13 internet. So let's
27:18 see and this is just one idea like this
27:20 is just an immediate idea that comes to
27:22 mind like what's going to what's going
27:24 to impact your you've got to be
27:26 physically there in person.
27:31 [Music]
27:35 The [ __ ] is going on
27:39 [Music]
27:46 there? Oh man. Come on. Hang on. My
27:50 machine was so dead. It it logged me out
27:51 of everything. I gotta I gotta do some
27:55 [ __ ] Hang on. Just I'll be there. Just
27:58 calm down. new pin while you fix that.
27:59 Okay. How do you leverage how to
28:02 leverage my skills mobile
28:04 engineer now with AI on the field? I
28:08 just got laid off due to
28:12 uh due to something due to Okay, Tik
28:15 Tok, fix your [ __ ]
28:18 interface. I just got laid off. Okay,
28:21 I'm a mobile engineer. I just got laid
28:23 off due to AI. Got it. Um
28:33 Hang on a
28:36 sec. I I'll tell you my favorite my
28:40 favorite app that someone built three
28:42 days ago. No, last
28:45 week
28:47 that complete sign on using your pass
28:50 key. Where's my pass key? Uh is it
28:59 Yes.
29:04 Done. I'm coming. Calm down. Calm down,
29:08 people. For the love of the
29:13 Lord. Continue with Google. Continue.
29:16 Come on.
29:28 Did it
29:32 work? I think it
29:36 worked. All right. This is not This is
29:38 not going well. Okay. So, we'll go to
29:39 mobile engineer first and then we'll go
29:41 to physical engineer second. So, we got
29:43 a bunch of engineers in here.
29:45 Awesome. And if you're wondering like
29:48 what are my qualifications? I have
29:51 none. I have none. I have a degree in
29:54 acting. I'm not an engineer. Um, but I
29:57 know
29:59 some. I've been an entrepreneur for 30
30:02 years. I was there for the early days of
30:04 the worldwide web. I started um
30:06 agency.com, which was one of the early
30:08 internet agencies. So, I built a lot of
30:10 the websites for the Fortune 500. So,
30:12 I've been around, right? I've seen a lot
30:14 of stuff. I've seen a technology like
30:17 the worldwide web take 30 years to
30:19 change everything. I think AI is going
30:21 to change everything in about five
30:22 years. And
30:24 and now if you're coming back at me, if
30:27 you're an engineer, you're probably
30:28 coming coming back at me going, "Well,
30:29 AI's been around for 50 years, young
30:32 man." Yeah, I know. The internet had
30:35 been around for seven decades or I don't
30:37 know, no, three or four decades. the the
30:40 foundation of it was earlier but you
30:42 know 70s Darpanet was late 60s early 70s
30:47 whatever it had been around for
30:50 decades and the worldwide web is the
30:54 thing that democratized it right because
30:56 it went from you had to do command line
30:58 [ __ ] which human beings don't do
31:00 engineers
31:01 do to being able to click it like this
31:04 right seems like a trivial difference
31:06 well this way it's relatively simple to
31:08 log in you just type words and then
31:11 click click well click click change the
31:14 world AI chat GPT generative AI starting
31:18 with chat GPT
31:21 primarily does for AI which has been
31:24 around for decades but the worldwide web
31:27 web did for the
31:29 internet so that's a little bit about my
31:32 background
31:33 um if you're a mobile engineer I'll tell
31:36 I'm going to tell you a really
31:38 stupid app that I think is going to
31:41 make, you know, $5 million in two months
31:44 and then never make another
31:48 penny.
31:50 Um, but but a couple of things I would
31:52 say. I would say get like insanely
31:56 curious about AI. Join communities like
31:59 the AI salon, which if you'd throw that
32:00 up, Brandon, um, go to the salon.ai and
32:04 say join our community.
32:08 Um, start vibe coding. I saw a great
32:12 post today. If you don't know what vibe
32:13 coding is, there are now apps out there
32:16 like lovable and replet agent and cursor
32:19 and all these things where you basically
32:20 just speak an app into existence and it
32:23 writes code for you and it p you makes
32:24 it publishable and stuff like that. Real
32:27 coders, engineers are mocking all of
32:30 those
32:31 people. and Rick Rubin who just came out
32:34 with a book on coding called the way of
32:36 code and it talks about vibe coding and
32:38 if you don't know who Rick Rubin is you
32:40 should look up Rick Rubin one of the
32:42 world's most
32:43 famous respected music producers that
32:46 doesn't write songs doesn't know how to
32:48 work the recording console doesn't sing
32:52 none of it he just has taste so anyway
32:56 he's a cool guy
32:58 um he did a thing today where he's like
33:01 okay it used to be if you wanted be a
33:02 real musician, you would go learn the
33:04 violin and you would join a symphony and
33:06 you would do your 10,000 hours. And then
33:09 in the 80s, punk rock came along and and
33:13 you only had to use, you know, two
33:14 fingers on two strings or one finger on
33:17 two strings and and and three chords and
33:20 you could be in a punk rock band. And he
33:23 said, "Vibe coding is the punk rock of
33:25 coding." And I love that idea. So
33:27 anyway, so I would get your ass vibe
33:29 coding if for no other reason understand
33:32 what it [ __ ] makes possible because
33:35 as a coder, as a mobile engineer, you're
33:38 going to understand what it does well
33:39 and what it doesn't. Which is going to
33:42 give you an advantage over a bunch of
33:44 clueless newbies that are like, I made a
33:47 program by just talking to it. And you
33:50 can go, yeah, but your security is [ __ ]
33:52 and I know how to fix that. Well, that's
33:54 a skill.
33:56 So, but what you're going to have to
33:58 give up to to experience vibe
34:03 coding is the ego of, well, I'm the best
34:06 coder in the world. Well, no, not
34:09 anymore. You might be better at some
34:11 things, but you're not as fast. And and
34:15 the the capability of these tools is
34:17 accelerating like nothing I've ever
34:19 seen. So anyway, so here's a stupid app
34:22 that someone came up with last week that
34:24 I think is going to make them a [ __ ]
34:27 fortune. And as a mobile developer,
34:30 you'll be like, "Well, that's relatively
34:32 trivial to make." I know. Which is why
34:34 it pisses me off that I didn't think of
34:37 it. You upload a picture of your dog.
34:41 Okay, this is an app. It probably costs
34:44 like five bucks on the app store. You
34:47 upload a picture of your dog and then
34:49 every day it emails
34:52 you a selfie photograph of your dog as
34:56 if your dog's traveling the world taking
34:58 selfies of
35:01 itself. So every day you get a dog
35:03 selfie of your dog, it's going to make a
35:06 [ __ ] fortune. Is it a novelty app?
35:08 Absolutely. Is he going to sell like,
35:11 you know, couple of million dollars
35:12 worth in two months? Probably. People
35:16 love
35:18 dogs.
35:21 So what what the acceleration of AI does
35:25 for for engineers in particular, I
35:27 think, is maybe you're really good at
35:29 engineering, but you're not good at
35:31 interface. Or you're really good at
35:32 interface and you're not good at
35:34 engineering. Or maybe you want to do
35:36 something with music and you're not
35:37 really a composer. Well, now you can
35:39 make songs, right?
35:44 Tik Tok pink. It's an AI rapper. Yeah.
35:46 But but so you know chat GPT is a large
35:50 language model rapper and large language
35:53 models are are you know math rappers
35:56 like
35:58 yes again you can put
36:02 down you can you can put down the fact
36:05 that that app is so simple and stupid
36:09 that it's trivial and it should be
36:11 mocked and made fun
36:13 of. But what's going to start
36:17 happening is
36:19 that coding is going to get so
36:22 democratized that
36:24 people who have no right creating
36:28 successful applications are going to
36:30 increasingly be creating successful
36:33 applications because precisely because
36:36 they don't know how to code. They're
36:39 gonna make
36:41 apps that don't follow any of the rules
36:45 and it is gonna piss you off if you know
36:48 the rules and you've been following the
36:50 [ __ ] rules for 30
36:52 years and some housewife from Pikipsy
36:56 comes along and makes a AAA video game
37:00 that becomes a hit. You're like, she
37:02 doesn't even know video games. How'd she
37:05 do that?
37:07 It's
37:09 like like get ready for that level of
37:12 dis
37:15 disruption and just because you
37:19 have feelings which listen I think are
37:22 appropriate like if you're pissed off at
37:24 this stuff if you're terrified of this
37:26 stuff if you like the mobile developer
37:29 just got laid off because of this stuff
37:32 be pissed off about it and [ __ ] jump
37:35 in and learn it. Learn what it makes
37:40 possible. Don't just sit on the
37:42 sidelines mocking people that are doing
37:44 rapper
37:46 apps. Do you know what just sold for $3
37:51 billion to Chat GPT to OpenAI?
37:55 Windsurf. It's a rapper app.
37:59 All they're doing is making calls to to
38:02 large language models to write code and
38:04 they're sort of packaging it in a clever
38:07 way. It's a rapper app, $3
38:11 billion. I think it was like a year and
38:13 a half
38:16 old. So,
38:19 yeah. Yeah. Exactly. Exactly. Hold on.
38:23 Off to make a selfie app of cats
38:25 traveling the world. Exactly.
38:28 Exactly. Go do that [ __ ] Okay, for the
38:30 other
38:31 engineer, do I have enough cord
38:35 here? Okay, here I am in chat GPT. Okay,
38:39 this is normal. Just chat
38:41 GPT. I might have an affiliate link.
38:44 What was the app called, by the way? I
38:46 don't [ __ ] remember. It pissed me off
38:48 so much. I saw it and I just I like as I
38:52 read the tweet, it was like upload
38:56 picture
38:58 And it was probably a 13-year-old of
39:00 probably, right? Upload your picture of
39:02 a dog and it sends you a selfie every
39:04 every day. My response in that moment
39:06 was
39:07 [ __ ]
39:11 [ __ ] just because I didn't think of it.
39:14 It is so stupid. I could have built that
39:16 app a year and a half ago and had my $5
39:19 million and been, you know, had an
39:22 actual studio rather than this shitty
39:23 home office. Which, by the way, if you
39:25 don't know the backstory, this is
39:26 actually a $20 million production
39:28 facility. We just make it look like it's
39:30 a shitty home
39:33 office. Anyway, okay. So, chat GPT. Now,
39:39 I'm going to go
39:41 into Hey Quinn, how's it going?
39:45 Hey, Kyle. I'm doing great, thanks. How
39:47 about you? Great. I'm going to turn on
39:48 the camera here and we're gonna you're
39:52 going to help me with some engineering.
39:53 Okay.
39:55 Absolutely. What kind of project are we
39:57 working on today? I don't know yet.
40:00 Um, let's see. Let's see. Oh, okay. Uh,
40:04 see that
40:06 thing? What is that?
40:09 It looks like a camera or some kind of
40:11 recording device on a stand. Are you
40:13 thinking about modifying it? No, it's a
40:14 gimbal.
40:16 Oh, got it. Are you looking to stabilize
40:19 your shots better or add some new
40:20 features to it? No, if but if I wanted
40:22 help troubleshooting the gimbal, could
40:25 you help me troubleshoot it?
40:28 Sure thing. What's the issue you're
40:29 running into with it? Um, I'm not an
40:32 engineer, but I want to play one on on
40:34 the internet. I want to I want to learn
40:36 how to do engineering.
40:38 Like here's a thing. See See this thing?
40:42 It's like a tripod thing with a phone
40:44 thing on it. See all that tape on it?
40:48 Yeah, I see it. Looks like you've got it
40:50 rigged up. Yeah, it's rigged. If I
40:53 wanted to fix that rather than buy a new
40:55 one and it it sort of snapped off at the
40:57 L joint, what what would I use to glue
40:59 it
41:01 to fix that joint? A strong epoxy glue
41:03 would be a good choice. Just make sure
41:05 the surfaces are clean and dry before
41:07 applying. Wait, but what's a good what's
41:10 a good epoxy? Like what's one strong
41:12 enough to actually hold it? Do you know
41:13 a brand or like is it a fast one, a slow
41:16 one? Is it like the the the yellow and
41:18 green one that turn, you know, turns
41:20 into clay? Like, which one should I use?
41:24 A two-part epoxy like JB Weld or Gorilla
41:27 Epoxy would be great. They're strong and
41:29 durable. The cure time varies, but So,
41:34 um, that's obviously if you're an
41:37 engineer, you're like, "Yeah, I'm
41:38 dealing with things bigger than duct
41:39 tape there, Bucky. I get it."
41:43 One of the things that you can do with
41:45 chat GPT is you can make things like
41:46 custom GPTs or if you work for a large
41:49 organization, they can create knowledge
41:53 bases that you can tap into, right? So
41:55 you could take all of the manuals of all
41:57 of the things and have them in your
42:00 knowledge base and then as you're
42:02 walking around, you know, point the
42:04 camera at things and it'll explain how
42:06 to do it. As we're as we're getting
42:10 farther down the road of augmented
42:12 reality, AI is going to allow you to
42:15 have, you know, those manuals are going
42:18 to be able to overlay instructions on
42:20 top of the actual device while you're
42:22 looking at it hands-free. I don't
42:24 remember what the name of that app was.
42:27 Um anyway, um I made my own
42:30 notetaker. So, no more pain uh for
42:33 fireflies otter. It's amazing. Wow.
42:35 There you go. See? See, there you have
42:38 it. All right, that was crazy. That was
42:40 a First of all, I don't know if the
42:43 person that that originally trolled and
42:46 fired me off um is still in here. I hope
42:48 you are. Um but thank you for for making
42:52 a comment
42:54 that let me know that there there were
42:57 some people in here who are not sure
42:59 about this
43:01 [ __ ] I get it. Listen, I have a degree
43:04 in [ __ ] acting. One of the very first
43:07 seminars that I gave about AI was to a
43:09 group of voice actors two years
43:12 ago and they got on the call and they
43:14 were like, "My my job's going to be
43:16 okay, isn't it?" And I was like,
43:19 "Uh, you should probably learn what's
43:21 coming." And actually, one of the things
43:23 I'm going to demonstrate
43:25 tonight is one of the reasons that
43:28 they're probably not working that much
43:32 and and they're probably not going to
43:33 work for much longer.
43:37 Every profession, every single
43:40 profession, whether you're an accountant
43:42 or a filmmaker or a writer or an art
43:47 director or a creative director, all of
43:50 it is going to
43:51 change. All of it. All of it. All of
43:55 it. No
43:59 exceptions. And then if you're like,
44:01 "Well, you seem excited about it. Why
44:03 are you excited about it? That sounds
44:04 [ __ ] awful."
44:08 Well, here's here's my here's my
44:18 thought. In the in the late
44:22 1800s, agriculture was all the thing. It
44:25 was all the
44:26 rage. We had figured out farming at
44:29 scale. Everyone had their horses and
44:32 their plows and their oxen.
44:36 And some [ __ ] [ __ ] probably from
44:39 Silicon Valley, tech bro from Silicon
44:42 Valley in like
44:46 1842, invents this stupid little thing
44:48 called the steam
44:51 engine. You you boil some water and it
44:55 makes pressure and you turn a thing. It
44:58 was noisy and it was loud and it was
45:00 stinky.
45:05 And then someone else went,
45:08 "Huh, wonder if I took that steam engine
45:11 and took this leather
45:15 belt, put it to the shaft, and stuck it
45:18 to the
45:19 wheel.
45:21 Huh? I could make like a a rolling thing
45:25 that I could roll into the fields
45:27 without having to have horses.
45:30 Well, let's test how strong this thing
45:32 is. Huh? It's as strong as like 20
45:38 horses. We still use horsepower today.
45:43 That [ __ ] [ __ ] from Silicon Valley
45:46 or the equivalent of it back then that
45:48 invented the steam
45:52 engine totally
45:55 destroyed the farm economy of the late
45:58 19th
45:59 century. Destroyed
46:02 it. And if you were someone who rose,
46:05 you know, raised horses or trained them
46:08 or created feed for them or built plows
46:13 that were designed to be pulled by
46:17 them, you looked at that [ __ ] thing
46:19 and you're like, "Well, [ __ ] that. It's
46:21 going to ruin my job."
46:25 Now, what no one could imagine then was
46:28 that all those displaced farm
46:31 workers left farm country because there
46:34 wasn't
46:35 work, right? They ate the horses or
46:38 turned them into glue, whatever it was.
46:41 They kept the fancy ones. The rich
46:42 people kept the fancy ones and they ran
46:44 them around
46:45 tracks. And all those people moved out
46:48 of farms and into the
46:51 cities and they invented the service
46:54 economy and they invented you know
46:56 factory work and right like entirely new
47:00 kinds of work that couldn't have been
47:02 imagined emerged out of that.
47:07 So what the steam engine did, what the
47:08 industrial revolution did was it
47:10 augmented human strength, our physical
47:15 strength. What AI is doing right now is
47:18 it's augmenting our
47:21 intellect. And why we're all so pissed
47:24 off is like we're the intellectual
47:27 equivalents of like the strong ass farm
47:29 hands in the 1800s.
47:32 Me like horses, me like
47:35 plow, right? And then the steam engine
47:38 comes along and there's like some skinny
47:40 guy riding the tractor. You're like, I
47:42 could destroy him. Well, yeah, except
47:45 his little machine is like 20 of your
47:47 horses and like 40 of you, you pin
47:53 head. Same thing's happening with our
47:55 intellect. But we have to let go
47:59 of us being the center of the equation.
48:04 Just like physical strength gave way to
48:08 machine
48:09 strength, physical intelligence,
48:12 intellect is going to give way to AI
48:15 intelligence, to artificial
48:18 intelligence. And it's going to just do
48:20 things different. And it's going to make
48:21 different things possible. Here's a
48:23 thing that it does that I find
48:24 tremendously exciting. It might stress
48:26 you the [ __ ] out,
48:29 but let's say you're really good at the
48:31 one thing you
48:33 do. Let's say it's a particular kind of
48:36 programming or it's a particular kind of
48:37 engineering. You got really good
48:38 expertise in
48:40 that. You now have capabilities in
48:45 things you never could have
48:47 imagined. Maybe you're not good at
48:49 poetry. You're good at code, but you're
48:50 not good at poetry. Well, now you're
48:52 good at poetry and now you're good at
48:54 creative writing and now you're good at
48:55 market marketing and now you're good at
48:58 illustrating and now you're good at
48:59 music and now you're good at film
49:01 making. Now, are you going to be as good
49:03 at film making as a filmmaker is at film
49:06 making? No. But are you going to be able
49:09 to make something that's going to make
49:10 you [ __ ] happier than [ __ ] and and
49:13 maybe realize I always wanted to tell
49:15 stories, but I'm really good at this
49:17 coding stuff, but now I can take my
49:19 coding skills and I can prompt this
49:23 thing in the logical way that I think
49:25 and out comes these magical movies that
49:27 I can now tell stories with. That's one
49:30 of the things that we experience here
49:32 night after night after night after
49:35 night is people rediscovering their
49:38 passion for [ __ ] they forgot they were
49:41 excited about they given up the dream
49:45 and all of a sudden they're like [ __ ]
49:47 maybe I can do it
49:50 now. I mean that's [ __ ]
49:54 insane. Anyway, all right, enough comedy
49:59 jokes. You can just say my name, Kyle.
50:02 Pate, what's happening,
50:05 Pate? Pate's, you know,
50:09 deep in what he does working at Google.
50:14 If he chose to, he could be doing
50:16 poetry, too. Kids
50:20 books. Okay, a few spotlights. Let's see
50:23 what we got going over
50:26 here. Some LinkedIn things. What is
50:29 going on? Hide
50:31 that. Okay. Start comment. Let's see.
50:35 I'm glad I was the first to say that
50:37 Kyle is the Rick Rubin of AI. Kyle
50:40 Rubin. Well, I you know what's funny is
50:43 I've been I've been talking up Rick
50:44 Rubin for a while, but I didn't know he
50:46 was into coding. I didn't know he was
50:48 into vibe
50:49 coding. And then he started posting on X
50:52 all these sort of code and AI
50:54 references. And I'm like, is that really
50:56 Rick Rubin? And it is. And he just he
50:58 just launched the way of code. Let's go
50:59 look at the way of
51:01 code. Um because it's it's super cool.
51:03 And there's some people in here who are
51:04 not
51:06 um well they they may have left. I may
51:09 have chased them
51:11 away. Um the wayofcode.com. So go to the
51:19 wayofcode.com
51:21 and what you're going to
51:26 see is a very very elegant very simple
51:29 website. This site very much very much
51:34 reminds me of of sites from like 1995
51:39 9697 where web developers no one had
51:42 figured anything out yet. So everyone
51:44 was experimenting. They were
51:45 experimenting with layout and design and
51:48 interactivity and how you could
51:51 incorporate it. And it was everything
51:52 was experimental. And this very much
51:55 feels like a site from that era
51:58 era. And what it is is it's 82 pages.
52:03 It's it's based on the um the the loud
52:07 sue or on on loud sue and uh wait black
52:12 bar. Hang on. And
52:14 pin. My friend lost her job and wrote 12
52:18 books and has recently applied for a
52:20 patent of a medical
52:23 device. There you
52:26 go. My friend was laid off. Listen to
52:29 that sentence in a in a Tik Tok comment
52:31 from Joy Party. My friend was laid off
52:34 and has written 12 books and applied for
52:37 a patent for a medical device.
52:43 That's what AI makes possible today.
52:47 Like, it's not even good right now. It's
52:49 getting there, but it's still jankier
52:52 than [ __ ] Anyway, back to this
52:56 thing. So, he he got inspired by this
52:59 work, someone who, you know, who's
53:01 inspired him for his whole life, and he
53:03 created this book, The Way of Code.
53:05 Every one of these
53:07 pages has a
53:09 poem and a piece of art of coded art of
53:15 animated interactive sometimes. Where is
53:18 my cursor? Um, coded
53:23 art. And Rick Rubin
53:26 vibecoded every one of these pieces of
53:30 art. Every one of
53:33 them. He vibecoded them on Claude.
53:41 Claude.ai. Oh yeah. Hey Khaki, if you're
53:44 still here, what number do you remember
53:46 what page the original art was that you
53:49 vibe coded that you did the remix
53:54 from? And so what happens here is this
53:58 goes one level deeper. So, not only are
54:00 these pieces of artwork cool and
54:02 interactive and beautiful, they all have
54:04 this kind of zen kind of aesthetic to
54:07 them. If you roll over the bottom of
54:09 them, there's this little button that
54:11 says modify page one. Okay,
54:15 great. So, khaki
54:19 k went to this page and there's this
54:22 little button here that says
54:24 modify. If you click it, it's a little
54:27 disconcerting. It takes you
54:31 away. Oh, what if it was
54:37 weightless? And it allows you to modify
54:40 that artifact that Rick Rubin created.
54:42 And so, let me show you the one that she
54:47 did. What if code was weightless is the
54:50 name of of that
54:52 thing. And then
54:56 She
54:57 made this variation of
55:01 it. What if it was weightless? Would it
55:04 drift, expand, dissolve, or transform?
55:07 Now what happens when structure is
55:08 released from gravity? Now pause. Let
55:12 your gaze rest on the image. Not to
55:14 analyze, but to arrive. There's no need
55:16 to act. No need to fix or solve. And
55:18 then if I
55:20 click thing makes music. Hang on. Let
55:23 me change how I'm
55:26 sharing. I got myself all horse. I was
55:29 on a rant. That was a good
55:31 [Laughter]
55:41 [Music]
55:42 one. We made up for taking off Monday.
55:45 Exactly.
55:47 So this is a variation of the way of
55:51 code thing that we were just looking
55:53 at of this one. Oops, I have to change
55:57 my share
56:01 here. And so for every one of them, you
56:04 can remix
56:05 them
56:07 and you can then publish them and then
56:11 you can then submit them to the project.
56:14 So, this is not only a book, not only an
56:16 interactive art installation, it's a
56:19 co-creation space where you can actually
56:23 collaborate in a timeshifted weirdass
56:26 way with Rick
56:27 Rubin.
56:30 Crazy. Crazy,
56:33 man.
56:37 Beautiful. And this is the kind of stuff
56:40 like this is very sort of edge, Casey.
56:42 No, very few people are going to pay
56:44 attention to
56:47 this. But if you start looking at what
56:49 this
56:51 is and if you start looking at the fact
56:53 that people from this channel because I
56:55 showed this the other
56:57 night, paid sponsor. Yeah, I
57:02 wish because I showed this the other
57:04 night and the people on this channel are
57:06 generous and curious and willing to
57:08 experiment and willing to share what
57:10 they learned. It's like the philosophy
57:12 of the of the AI salon is play
57:15 first, mindfully create, generously
57:18 lead. And generously lead looks like
57:21 learn out loud. Come here, play with
57:24 this, go mix something up, share it with
57:27 the salon, share it with the other
57:29 people. Because in doing that, you
57:31 establish yourself as a leader, but you
57:33 also bring people along and get them
57:35 inspired. The reason the salon has 3,000
57:38 AI optimists in it that are just playing
57:41 and exploring and supporting one
57:45 another is because people are willing to
57:48 be
57:50 vulnerable and and share something that
57:53 they're like, I I don't really know how
57:54 to code, but I did
57:56 this. I don't know really how to make
57:58 movies, but I did this.
58:01 I'm not good at marketing, but here's a
58:03 whole [ __ ] sophisticated marketing
58:05 campaign that kicks the [ __ ] out of what
58:07 the agency gave
58:11 me. Oh
58:15 man. I'm just curious how much Rick was
58:18 actually involved in this. I think he
58:20 was decently involved. Like I said, he's
58:22 been he's been tweeting for the past
58:24 four or five months a lot about AI. Um,
58:27 and he just put out this video. Let let
58:29 me find the punk rock one because it's
58:31 it's really
58:33 good. We'll watch a little movie
58:35 together, a little interview, a little
58:38 portion of an interview.
58:41 Rick
58:45 Ruben vibe
58:48 punk tabs. Yes. Yes. What you said. Here
58:53 it is. This is such a good thing.
58:56 Whoever filmed this had [ __ ]
58:57 fingerprints on their on their camera
59:02 lens.
59:04 H amateurs. Amateurs, I tell you. All
59:09 right. Conservatory and study. So, in
59:13 the past for music, you had to go to the
59:14 conservatory and study years and years
59:17 and then someday you could play in a
59:19 symphony. And then when punk rock came
59:23 along, you could maybe learn three
59:25 chords in a day. Then there were all
59:29 these bands and that made it everybody.
59:31 And that was how I started in music was
59:32 punk rock. If you had something to say,
59:36 you could say it. You didn't need the
59:40 the uh expert. That's
59:43 important. If you had something to say,
59:46 you could say
59:48 it, right? Three chords in a dream.
59:51 Three chords in the truth. Right. Punk
59:53 rock is just country with electric
59:55 guitars. Three chords in the truth. If
59:58 you had something to say, you could say
1:00:01 it. Punk rock
1:00:04 democratized musical
1:00:07 expression tease or skill set other than
1:00:12 your idea and your ability to convey it.
1:00:15 And wide coding is the same thing. It's
1:00:17 it's the punk rock of uh of coding. I
1:00:22 think the biggest uh disconnect that
1:00:24 that I feel myself is that it's such a
1:00:26 strong tool that can do so much.
1:00:30 We need some examples of some of the
1:00:32 different things it can do. Now, we'll
1:00:34 see it can make uh it can make animation
1:00:37 that looks like your favorite cartoon
1:00:39 and then you see a million people doing
1:00:41 that. Yeah, that's one idea. I want to
1:00:45 see all of the things it could do to
1:00:47 understand what's possible instead of
1:00:50 just I'm going to get it to do the same
1:00:52 thing everyone else is getting it to do.
1:00:54 I think it's beyond our scope to
1:00:56 understand what it actually can do. And
1:00:59 I'm looking that's important too. I'll
1:01:03 play that again.
1:01:06 Is getting it to do. I think it's beyond
1:01:09 our scope to understand what it actually
1:01:11 can do. It's beyond our scope to
1:01:14 understand what it can actually
1:01:16 do. One of the reasons I say in here
1:01:19 that there are no experts. I'm not an
1:01:22 expert. Anyone who tells you there's an
1:01:23 exp they're an expert at this stuff is
1:01:25 full of [ __ ] I don't think we have a
1:01:29 clue what these tools make possible. I
1:01:32 don't think we'll have a clue what these
1:01:34 tools really make possible for probably
1:01:36 10 or 20
1:01:39 years because it is so far
1:01:43 beyond what any of us can do
1:01:46 individually.
1:01:48 But the purpose of coming to lives like
1:01:53 this, the purpose of joining groups like
1:01:55 the AI salon or she leads AI or or
1:01:58 communities that are focused on
1:02:00 exploring AI. The reason to do that is
1:02:04 that the more people that you're around
1:02:06 that are curious and exploring
1:02:08 AI, they're going to teach you what this
1:02:11 stuff makes possible in ways that you
1:02:13 could never learn on your own. Because I
1:02:16 guarantee you the way Jim Ross thinks
1:02:18 about AI in relation to his self-
1:02:21 storage business, self- storage
1:02:23 management
1:02:24 business is going to be completely
1:02:27 different than anything you're thinking
1:02:30 about. And then when he shows what he
1:02:32 did to him, what he did with AI is this
1:02:37 humble little, oh, I just did this
1:02:40 thing, but someone else sees it and
1:02:42 they're like, oh my god, I didn't know
1:02:44 you could do that with it because they'd
1:02:46 never thought about the problem that he
1:02:48 was solving because they're in a
1:02:50 completely different sphere.
1:02:53 When we're at the early stages, we're at
1:02:55 the p primordial soup stage of
1:02:58 AI. There are thousands of tools, many
1:03:01 of them
1:03:02 rappers, GPT rappers. Thousands of
1:03:06 them. Eight are going to survive, right?
1:03:10 We're in the primordial soup of just
1:03:12 bloop bloop
1:03:14 biology just spitting out organism after
1:03:17 organism after
1:03:20 organism. Just spitting them out.
1:03:23 What do those organisms do? I don't
1:03:25 know. We don't know. You can't possibly
1:03:28 learn them all. And by the time you
1:03:31 learn all of them, when you go back to
1:03:32 the first one, it's completely different
1:03:34 or
1:03:36 gone. So, so that's an impossibility. So
1:03:39 the only thing you can
1:03:41 do is figure out your point of view,
1:03:45 your mission, your how you want to
1:03:48 affect people, how you want to change
1:03:50 the world, how you want to solve a
1:03:52 problem, and then just go learn these
1:03:54 tools well enough to go play in that
1:03:56 space and get your ass in a
1:04:00 community. Tom Nodler, biology, natural
1:04:03 selection, survival of the fittest. Now
1:04:05 you're talking my language, Kyle. There
1:04:07 we go.
1:04:09 Tik Tok pin. I worked on my movie All
1:04:12 Day with AI, King Arthur. This
1:04:14 afternoon, I created a little short
1:04:17 about Christmas. Perfect.
1:04:21 Beautiful.
1:04:24 Beautiful. Next Tuesday, um, at the AI
1:04:28 salon, we've got, um, so the community
1:04:30 that I said you should get your ass in,
1:04:32 it's at if you go to the
1:04:34 salon.ai and join. Click on the button.
1:04:37 you scroll down and say join our
1:04:39 community. Join it and then go to events
1:04:41 and next
1:04:43 Tuesday, June the 5th, maybe 3rd, I
1:04:46 don't
1:04:48 know. Uh we're having Danny Kravitz who
1:04:51 teaches uh screenwriting and creative
1:04:54 writing at the University of Chicago, I
1:04:58 think, or University of Illinois.
1:05:01 Chicago, I don't know, somewhere in
1:05:02 Chicago. He's teaching people about
1:05:04 storytelling and creativity and he's
1:05:06 written movies. He's a produced
1:05:10 screenwriter and he's going to come talk
1:05:12 about
1:05:15 creativity
1:05:19 and
1:05:22 opening yourself up to the
1:05:24 possibilities of what's there. Let me
1:05:26 tell you another thing again. I don't
1:05:29 Are any is anyone still here?
1:05:32 When I go on these rants, I assume just
1:05:34 everyone just [ __ ] leaves. They're
1:05:35 like, "I'm
1:05:36 [Laughter]
1:05:39 good.
1:05:43 Um, oh, where was I going to go? I lost
1:05:46 my train of
1:05:49 thought." Danny Kravitz,
1:05:53 possibilities. I lost it.
1:05:56 Whatever. Whatever. We'll make more. Oh,
1:05:59 this was a thing.
1:06:04 One of the reasons I'm so bullish on
1:06:08 AI and why I think it's going to make
1:06:10 the world a dramatically better
1:06:16 place is that we're all
1:06:20 human and one of the things that humans
1:06:23 do I've been I've been on the planet for
1:06:25 a while. I've been to a lot of self-help
1:06:27 sort of things. And one of the things
1:06:29 you learn in in a lot of those self-help
1:06:33 programs is
1:06:35 that at some point in our
1:06:40 childhood, we
1:06:44 learn that the world is a dangerous
1:06:46 place.
1:06:50 Maybe you put on the pink tutu and
1:06:53 you're a little boy and your stepdaddy
1:06:55 called you a
1:06:56 [ __ ]
1:06:58 Oh, world's not safe. Maybe you were
1:07:02 making a picture on
1:07:03 the dining room table and your mom said,
1:07:06 "That's a shitty drawing. You have no
1:07:08 talent. Oh, world's a dangerous place."
1:07:13 And at some point in our childhoods, we
1:07:15 shut down a part of
1:07:18 us. We shut down a dream. We shut down a
1:07:22 possibility. We
1:07:24 define who we
1:07:28 are. I I've got one I can remember like
1:07:31 it was [ __ ] yesterday.
1:07:36 And I have designed my life to
1:07:40 compensate for this one
1:07:44 moment.
1:07:46 So my parents got divorced when I was
1:07:48 young. And when I was
1:07:51 in third grade going into fourth grade,
1:07:54 I said I wanted to go live with my dad.
1:07:56 He was at Scott Air Force Base in
1:07:58 Illinois. And I was going I was going to
1:08:00 go into Mosscuda Elementary School.
1:08:03 Mscuda,
1:08:06 Illinois. And my mom didn't have it
1:08:08 together and she didn't have my records,
1:08:11 my school records. So they had to give
1:08:13 me like the test, whatever the [ __ ]
1:08:16 test was, the entrance test for a
1:08:19 [ __ ] fourth grader, right? And so I
1:08:22 take the test and one of the
1:08:24 tests was this grid of
1:08:28 like five squares by five squares, like
1:08:30 a 25 square grid. And you had colored
1:08:33 pencils and you were supposed to color
1:08:34 in the
1:08:36 grid. This was the late 70s and my mom
1:08:39 was really into bargello which was a
1:08:42 geometric needle point
1:08:45 thing technique whatever
1:08:49 practice style I don't [ __ ]
1:08:52 know. And so I looked at the grid and I
1:08:55 and I saw like her bargello things in my
1:08:58 head. And so I made this really cool
1:09:00 geometric pattern and I was really proud
1:09:02 of it. And so then I was I was in the
1:09:05 counselor's office where they were like
1:09:07 looking over my test like the the adult
1:09:10 who was monitoring the test was looking
1:09:13 at it. And at one point they got called
1:09:15 out of the
1:09:17 room and they left the answer book
1:09:21 open and I'm sitting there by myself and
1:09:23 the answer book's open and then it was
1:09:26 the the upper left page of the answer
1:09:30 book was the description of the grid the
1:09:34 grid drawing and it said if the child's
1:09:38 drawing looks like this and it looked
1:09:41 exactly like mine it was geome
1:09:45 metric. It said it's a sign of a less
1:09:49 creative
1:09:51 mind than a child's who looks like this,
1:09:54 which was chaos. It was just colors all
1:09:56 over the
1:09:59 place. And in that moment, my heart
1:10:03 sunk because what that book told me was
1:10:06 that I wasn't
1:10:08 creative. And I knew in that
1:10:11 moment that I wasn't creative. And it
1:10:14 broke my [ __ ] heart because I was so
1:10:16 proud of that
1:10:20 thing. And I spent my life trying to
1:10:25 compensate for not being
1:10:27 creative by being overly
1:10:31 creative. But as you learn in these
1:10:35 programs, that overcompensation is
1:10:37 hollow because every time you succeed at
1:10:41 being
1:10:42 creative, all it does is remind you of
1:10:45 who you really are, the non-creative
1:10:49 kid. And then what you learn at some
1:10:52 point
1:10:53 is you made a choice back then that
1:10:57 you're not creative.
1:10:59 But the nice thing about a choice is you
1:11:01 can take the other
1:11:04 path and you can be creative. You can be
1:11:07 an artist. And so one of the things that
1:11:10 AI
1:11:11 does that just [ __ ] lights me on fire
1:11:15 is it reintroduces human beings to that
1:11:20 moment where they told themselves they
1:11:23 couldn't do this. They weren't smart
1:11:24 enough to do this. They didn't go to the
1:11:26 right school for this. They can't do
1:11:28 math well enough to do this. The thing
1:11:30 that they really wanted to do that got
1:11:32 [ __ ] snuffed out when they were a
1:11:36 kid. And they're back like a [ __ ]
1:11:39 seven-year-old like, "Oh my god, I can
1:11:41 do it
1:11:43 now. It's
1:11:46 insane. It's
1:11:50 crazy." And yeah, it was probably
1:11:52 trained on some living artist's artwork
1:11:54 and then they're not compensating him
1:11:56 for it.
1:11:59 Both things can be
1:12:01 true. Ah, people with AD AD are the most
1:12:05 creative. Look how many famous people
1:12:07 have it. Nice. Precisely. I've always
1:12:10 told my three sons that they were the
1:12:12 greatest. They turned out great.
1:12:15 Amazingly. Yeah.
1:12:17 Exactly. Exactly. The person who
1:12:20 developed that test was not creative.
1:12:23 Exactly. Exactly. like, you know, I
1:12:26 think back on it
1:12:27 now and it's like I get what they were
1:12:30 trying to do, but it was just the way it
1:12:32 was
1:12:33 phrased and it was the fact that I had I
1:12:36 was so proud of it, right? It was like I
1:12:38 knew I did really good. I kept within
1:12:41 the lines because normally I'm all over
1:12:42 the [ __ ]
1:12:45 place. So
1:12:50 anyway, so Oh, he teaches at Columbia.
1:12:54 Okay, cool. Danny Kravitz. Yeah. So,
1:12:56 next Tuesday from 5:00 to 7:00 p.m. AI
1:12:59 Salon presents. We're going to have
1:13:01 Danny Kravitz and he's going to talk
1:13:02 about um creativity and unlocking your
1:13:06 creativity. He's got a whole methodology
1:13:07 for
1:13:08 it. Um and you should
1:13:11 come. One of the things I've long
1:13:13 believed, you know, when I when I was
1:13:15 the chief creative officer of
1:13:19 agency.com, I remember we we had a
1:13:22 creative group. We had a project
1:13:24 management group. We had an account
1:13:26 group. We had builders, website
1:13:30 builders, technologists,
1:13:32 engineers. And I remember there was a a
1:13:35 rift that came into the
1:13:38 company. And it was when the
1:13:41 creative
1:13:44 group was starting to get a lot of
1:13:46 attention. They were starting to do work
1:13:48 that was people were talking about and
1:13:50 they started to get egos to match.
1:13:53 They're like, "Well, we're
1:13:55 creative." And when they said, "We're
1:13:57 creative." Right? They're part of the
1:13:59 creative group and they create these
1:14:01 creative products. And they walked
1:14:04 around the office going, "We're
1:14:05 creative." And it got more and more and
1:14:08 more
1:14:09 pretentious. And it started to make all
1:14:11 the other people who were not in the
1:14:12 creative group resent the [ __ ] out of
1:14:15 them.
1:14:18 Why? Because project managers can solve
1:14:23 problems in very creative ways. Account
1:14:26 managers can be creative in how they
1:14:27 manage the relationship. Builders can be
1:14:30 creative in how they architect a
1:14:32 technology or how they build
1:14:35 code. And what struck me was that
1:14:38 there's actually there's actually a
1:14:39 difference between creative
1:14:42 product and creative process.
1:14:45 Everyone can be
1:14:48 creative. You just might be being
1:14:50 creative and not be creating a creative
1:14:53 product. You might be being creative in
1:14:56 how you create formulas in a spreadsheet
1:14:59 that's [ __ ] elegant and beautiful and
1:15:02 every bit as creative as someone that
1:15:05 did something over here that looks like
1:15:06 a
1:15:09 cartoon. And so again, what's that got
1:15:12 to do with AI? Well, AI lets you take
1:15:15 any idea you have and express it in all
1:15:19 these different
1:15:20 outputs. You get to be creative across
1:15:24 disciplines. It's [ __ ]
1:15:28 magical.
1:15:29 Magical. All
1:15:33 right. I don't even know what the [ __ ]
1:15:35 we were talking about.
1:15:44 Ah, all right. Yeah, let's go do the
1:15:46 Hume thing because the Hume thing's kind
1:15:48 of
1:15:50 cool. Or do we want to do Korea? What
1:15:52 time is it? It's 9:18. Nah, we'll do
1:15:54 we'll do Korea tomorrow. Let's do Korea
1:15:58 tomorrow. Um, all right. So, I got to
1:16:02 switch up my my what you call
1:16:05 it. Where is him? There it is. Okay.
1:16:16 Are any of my AI haters still in the
1:16:21 house? Did I convince anyone to at least
1:16:23 give it a [ __ ]
1:16:27 shot? Anyone? Is anyone still here?
1:16:31 Hello. Hello. All of my songs are on
1:16:34 Soundcloud. My book is on Amazon.
1:16:38 You can do anything you want to do.
1:16:41 Exactly.
1:16:46 Exactly. I I'm here and right now I hate
1:16:49 AI. That's Pate. Listen, here's the
1:16:52 deal. I
1:16:54 1,000% respect Pate and his opinion on
1:16:57 AI. And some days he [ __ ] hates it.
1:16:59 Some days he actually Well, I don't know
1:17:01 that he ever loves it. Some days he
1:17:03 thinks it's quite useful, which for Pate
1:17:05 that's like [ __ ] gargantuan love.
1:17:08 Here's the thing about Pate and and just
1:17:10 anyone like
1:17:12 that. I don't If you've got an opinion
1:17:15 about AI and you know it deeply and you
1:17:18 explore the [ __ ] out of it and you still
1:17:19 hate it and you still take a stand
1:17:21 against it, total respect for that. Like
1:17:25 1,000% respect for that. The thing that
1:17:27 grinds my [ __ ] gears is people that
1:17:31 assume AI is something that it's not and
1:17:34 have an opinion based on nothing. That
1:17:37 drives me [ __ ]
1:17:39 baddy because it's only going to hurt
1:17:41 them. All right.
1:17:45 Anyway, so today, if you don't know
1:17:47 Hume, Hume's been around for, I don't
1:17:49 know, three or four years. I've known
1:17:51 about them for, I don't know, two and a
1:17:52 half, three years.
1:17:54 They've always had a very interesting
1:17:57 kind of product where it's a it's a text
1:18:02 to voice and voiceto text synthesis
1:18:06 model. And one of the things that
1:18:08 they've always done is as you talk, it's
1:18:11 measuring
1:18:12 emotional attributes of your voice in
1:18:15 real time. So, is he angry? Is he happy?
1:18:18 Is he curious? Is he, you know,
1:18:20 annoying? Yes, he's annoying. um and
1:18:23 they're sort of measuring it in real
1:18:25 time. And then when their voice responds
1:18:27 back to you, it's it's responding in
1:18:30 kind, right? So if you're really angry,
1:18:32 it it either will get angry with you or
1:18:34 maybe it decides it needs to get calm,
1:18:36 but it it will respond to the the
1:18:38 emotional content of your voice as much
1:18:41 as the words. And so they've been, you
1:18:45 know, doing different versions and
1:18:47 things like this. So the new version
1:18:49 that they just launched today is called
1:18:50 EV3. And you can design a voice. And you
1:18:53 can design a voice right now in 11 Labs.
1:18:56 You can sort of do a little bit of voice
1:18:58 design in Cartisia. I'm sure there's a
1:19:00 couple of others that you can do it. Um
1:19:03 11 Labs is the most famous. What what
1:19:05 they did here that's really clever is
1:19:06 you actually just talk
1:19:08 to an agent, a
1:19:12 voice that guides you through designing
1:19:14 another voice. So So let's go design
1:19:17 another voice. All
1:19:18 right. I'm the coal shoveler at the
1:19:21 bottom of the Titanic. You all enjoy
1:19:23 your
1:19:24 sandwiches. Okay, Tate, you're the
1:19:27 [Laughter]
1:19:29 best. Someone's got to shovel the coal.
1:19:32 It's true,
1:19:33 man. Oh my god, that's awesome. Okay, I
1:19:38 don't think my accent I think my accent
1:19:39 would confuse the hell out of Hume. It
1:19:41 might not. It's worth trying it. By the
1:19:42 way, I'm I'm at Hume AI and then
1:19:44 introducing EV3 or EV3. The next
1:19:48 iteration of I got to change glasses so
1:19:50 I can actually read the screen. The next
1:19:52 iteration of um voice interface from
1:19:56 Humei is here. Okay, so we'll create a
1:19:58 voice that matches your personality and
1:20:01 style. This is cool.
1:20:07 Hey there, welcome. I'm E3 from Hume AI.
1:20:12 Think of me as your personal guide to
1:20:14 creating a custom AI voice and
1:20:16 personality. Something designed exactly
1:20:18 around your preferences. So tell me, do
1:20:22 you have a really strong idea? Sounds
1:20:24 like crap on YouTube. Hang on, let me
1:20:26 change tabs. Maybe some Maybe I did
1:20:28 something
1:20:37 wrong. Hey there. Welcome to EV3. I'm
1:20:41 here to help you create your Okay. It's
1:20:44 Steve Carter. No, no, I just said, "Is
1:20:47 it still bad?"
1:20:50 Okay. Okay. All right. Okay. Oh, well,
1:20:54 tough. This is how it is. Sorry. Yeah.
1:20:57 Weird. All right. Uh I'll I won't do
1:20:59 this that long then. Um Okay. So, I want
1:21:01 to design a voice with you. I want to
1:21:03 design a voice that's kind of like Yoda,
1:21:07 kind of like Boba Fett, and kind of like
1:21:12 um Joe Peshy if they all mix together.
1:21:17 Okay. Steve Carter, Yoda, Boba Fett, and
1:21:21 Joe Peshy. Yeah, but not Steve Carter. I
1:21:23 don't know who that is. You just
1:21:24 misheard me about something else.
1:21:28 Okay, Steve Carter. I have what I need
1:21:30 for now. Hang tight and we'll be back in
1:21:33 touch soon to get started on creating
1:21:35 that unique voice just for you.
1:21:41 All right, they're off making the voice
1:21:43 now. Experiencing high load. So, uh,
1:21:48 you're here, right? Well, hello there.
1:21:52 Name's Well, they call me EV3. Built by
1:21:55 Hume AI, apparently. Yeah, but can you
1:21:58 say some I want you to speak as if
1:22:00 you're Yoda. Just everything you say, I
1:22:02 want you to say it with Yoda grammar.
1:22:06 So, you wish for me to speak like this
1:22:09 with the words all jumbled around? Yes.
1:22:12 An interesting path you lead. All right,
1:22:14 I shall try. What is it you want to
1:22:16 discuss? H I want to discuss turtles and
1:22:20 why they're awesome.
1:22:22 Turtles, you say? H powerful little
1:22:25 creatures. They are slower than a
1:22:28 speeder perhaps, but tenacious. So
1:22:31 anyway, there you go. So, it's
1:22:33 relatively easy. You just basically talk
1:22:35 a voice into existence. Now, there's all
1:22:38 sorts of other ways you can tweak things
1:22:39 here. I don't know exactly what you can
1:22:41 do on the site. I haven't been on it in
1:22:42 a while. Um, but I just thought that was
1:22:44 cle clever. I thought it was something
1:22:46 worth uh worth showing. Um,
1:22:53 [Music]
1:23:00 Tabs and Perplexity Labs or
1:23:05 Korea. Let's go look at I want to look
1:23:07 at Korea context.
1:23:09 Okay. So, if you don't know Korea, Korea
1:23:15 AI. All right. There's
1:23:22 There there are a number of sites out
1:23:24 there that
1:23:26 really really work hard at giving you
1:23:30 access to lots of different
1:23:33 models. Like on the large language model
1:23:35 side, there's a site called PO POE. If
1:23:38 you go to po.com, it lets you play with
1:23:40 all sorts of different models, different
1:23:42 video models and LLMs and things like
1:23:44 that. Korea does that for video and
1:23:47 images. There's lots and lots of things
1:23:49 that you can do here. Lots.
1:23:53 Um, as someone said earlier, h it's a
1:23:56 wrapper app. They they may they may have
1:23:58 their own models, but but they're doing
1:24:00 a lot of interface wrapping around API
1:24:03 calls to open- source models and
1:24:06 commercial models that you know that
1:24:07 they're they're you're paying for access
1:24:09 to multiple models. So, it's a it's a
1:24:12 really interesting site. It's an
1:24:13 interesting model. The downside of that
1:24:16 approach in my opinion is that it's
1:24:18 these sites can be completely
1:24:20 overwhelming, right? They're they're
1:24:21 just there's access to too many models.
1:24:23 There's access to too much [ __ ] As as
1:24:25 the tools are proliferating, they just
1:24:27 get they just get very Byzantine. Um and
1:24:30 they're kind of a mess. But there's this
1:24:34 new thing called context edit with
1:24:37 context. So in chat GPT right now, in
1:24:41 Gemini right now, they have these things
1:24:44 that are native image generation models
1:24:46 where you can upload an image of
1:24:48 yourself and say, "Hey, put me in a blue
1:24:49 sweater." And it'll put you in a blue
1:24:51 sweater. And then you're like, "Well,
1:24:53 put a hat on me." It'll put a hat on
1:24:54 you. And then you'll say, "Have me
1:24:56 holding a telephone receiver." And it'll
1:24:59 have you holding a telephone receiver.
1:25:01 And every time it does it, your face
1:25:03 slightly changes. So, by about the fifth
1:25:05 time you've done this, you don't look
1:25:07 like you, right? Everything's completely
1:25:10 different. It's all discombobulated.
1:25:12 It's a [ __ ]
1:25:14 mess. Context apparently retains
1:25:18 anything that it's not changing, which
1:25:21 is really cool. Not quite sure how they
1:25:22 did it. I I assume they're doing some
1:25:24 layering um dynamically, but let's go
1:25:28 play. Log in.
1:25:34 [Music]
1:25:42 Tik Tok pin Joy Party Hedra reached out
1:25:44 to me and offered me a free month if I
1:25:47 engaged with their launch post. Awesome.
1:25:50 Congrats. That's great. Tell them a
1:25:53 month free. That feels a little chintzy.
1:25:55 Why don't you write back to them and go,
1:25:56 "I'm joy [ __ ] pretty. Why don't you
1:25:59 give me three months for free if I
1:26:00 mention you at all?
1:26:05 [ __ ] She would never say that cuz
1:26:08 she's a sweet person. I'm the
1:26:14 [ __ ] All
1:26:17 right. All right. Fantastic.
1:26:20 Fantastic. Oh, it was Hey,
1:26:24 Jen. Oh, I got I got a note from Hey
1:26:27 Jen. They were like, "Hey, we're doing
1:26:28 this launch on the 3rd. Do you want do
1:26:30 you want
1:26:31 to do something?" And I said, "Yeah,
1:26:33 what's the story?" And then they never
1:26:34 got back to me. So, they like you better
1:26:37 than they like me. Wish they should.
1:26:40 Okay. So, we're going to upload an
1:26:42 image. Which one are we going to do? Oh,
1:26:46 that's a good one. Yeah, let's do that
1:26:48 one. All
1:26:50 right. Okay. So now we could say make
1:26:56 his bow
1:26:58 tie
1:27:00 yellow. You can all see this,
1:27:05 right? And in
1:27:15 theory, where do we submit our Rick
1:27:18 Rubin recodes? Okay, Vicki, when you
1:27:21 when you remix one of the Rick Rubin
1:27:23 things from the way of code, you'll
1:27:25 you'll go to Gemini and then when you
1:27:26 publish it, well, there there's two
1:27:30 places. When you do a remix, it'll sort
1:27:33 of write out in the lefth hand pane what
1:27:36 it's doing. And in that text, there's
1:27:39 there's for me there was information
1:27:41 talking about you can publish it and
1:27:43 there's a Google form there, but I think
1:27:45 it also happens when you hit the publish
1:27:47 button. Oh yeah, look. It made his bow
1:27:48 tie yellow. Okay, great. Um, make the
1:27:55 theater seats
1:28:01 blue. Hedra is expensive. Hey, Jen is
1:28:05 okay. Just not all I need at the moment,
1:28:07 but good. Hedra is the best one. Their
1:28:10 character animation's really good. Like
1:28:12 the ones that I think are good right now
1:28:14 are V3 is amazing but expensive. There's
1:28:18 the blue seats and it makes it made his
1:28:20 hat blue. Let's say we don't want Let's
1:28:22 Let's give him a different hat. Let's
1:28:23 say
1:28:25 um put on a uh change his hat change his
1:28:31 hat to a bowler.
1:28:38 This is one of those annoying input
1:28:39 boxes that when you hit return, it just
1:28:42 does a line break and doesn't submit it.
1:28:44 Oh [ __ ] I closed the page. Ah, that's
1:28:46 No, you should be able to go back to it.
1:28:48 Vicki, if you go back to your If you go
1:28:51 into Claude, you should have a a chat
1:28:54 there that you should be able to go back
1:28:56 into and spin up the the artifact window
1:28:58 and then have your chat history still
1:29:00 there. It should still be in
1:29:02 there. There he's got the bowler. Uh,
1:29:05 let's say um let's give him a
1:29:11 um tux
1:29:21 jacket. Uh oh, my phone is about to die.
1:29:25 That's always fun. People like, "Where
1:29:28 did Kyle go? Why did he leave without
1:29:30 saying goodbye? That was rude.
1:29:35 Wow. Okay. Um, let's see. Um, red bow
1:29:40 tie. This is quite
1:29:43 impressive. This is quite
1:29:53 impressive. And then we're going to do
1:29:57 bright blue eyes.
1:30:01 I'm going to do electric blue eyes.
1:30:04 Electric blue
1:30:07 eyes. Hazel
1:30:09 eyes. Thank you. I got it. Good.
1:30:11 Perfect, Vicki.
1:30:14 Yay. Hours of frustration
1:30:16 [Laughter]
1:30:23 saved. That's the one thing the way of
1:30:26 code site. The one thing they did really
1:30:28 bad in my opinion is the word modify. So
1:30:32 when you've got the artwork there and it
1:30:34 says modify, they could have come up
1:30:36 with a more creative word than modify.
1:30:38 Oh, look at his eyes. Those are great.
1:30:39 Um, they could have come up with a
1:30:41 better word than modify. And then when
1:30:43 you click on it, it doesn't tell you
1:30:45 what's happening. It literally just
1:30:47 launches you over to Claude, which if
1:30:49 you've never used Claude before, you're
1:30:51 like, "What the [ __ ] is this?"
1:30:54 And then you you you maybe type in
1:30:56 something and hit return and it starts
1:30:58 writing code. If you're like if you're
1:31:00 just clicking on [ __ ] it's very very
1:31:02 disconcerting. So
1:31:05 um or even next to him. Wait, what was
1:31:07 that? Can you add another person? Yeah.
1:31:09 Let's say um add
1:31:15 some
1:31:17 happy children in the
1:31:21 seats
1:31:23 behind
1:31:26 him. Oh my god, that return thing is
1:31:29 driving me crazy. If I hold down the
1:31:31 command key when I hit return, it should
1:31:35 it should do it. I just always forget to
1:31:36 do that. This new feature in Korea is
1:31:39 wonderful. Love restyling video and VO3
1:31:44 will be in there too. Yeah. Favorite
1:31:46 tool? Crea unlimited 30 bucks a month.
1:31:48 Live drawing is magic. Cool. So Rick
1:31:52 McCaulay knows his stuff. He's Rick Rick
1:31:54 is a professor that dives deep into
1:31:56 these tools and is is really good at
1:31:59 this stuff. So he is if he is saying
1:32:01 Korea is worth your time, it's probably
1:32:04 worth your time. I don't spend enough
1:32:06 time with it to know. I just know that
1:32:08 there's a lot in it. So, like if you
1:32:10 want to Oh, look at that. If you want a
1:32:12 rabbit hole to run down and then let's
1:32:14 say um put a scary alien with green
1:32:25 eyes behind the little
1:32:28 girl on the left.
1:32:36 I'm such a
1:32:37 [Laughter]
1:32:40 douche. Robots in the other seat.
1:32:43 Aliens, robots, and we were in the
1:32:47 neighborhood. Oh
1:32:54 man. No, it replaced the
1:32:57 girl. Well, that's okay. That's pretty
1:33:00 good. The kids are happy.
1:33:03 Let's see. Uh, let's see.
1:33:08 Um, what are we going to do? Oh,
1:33:12 give the man a
1:33:15 handle bar
1:33:22 mustache. He's on her lap. He ate her.
1:33:37 But like the fact that we've done I
1:33:39 don't know whatever I've done here seven
1:33:40 or eight
1:33:42 generations and his face still looks
1:33:44 like his face and the theater's still
1:33:46 there and the composition hasn't
1:33:48 changed. Isn't it interesting how many
1:33:50 times the alien in a prompt is a
1:33:52 humanoid? Yeah, that's a good point.
1:34:03 Huh? So, can I download this? Yes,
1:34:09 download. So, let's
1:34:19 see. So, there's that. And then let's go
1:34:22 open this and
1:34:31 go. I think it was on desktop, wasn't
1:34:44 it? Where was that dude?
1:34:53 There it
1:35:04 is. So, it did change his skin
1:35:08 tone. That's interesting.
1:35:20 He's got these vertical lines in his
1:35:23 skin. So, not
1:35:25 perfect, but still pretty
1:35:31 good. Yeah, that's pretty
1:35:34 crazy. Pretty crazy. What are we up to?
1:35:36 So, we're at Korea. Korea's got this new
1:35:38 tool in called
1:35:41 context. Korea.ai. And Rick McCauley was
1:35:45 saying that he loves Korea. It's 30
1:35:47 bucks a month and you can do all sorts
1:35:48 of [ __ ] on it. Um, it's one of those
1:35:51 Swiss Army knife tools. So, yeah, you
1:35:53 got to do a lot. You got there there's a
1:35:55 lot to do and play with. If you if you
1:35:57 want to run down some rabbit holes,
1:35:58 Korea is a good one. All right. And that
1:36:00 tool is pretty good. So, what you can do
1:36:02 here is you can change elements in an
1:36:05 image without [ __ ] up the original.
1:36:07 Too bad.
1:36:09 up up resolution now to 100,000 pixels
1:36:12 uses Topaz Paz's upscaling enhanced
1:36:14 version. Holy [ __ ] Like, I don't know
1:36:17 why you would need a 100,000 pixel
1:36:19 upscale, but but I guess more is
1:36:24 better. All right, let's let's get on
1:36:27 the [ __ ] out of here. Um, tomorrow's
1:36:30 Friday, so on Fridays, a couple of
1:36:33 things
1:36:33 happen. Um, at 11:00 am Mountain time on
1:36:37 LinkedIn, I go live with this thing
1:36:39 called AI office hours. And if you
1:36:41 haven't been to an office hours, you owe
1:36:43 it to yourself to come to one. So, how
1:36:45 you find it? Go to my LinkedIn profile.
1:36:47 It's Kyle Shannon on LinkedIn. And just
1:36:50 go to any one of my events and there's
1:36:52 you'll see it in there. Um, it's office
1:36:55 hours. I always use the same um the same
1:36:58 URL and so just pick any of the old ones
1:37:00 and it's at 11:00 a.m. and I'll put out
1:37:02 a reminder tomorrow morning for it. Uh
1:37:06 and it's a it's a really good group of
1:37:07 people and we just talk about like what
1:37:10 people are up to with their work with AI
1:37:13 with we'll talk about new technology and
1:37:15 what are the implications of it and it's
1:37:17 it's a really smart bunch of people and
1:37:19 very generous very um just warm and
1:37:23 curious and good people. So um so do
1:37:27 that. Tomorrow night, Friday night date
1:37:29 night. Tomorrow is a first Friday, so
1:37:31 there's probably a high likelihood I'm
1:37:33 going to be going, you know, gallery
1:37:35 hopping with the misses. So, um, so
1:37:39 we'll probably start 8:30 or 9 Mountain
1:37:41 time tomorrow
1:37:43 night. Oh, it's last Friday.
1:37:46 Oh, so it's not first Friday, so we'll
1:37:48 probably be on time. 8:00 PM Mountain
1:37:50 time. Fun. So, just deal with it. Get
1:37:53 your nachos ready. Get your Hot Pockets
1:37:56 ready. Get your dirty microwave ready.
1:38:03 Check the home screen to see all the the
1:38:05 features. Yeah, I know. There's there's
1:38:07 a [ __ ] ton of features on this
1:38:11 thing. All right, look what I made
1:38:13 before you leave. Okay,
1:38:16 cool. Yeah. So, announcing lip syncing.
1:38:19 Let's let's see all their featured
1:38:21 featured things.
1:38:22 New enhanced models. Upscale and enhance
1:38:25 free Topaz
1:38:26 Labs. Visual
1:38:31 compositing. Chat GPT plus Ca. Cool. New
1:38:36 Crea announcing our series B. They
1:38:39 raised some money. Good for them. Video
1:38:41 restyle. 3D
1:38:43 objects. Chat image editing which I
1:38:46 guess is different than context. Style
1:38:48 of the week. Camera model. camera
1:38:51 controls, create a chat, and then yeah,
1:38:54 just here's all the things people are
1:38:56 making. And then here's just tons and
1:39:00 tons of things you can
1:39:03 do. Assets, gallery, chat, video, lip
1:39:06 sync, real time, create real time is a
1:39:10 blast. Stage is an experiment, which is
1:39:13 a big
1:39:15 infinite canvas. So yeah, lots of [ __ ]
1:39:18 there.
1:39:20 All right, good people. Fash, I hope you
1:39:23 have fun tonight. Sorry I went off on a
1:39:26 little rant there, but you know, I get
1:39:27 passionate about the
1:39:29 AI. I love people asking and challenging
1:39:34 me.
1:39:36 Um, but I but I'm you know what? I'm not
1:39:39 going to fight the fight of is AI good
1:39:41 or not. It's it's
1:39:44 irrelevant. What your opinion of it is
1:39:47 is irrelevant. It's here. Ain't going
1:39:50 away. All right, look what I
1:40:05 made. Fun little trippy generator. Ran
1:40:07 out of context limit before I could go
1:40:10 further. Was this from the from the way
1:40:13 of code? What else do we have in here?
1:40:15 Story vineg. That's so
1:40:18 cute. Kyle in his $20 million studio.
1:40:22 There it is. See, that's exactly what it
1:40:24 looks like,
1:40:29 too. Mr. It coming in hot with an iPad.
1:40:33 I like
1:40:34 it. All right. What are we looking at
1:40:36 here from Archetypal Architect? Let's
1:40:38 see what he
1:40:44 created. Wonder if it has audio. Oh,
1:40:46 cool. Oh, no. This is this is not a
1:40:48 remix of Rick Rubin thing. Okay, so the
1:40:51 size. So, you change the size of the
1:40:54 things. Very cool. Here, hang
1:40:57 on. Yeah, this is It's so cool that
1:41:00 we're in a world we can just do [ __ ]
1:41:01 like this. Shapes. Oh, number of
1:41:06 shapes. Those are geometric
1:41:09 shapes. Those are g geometric shapes.
1:41:12 Must not be very creative. Exactly.
1:41:17 [ __ ]
1:41:20 [ __ ]
1:41:21 Circles, spirals,
1:41:25 mixed
1:41:27 polygons. Right. Right. Right. Color
1:41:34 palette. Very
1:41:37 cool. Well, that one's
1:41:41 nice. You can just make [ __ ] now.
1:41:45 I don't know if archetypal architect was
1:41:47 a
1:41:48 coder. He's an architect. He's probably
1:41:50 good with CAD and [ __ ] like that, but
1:41:53 still. Come on. Look at
1:41:56 that. Gorgeous. Gorgeous. Gorgeous, I
1:42:00 tell you. Fantasy Bob. Hey, Marge. Hey,
1:42:04 Marge. Yeah. Listen. Hey. Hey. And I
1:42:08 don't know. How do I turn off the
1:42:10 YouTube?
1:42:13 You
1:42:13 what? Yeah, I know. I know. Was it a
1:42:17 marathon tonight? You're You're gonna
1:42:20 watch the wheel till till the news comes
1:42:22 on. Oh my, that's exciting. Um how But
1:42:27 figure it out myself. Yeah. No, I I
1:42:29 will. I I'm good. I'm getting better
1:42:32 with No, I know. I know. I'm not good
1:42:34 with the computers. Yeah, I know. I
1:42:36 know. March. Hey, do do you want some
1:42:39 more oatmeal? No, you're good. Okay.
1:42:41 Okay, hun. Back back to the wheel. That
1:42:45 Marge, she's the
1:42:52 best. All right. Fantastic. Everybody,
1:42:56 I'll see you later. See you later. So,
1:42:59 tomorrow, uh, office hours at 11 on
1:43:02 LinkedIn and then right back here 8:00
1:43:04 PM for Friday night date night. Deal.
1:43:07 All right. Deal. Peace out, everybody.
1:43:10 Bye. Bye. W up.