9/24/2025 - AI Learning Lab: Exploring Generative AI, MCP Servers, and Vibe Coding

Description

Chapters

Transcript

0:00 You ready? You ready for some action?
0:03 Are you ready for some football?
0:10 [Music]
0:22 Hey.
0:26 [Music]
0:33 [Music]
0:50 Woohoo.
0:53 [Music]
1:01 [Applause]
1:03 [Music]
1:11 There's been something, baby, I've been
1:13 trying to say
1:15 [Music]
1:17 for an age and it seems I don't know Now
1:23 with a past and a future now surrounding
1:26 me.
1:30 Surrender to whatever cheap can be
1:32 found.
1:34 [Applause]
1:35 There's been a little trouble
1:39 since you came to my rescue.
1:41 [Applause]
1:42 [Music]
1:46 And if you were like all of the rest, I
1:49 would have quit you long ago. But I
1:52 couldn't do that.
1:55 [Music]
2:10 Good evening, good people. What's
2:12 happening? What's going on with the
2:13 hair? I we have hair management issues
2:16 tonight. We're just going to deal with
2:17 it. It's not as bad as you all make it
2:20 out to be. All right. I understand
2:22 there's issues.
2:33 [Music]
2:35 Um
2:37 [Music]
2:41 we played with Suno V5 last night. That
2:43 was pretty impressive. We also did a
2:45 little vibe coding last night.
2:48 Um,
2:50 we played with Google Mixboard. I don't
2:52 think anything major came out today, so
2:54 that's good.
2:56 Um, so we can do whatever you want to
2:58 do.
2:59 Hair's not that bad. Hey, at least I
3:01 [ __ ] have hair, right?
3:04 [Music]
3:08 [Applause]
3:09 [Music]
3:16 Hey champ, what is it about guitars that
3:18 you like? What is it? Hey, producer
3:21 Brandon.
3:26 [Music]
3:31 Producer Brandon going. He said I said
3:34 at least I have hair. Producer Brandon's
3:36 like, "Hey,
3:38 sorry about that. didn't even take that
3:40 in.
3:42 [Music]
3:43 My dad at this age was was way more bald
3:46 than I am. So
3:50 [Music]
4:16 [Applause]
4:19 [Music]
4:26 [Music]
4:33 I there was zone.
4:36 [Applause]
4:36 [Music]
4:38 David played at a place of love.
4:43 Do you don't really care for music to y
4:52 goes like this. The fourth, the fifth
4:56 man falling majorly.
5:01 Therefore, king composing.
5:03 [Music]
5:04 Hallelujah.
5:09 Hallelujah.
5:13 Hallelujah.
5:16 [Music]
5:18 Hallelujah.
5:19 [Music]
5:22 Al
5:25 [Applause]
5:28 yeah
5:29 [Applause]
5:30 [Music]
5:35 [Applause]
5:37 you're feeling strong but you need a
5:39 proof.
5:41 You saw her baby on the roof.
5:47 Beauty and a moonlight
5:49 overthrew you.
5:54 She tied you to the kitchen chair. She
5:58 smashed your throne and cut your hair.
6:02 And from your lips she drew.
6:06 Hallelujah.
6:10 Hallelujah.
6:15 Hallelujah.
6:17 [Applause]
6:20 Hallelujah.
6:21 [Music]
6:24 Hallelujah.
6:26 [Music]
6:28 Oh yeah.
6:31 [Music]
6:42 Whoa.
6:44 Wow. That dog is a good dog. All right.
6:49 Share the live. Boom. There's that. Do
6:52 the black bar thing. Boom. There's that.
6:56 And uh welcome everybody. If you're new
7:00 here, this is the AI learning lab. We
7:04 occasionally learn something about AI.
7:08 [Music]
7:40 [Applause]
7:40 [Music]
7:55 [Applause]
7:57 [Music]
8:14 [Applause]
8:16 [Music]
8:21 [Applause]
8:23 Heat. Heat.
8:25 [Music]
8:27 [Applause]
8:28 [Music]
8:36 She came on to him
8:38 like slow moving cold front.
8:41 [Music]
8:45 His beer was warmer than a look in her
8:48 eye.
8:53 All right, enough.
8:56 guitar.
8:58 Enough musical instruments. What do you
9:01 want to talk about tonight? What do you
9:03 want to talk about tonight?
9:05 Let's talk about this AI stuff.
9:08 Who needs suno when we have Kyle?
9:12 MCP servers. So, here's the deal with
9:14 MCP servers. If you don't know what an
9:17 MCP server is, stands for model context
9:20 protocol. It was created by a company
9:23 called Enthropic.
9:26 If you know what an API is, it's similar
9:28 to an API.
9:30 It basically allows you to point your
9:34 large language model like chat GPT or
9:36 Claude at a website and the MCP part is
9:41 where you define what actions you can
9:43 take on that website. So it might be
9:46 Google Maps and you could have an MCP
9:48 server to Google Maps and you could go
9:50 find [ __ ] on Google Maps through a chat.
9:54 So it lets you use tools through your
9:57 through your chat. Um
10:01 I haven't played with them for a couple
10:03 of reasons. One is they're ve it's very
10:05 new. You still need to be fairly geeky
10:07 to set them up. They're not super
10:09 intuitive. They're not super supported,
10:12 although OpenAI now supports them. Um,
10:16 Enthropic obviously supports them. They
10:19 created the protocol.
10:21 Um,
10:23 there they are any anything right now
10:26 where large language models are using
10:29 tools
10:31 are replete with security and privacy
10:33 challenges. So, I'm just kind of
10:36 waiting. I'm just kind of waiting on
10:37 them to get a little more baked and a
10:39 little more easy to use and a little
10:41 more secure. And they will they'll get
10:43 there whether it's it's MCPS or some
10:46 other flavor of them that comes up.
10:49 Basically, we need in order for
10:53 large language models to go truly
10:55 agentic. You you hear a lot about agents
10:58 right now. A lot of people are talking
10:59 about agents, but everyone has their
11:00 head up their ass and everyone's calling
11:02 everything an agent. An agent is is a is
11:08 a large language model that can reason
11:11 and take autonomous action. So it can do
11:14 things without you babysitting it. And
11:17 it can use tools.
11:19 And so using tools might be I go out to
11:22 this website and I do some sort of
11:23 transaction and then I come back and I
11:25 tell the thing what I learned or what I
11:28 did and then it thinks some more and
11:30 adds that to the reasoning path and goes
11:31 out and does something else. Those are
11:34 coming. We've got little hints of them
11:37 right now with things like Gen Spark and
11:39 Manis. Um,
11:42 but it's very, very early days. So, I
11:45 would say unless you're super geeky,
11:48 don't worry about it right now. It'll
11:50 come. It'll come soon enough. And quite
11:52 frankly, I wouldn't get too super
11:55 excited about agentic stuff because at
11:58 some point uh they're going to get good
12:00 enough that those are going to be the
12:01 things that do your job that get a bunch
12:03 of people fired. So, so cool out on
12:07 wanting those things to get really good.
12:09 Ann Murphy, what's happening? What is
12:11 happening?
12:14 Um, we got Ann Murphy in the house. We
12:15 got Dr. Jay ion on China sad hustle mimi
12:21 shaking it up in here source camp of
12:24 course Tobias what's shaking Tobias
12:26 probably calling in from the hot tub
12:27 with a wine
12:29 you know why cuz he knows how to live
12:31 life I'll tell you that right now
12:38 I remember the night what was it you you
12:40 were doing something and you dropped
12:42 your phone in the hot tub and kicked you
12:50 Oh man,
12:52 Jetet Nick is in the house. What's
12:54 happening, Jet Setter? Um Sharon
12:57 Crawford over on YouTube. How y'all
12:59 doing? Um so here's the deal. If you're
13:01 new here, my name is Kyle Shannon. This
13:03 is the AI Learning Lab. We're just
13:04 getting started.
13:06 If you have questions about AI, if you
13:08 have opinions or thoughts about it, um
13:11 if you want me to demo something, I'm
13:12 happy to demo something. Um I tend to
13:15 focus more on generative AI than than
13:17 you know sort of building AI or doing
13:19 coding kind of stuff.
13:22 Um the reason for that is because my
13:25 background is in storytelling and
13:28 self-expression not in in code. Um I do
13:31 have a tech company. Um but I'm not a
13:34 I'm not a coder. I'm more of a product
13:35 person than a coder. And uh and I the
13:39 thing that excites me the most about AI
13:42 is the fact that non-engineers are now
13:45 able to harness the power of this thing.
13:47 And so that's kind of the focus of this
13:48 channel. But I welcome engineers in
13:51 here. It's just if you're an engineer
13:53 and you have engineery kind of
13:54 questions, I'm probably going to be a
13:56 little disappointing to you.
14:02 And I have become one with my with my
14:04 disappointment.
14:05 [Laughter]
14:08 Oh man.
14:11 All right. So
14:16 Vickiy's in the house. Hope you're
14:18 feeling well.
14:20 Yeah, actually that's a good point,
14:22 Brandon. If you are geeky and you want
14:24 to you want to um connect with someone
14:26 who's really
14:28 deep into the weeds on this stuff, Pton
14:30 M um is a member of the AI salon. If you
14:34 join the AI salon at
14:35 community.thesalon.ai,
14:38 if you could pull up that URL, Brandon.
14:41 Um, that is
14:44 um,
14:46 how is Vicki? Vicki is good. Well, Vicky
14:49 Vicky had um had some some achy joints.
14:54 Deachd
14:56 achy tonight and tired. Yeah, exactly.
14:59 to to deach your joints, you have to go
15:01 through a really achy period.
15:05 So, I think she's in the achy period,
15:08 which probably sucks. So, welcome. I'm
15:10 glad you're here. Um,
15:13 so, so speaking of Ann Murphy, since
15:15 she's in the house, so if you are a
15:18 woman in tech, if you are someone who is
15:20 trying to figure this AI stuff out, the
15:23 create conference is happening October
15:25 11th to October 14th in Salt Lake City,
15:29 put on by She Leads AI. So, I do the AI
15:32 readiness um project podcast with an on
15:35 Wednesdays and um that happened today.
15:39 So, I hope you saw it. Uh, if not, shame
15:42 on you. We We are not above shaming
15:44 people here. We'll do it. We will We
15:46 will cast stones. We will cast slurs.
15:49 We'll do it. We We'll go there. Anyway,
15:53 um go to sheleadsai.ai
15:56 and check out the create conference and
15:58 uh if you have not signed up for it, you
16:00 should go do so. So, with that, um what
16:05 was that? Fortune.
16:07 Oh, cool.
16:10 So, breaking here tonight, this is an AI
16:13 learning lab exclusive.
16:16 If you go to sheleadsai.ai
16:18 and you want to go to the create
16:21 conference, if you put in the code
16:23 fortune 500, all caps fortune 500, you
16:27 get 500 bucks off
16:30 straight from man Murphy. There's a
16:32 deal. All right, so go sign up. Let's
16:35 get some people there. All right.
16:38 Fantastic. Fantastic. Fantastic.
16:42 I tried to join and couldn't find it.
16:45 Tried to join what
16:48 the the uh the create conference.
16:52 Let's see. Let me jump over there right
16:55 now.
16:57 She leads
16:58 AI.ai.
17:00 Maybe you're talking about something
17:02 else.
17:04 Yeah, it was right there on the right
17:06 there on the homepage. It was there.
17:10 Oh, no. The podcast. Oh. Oh, the
17:13 podcast. Yes. So, the podcast is now
17:16 happening on
17:19 uh a standalone YouTube channel. So, I
17:21 think I need to update and I've been a
17:24 bad boy. I have not updated our uh
17:28 calendar invites in the salon. That's
17:30 that's my problem. Uh Vicki, that's
17:34 something I've got to fix. Thank you for
17:35 the reminder for that. Um okay, Groovy,
17:38 um can you tell me some advice in
17:40 building an AI SAS app? Tools and advice
17:45 on the process. Well,
17:49 I need to know a bit more about you.
17:51 Let's see what's your first name there.
17:53 I can't really see it with this. Um,
17:56 someone someone M38
18:01 someone M38. Listen, someone listen,
18:05 you're never going to be somebody until
18:06 you're anybody. I am someone. Who's on
18:10 first? Okay.
18:13 Um, so
18:16 I ass I assume by how you asked the
18:18 question, you may not be technical.
18:21 Um, and if you're thinking of how do I
18:24 vibe code uh a SAS app uh and get that
18:28 out there because everybody's posting on
18:30 X that they vibe coded a SAS app and now
18:33 they're making $20,000
18:35 MR um overnight. Um that you're
18:40 >> you can make money with
18:42 >> you're uh you're you're in that category
18:44 of of people wanting to want to do
18:46 things. Now so a couple of things. If
18:49 you really know what you want to do, if
18:50 you if you know what you want to do,
18:53 vibe coding is an option. But there's a
18:56 couple of things you need to know. If
18:58 you're new to software development, you
19:01 need to understand that in order to have
19:03 a functional application
19:06 that you're going to charge money for,
19:08 you need
19:10 a number of different components. You
19:12 need the application itself. That's a
19:15 relatively straightforward thing to do.
19:17 Then you need a database. You need a web
19:19 server. You need authentication like
19:22 people need to be to be able to sign up
19:24 for it, right? And then you need uh sort
19:27 of a transaction piece. You need to be
19:29 able to charge for it and all of those.
19:32 And and then the I guess the fifth one
19:34 is you need to have it that all set up
19:36 correctly so that you don't get your
19:38 [ __ ] hacked.
19:41 Um and and what I would say is this.
19:44 you can vibe code together a prototype
19:47 of a SAS app very very quickly.
19:51 Going from
19:53 prototype level to production level is
19:56 is a much heavier lift. Um and and you
19:59 probably need to learn about some of
20:01 those other components. Now what I can
20:03 tell you is some of the vibe coding
20:06 platforms are getting better at this.
20:08 Like Lovable for example has a deal with
20:12 um Supabbase for the database component
20:15 of it and they've got an authentication
20:17 thing and they're in beta right now of
20:19 having their own database backend and
20:22 their own authentication backend. So
20:25 Lovable applications are going to handle
20:29 a lot of the heavy lifting of that
20:32 stuff. So it's lovable.dev. If you
20:34 haven't vibecoded something, you should
20:36 go do that.
20:39 Um,
20:42 you know, some of the other ones are you
20:44 can do this in Claude, you can do this
20:46 in chatgpt in the canvas, you can do it
20:49 in Gemini in the canvas,
20:53 you can go to
20:55 astudio.google.comapps
20:58 and that's a whole vibe coding area
21:00 there. There are a bunch of different
21:01 places to do it. Replet's got an agent.
21:05 Um uh uh uh Cursor is an app that that
21:09 is a vibe coding app that integrates
21:12 with with um software development
21:15 environments.
21:16 Um there's a bunch of them out there,
21:18 but just make sure that you're going
21:21 into it with eyes wide open to know that
21:24 you need those components. And if you
21:26 don't have those components bulletproof,
21:28 you're going to get hacked. Because
21:30 what's happening is that the software
21:33 development community is
21:38 they're they're not quite as cranky as
21:40 the creative community is with with
21:43 things like music generation software
21:46 and video generation software, but the
21:48 the people who've been developers for a
21:51 long time are decently cranky about
21:54 people vibe coding apps together. So, so
21:57 what they're doing is they're letting
21:59 people vibe code a SAS platform, launch
22:02 it, and then they go exploit it and hack
22:04 it and do things like that and teach
22:05 them a lesson. So, just just be careful.
22:08 Be careful out there.
22:13 Lovable to do a 2D game. Absolutely. I
22:16 Well, let's go. Let's go make a game.
22:18 Let's go. Uh, black screen, please. I
22:22 think I do have the the black screen up.
22:24 Oh, but you know what I need to do? So,
22:26 I need to get rid of my bookmarks here.
22:30 Always show bookmarks. That helps. That
22:31 gives us more black.
22:34 And then I can bring this up like this.
22:36 There. That's better. Is that better? I
22:39 think that's a little better. All right,
22:44 let's go make a 2D game. Um,
22:49 [Music]
22:52 do I want to do it in lovable? Yeah,
22:54 let's do it in lovable. Well, I haven't
22:55 played in Lovable in a while.
23:02 MRR monthly recurring revenue. Yes, ARR
23:06 annual recurring revenue. Yeah, I just
23:09 saw I just saw an ex post today that
23:11 some guy was like a thing I completely
23:13 vibe coded just hit 20 20,000 MRR, which
23:17 you know that's not bad.
23:20 It's 240 grand a year, but you know,
23:22 it's also it's 20,000 MRR probably, you
23:26 know, 6 weeks after he made it. So, is
23:29 that is he going to grow this? Are
23:31 people actually using it? Like there's
23:33 the the thing about creating a SAS app
23:36 is that your competition is increasing
23:39 exponentially
23:41 and
23:44 everyone else's ability to create a SAS
23:46 app is is going to be easier and easier
23:49 and easier. So your competition is is
23:52 going to be accelerating around you. Um
23:55 so make sure that if you're going to
23:57 start something like that, don't don't
23:59 go into it thinking like it's just going
24:00 to be easy. like you've got to like
24:02 market to customers and understand
24:05 things like customer acquisition cost
24:07 and and churn rate and onboarding and
24:10 just there's all sorts of all sorts of
24:12 [ __ ] you got to contend with if you're
24:15 doing a SAS app. All right. I'm willing
24:16 to learn cyber security. Wait, this was
24:19 good. Um, who do you recommend me doing
24:23 to protect it? I Oh. Oh. What do you do
24:27 recommend
24:28 me doing to protect it? So, I'll tell
24:32 you what you could do is
24:35 you could just go on X. Actually,
24:36 actually, you know what? Maybe rather
24:38 than vibe coding right now, let's let's
24:39 hop over to X for a second. Let me share
24:41 this tab.
24:47 Um, I'm on X right now and I'm going to
24:50 go to Grock. So, so one of the nice
24:52 things about Grock, if you've not used
24:54 Grock,
24:56 is that Grock can tell you about
25:02 Oh, what's a SAS app? Okay, I'll get to
25:04 that in a second. Um,
25:07 one of the nice things about Grock is
25:09 that because
25:11 Elon Musk shut down the API to Twitter,
25:16 Grock is is the only large language
25:18 model that can actually search Twitter,
25:21 right? And and and consolidate what's
25:23 going on there. So you can go to Grock
25:25 and you can say
25:28 um I am going to vibe code a SAS
25:37 app
25:39 and want to launch it commercially.
25:45 Um,
25:46 I want to better understand
25:51 the security and configuration
25:57 issues
25:58 that
26:00 beginner
26:02 vibe coders
26:07 are overlooking
26:11 and I want to get
26:14 better at my craft.
26:19 So, my app is bulletproof.
26:24 Okay. So, that's going to go off and do
26:26 that. It's going to go think the request
26:29 involves vibe coding a SAS app for
26:31 commercial launch.
26:36 Uh,
26:38 check in. What does SAS stand for? SAS
26:40 stands for software as a service. And so
26:43 the the sort of classic the the classic
26:47 SAS model is Salesforce, right? So you
26:51 could create your own database of, you
26:54 know, uh customer records, right? Say,
26:56 you know, sales prospects
26:59 or you can pay Salesforce a monthly fee
27:03 for them to build the application for
27:05 you to, you know, for your CRM. And so
27:08 that's so the classic model is you build
27:11 a website that people log into and then
27:14 they pay a monthly fee per seat, right?
27:16 So per user and then you make buckets of
27:19 money because you don't really need that
27:21 many staff because they're just logging
27:23 into this piece of software and using
27:24 it. But everyone needs a different
27:28 feature. And like one so um Salesforce
27:31 is a classic SAS app where it's got so
27:34 much feature bloat and so much um sort
27:37 of overkill on features that it's
27:40 completely unusable at this point, but
27:43 no one can leave it because it's got all
27:45 their customer data in it, right? Like
27:47 the the the cost of of transitioning
27:50 from Salesforce to some other CR CRM is
27:54 ridiculous. And so they they're
27:56 basically just printing money. So the
27:59 the ideal thing that you want when you
28:01 create a SAS app is what's called a
28:03 sticky SAS app where once someone starts
28:06 using it, they can't unuse it. Like chat
28:08 GPT is a good example of that. Like once
28:11 you've used a large language model, you
28:13 don't want to not use one again, right?
28:16 So
28:20 uh not to be confused with SAS, which is
28:22 what I have exactly. All right. Building
28:24 a SAS app using no code or low code. Um
28:29 I I I've structured this table
28:33 for clarity focusing on key categories
28:35 of risk.
28:37 Account impersonation and authentication
28:40 failures. Platforms often default to
28:42 simple login or use builder credentials
28:45 for everything and new devs assume
28:47 build-in off is good enough without
28:50 testing edge cases. attackers can hijack
28:53 sessions, right? So, here's a big big
28:56 longass table that Grock just put
28:59 together. So, so this is good. I mean,
29:02 what what's nice about Grock
29:05 looking at X, what people are talking
29:08 about on X, X is where all of the the AI
29:12 and dev communities are sort of
29:14 clashing, right? And so, you've got vibe
29:17 coders going, "Look, ma, look, I made my
29:20 first app. And then the developers are
29:23 going, "Hey, Numnuts, you know, you left
29:25 the back door open, and by the way, I
29:27 went in there and got all your, you
29:28 know, customers credit card information.
29:30 You might want to close that back up."
29:32 So, you have these these conversations
29:34 going back and forth.
29:36 The vibe coding
29:38 platform community is well aware of all
29:42 these issues and and there there's a
29:44 race right now in my opinion for the
29:47 vibe coding platforms. There's two basic
29:50 audience for for AI coding platforms.
29:53 One is professional engineers, right?
29:56 And so things like cursor and codecs and
29:59 and um and and those sorts of tools are
30:03 geared at existing engineers doing their
30:05 work more efficiently. Then you've got
30:07 things like lovable um and and now and
30:11 and the the sort of uh the canvas
30:14 features in um Claude and Chatg GBT and
30:19 now in Gemini are aimed more at think of
30:22 it as casual developers, vibe coding
30:25 developers. So two really distinct
30:27 audiences.
30:29 The professionals they know how to do
30:31 this [ __ ] for the most part. So, so
30:33 they're going to do all their
30:34 configuration, you know, how they do
30:36 their configuration and they're going to
30:38 get that locked down. The casual vibe
30:40 coders, that's where there's a race
30:43 right now to make it as simple as
30:45 possible to literally just ask for an
30:48 app, have it build it, and have it
30:50 automatically set up all of these things
30:52 like authentication and databases and
30:55 commerce and security and create a
30:58 bulletproof app. there's not one out
31:00 there that that does it, you know, super
31:02 simply right now. Um, but I would say
31:05 probably within a year there will be.
31:07 So, like right now is a really good time
31:08 to do it because there's good
31:10 opportunity. Um, because once it's
31:12 trivial for anyone to make an app, then
31:14 you're going to be competing with 98.5%
31:17 of people that aren't coders. So, hope
31:19 that helps.
31:22 Erica Hana, yeah, also how much ad spend
31:25 to promote it? Um, people never disclose
31:28 that. Yeah, check in. What does SAS
31:30 stand for? Oh, we did that already. Um,
31:35 yeah, the the um Erica's point's a good
31:38 one. So, I talked about customer
31:40 acquisition cost or CAC as they call it.
31:43 If if you're talking to a VC and you're
31:46 like, "Hey, I got a SAS company."
31:48 They're probably going to go, "What's
31:49 your CAC?" You're going to be like,
31:52 "Excuse me, did I stutter?
31:55 What is CAC?" customer acquisition cost.
31:59 So, you want to know your customer
32:00 acquisition cost. So, if if you run let
32:04 let's say um
32:08 you you have a a six-month churn, right?
32:10 And you get a customer
32:13 and you're charging them 20 bucks a
32:15 month and on average your customers last
32:17 for six months. Ideally, they last
32:19 longer than that, but let's say on
32:20 average they last for six months. the
32:23 lifetime value of that customer or the I
32:26 guess the annual value of that customer
32:29 will be $120, right? Six months times 20
32:31 bucks a month if you're charging 20
32:33 bucks a month. So you're going to make
32:35 $120 on that customer.
32:38 If you spend $200 in advertising to get
32:42 every customer that earns you $120,
32:46 you're not making money.
32:49 But if you spend $20 per customer to
32:51 acquire them and you're making $120,
32:54 then you're making $100 per customer and
32:56 you might be making money. So So
33:01 the here's the other thing about about
33:03 SAS applications and I where did I hear
33:06 this?
33:09 I think it was just a Tik Tok channel.
33:13 And I forget I forget who said it. I'
33:15 I'd like to credit them with this, but
33:17 they basically said, "Listen, one of the
33:19 challenges with vibe coding is you vibe
33:22 code up an app you think is really cool
33:24 and then you put it out in the world and
33:26 you wonder why no one comes and and uses
33:29 it." Well, does it actually solve a
33:32 problem that people have? Does it solve
33:34 a problem that people have that they
33:36 there's enough urgency around that
33:38 they're actually willing to pay for a
33:39 solution? And then can they find you?
33:43 And so all of these things are really
33:44 important. So I would say a a big
33:49 thing that you should do before you
33:50 start vibe coding
33:53 is a phase that I'm calling in my new
33:56 book validation. So you've got your
33:58 ideiation phase where you're like, "Oh,
34:00 I got an idea for a SAS app. Cool."
34:04 The validation phase before you get the
34:07 planning and building what you're going
34:08 to go do. The validation phase is like
34:12 go use deep research. Go use GenSpark or
34:15 Manis or ChatgPT Deep Research or
34:18 Perplexity and go say, "Hey, I've got an
34:21 idea for an app and go find me any
34:24 competitors. Find out how much they're
34:26 charging. Find out if people are
34:28 actually signing up for this. Find out
34:30 what the churn rate is for for this
34:32 category. Find out what the customer
34:34 acquisition cost is for this category.
34:38 Go find out if anyone actually wants
34:40 this thing. Right? Very often what
34:44 you'll find is there are already two or
34:46 three major players doing what you want
34:47 to do. It doesn't mean you can't do it
34:50 because execution's hard. You might be
34:52 able to do it better than they do it.
34:53 But
34:56 if you build it and just assume you're
34:58 like the only one out there, that's just
35:00 [ __ ] stupid because we now have deep
35:02 research, right? We now have agents that
35:04 will go out and find out all this [ __ ]
35:07 for you,
35:09 right? So do that first.
35:13 All right.
35:16 Uh uh uh new book. Um I check out for a
35:20 week and there's a new book.
35:23 Let's do a deal for for a launch party.
35:25 Yeah, let's do it. Um it's it's uh my
35:29 new book's called 10person team and it's
35:32 about if you um learn to feed your
35:35 prompt like a producer. You put yourself
35:37 in the producer role, then you can sort
35:40 of spin up a 10person team that helps
35:43 you take any idea that you have and 10x
35:46 its impact and 10x its power and things
35:49 like that. So, I've been thinking a lot
35:50 about the the progression of how you
35:53 move from an idea to really launching
35:55 something, launching and growing
35:57 something. And you know, and the phases
36:00 are ideation, validation,
36:02 validate your idea, right? If you're
36:05 going to put time into this thing,
36:07 [ __ ] figure it out, right? Then
36:10 planning and then building and then
36:12 launch and growth. So, anyway,
36:18 yes, yes, yes. I like it. It's a vibe.
36:20 Thank you. Yeah, I'm excited about it.
36:21 I'm really excited about it. And you
36:23 know, the the thing the thing that that
36:26 the 10erson team thing does
36:29 is
36:33 one of the things I've been thinking a
36:34 lot about is every one of us puts limits
36:37 on oursel. Let me switch my
36:41 view here for a second. Every one of us
36:44 puts limits on oursel and there's
36:46 self-imposed limits and then there are
36:48 actual just limits of you might not have
36:51 skills, you might not have knowledge.
36:53 Some of the self-imposed limits are you
36:56 might not have confidence to like I'm
36:58 not really a business person. I don't
37:00 really know a lot about business but I
37:02 kind of have this idea for business.
37:05 Um,
37:08 so our limits are defined by our gaps in
37:13 knowledge and skill and confidence, I
37:16 guess, is the third one.
37:18 What AI allows you to do is to fill in
37:21 those gaps, right? If you're like,
37:23 "Well, I can't be an artist. I'm not
37:25 creative." Well, now you are.
37:27 Well, I'm not really a writer. Well, now
37:29 you are. Well, I don't know how to write
37:30 a business plan. Well, now you do.
37:33 Right? uh I can't do um research and
37:36 analysis. Oh yeah, actually you can now,
37:39 right? You can spin up a team member to
37:40 do that. So what you can do is you can
37:43 just put yourself in the role of the
37:45 producer where your only job is to have
37:47 the idea and hold the idea and like you
37:52 get to you get to define what good looks
37:54 like, right? And then AI fills in all
37:58 these gaps that you don't have and you
37:59 can just sit there and be like the
38:01 orchestra conductor. So that's the basic
38:04 idea. So I'm super excited.
38:08 Under and overhyped. Oh, I like that.
38:11 Yeah, that's actually a really good
38:12 point. I say that AI is paradoxically
38:15 under and overhyped. Yeah, I think
38:18 everyone talks about it being overhyped.
38:20 I think it's actually underhyped. But I
38:23 will agree with you, Pate. I gotta tell
38:24 you, man. I try I took on a video job
38:27 over the weekend and and like every
38:29 [ __ ] time I try to take something
38:33 beyond the prototyping stage, try to
38:37 make something professional, like
38:40 complete. Oh my god, is it a nightmare,
38:43 right? Because it's like AI is really
38:45 good up to about 80 90%.
38:48 But it's the 10% that makes a project
38:51 [ __ ] viable.
38:53 And it's just like ah it's maddening.
38:57 So, and you'll find out with vibe
38:59 coding, too. Like, when you start vibe
39:01 coding your SAS app, it it will
39:04 absolutely blow your [ __ ] mind how
39:06 quickly you can get that thing to like
39:09 80. You're like, "Holy [ __ ] this thing
39:11 works." And then you're like, "Oh, well,
39:13 wait. It's creating two things here, and
39:15 it it should only create one. Let me
39:18 just go fix that." And then you go try
39:19 to fix that, and it breaks everything
39:21 else.
39:22 It's a [ __ ] nightmare.
39:26 Archetypo, what's happening?
39:29 Oh, man. All right. Um,
39:33 [Music]
39:39 is there an AI bubble?
39:42 Au. Oh, yes, definitely. Um, well,
39:47 so, so, so there's a couple of things
39:49 going on. I think we're gonna see waves
39:52 of AI bubbles. Um, the fact that
39:56 Meta was hiring 24 year old developers
40:00 with $200 million signing bonuses,
40:04 there's a sign.
40:07 there's a sign that we might be in a
40:08 bubble that if you're that desperate for
40:12 talent that you're willing to pay $200
40:15 million to get an engineer um for your
40:18 AI initiative um that might be an
40:21 indication we're in a bubble and
40:25 I think as the tools get better and
40:27 better and better and as more and more
40:29 compute comes online TPUs in Google's
40:32 case and GPUs in Nvidia's case Um,
40:38 we're going to see new capabilities
40:40 coming and we're going to see as as um,
40:42 agentic AI gets stronger and stronger, I
40:46 think we'll see waves of bubbles where
40:47 we're like, it's absolutely ridiculous.
40:49 It's going to crash and it's going to
40:51 sort of it's going to build
40:53 ridiculously, then it's going to plateau
40:56 and and the theory is that at some point
40:58 the bubble will crash and will come
40:59 crashing back down. I think what's going
41:01 to happen is it's it's going to build
41:03 ridiculously plateau and then some new
41:06 breakthrough is going to happen. It's
41:07 going to build again and then that'll
41:09 plateau and then it's going to build
41:10 again. I think we're probably five years
41:12 out from this thing getting really
41:15 frothy.
41:16 Um, metaphorically, if I compare this to
41:20 the early days of the worldwide web,
41:22 we're probably more like 96 97 than we
41:26 are 1999.
41:28 And the bubble burst in 2000. So 1997,
41:33 1998, 1999, it the worldwide web got
41:37 absolutely [ __ ] ridiculous. There
41:39 were things being funded that shouldn't
41:41 have been funded. It was just stupid.
41:43 There was just stupid money in it and it
41:46 needed to crash. I don't think we're
41:48 there yet with AI. I think I think there
41:50 are some areas of AI that are like that,
41:53 but I think it's just going to keep
41:54 rising.
41:56 So anyway, um it's like the.com bubble.
42:00 A lot of businesses are going to die
42:02 soon. Yep, I agree with that. And then
42:04 and then new ones are going to roll out.
42:06 The difference between this and the.com
42:08 boom is that the core feature of the
42:13 technology of the worldwide web. It was
42:16 a single feature.
42:18 It was the hyperlink was the single
42:21 thing
42:22 that was the foundation for all of the
42:27 worldwide web and all of the internet
42:29 and all of us being connected all the
42:31 time and commerce all that sort of
42:33 stuff. There were things that were
42:35 layered on top of that, but the core
42:36 breakthrough technology was the
42:38 hyperlink. Very simple technology.
42:40 AI is much more complicated, much more
42:43 sophisticated, much more capable than
42:47 just that. So, so I think where you're
42:49 going to see growth and evolution is
42:52 just going to be it's going to just be
42:54 way more complicated than than the
42:57 worldwide web. Um,
43:00 YouTube comment, David, but then again,
43:03 Zuck wasted hundreds of million dollars
43:05 on metaverse and didn't get anything
43:06 from it really. Well, no, that's that's
43:08 my point is that I I listen, I know
43:12 there's some talented 24 year olds out
43:14 there. Um I don't I can't imagine one
43:17 that's worth $200 million. I mean, if
43:20 you back a a trusted founder and they
43:24 have a $200 million exit 10 years from
43:27 now, that's a that's a good startup. So
43:30 to invest $200 million just as a signing
43:33 bonus to a 24 year old just as an
43:36 employee that they better be doing some
43:39 pretty [ __ ] remarkable work. Um yeah,
43:42 that's that's some of the bubble
43:43 behavior that's out there right now.
43:46 [ __ ] sucks. Exactly.
43:53 Hopefully they'll shop at the AI salon
43:55 mall. That's great. the smart VCs will
43:58 stop will start shopping and
44:00 consolidating talent. Yeah, I'll tell
44:02 you. I'll tell you here's one that
44:06 I I think is
44:09 if if you want an area to play in
44:13 um where I think there's massive
44:15 opportunity. This is sort of Gary Vee
44:17 level. You know, you wanna you want to
44:20 make a name for yourself right now?
44:24 Get yourself
44:26 turn yourself into a content creator, an
44:29 AI content creator. Make AI videos, AI
44:33 images, AI songs. Get really good at
44:35 content creation. And don't just, and I
44:38 don't just mean like I can prompt and
44:40 make a cute image. I mean put together
44:43 projects. Put together two, three minute
44:47 short films. Put together music videos.
44:49 Put together
44:51 um art projects that that take images
44:54 and and do something with them. 3D space
44:57 explorable 3D spaces.
45:00 If you can establish yourself as a
45:02 talented content creator right now, one
45:04 of the things that I'm seeing and I'm
45:06 seeing it more and more and more and
45:08 more is that the content creation
45:11 platforms are turning into media
45:13 companies. They're turning into content
45:15 companies. They're turning into film
45:17 studios.
45:20 And who they're going to hire first and
45:22 who they're going to feature first on
45:24 their TV channels. Like Runway um Runway
45:29 ML right now just launched a TV channel.
45:31 What's it called, Brandon? Do you
45:32 remember what it's called? They have a
45:35 they have a a channel now. Sunno has a
45:39 radio station and they've now got this
45:41 thing called Hooks, which is like Tik
45:43 Tok. Yeah, Flume. Let's Let's go check
45:45 out Flume.
45:47 Um,
45:52 runwayml.com.
45:58 And if you're old enough to remember,
46:01 get started. Where's Flume? Is it
46:04 runwayml.comflume?
46:08 and go to dashboard
46:13 now.
46:16 Let's see.
46:19 RunwayML.comflume.
46:26 This page does not exist.
46:29 Take me home.
46:32 [Music]
46:35 URL. Anybody?
46:38 A Sununo creator in
46:40 Oh, yeah. Auno creator in Mississippi
46:42 just got a $3 million recording
46:44 contract. Yeah. Yeah. She created the
46:47 virtual her the like the the voice, the
46:51 look and feel. She's a producer. She's
46:53 created the songs using Suno. She's got
46:55 a $3 million a $3 million deal. Um,
47:03 it's going to seem weird right now, but
47:05 I'll tell you what, when MTV first came
47:07 out, it seemed weird. People are like,
47:11 "Wait, what? You're watching songs.
47:15 This is a channel about watching songs."
47:18 And it became a juggernaut when Netflix
47:21 went from delivering physical DVDs to
47:26 streaming to then producing original
47:28 content. People were like, "Wait, a
47:31 streaming service is making content."
47:33 And then they won some Emmys and some
47:34 Oscars and holy [ __ ] all of a sudden
47:37 they were taken seriously. That's going
47:39 to start happening with these content
47:41 creation. Boom. But what's the URL? I
47:44 guess I could go Google it, couldn't I?
47:47 Runway. Boom.
47:51 [Music]
47:55 LA.
47:56 [Music]
47:58 Where is it? I just don't see it.
48:06 I don't I don't see it in in Google
48:08 search. Or did I not say Yeah, I said
48:11 fume. Runway fume. Runway mlfoom.
48:17 App dot watchfume.com. Okay, here we go.
48:21 watchfume.com.
48:31 For kind of context, I was working on a
48:34 project in the Netherlands where I
48:36 wanted to kind of generate a tulip using
48:40 a GAN. So, I took 10,000 photographs of
48:44 tulips, stripping the tulips, holding
48:47 the tulips, photographing the tulips.
48:51 But after I made this data set, I
48:54 realized that it was important for me to
48:56 present it as an artwork in and of
48:58 itself.
49:00 [Music]
49:05 And this is now a massive installation,
49:09 >> right? So, here's
49:11 >> you understand.
49:12 >> Here's a documentary about someone
49:14 making data sets, generating content,
49:17 and then there's other things here that
49:19 are runway ML created
49:22 um
49:24 films,
49:26 right? And these things are going to go
49:28 from curiosities
49:30 and weirdnesses to all of a sudden
49:34 Paramount is going to buy one of these
49:38 media properties, right? or they're
49:39 going to acquire them or something like
49:41 that or these things are going to spin
49:43 up and become as powerful as Netflix.
49:46 Um, so like and and that's going to
49:48 happen quickly. That's going to happen
49:50 probably over the next three years
49:52 because what's going to happen is right
49:54 now the technology is still bad enough
49:58 that professional filmmakers can point
50:00 to it and go, "See, it's still shitty."
50:06 And they and and they'll point at how
50:08 amateur-ish anyone who's making films
50:10 with this is and how bad the physics are
50:13 and how bad the editing is and how bad
50:14 the continuity is and there's no
50:16 character consistency.
50:19 And in the past three months, character
50:21 consistency has gone from unachievable
50:23 to it's pretty close and three months
50:26 from now it'll be really close, right?
50:29 And so what's going to happen is you're
50:31 going to have existing story makers that
50:34 ignore AI and resist it. You're going to
50:36 have new people that embrace AI and get
50:39 really good at it, but they don't have
50:41 storytelling chops. And then you're
50:42 going to have existing filmmakers that
50:44 embrace AI and take their really good
50:46 storytelling chops. and and those latter
50:48 two categories, pure AI storytellers or
50:52 existing AI storytellers that level up
50:55 with AI, they're both going to start
50:58 rising above the noise and the
51:00 traditional filmmakers are are going to
51:03 be [ __ ] in a world of hurt and it
51:05 sucks
51:07 be trying to catch up and there's
51:09 there's there's a huge opportunity right
51:11 now. Huge opportunity.
51:15 All right.
51:18 Continuity is the biggest thing. Where
51:20 are you watching this? The doc the
51:22 documentary is at watchfoom.com.
51:24 W Ah foom m. And that's runway ML's new
51:30 like MTV, I guess.
51:34 It's crazy.
51:36 Absolutely crazy.
51:39 Um
51:42 Bam.
51:45 Let's go make music.
51:48 Um, so last night,
51:52 so if you don't know it, SNO, which is
51:55 the new
51:57 or not the new
51:59 um, which is the music generation tool
52:03 that's kind of the industry leader. It
52:05 used to be Sunno and Yo were duking it
52:08 out and UDIO's just sort of fallen off
52:10 the face of the planet. Um, 11 Labs now
52:13 has a music generation tool, but 11 Labs
52:16 music generation tool is probably like
52:18 Sunno's 3.5, which was a year ago, year
52:21 and a half ago. Sunno just dropped
52:26 um V5
52:28 and we made a we made a song last night
52:31 um
52:33 and I'll play it for you. It's really
52:34 good.
52:41 [Music]
52:51 A cafe door,
52:54 a fleeting face,
52:56 a matchbook offered. Time misplaced.
53:02 She slipped it deep.
53:04 Her coat held tight. No flame was born,
53:10 just endless night.
53:13 Years roll by like tides on stone.
53:19 But sun lit still chills the batch
53:25 of memory never struck a spark that
53:31 faithful got to pluck. A stranger's gift
53:37 a fleeting glance a fire lost to
53:42 circumstance.
53:47 Ooh,
53:52 [Music]
54:00 the scent of rain.
54:06 >> A thousand lives. She's lost in one.
54:12 Her pocket holds a quiet plea.
54:17 a relic from what I could not be.
54:23 What if the mag kiss the whip? What if
54:26 the clock had passed it tick?
54:30 [Music]
54:34 A tiny blaze of fleeting glow. Will she
54:37 have stayed? We'll never know.
54:45 book memory and never struck a smart
54:50 faithful God.
54:54 A stranger's gift a bleeding glance a
54:59 fight a lost circumstances.
55:05 [Music]
55:10 Hey.
55:16 [Applause]
55:17 [Music]
55:18 [Applause]
55:21 [Music]
55:35 Hey.
55:39 [Music]
55:42 Nice. Um, interesting comments while
55:45 that was playing. Um, if that if that
55:48 elementary school student makes a real
55:51 artist feel scared, it says more about
55:53 the artist than the medium. Well, I
55:54 mean, listen, here's the thing.
55:57 I don't trivialize the fact that this
55:59 that this stuff is scary.
56:02 Like if you've been a musician for 20
56:04 years or 30 years and all of a sudden
56:07 this thing comes along that can do what
56:09 you do better than you can do it or or
56:11 even approaching the level that you can
56:13 do it. That's [ __ ] terrifying. Like
56:15 that I get. Um Ashu asked the question,
56:18 "What do you say to the cohort of people
56:20 that want to see AI for the public at
56:23 least crash and burn?" Um
56:29 what what I say to them is I absolutely
56:33 understand.
56:36 I understand why you want it to crash
56:37 and burn
56:39 and I would encourage you
56:42 to to hold the hate of AI
56:48 in parallel with the fact that you
56:50 actually try it.
56:54 like you can you don't Ann Murphy made a
56:56 video about this today that I thought
56:58 was really good. So someone said how how
57:00 do how can you help me not hate AI and
57:04 and I her her response was something in
57:07 the neighborhood of you know what you
57:09 can hate AI but I want you to try it.
57:12 You can you can keep hating it. What
57:14 what what my experience has been is that
57:17 the people that vehemently hate AI have
57:21 never actually [ __ ] used it.
57:26 Because the minute you use it and the
57:28 minute you shift your mindset from AI is
57:32 this tool that's going to replace me
57:35 to AI is this tool that I can strap on
57:37 like a jetpack and augment my ideas and
57:40 amplify my ideas and create this
57:42 10person AI super team that can fill in
57:47 the gaps where where I don't have
57:49 knowledge and I don't have skills.
57:51 All of a sudden AI starts to get pretty
57:53 [ __ ] exciting.
57:57 So, so hating AI in the absence of
58:00 actually understanding it, that's the
58:03 thing that is it's the ironic thing is
58:06 it's going to hurt them the most,
58:10 right? It's a self-fulfilling prophecy
58:12 that if you just if if you use your hate
58:15 of AI as the justification for sitting
58:18 on the sidelines going, I'll never use
58:21 it.
58:24 then AI happens to you
58:28 and it's going to be [ __ ] miserable.
58:32 But if you're like, I hate AI. I hate
58:35 everything it stands for, but you know
58:36 what? I'm pretty sure it's not going
58:39 away. So, I'm going to go hang out at
58:41 the AI learning lab or I'm going to go
58:43 to join the AI salon or I'm going to go
58:45 to the she leads AI create conference
58:48 and I'm going to sit there pissed off
58:49 and hating it, but I'm going to learn
58:51 about it and I'm going to try it and I'm
58:54 going to see what it's like. There's
58:56 we're we're trying to get we're trying
58:58 to arrange right now. There's a musician
58:59 named Gabriel. I forget his last name. I
59:03 don't know if you remember it, Brandon.
59:05 Um, he's been like a a songwriting
59:09 producer for I don't know 25 years. Like
59:13 he's an older Gen X dude.
59:16 And someone told him about Sunno that
59:18 you can upload your songs into Souno and
59:21 it it will produce them and you're you
59:23 can just like record like a song sketch
59:26 into Sunno and it'll produce it. And
59:28 there's these videos of him
59:31 where he's like
59:34 he's pissed off
59:36 and intrigued at the same time. Hang on.
59:40 Got to do a Tik Tok physical dexterity
59:43 challenge, which I failed. Damn it.
59:49 Um.
59:51 Uh oh. What happened? All right, there
59:53 we go. Um,
59:55 and and he made he he thankfully made
59:59 these Tik Tok videos of himself playing
1:00:02 his song into
1:00:04 Sunno and then hitting the generate
1:00:07 button and and like 10 seconds later,
1:00:10 two versions of his song come out that
1:00:12 are fully produced. And you can watch
1:00:14 his face melting and you can watch him
1:00:17 trying to process what this actually
1:00:19 means.
1:00:20 And then over the next week or two, he
1:00:24 takes his most complicated song and then
1:00:26 his his his favorite song and then a a
1:00:29 song that he wrote for a teacher when he
1:00:30 was 17
1:00:32 and he puts them all in. And you just
1:00:34 witness him
1:00:36 co-processing
1:00:38 how pissed off he is that everything
1:00:41 he's done in his life is now up for
1:00:43 grabs
1:00:45 and going, "These tools are remarkable
1:00:47 and I'm going to learn how to use them
1:00:49 because he's smart enough to recognize
1:00:51 that this shit's not going away." So
1:00:53 that if he wants to be a relevant music
1:00:55 producer, he's at least got to
1:00:57 understand what he's [ __ ] competing
1:00:59 with, right?
1:01:02 That that kind of attitude is what I
1:01:05 would say to the haters. Hate it all you
1:01:08 want, but [ __ ] understand it.
1:01:12 Like understand it from the inside.
1:01:14 Understand it from using it. Because
1:01:17 what you may find is that once you
1:01:20 actually understand it, you're like,
1:01:22 "Huh, I actually kind of like this. This
1:01:24 can actually make a difference for me.
1:01:26 It can make a difference for other
1:01:27 people."
1:01:28 Um Jeff Flanigan on YouTube spotlight.
1:01:33 Um hate it but firsthand. Well, but what
1:01:36 I would say is hold hold the hate,
1:01:39 right? Just put the hate on pause. You
1:01:41 can come back to the hate. Just pause it
1:01:43 for a second. Try on AI.
1:01:47 Oh, I still hate it. I hate the way it
1:01:49 was trained. I hate the way it was
1:01:50 trained, too. And I also know that we
1:01:54 would not have the AI we have today if
1:01:57 they had asked permission to do that. I
1:01:59 know that,
1:02:01 right? And they're paying the price.
1:02:03 Anthropic just got slapped with a $ 1.5
1:02:05 billion settlement.
1:02:08 Now, luckily, they raised $10 billion,
1:02:10 so they're going to spend 1.5 of that
1:02:13 paying book authors. Great.
1:02:16 And now that it's trained, we have this
1:02:19 remarkable [ __ ] tool, right? And so
1:02:22 just pause the hate for a moment, try it
1:02:25 on, see where it is, and then think
1:02:29 critically
1:02:31 about where where you stand with it. You
1:02:33 can return to the hate. You can return
1:02:34 and say, "Hey, I did it. I went in. I
1:02:36 had a couple of Kevin Mallister moments.
1:02:39 I'm going to choose to never use it
1:02:40 again. I'm getting a cabin in the woods.
1:02:43 Peace out. I'm checking out
1:02:45 um Erica, the same people hated email um
1:02:49 and didn't trust the internet. But uh so
1:02:51 this is the other thing. So so the one
1:02:53 thing is I don't think that you can
1:02:57 I don't think it's realistic for
1:03:00 AI optimists like myself to evangelize
1:03:05 people that hate AI not to hate it. But
1:03:08 what I can evang evangelize is give it a
1:03:11 [ __ ] shot. Don't sit on the
1:03:13 sidelines.
1:03:15 Right.
1:03:17 Here's the thing about the haters. The
1:03:19 haters have always hated.
1:03:22 Every single every single technological
1:03:26 revolution in human history has been met
1:03:30 with people that say this new technology
1:03:33 will ruin the world.
1:03:36 It happened with the steam powered loom.
1:03:39 It happened with the printing press. I'm
1:03:41 sure it happened with fire.
1:03:43 You realize I could burn down the
1:03:45 village.
1:03:47 Don't you dare bring fire near me, you
1:03:50 heathen.
1:03:53 Right.
1:03:55 Wheels. Can you imagine how many people
1:03:57 got rolled over by those big giant stone
1:04:00 wheels and got crushed to death? I'm
1:04:02 sure there were the equivalent of
1:04:05 posters, anti-wheel posters in the wheel
1:04:08 community.
1:04:11 Let's go back to dragging things.
1:04:14 [Laughter]
1:04:21 Right.
1:04:23 The [ __ ] steam engine. Think about
1:04:25 this. Over a 40 or 50 year period from
1:04:28 when the steam engine was invented,
1:04:32 80%
1:04:34 80%
1:04:36 of the jobs that people had were
1:04:39 eliminated.
1:04:44 80%.
1:04:45 Now it took 40 or 50 years.
1:04:49 We are likely going to live through
1:04:52 something like that happening over five
1:04:54 or 10 years. Probably more like 10 cuz
1:04:57 [ __ ] takes longer than you think it's
1:04:59 going to. But probably over 10 years,
1:05:02 80% of the current jobs are going to be
1:05:04 gone. And and then you're like, well,
1:05:07 wait, what happened to those 80% of the
1:05:09 people? Well, those 80% of the people
1:05:10 all worked on farms. We were an agrarian
1:05:14 society. The steam engine came along and
1:05:17 it it had better muscles than men and
1:05:19 horses did, right?
1:05:23 10 horsepower. Wait, we could get rid of
1:05:26 10 horses,
1:05:28 right? How many people power is that?
1:05:30 Oh, that's like 50. Wait, 50 people are
1:05:33 going to lose their job? Yep.
1:05:36 What happened to those people? Did they
1:05:38 just dig a hole and crawl in it and die?
1:05:40 No. We're human beings. We adapt.
1:05:45 They they left the farm. Where did they
1:05:48 go? They went to New York City. They
1:05:49 went to Philadelphia. They went to
1:05:51 Cleveland. They went to cities. They
1:05:54 created cities. They created the service
1:05:56 economy.
1:05:58 They created restaurants. That they
1:06:00 created jobs that when they worked on a
1:06:02 farm, you couldn't imagine that that
1:06:04 would be a job. Wait a minute. You can
1:06:07 you can put a cart full of meat on the
1:06:12 street and people will walk by and they
1:06:14 will give you money to cook them
1:06:17 meat on the street. Yeah. How How do you
1:06:22 do that? I I made a cart and I have hot
1:06:24 water in it and I stick the hot dogs in
1:06:26 the water and then I have mustard in a
1:06:29 little thing here.
1:06:31 What happens when it gets sunny? Oh, I
1:06:32 have an umbrella.
1:06:36 And I painted my umbrella yellow. And
1:06:38 people like my yellow umbrella, so they
1:06:39 come buy my hot dogs.
1:06:42 That's a job.
1:06:44 Yeah. Yeah. That's a job. That's still a
1:06:48 job today, right?
1:06:52 So, we don't know what the next jobs are
1:06:55 going to be,
1:06:57 but but the the fear is still very real
1:07:00 and the steer the fear is still very
1:07:02 palpable.
1:07:06 What what makes it particularly
1:07:12 um
1:07:15 pointed right now
1:07:19 is that
1:07:24 all of the well-educated people, the
1:07:26 knowledge workers
1:07:28 assumed
1:07:30 that they were immune
1:07:34 from this change.
1:07:37 And they're not,
1:07:40 you know, frankly, I don't think
1:07:41 anyone's immune. I think if you're a
1:07:43 plumber and if you're paying attention
1:07:45 to what's going on with humanoid robots,
1:07:48 at some point we're going to have
1:07:51 humanoid robots coming over to the house
1:07:53 to fix the plumbing.
1:07:55 Now, is that, you know, this year next?
1:07:58 No. Is it five years from now? Probably.
1:08:02 What the [ __ ] What are the what the
1:08:03 [ __ ] are we supposed to do with that? I
1:08:06 I don't know. I don't know. But but what
1:08:10 I do know is if you're sitting on the
1:08:13 sidelines just watching the spectacle,
1:08:17 it's going to suck for you. It's going
1:08:19 to suck bad.
1:08:25 So, you might as well [ __ ] get, you
1:08:27 know, grab a surfboard. I hate the
1:08:30 water. Okay. How do you like waves
1:08:32 smashing you in the face? Oh, that's not
1:08:34 good. All right. Well, paddle,
1:08:38 pick up a surfboard, paddle out.
1:08:42 Two choices.
1:08:45 Deal with it or not. Anyway,
1:08:53 who will be the first to go?
1:08:57 Selling hot dogs is a safer job than
1:08:59 being a parallegal. It is.
1:09:01 It is. And scene. Yeah, exactly. Um,
1:09:04 let's go. Let's go make some songs.
1:09:09 All right, we're in create. Okay, so
1:09:11 we're going to do this. We're going to
1:09:12 do this interactive style.
1:09:15 Um,
1:09:18 so I want you can either shout out on
1:09:21 Wait, creatives outside of the ones that
1:09:23 use see themselves on the side of
1:09:28 soul versus soulless. Yeah, but Ashu,
1:09:30 here's here's the Okay, as as someone
1:09:33 listen, I have a degree in [ __ ]
1:09:35 acting. Sorry. Sorry. Back on the soap
1:09:37 box, Brandon. Thank you. I have a degree
1:09:40 in [ __ ] acting, right? I've been a
1:09:41 creative a creative professional my
1:09:44 whole life.
1:09:47 The trope
1:09:49 that creatives have right now that AI
1:09:52 can't be creative is simply wrong
1:09:55 because they're missing a huge part of
1:09:58 the equation.
1:10:00 These AI tools are not prompting
1:10:03 themselves.
1:10:04 There's a human being there sitting in
1:10:06 the producer role.
1:10:11 Now, how these tools produce content is
1:10:15 completely different than how it's
1:10:16 produced today. And it may feel like
1:10:18 cheating, and it may feel um unfair, and
1:10:23 it might feel like I spent 20 years
1:10:25 learning music theory just to have some
1:10:27 stupid [ __ ] machine be able to do
1:10:29 this. It's where we are.
1:10:33 And what what the creatives are
1:10:35 discounting, the soul versus soulless
1:10:37 thing. I agree that the technology on
1:10:39 its own is soulless,
1:10:41 but the technology when you when you
1:10:43 strap AI on and you you sit in the
1:10:46 producer role and I say, "Hey, I've got
1:10:49 an idea for this song called Weird
1:10:51 Mary."
1:10:54 That was not my idea.
1:10:57 I sort of said, "I wanted some
1:10:59 non-traditional ideas."
1:11:01 Chat GPT came up with that idea and it
1:11:04 sounded weird and I liked just how it
1:11:06 sounded. And then I stepped into the
1:11:08 producer role and I said, I'm going to
1:11:10 take this idea of weird Mary and I'm
1:11:11 going to turn this into something that I
1:11:13 like.
1:11:15 And then over the next three hours on
1:11:17 this live,
1:11:19 we crafted the lyrics. We co-wrote them.
1:11:21 The irregulars, the people that were
1:11:23 here co-wrote them with me. We put them
1:11:26 into Sunno. I created, I don't know, 15,
1:11:28 20, 30 variations of the song until we
1:11:31 got one that was like, "Oh, that's
1:11:32 weird. That's kind of cool."
1:11:36 And then we tweaked it and tweaked it
1:11:37 and tweaked it. And over three hours, we
1:11:40 generated this thing that I was super
1:11:42 happy with.
1:11:46 That's the soul is me.
1:11:49 The AI was just the tool that I used to
1:11:53 take some nebulous idea and bring it to
1:11:56 [ __ ] life.
1:11:58 So, it's not soulless. It's a human
1:12:01 being using a tool that's incredibly
1:12:04 powerful. But again, if you're sitting
1:12:07 on the outside of it and you assume that
1:12:10 AI is this,
1:12:11 >> you can make money with
1:12:13 >> you push a button and out squirts a a
1:12:15 musical,
1:12:17 then yeah, of course you're going to
1:12:19 resent it. Of course, you're going to
1:12:20 say it can never be creative. It's
1:12:22 soulless. It will never have soul.
1:12:26 They're absolutely right in their
1:12:28 conviction, but they're not looking at
1:12:30 the actual reality of how these tools
1:12:32 are being used. So,
1:12:40 you want to go listen to Weird Mary?
1:12:43 Have you all heard Weird Mary? I know
1:12:45 the irregulars like, "Oh, Weird Mary,
1:12:47 Kyle, could we could we not
1:12:49 um
1:12:53 share this tab instead?" Weird Mary
1:12:59 from Cedar Hill.
1:13:09 [Music]
1:13:28 Can't be singing.
1:13:33 [Music]
1:13:42 Cedar Hill is a quiet town
1:13:46 where the oak trees sway.
1:13:51 >> The visuals here, Kelly Bosch did the
1:13:53 visuals for this. She used Chat GPT to
1:13:56 make the images and then I think she
1:13:58 used Lumal Labs to animate them.
1:14:01 >> The whistle of the train echoes through
1:14:04 the day.
1:14:08 People greet you with a nod.
1:14:12 Nothing ever seems to change,
1:14:16 but the stories still unfold.
1:14:20 Each unique and strange.
1:14:23 [Music]
1:14:24 She grew up tough
1:14:27 in the town we called home.
1:14:32 Folks never took to her, no matter her
1:14:37 tone.
1:14:41 longing to be on the inside.
1:14:45 Longing for love just a little.
1:14:49 She found herself as a guest in the
1:14:52 nervous hospital.
1:14:54 >> Like like that that line right there.
1:14:57 She found herself as a guest in the
1:14:59 nervous hospital. Nervous hospital was a
1:15:01 line from Joker. Like throughout this
1:15:04 song, there are things that people
1:15:07 people humans with souls contributed to
1:15:10 this creative process,
1:15:12 right? Shadow lighting in the direction
1:15:15 of these visuals is chef's kiss, isn't
1:15:16 it? And this is like a year and a half
1:15:18 old. That song might benefit. You know
1:15:21 what's funny? Archetypal, I was just
1:15:23 thinking that. Um, let's go. Let's go
1:15:26 find Weird Mary. Did we do Weird Mary
1:15:28 and Sunno? I think we did. Oh, no. We
1:15:30 did it in Udo. That's okay. Um,
1:15:35 all right. Let me see if I can go find
1:15:37 it. Hang on a sec.
1:15:40 Weird Mary. We'll we'll upload it into
1:15:42 Sunno and remaster it. We did this in
1:15:45 UIO because Oh, that's right. Because
1:15:47 when we did this UDIO,
1:15:52 so
1:15:54 when it when it was in version three had
1:15:57 really bad um compression artifacts.
1:16:01 Everything sounded like autotune. It was
1:16:03 really bad. And Udo was clean. So when
1:16:07 when we made Weird Mary, Udo was my tool
1:16:09 of choice. And then when
1:16:12 Sunno 4 came out, it got better. And
1:16:15 then when 4.5 came out, it got good. And
1:16:18 then 4.5 plus came out, it was really
1:16:21 good. And now five's great. Um, let me
1:16:24 see if I can find Weird Mary. Weird Mary
1:16:28 Wave.
1:16:36 Okay, there it is.
1:16:39 That's it, right? Yeah. Okay. So, let's
1:16:44 go to
1:16:50 uh uh uh uh uh uh uh. So, we're going to
1:16:52 go plus audio. We're going to upload
1:16:54 some audio. Am I sharing this? I'm not.
1:16:57 Hang on.
1:16:58 There you go. Grab weird Mary. Weird
1:17:01 Mary.
1:17:20 All right, that's Flack. I guess I can
1:17:21 open Can it Can it open that? Yeah. All
1:17:25 right. Save.
1:17:28 All right. So, it's uploading that clip.
1:17:30 Let me share this tab. Okay.
1:17:33 So, here we are. So, we're taking So,
1:17:35 here's what we're going to do. Oh,
1:17:36 Erica's out of here. I got to jump. See
1:17:38 y'all tomorrow. All right.
1:17:41 Peace out, Erica. Good to see you.
1:17:44 Um
1:17:46 All right. So, we're So, we're uploading
1:17:48 the Weird Mary clip, and then we're
1:17:50 going to do a
1:17:54 We're going to do a remaster of this,
1:17:59 which should be interesting.
1:18:05 I've made thousands on my soulless
1:18:07 writing. Yeah, there you go. I think
1:18:10 that you are doing the job.
1:18:13 Wait. Oh, thinking that you are doing
1:18:16 the job then realizing AI is
1:18:20 mowing and how you create believers in
1:18:24 AI. Oh, how wait how it's moving.
1:18:28 Oh, how it's mowing and how you create
1:18:30 believers in AI. Listen, I
1:18:34 I am I am I am personally a technology
1:18:37 optimist and I am personally
1:18:42 like one of my passions in life, the
1:18:44 through line in my career
1:18:46 is finding ways to use new technology to
1:18:52 as a tool of self-expression for human
1:18:54 beings. Oh, wait. Continue. Okay.
1:18:58 Um,
1:19:00 so it's something I've been super
1:19:02 passionate about my whole life,
1:19:07 but I don't actually have a I don't love
1:19:11 or hate AI.
1:19:13 I'm excited to under
1:19:41 Um, this thing's not uploading right.
1:19:43 Something's broken here.
1:19:46 We're losing you. Oh, yeah. I thought
1:19:50 so. All right.
1:19:52 Um,
1:19:55 let me try something here. Open with
1:19:59 QuickTime player
1:20:03 and then let me export this as audio
1:20:06 only.
1:20:08 We'll do it to the desktop.
1:20:13 Okay,
1:20:16 coming back now. Yeah, something was
1:20:18 weird with that upload. It It was a
1:20:20 flack file and that's I think that was
1:20:23 bad.
1:20:25 Okay. So now now I've got it as a M4A
1:20:28 file.
1:20:30 Save.
1:20:33 Continue.
1:20:38 I don't know. Maybe it's maybe it's you
1:20:41 might lose me again here in a second.
1:20:46 was searching for this comment I left on
1:20:50 LinkedIn recently.
1:20:54 Uh I thought it was me. How are Okay.
1:20:56 Did that upload? Yeah, it did. Okay,
1:20:59 great. So now
1:21:03 we're going to go create remaster
1:21:10 variation strength normal or high.
1:21:15 Um,
1:21:18 we'll go normal.
1:21:21 Remastering clip.
1:21:23 The The other thing that's amazing about
1:21:25 this is how quick it does this. So, we
1:21:27 uploaded Weird Mary, and now here we
1:21:29 are. We're done.
1:21:36 There are only so many letters, but it's
1:21:37 how you arrange them. There are only so
1:21:40 so many chords, but it's how you play
1:21:41 them. That picture of the sunset you
1:21:43 took, it's a replica. It's not real,
1:21:45 right? Yeah. Exactly. Like
1:21:52 one of one of the one of the questions
1:21:55 that bugs the [ __ ] [ __ ] out of me
1:21:57 right now is, "Oh, is that AI?"
1:22:01 Because if your answer is, "Yes, I used
1:22:03 AI for that," then they completely
1:22:05 dismiss the work.
1:22:08 Even if you only used AI for some small
1:22:11 portion of it,
1:22:14 it's it just it just drives me [ __ ]
1:22:17 crazy. Anyway, let's listen to let's
1:22:19 listen to AI. How AI ruined this AI
1:22:21 song.
1:22:25 [Music]
1:22:28 That's cleaner.
1:22:32 Well, that's wild.
1:22:35 [Music]
1:22:46 [Applause]
1:22:50 [Music]
1:22:52 That's bad. That muddied it up.
1:22:57 [Music]
1:22:59 [Applause]
1:23:00 It's weird. It's like taking the
1:23:02 artifacts that were like bad audio and
1:23:05 turning those into
1:23:08 bad instruments.
1:23:14 [Music]
1:23:16 Yo, Andy, what's happening? You're
1:23:18 watching the live. We're having I don't
1:23:20 know how long you've been watching. I've
1:23:22 been ranting for about an hour.
1:23:25 [Music]
1:23:25 [Applause]
1:23:30 [Music]
1:23:47 Cedar Hill.
1:23:50 [Music]
1:23:53 >> Yeah, I don't like it. Okay, so let's
1:23:55 let's go back to this one. Let's Let's
1:23:59 experiment here. So, we're going to go
1:24:00 remaster. I'm going to go I'm going to
1:24:02 make this one subtle. So, the variation
1:24:05 strength here, I'm going to make it low.
1:24:08 And then I'm going to go down and I'm
1:24:11 going to remaster it. I'm going to make
1:24:12 it high.
1:24:14 And then we're going to do a cover of
1:24:16 it.
1:24:19 So, I'm going to go down and I'm going
1:24:21 to take the remix edit. I'm going to do
1:24:24 a cover.
1:24:29 And then the cover I'm going to do,
1:24:31 let's see, acoustic folk rock style in a
1:24:33 minor key with a male vocalist.
1:24:38 Oh, that's interesting.
1:24:41 So, when it imported this,
1:24:45 it wrote a description of the music.
1:24:52 Huh. All right. Right. So, I'm going to
1:24:54 do a version where I just leave that the
1:24:56 same
1:24:57 cover.
1:25:01 And then I'm going to do a version where
1:25:03 I just put
1:25:05 um Americana
1:25:09 um rock
1:25:13 with um
1:25:20 acoustic and electric
1:25:24 guitars.
1:25:28 Um, a solid beat. Wait, a solid beat
1:25:35 and uh,
1:25:38 let's see. And, um,
1:25:42 three-part harmonies
1:25:46 and we'll create that. Okay. So, let's
1:25:48 go see what the [ __ ] we've just created.
1:25:50 Okay. So, we had two versions that
1:25:53 sucked. So, these are the these are the
1:25:54 two. The next two here are
1:25:57 um low
1:26:00 low variation.
1:26:03 So this should sound like what we had.
1:26:07 [Music]
1:26:10 Yeah, that sounds more like it.
1:26:13 [Music]
1:26:38 Cedar Hill is a quiet town
1:26:42 where the oak trees.
1:26:45 >> That's not bad. That's that.
1:26:49 So there's so tomorrow, starting
1:26:51 tomorrow, there's a new
1:26:55 um there's a new thing coming from UDI
1:26:58 or from Sunno called Sunno Studio. And I
1:27:01 think premier members I think I'm a
1:27:03 premier premier member. Premiere members
1:27:05 get to play with this new studio. So,
1:27:08 what I'm going to be able to do is like
1:27:10 right now I feel like the mix in that
1:27:12 that early instrumental thing is is
1:27:14 weird, but I'm going to be able to go
1:27:16 into it and take like the guitar levels
1:27:18 and turn them down a bit.
1:27:21 So, like this, we might be able to go in
1:27:23 and actually edit these songs like you
1:27:25 know the way you edit a real song.
1:27:29 So, this is the second version of that
1:27:31 that low variation.
1:27:34 [Music]
1:27:55 [Applause]
1:28:00 [Music]
1:28:03 to the
1:28:06 Cedar Hill is a quiet town
1:28:10 where the oak trees sway.
1:28:15 The whistle of the train.
1:28:17 [Music]
1:28:19 >> All right. It's I like I can hear it.
1:28:21 It's good, but it's like
1:28:24 like I I feel like the original is just
1:28:27 fine. But I'm going to give that one a
1:28:29 thumbs up. Like that's that's a decent
1:28:30 one. Okay. Okay, so these next two are
1:28:33 with high variations. So these these
1:28:36 might get weird, but they might get
1:28:37 better.
1:28:42 [Music]
1:28:50 [Applause]
1:28:51 [Music]
1:29:01 That's cool.
1:29:03 [Music]
1:29:16 >> That's [ __ ] awful.
1:29:21 [Music]
1:29:26 Oh, that's bad.
1:29:28 [Applause]
1:29:29 [Music]
1:29:40 All right. Hate it. So now these are the
1:29:42 covers. So the first two the first two
1:29:45 are the covers where we left the
1:29:47 description what it interpreted out of
1:29:50 the song.
1:29:52 [Music]
1:30:02 Okay. So, so archetypal. Yeah. Kind of
1:30:05 feels like the mediocre quality was part
1:30:08 of the charm. It felt more vintage. I
1:30:09 agree with you. But listen to this one.
1:30:12 So, what we were doing with that other
1:30:15 ones was remastering it. So, it was
1:30:17 taking that effectively taking that
1:30:19 existing recording and trying to
1:30:21 recreate it. These two things are
1:30:23 covers. So, like I can already hear this
1:30:26 one. This one feels much cleaner to me.
1:30:28 And this feels like someone actually
1:30:30 covered covered the song.
1:30:35 [Music]
1:30:54 Cedar Hill is a quiet town.
1:30:57 >> This is good.
1:30:59 >> Where the oak trees sway.
1:31:03 The whistle of the train echo through
1:31:06 the day.
1:31:09 People
1:31:12 greet you with a All right, let's hear
1:31:15 this other one.
1:31:17 [Music]
1:31:48 Cedar Hill is a quiet town.
1:31:52 >> I love that melody variation. Cedar Hill
1:31:56 is a quiet town. Where the oak trees
1:31:59 sway,
1:32:02 the whistle of the train echoes through
1:32:05 the day.
1:32:08 >> Yeah, this one's good.
1:32:11 >> People greet you with a knot.
1:32:15 Nothing ever seems to change,
1:32:19 but the story still unfolds. Each unique
1:32:23 and strange.
1:32:24 >> This is really good.
1:32:27 She grew up tough
1:32:29 in the town we call home.
1:32:35 Folks never took to her no matter.
1:32:42 Longing to be on the inside,
1:32:46 longing for love just a little.
1:32:50 She found herself as a guest in the
1:32:53 nervous hospital.
1:32:56 This one's really good. This one's
1:32:58 really good. This is a really good
1:32:59 That's a really good version of the
1:33:01 song. All right. Now, let's hear. So, so
1:33:05 those two were both It basically took
1:33:08 the it analysis of the musical style and
1:33:12 applied it. So, this next one is I just
1:33:14 did a much more simple Americana rock
1:33:17 with acoustic and electric guitars,
1:33:20 solid beat, three-part harmonies. So,
1:33:22 these should be even more different. And
1:33:25 then we could also do this as a funk
1:33:27 song that maybe we'll actually let's do
1:33:29 that. We'll do um we'll do um
1:33:33 uh we'll do 70s funk.
1:33:37 Um uh
1:33:41 just 70s funk. We'll just do it as 70s
1:33:43 funk. We'll do it really simple. All
1:33:44 right. So here's the here's the
1:33:46 Americana version.
1:33:51 Listen
1:33:55 [Music]
1:34:10 to the drum. That That drum is just
1:34:12 That's a much more like straightforward
1:34:15 [Music]
1:34:24 Cedar Hill is a quiet town
1:34:28 where the oak trees sway.
1:34:32 The whistle of the train echoes through
1:34:35 the day.
1:34:37 [Music]
1:34:40 People greet you with a knot.
1:34:43 Nothing seems to change.
1:34:47 But the story still unfolds, each unique
1:34:51 and strange.
1:34:55 She grew up tough
1:34:58 in the town we called home.
1:35:03 Folks never took to her, no mattering.
1:35:09 [Applause]
1:35:09 [Music]
1:35:13 >> There's there's a three-part harmony.
1:35:17 [Music]
1:35:18 That's two-part harmony, but that's
1:35:19 cool. All right, here's the other one.
1:35:23 [Music]
1:35:24 >> I like this. I like this a lot.
1:35:29 [Music]
1:35:37 [Applause]
1:35:39 [Music]
1:35:56 Cedar Hill is a quiet town.
1:35:59 >> This one I like a lot.
1:36:02 Where the oak trees sway,
1:36:06 the whistle of the train echoes through
1:36:09 the day.
1:36:14 People greet you with a knot.
1:36:18 Nothing ever seems to change,
1:36:22 but the story still unfolds. Each unique
1:36:26 and strange.
1:36:28 [Music]
1:36:30 She grew up tough
1:36:33 in the town we call home.
1:36:38 Folks never took to her no matter her
1:36:43 tone.
1:36:44 >> All right. All right. That one's cool.
1:36:46 Um All right. Let's listen to the funk
1:36:48 one. And then I also did one while that
1:36:51 other one was playing with a female
1:36:52 vocal.
1:36:54 So that should be fun.
1:36:59 So this is 70s funk, right? Yeah.
1:37:03 [Music]
1:37:35 That's not 70s funk, but whatever.
1:37:41 [Music]
1:37:53 Yeah, neither of those really worked.
1:37:54 All right, let me let me try something
1:37:56 here.
1:37:58 I'm going to go We'll do rap jazz plus
1:38:02 uh studio recorded plus
1:38:06 [Music]
1:38:08 R&B.
1:38:10 All right, so there's that one. All
1:38:12 right. Now, let's go listen to I did two
1:38:14 with female female leads.
1:38:19 [Music]
1:38:25 And this was the the uh style here is
1:38:27 just Americana rock. That Americana
1:38:29 thing I wrote.
1:38:34 [Music]
1:38:40 Now, I'm I'm not remixing right now,
1:38:42 Gareth. I'm the These are all covers. I
1:38:45 found I didn't like the remixes. I The
1:38:47 the original recording of it had too
1:38:49 much like sonic mud in it that that the
1:38:53 remixes are trying to recreate and just
1:38:56 made it messy. So, the cover the covers
1:38:58 I like better.
1:39:03 [Music]
1:39:14 Cedar Hill is a quiet town.
1:39:16 >> I like it.
1:39:19 [Music]
1:39:26 >> She sounds like Chrissy Hines.
1:39:31 People greet you with a nod.
1:39:35 Nothing seems to change,
1:39:39 but the story still unfolds.
1:39:45 >> It's not Chrissy Hines. That's not how
1:39:47 these things work. But anyway,
1:39:51 [Music]
1:39:57 Oh yeah. Love that.
1:40:04 [Music]
1:40:14 [Music]
1:40:23 Cedar Hill.
1:40:25 [Music]
1:40:29 This sounds like Cheryl Crow.
1:40:34 [Music]
1:40:40 >> People greet you with a nod.
1:40:44 Nothing ever seems to change,
1:40:48 but the story still unfolds.
1:40:53 [Music]
1:40:56 She grew up.
1:40:58 >> Oh yeah.
1:40:59 >> In the town we call home.
1:41:04 >> Folks never took to her no matter her to
1:41:12 longing to be on the inside.
1:41:16 Longing for love just a little.
1:41:20 She found a guest.
1:41:26 >> That's a really good That's a good
1:41:27 version of that song. All right, two
1:41:29 more and then we'll call it a night. Um,
1:41:33 this is rap jazz studio recorded R&B.
1:41:44 [Music]
1:41:52 like this.
1:41:55 >> Oh, love that.
1:42:00 [Music]
1:42:06 [Music]
1:42:12 City Hill is a quiet town
1:42:16 where the oak trees sway.
1:42:19 [Music]
1:42:20 The whistle of the train echoes through
1:42:24 the day.
1:42:28 People greet you whether or not.
1:42:32 Nothing never seems to change,
1:42:36 but the story still unfolds.
1:42:39 [Music]
1:42:41 >> That's [ __ ] hot. I love this.
1:42:48 [Music]
1:42:49 >> See, it [ __ ] that up. It [ __ ] up the
1:42:51 timing there.
1:42:54 [Music]
1:43:00 Yeah, that's not usable. I mean, you get
1:43:03 what's cool about Sununo is you can go
1:43:05 in and you can just rerender that
1:43:07 section of audio,
1:43:09 which might be, but let's let's hear
1:43:10 what the other one sounds like.
1:43:13 Like that.
1:43:16 [Music]
1:43:46 Dude, it did a sack solo of the of the
1:43:49 [ __ ] melody. Oh my god.
1:43:54 [Music]
1:44:03 Sist
1:44:06 [Music]
1:44:17 good side hustle, Mimi. Okay. So, wait.
1:44:20 So, we're going to we're going to keep
1:44:21 all of these because I like this overall
1:44:23 sound. So, rap jazz studio recorded R&B,
1:44:26 but then we're going to put in
1:44:27 Appalachian
1:44:29 influence. That's really good.
1:44:30 Appalachian.
1:44:34 I'm glad you spelled that influence.
1:44:38 Okay. Appalachian.
1:44:41 Oh, I spelled it wrong still. There. Oh,
1:44:43 I needed a capital. Okay. Um, so so
1:44:46 we'll do two more. All right. Two more.
1:44:47 But let's keep listening to this one
1:44:49 because this this one's really tight.
1:44:52 [Music]
1:44:58 >> Nothing seems to change,
1:45:02 but the story still unfolds each unique
1:45:05 and strange.
1:45:10 >> She grew up
1:45:12 in the town called home.
1:45:16 >> Yeah.
1:45:18 folks.
1:45:22 >> Listen to her. She's in the [ __ ]
1:45:24 pocket here.
1:45:28 Longing to be on the inside,
1:45:32 longing for love just a little.
1:45:36 She found herself as a guest in the
1:45:39 nervous hospital
1:45:41 [Music]
1:45:44 with Mary from Cedar Hill.
1:45:48 She's
1:45:50 blue
1:45:52 sadness and shame.
1:45:56 All right. I think that one's my
1:45:58 favorite one so far. That one's [ __ ]
1:45:59 sick. All right, that one's not
1:46:01 rendering. Oh, yeah. Rendered. It's just
1:46:03 interface problems. Let's try that one.
1:46:10 No. No, it's dead. All right, that one's
1:46:14 dead.
1:46:16 This one's alive.
1:46:19 All right, this is
1:46:22 this is the one where we added
1:46:23 Appalachian Influence at the end.
1:46:26 So, rap jazz studio recorded R&B
1:46:30 Appalachian influence. Let's see how it
1:46:32 does.
1:46:36 [Music]
1:46:48 Oh, nice
1:46:51 horn section.
1:46:55 [Music]
1:46:59 Hill.
1:47:02 >> Damn.
1:47:06 [Music]
1:47:09 >> That just had a weird bad chord.
1:47:15 [Music]
1:47:18 >> That's bad.
1:47:21 [Music]
1:47:27 All right, I'm already getting the
1:47:28 thumbs up.
1:47:32 [Music]
1:47:38 [Applause]
1:47:38 [Music]
1:47:50 This one's making champ.
1:47:52 [Music]
1:48:09 >> Damn.
1:48:12 People greet you with a
1:48:17 seems to change.
1:48:20 >> This is kind of dirty. I like it.
1:48:27 >> She going to break your [ __ ] heart.
1:48:30 [Music]
1:48:36 >> Yeah. The these two are both the ones
1:48:38 with Appalachian influence.
1:48:41 home.
1:48:46 >> Yeah, [ __ ] that. That that [ __ ] up the
1:48:47 timing pretty bad.
1:48:50 [Music]
1:48:52 >> Just a little
1:48:56 strange.
1:48:59 She grew up in the town called home.
1:49:04 Folks
1:49:05 >> Yeah, that like that one's not usable.
1:49:07 That that's I'm gonna give that a thumbs
1:49:08 down actually.
1:49:11 Well, which made it disappear. That's
1:49:12 fascinating. Okay,
1:49:15 [Music]
1:49:16 this is the one with Appalachian
1:49:17 influence. I think this one's pretty
1:49:19 good.
1:49:22 [Music]
1:49:34 Oh, yeah. That's one of the other things
1:49:36 you can do in Sunno. And I haven't
1:49:38 played with this much.
1:49:39 side hustle. Mimi said she she liked the
1:49:42 Oh, bye Cam Katkin. Um,
1:49:45 she liked that voice. You can take any
1:49:47 of these voices and turn them into a
1:49:50 character, a persona, and then use that
1:49:52 voice in other songs. So, that might be
1:49:54 something to play with, too.
1:49:58 [Music]
1:50:02 Love that.
1:50:07 [Music]
1:50:11 Hill is a quiet town
1:50:15 where the oak trees
1:50:17 [Music]
1:50:21 >> one's got that weird chord. All right.
1:50:23 So, which was our favorite? I think it
1:50:24 was this one.
1:50:28 [Music]
1:50:33 [Applause]
1:50:38 But the story still unfolds.
1:50:44 The story still unfolds.
1:50:46 [Music]
1:51:09 [Music]
1:51:17 Cedar Hill is a quiet town
1:51:21 where the oak trees sway.
1:51:24 [Music]
1:51:26 The whistle of the train.
1:51:30 >> All right. Well, we could do this all
1:51:32 night. Anyway, um yeah, this is
1:51:37 this is fascinating. And you know,
1:51:40 again, it's like I go back to,
1:51:44 you could absolutely argue that this is
1:51:46 soulless music production, but I'm going
1:51:48 to go back to, well, we took a song we
1:51:51 put a lot of time into, we produced,
1:51:54 there was a lot of ideas and inputs in
1:51:56 it. It's a song we actually know and
1:51:59 love because it was a good song
1:52:01 independent of how it was made. And now
1:52:03 we're tweaking it, modifying it, pushing
1:52:07 it into different areas.
1:52:09 this is music production. Like I don't
1:52:12 the fact that
1:52:13 we don't have musicians,
1:52:16 you know, running loops in a DAW doesn't
1:52:19 mean it's not music. So anyway,
1:52:21 whatever.
1:52:24 That's a that's again there's like
1:52:28 the theory of what this is you can be
1:52:32 enraged at when you actually use these
1:52:34 tools.
1:52:36 The fact that it [ __ ] up some of the
1:52:38 chords and it [ __ ] up the timing of
1:52:40 the things. Like that's like that's just
1:52:43 like a musician [ __ ] that up. It's
1:52:44 like were we gonna save the recording? I
1:52:46 don't know. Was it good enough for us to
1:52:47 save? Maybe, maybe not. Maybe we go try
1:52:49 something else. Anyway, crazy. Yeah, we
1:52:52 have no no union scale.
1:52:55 We just have my time. Uh and all of
1:52:57 yours. Okay, cool. Um why aren't
1:53:00 musicians cloning their voice to
1:53:02 prototype songs? I'm sure some are. I
1:53:04 mean, listen, some, you know, some
1:53:07 people are jumping on this stuff. There
1:53:08 there's producers out there. Timberland
1:53:11 is a producer that's all over this right
1:53:13 now. Um, so anyway, all right, I got to
1:53:16 get out of here. It is Wednesday.
1:53:18 Tomorrow's Thursday. I don't think
1:53:19 anything's going on. I'll see you at 8
1:53:21 o'clock tomorrow night. Have a fantastic
1:53:24 evening. Peace out.
1:53:28 [Music]