10/6/2025 - Exploring OpenAI's New App Integrations and The Rise of The Everything App

Description

Chapters

Transcript

0:08 [Music]
0:11 Yay!
0:19 [Music]
0:29 [Music]
0:33 It's not simp.
0:40 [Music]
0:51 It's not easy to know
0:54 I'm not anything
0:56 like I used to be. Although it's true. I
1:00 was never attention sweet Santa. I still
1:04 remember that girl.
1:07 She's imperfect but she tries.
1:11 She is good but she lies.
1:15 She is hard on herself.
1:19 She is broken and won't ask for help.
1:23 She is messy. She's kind.
1:28 She's lonely.
1:30 Most of the time she's all this mixed
1:34 up.
1:36 Beautiful.
1:38 She's gone, but she used to be mine.
1:42 [Music]
1:53 [Music]
1:57 Happy Monday, good people of the world.
2:01 Good lord, was I busy today. I am a
2:04 quivering mess of jello. I uh do not
2:08 expect me to be coherent tonight
2:11 or have patience
2:18 and none of it. I will have none of it.
2:21 I will have none of it.
2:25 Um
2:27 I got to watch the OpenAI announcement
2:31 today
2:32 without listening to it. Well,
2:34 technically I had it on half volume in
2:37 the
2:39 um presentation I was supposed to be
2:40 paying attention to at half volume. So,
2:42 I paid attention to neither of them and
2:45 I got a headache
2:47 because even for my ADHD brain, it was
2:50 too much.
2:53 It was hilarious.
2:56 At some point, I just I just checked out
2:58 and the headache went away.
3:04 Oh man.
3:07 [Music]
3:23 10,000 words swarm round my head. The
3:26 million more in books written beneath my
3:30 bed.
3:31 [Music]
3:35 I wrote or read them all. Went searching
3:38 in the swamp. Still can't find how to
3:42 hold my hands.
3:47 And I know you need me in the next room
3:51 over. I'm stuck in here on paralyzed.
3:56 [Music]
4:00 For months I got myself in in wait. For
4:04 months I got myself in too mucha.
4:11 I don't remember the words.
4:16 Ain't it like most people? I'm no
4:18 different. Like to talk on things we
4:20 don't know about.
4:22 [Music]
4:24 [Applause]
4:26 Well, ain't it like most people? I'm no
4:28 different. I like to talk on things they
4:31 don't know about.
4:32 [Music]
4:48 [Applause]
4:51 [Music]
4:53 [Applause]
4:55 [Music]
5:13 [Applause]
5:15 Ah, good lordy. Good lord. Good lordy.
5:19 Good lordy. Good lordy.
5:24 [Music]
5:45 [Applause]
5:45 [Music]
6:00 Good morning. Good lord. Good lord. Good
6:03 lord. Guess I will stop making Sora
6:06 videos to watch Kyle. No, don't do it,
6:08 Tobias. Just keep making Sora videos.
6:14 That is the much better choice in life.
6:17 You can make videos of me if you want to
6:19 take it. If you want to go meta,
6:27 do me a favor. If you're on TikTok,
6:29 share the live so we can get some more
6:31 folks in here. Um,
6:38 oh man. Um, if you're new here, my name
6:41 is Kyle Shannon.
6:43 Hal K Y L E.
6:48 [Music]
6:55 You know, it just hit me when I said
6:57 cow. My name is Cal.
7:00 Um I have I have lived in places and I
7:03 have been to places that when they hear
7:05 my name,
7:07 they're like,
7:10 "We don't know how to say that.
7:14 Cow. Cal. It's Kyle. Cow. No. Kyle.
7:22 Cow.
7:26 Like, do you think it's like a genetic
7:28 thing where some people's tongues just
7:29 can't say that? Tik Tok question.
7:34 You get notifications when someone uses
7:36 your cameo on Sora. Yes. Well, you don't
7:39 so much get notification. you might get
7:41 a notification. I think there might be
7:43 something in in there. Um,
7:47 you get notifications if someone
7:48 commented on a video that you've been
7:50 cameoed in. But if you go to your
7:52 profile, you've got your videos and then
7:55 you've got cameos.
7:57 And the cameos button lets you see here.
8:00 Let me flip flip my cameras here. Let me
8:03 flip my little cameras over. the uh the
8:06 cameo button
8:08 in your profile shows you all the cameos
8:11 you've done of yourself, but also cameos
8:14 that other people have done of you.
8:17 Um,
8:21 you know, it it's it's kind of a funny
8:23 it's a funny kind of thing because I I
8:25 put myself out there a lot. I'm I'm used
8:27 to seeing my stupid fat face and and
8:30 seeing like, you know, my my initial AI
8:33 project was called Kyle Shannon Dreams
8:35 and it was kind of like this twisted
8:37 self-portrait sort of thing. So, I'm
8:39 used to seeing weird versions of me.
8:43 But when people from in here largely
8:48 from the salon and from in here started
8:50 making cameos of me. Um
8:54 there's kind of a cool kind of
8:58 when like when you give people
8:59 permission to do that. I mean I suppose
9:02 it could be used nefariously. I guess
9:04 I'm I'm leaning on there there's two
9:07 there's two places my head are. If Sam
9:08 Alman can do it like no one gives a [ __ ]
9:10 about me. So, um
9:14 it's a it's kind of it's a kind of
9:16 interesting brand extension.
9:19 Um we created about six months ago in
9:22 the AI salon. We created a new branding
9:24 look and feel like a new style for our
9:27 images
9:28 and we did that in midjourney and we we
9:32 came up with a specific mood board for
9:35 that image style and we gave that mood
9:38 board out to the community and it's it's
9:41 the first time it really hit me that
9:44 there's a lot of times comp no most
9:47 times companies that have brands will
9:50 have a brand book and the brand book is
9:52 here's how you use our brand
9:55 and it's like, you know, here's the
9:56 colors, here's the this, here's the
9:57 photos, whatever. And and then maybe
9:59 it's even like, you know, if you're
10:01 going to use stock photos, here's the
10:02 rules.
10:05 But
10:06 what giving people that code allowed
10:09 them to do or allows them to do in
10:11 present tense is to create imagery
10:14 that's on brand for the AI salon. And I
10:17 kind of feel like these cameos are kind
10:19 of a similar kind of thing where in in a
10:22 great sense we're co-creating content,
10:25 right? Like I've said, feel free to use
10:27 my my visage.
10:32 He's carry on. Feel free to use my vage.
10:37 My visage
10:40 is available for you to cameo at your
10:43 leisure.
10:45 And uh
10:47 and so it's a weird in a weird way it's
10:49 kind of a brand extension and it's kind
10:51 of it's kind of fun watching them like
10:53 you know my my request I put out a video
10:56 that said feel free to use my [ __ ] Um
10:58 and I basically said just be creative
11:00 with it right come up with interesting
11:02 stuff. Don't just be derivative like
11:04 everyone else is being
11:07 [Music]
11:10 source camp
11:12 totally lost track of time. What were
11:14 you building? New agents in the new
11:15 OpenAI agent builder.
11:18 I don't know if y'all saw that.
11:21 At least you don't work at Zapier
11:30 or NAN.
11:32 [Music]
12:09 Um,
12:11 let's see. Kyle, now that my siblings
12:14 are on Sora, I control them now. Not
12:17 just you. That's good. Thank goodness.
12:21 Um, but I kind of I kind of really dig
12:24 the uh I dig the the the uh the things.
12:27 Let me actually jump over to Sora for a
12:29 second. Let me see if they made their
12:31 maybe they made their website more
12:33 stable. I doubt it because they were
12:36 probably busy getting ready for
12:37 developer day. Um I saw a really
12:39 interesting um what you call it? I saw a
12:44 really interesting
12:47 um
12:50 tweet today
12:54 that was
12:56 um
12:58 it was some version of like when is open
13:01 AI going to realize that developers are
13:03 not their core target audience
13:07 and
13:09 and so I replied to it and I said I said
13:12 yeah they you know the the day after
13:15 develop er day should be AI for the rest
13:17 of us day, right? So I leaned into AI
13:19 festivist branding there. Um
13:24 but it's actually a really interesting
13:26 point and you know we could wait for
13:29 open AI.
13:31 Actually I'm I'm having a brainstorm in
13:34 real time right now. Here's an idea.
13:38 What if we as the AI salon and producer
13:41 Brandon, grab grab this idea because I
13:44 think this is something worth worth
13:46 exploring,
13:48 which is what if we put together a
13:55 Are you viewing that? Yeah, I think
13:57 you're viewing that. Okay. What if you
13:59 put together a There's There's Danielle
14:01 and Vicki in the upper leftand corner.
14:03 That's awesome.
14:05 Um,
14:07 what if we put together an AI salon
14:11 branded AI for the rest of us day in
14:14 conjunction with developer days? So,
14:16 it's either on the same days as
14:18 developer days or it's like the it's
14:19 like the day after or something like
14:21 that
14:23 because
14:24 I I think
14:26 [Music]
14:28 I think how we use these tools is as
14:31 important as as uh how we program with
14:34 these tools. It's not just about
14:36 developing applications. And with vibe
14:38 coding getting more and more accessible,
14:42 um the distinction between who's a
14:45 developer and who's a not not is getting
14:47 really blurry. Um so there's going to be
14:49 a lot of people trying to figure out,
14:52 you know, how to how to do stuff. I
14:54 don't know. Just thinking in real time.
14:56 Just thinking in real time here, people.
14:58 But here's what struck me today about
15:01 the
15:04 the visual agent builder. First of all,
15:09 I still can't quite tell
15:12 because I didn't get a chance to
15:13 actually watch the thing and I haven't
15:14 had a chance to play with it because I
15:16 had a really busy day today. Um,
15:21 I can't quite tell
15:24 if what the agent builder from it's
15:28 called agent kit. if agent kit from
15:30 OpenAI is an agent builder or it's a an
15:33 automation builder. Um, so that'll be
15:38 curious to find out. And the the
15:39 difference I I assume it's an agent
15:41 builder because they talk about agents
15:43 all the time. And an agent is an
15:45 autonomous, you know, thing that can
15:47 autonomously run and use tools. And I
15:49 know that you can you can call tools
15:51 from within it. So I assume it's truly
15:54 an agent builder, not just an automation
15:56 builder. Um, but I don't know. But
15:59 here's my instinct. 11 Labs today also
16:02 came out with a visual
16:05 um workflow builder for 11 Labs.
16:09 And
16:11 while I tend I tend to not lean into um
16:15 developer stuff on this channel, my gut
16:18 is telling me we need these skills. Like
16:21 we as non-developers
16:24 probably need to if we haven't already
16:26 like I I know there's a fair amount of
16:27 people in the AI salon that that you
16:29 know do a lot of automations and stuff.
16:30 I've done a fair amount of automations.
16:33 Um they're just not fun to watch build.
16:37 But that said it kind of this is feeling
16:40 like an important moment.
16:43 [Music]
16:45 It may not be
16:47 [Music]
16:56 but I don't know
16:58 [Music]
17:01 the other thing that they announced
17:02 today we'll play with a little bit is uh
17:07 is app app app app app access within
17:11 chat GPT T. And the minute I saw it, the
17:17 minute I saw it,
17:21 what popped in my head is, well,
17:25 we're not going to need uh
17:28 we're not we're not going to need uh
17:31 we're not going to need websites for
17:33 much longer now. We'll still have
17:36 websites just like we still have the
17:38 theater, even though we can do
17:40 storytelling with film and TV now. and
17:42 YouTube and Netflix. We don't need the
17:44 theater. I think that's how websites are
17:46 going to go. We don't need websites
17:48 other than as a source of data for
17:50 things like what they just built. Um
17:56 I think what all all of the uh
17:59 all of the frontier model companies
18:01 right now are racing,
18:04 they're not racing to a ASI, artificial
18:07 super intelligence.
18:09 What they're racing for is to build all
18:12 of the components
18:14 of the everything app. And the
18:15 everything app will be you log into
18:17 basically what they're trying to create
18:19 is their their own AOL.
18:22 They want to create their own America
18:24 online from the early 90s where the
18:28 entire internet existed within this
18:30 walled garden called AOL and then the
18:33 real internet sort of burst out through
18:35 it and and you know took off on its own.
18:38 Um, but I feel like we're going in the
18:39 reverse direction where they're trying
18:40 to build all these components where
18:42 we're just going to do the stuff within
18:43 their little walled garden. Um, it's
18:47 going to be really interesting to see
18:48 how that played out. Um, there was a Tik
18:50 Tok pin there that I missed. Can someone
18:52 repin that?
18:57 Side hustle Mimi, I agree. I think it
19:00 will be inevitable to learn to swim in
19:02 these waters. Yeah. Well enough to be
19:05 useful. Yeah, that's that's the that's
19:07 the that's exactly my instinct side
19:09 hustle is um
19:12 like we don't need to go deep. Th this
19:14 is not about turning this channel into a
19:17 developer channel or like you know like
19:19 like Riley Brown now, you know, went
19:22 from talking about apps to building
19:25 automations to vibe coding apps and now
19:28 he's vibe coded a vibe coding app
19:32 and and now he's hawking he's hawking
19:34 some new piece of software they're about
19:36 to launch. Like he started a software
19:38 company. I'm not suggesting we go there,
19:41 but
19:43 I am suggesting that it's
19:47 it's making some really powerful tools
19:50 easy enough to access and and to
19:53 develop, you know, something truly
19:55 interesting with that it's it's probably
19:56 worth our while to to do it. Um,
20:01 archetypal architect. I'm starting to
20:03 get that feeling, too. The broader
20:04 development skills might be necessary.
20:07 Yeah, just yeah, just even awareness
20:09 awareness of other skills um
20:13 and like what's needed and and and you
20:16 know what it is archetypal is it's
20:18 almost like um
20:23 for me personally I want to know where
20:25 the boundaries are of my interest. I
20:28 want to know how far back I want to go.
20:30 So so what happens is this. You've got
20:33 you've got software users out here,
20:35 right? You've got developers back here
20:38 and historically that was a pretty broad
20:40 chasm, right? If you had an idea for a
20:42 business that involved software, you
20:45 needed to find a technical co-founder.
20:47 Now with Vibe Coding, you can have an
20:49 idea for a business and you can kind of
20:52 develop your own thing. But then there's
20:54 a point at which if you want it to be a
20:56 real scalable app, you still need a
20:58 developer there. these worlds are coming
21:01 together
21:02 and I kind of want to figure out for
21:04 myself how far back do I want to go and
21:07 then how far
21:09 towards me is the development world
21:12 coming right like lovable just took a
21:15 very strong step in the direction of
21:19 anyone can build an app by creating
21:21 their own backend their own cloud
21:24 service right
21:26 um Gemini is free and lovable for
21:29 another week. Okay, that's great. You
21:31 know what I don't quite get about that,
21:33 Brandon, is I get that you can develop
21:35 an app with it,
21:38 but they still don't have the ability to
21:40 monetize built into Lovable. So, if you
21:43 wanted to turn your Lovable app into a
21:46 an e-commerce app, you're still on your
21:48 own for integrating that. So, having
21:50 that [ __ ] for free is good for
21:52 developing the app, but you can't really
21:53 then put it in the world. So, I I don't
21:55 know. I don't get it.
21:57 Um, okay. You've got mail. Exactly.
22:03 [Music]
22:06 On your screen. That's my brother having
22:08 his bo his face painted by Bob Ross.
22:11 That's awesome.
22:17 H. Easy enough to integrate with Stripe
22:20 is what producer Brandon just said.
22:22 Here's where I'll push back. It's easy
22:24 enough to integrate with Stripe if
22:27 you've got a developer mindset, if
22:29 you're used to thinking in developery
22:31 kind of stuff.
22:33 But if but if what if what lovable just
22:35 told me is you can come in here and you
22:37 can build and launch a a you know a
22:41 usable you know application
22:44 and then at at some point I get to I've
22:47 gota I've got to figure out how to
22:49 connect that to Stripe. It's just it's
22:52 like a it's that extra step that extra
22:55 well you just go over here and you sign
22:57 up for an account and you configure it
22:59 and you do the
23:02 I mean if you've ever tried to set up a
23:04 Stripe account it's there's a lot you
23:07 got to set up tax [ __ ] and banking
23:09 account [ __ ]
23:13 and like that's enough that it'll
23:14 that'll break someone's brain who's not
23:16 used to being a developer. So again, I'm
23:20 waiting for the first of these vibe
23:22 coding apps to show up to be actually
23:24 really good. Like really truly
23:27 launch a sellable application, either a
23:31 mobile app or a web app
23:34 or start a SAS SAS company um with just
23:38 an idea. That's not there yet. So, but
23:41 anyway, that's where I think we should
23:43 play. Okay. Anyway, so I'm on the Sora
23:45 website. We'll see if this crashes
23:46 within five minutes because that's been
23:49 the historical trend. And it looks like
23:51 it's already crashed just sitting here.
23:56 Yep. Oh, you can't see that. You can't
23:59 see the page unresponsive error. You can
24:03 wait until the page becomes responsive
24:05 or exit the page. I'm gonna exit the
24:08 page.
24:09 Oh, snap.
24:15 Oh, open AI, what are we gonna do with
24:18 you?
24:20 Okay, let's go look at
24:24 Okay, so I'm on
24:30 So, I'm on my um my page, my profile
24:34 page.
24:37 >> I've been This is
24:44 If I go to cameos
24:47 >> London after midnight.
24:49 >> See like I didn't make
24:53 >> this.
24:55 I didn't make this. I didn't make this.
24:57 So these are all cameos people have made
24:59 of me. I haven't seen most of these
25:00 actually.
25:02 >> Switch to. Looks innocent, right? This
25:04 is the new paper cup they made us switch
25:06 to. Looks innocent, right? Watch what
25:07 happens after 5 minutes. Burns your
25:09 hand. It caves in. It leaks. It hurts.
25:11 Remember this?
25:13 >> This is the new paper cup. They made a
25:15 switch.
25:16 >> Steer through thoughts and across this
25:18 glassy blue I ride with engines hum and
25:20 open tide. Each wave of verse, each
25:22 breeze a rhyme. I steer through thoughts
25:24 and capture time. A heart that's
25:26 grounded yet. Across this glassy blue I
25:29 ride with engines hum and open tide.
25:31 Each wave of verse. Each breeze.
25:36 >> Hi,
25:37 I'm Kyle.
25:39 I
25:41 love
25:43 AI.
25:45 Hi,
25:47 I'm Kyle.
25:49 Across this glass comes up.
25:50 >> That's incredible. All right, let's see
25:52 what kind of beauty is hiding in here.
25:55 Whoa, look at them. They turned into
25:56 butterflies. That's incredible.
26:01 >> And I let it be just a simple song. It's
26:03 enough for me. Every string I play
26:05 carries little hope to cope after
26:07 midnight. London after midnight. The air
26:09 is thick enough to smear on a window and
26:11 the stones shine like polished slate.
26:12 The man ahead of me, long coat, brim
26:14 pulled low. He's the thread and I'm not
26:16 letting go. You feed the trail with p
26:17 London after midnight.
26:18 >> HD at HD's got me doing uh lots of uh
26:23 >> like a race now. They're more like a
26:26 tuneup. A few quiet minutes to get the
26:31 >> go back to cameos.
26:34 So these are all mine. You can make
26:36 money with chat GPT. Yeah. Yeah. Make
26:40 money with Chad GPT. Turn that idea into
26:43 gold.
26:44 >> It is kind of uncanny. And then I can go
26:46 into drafts. So drafts featuring your
26:49 cameo.
26:51 >> Stuff people have made that they haven't
26:53 launched yet.
26:53 >> Now they're more like a
26:55 >> There's Who did this one? Todd made
26:57 something with Vicki.
27:01 >> This is insane. You said they were
27:03 friendly.
27:03 >> I said they had good energy. I did not
27:05 say they'd steal a parade float.
27:06 >> Forget the float. They don't know how TO
27:08 DRIVE A STICK.
27:09 >> BEST ROAD TRIP EVER. WOO! WE'RE FAMOUS.
27:11 >> This is insane. You said they were
27:13 friendly.
27:13 >> I said they had good energy. I did not
27:15 say they'd steal a parade float.
27:16 >> Forget the float. They don't know how TO
27:17 DRIVE A STICK.
27:18 >> BEST ROAD TRIP EVER. WOO! We're famous.
27:21 >> Hang on a sec.
27:24 By the way, Jake Paul has a cameo. You
27:26 can you can do cameos with Jake Paul
27:28 now. Those are kind of funny. Oh, and we
27:32 crashed. Did we crash? Maybe.
27:41 Oh, this looks funny.
27:45 Come on.
27:47 Mornings used to feel like a race. Now
27:49 they're more like a tuneup. A few quiet
27:52 minutes to get the engine running again.
27:57 Mornings used to feel like a race. I'm
27:59 going to show you how I finish this
28:00 plate. You start with warm sushi rice
28:01 shaped into ngiri, salmon, tuna, shrimp,
28:03 and a sweet omelette. Set them in a
28:04 little curve so every piece is easy to
28:05 pick up.
28:05 >> Beautiful. How do you keep the fish so
28:06 shiny?
28:07 >> I brush it with a touch of nigiri soy.
28:14 >> Drag me from the ground. Feel the rhythm
28:17 in my bones. But moving
28:21 every
28:23 broken
28:28 [Music]
28:30 Anyway, that's that. The this site is
28:34 still so buggy that it's it's really
28:36 hard to demo [ __ ] because it's just it's
28:37 really bad.
28:40 But uh here's here's here's one I made
28:43 of my writing partner. He's not too just
28:45 crashed.
28:46 >> Let the fellow folks whisper.
28:49 >> Let's see.
28:52 Yeah, she's she's gone ahead and
28:54 crashed.
28:59 All right. Um, enough of that comedy
29:02 nonsense.
29:04 Um, thoughts on that. The app is more
29:06 stable. The app is definitely more
29:08 stable.
29:09 It never gets my accent and other
29:11 people's cameos. Yeah, there's there's a
29:14 bunch of weird stuff. In fact, I saw
29:15 someone made a Sora video that was
29:19 someone telling Sam Alman that it just
29:21 can't get the voices. It's it assigns
29:23 the voices to the wrong people all the
29:25 time. And
29:27 I don't I don't know if you've
29:28 experienced this, but one of the things
29:31 that I've been experiencing is I'll be
29:34 flipping through Sora videos
29:38 and mostly they're just stupid. Mostly
29:40 I'm not watching them, but then
29:41 occasionally one will come up and it
29:45 it's realistic enough that for a moment
29:48 I'm like like I take it in as real
29:51 content and then like I'll scroll to the
29:54 next video and I'll realize oh right I'm
29:56 not in Tik Tok I'm in Sora. I lose track
29:59 of the fact that I'm looking at AI
30:02 generated video
30:05 because I'll see like, you know, some
30:06 something with someone familiar to me in
30:08 it and then I'll see like Sam Alman
30:11 saying something or someone else talking
30:12 to Sam Alman and it just like flips my
30:15 brain into like this is real mode. It's
30:17 bizarre.
30:20 It really is bizarre.
30:24 Yeah, it's
30:28 we're gonna So, here's a really
30:31 interesting Here's a really interesting
30:35 um
30:38 Well, I don't know if it's really
30:39 interesting. It's really interesting to
30:41 me. A really interesting
30:45 transition that we're about to make. So
30:47 I don't know when it was when when when
30:50 painters,
30:52 you know, painted, you know, hyper
30:54 realistic stuff. It was the Renaissance,
30:56 right? When they started doing like like
30:58 m, you know, forced perspective and and
31:00 and
31:03 they tried to paint reality
31:06 and and that remained so for a good long
31:09 time. And then and then the camera came
31:10 out
31:12 and it allowed painters to not have to
31:14 capture reality.
31:16 and and then they shifted into,
31:21 you know, abstract and and, you know,
31:25 surrealism and and and things that
31:28 weren't about capturing reality because
31:30 they didn't have to capture reality
31:31 anymore.
31:34 And I'm sure the painters, you know, the
31:36 the the painters of the of the 16 and
31:40 1700s, if they saw, you know, Jackson
31:43 Pollock or Andy Warhol or, you know, or,
31:47 you know, any any kind of the modern
31:50 abstract impressionists,
31:55 like
31:57 they would just be like, "That's
31:58 garbage. That's not painting."
32:01 And I kind of feel like that's where we
32:04 are with with video right now where like
32:09 we have films, right, that we know are
32:13 fiction. We we have books that are
32:15 non-fiction and we have books that are
32:17 fiction. And there's like and we can
32:20 coexist in that world and it doesn't
32:21 seem to bother us
32:24 when we read Lord of the Rings that it's
32:26 not real. We we actually kind of like
32:29 it.
32:31 And we can go to a movie about a man
32:33 that flies through the sky and he's he's
32:35 a reporter during the day and he saves
32:37 people in the afternoon. We we can hold
32:40 that in our consciousness. But somehow
32:43 right now we're in this weird thing
32:45 where people are like, "Well, AI video
32:47 like that's that's you know, it's just
32:50 deep fakes." No, it's actually not. I
32:53 mean, it can be.
32:55 It can be a deep fake, but it could also
32:58 just be entertainment. It could also
33:01 just be,
33:03 you know, um a fictitious story that
33:06 happens to have someone that looks like
33:09 you.
33:11 And that was that was actually what I
33:12 kind of explored explored.
33:16 Okay. AI Okay. Jay Hamilton. AI
33:20 generated content should be forced to be
33:23 labeled as AI. Why?
33:26 Why?
33:29 Here's what I would argue with you.
33:33 Content that is designed to deceive
33:36 someone should be labeled as deceptive.
33:40 That I'll give you.
33:43 But just because it was created with an
33:44 AI tool, why should that be labeled?
33:47 I can make I can make a deceptive video
33:50 without using AI.
33:53 And that would be called fraud or
33:55 defamation.
33:58 And if if you think that we're going to
34:01 be able to tell the difference,
34:04 and if you think that every video that
34:06 you see is either going to be purely AI
34:10 or purely not AI, that's not even the
34:13 case today.
34:16 Like if you want to actually do
34:18 something with video today and you use
34:22 generative AI tools, chances are really
34:24 good you're going to be going into
34:25 Photoshop to to to modify the images
34:29 that are used as the source. You're
34:31 probably going to be working within
34:33 Premiere or After Effects to tie things
34:37 together.
34:39 So I might use AI in a video output, but
34:43 that doesn't mean it's an AI video. I
34:45 think what you're getting at the fear I
34:48 understand
34:50 and relate to, but it's it's not about
34:53 the tool, it's about the use,
34:56 right?
34:58 Why should it be labeled?
35:01 Because of it being misused as
35:03 deceptive. But yeah, but but you're but
35:06 but but your your logic is is skewed to
35:11 assume that if it's AI, it's going to be
35:15 nefarious.
35:16 That's [ __ ] Hollywood.
35:19 That's what Hollywood has taught us for
35:22 50 [ __ ] years. The robots are going
35:23 to kill us. AI is evil,
35:26 right? Don't fall in love with a robot.
35:31 Just because it's AI doesn't mean it's
35:33 evil. If it's evil,
35:36 it should be prosecuted. Or if it's set
35:39 to deceive, it should be called out.
35:42 Or if it's set to be like like, you
35:45 know, if it's said to be parody,
35:49 why can I not make fun of someone?
35:52 Saturday Night Live for 50 years has
35:54 been
35:56 making it look like real people do
35:58 ridiculous things. That's parody. That's
36:00 covered. That's satire. That's covered
36:03 under law. What the [ __ ] does it matter
36:05 if you're using an AI tool or not? If it
36:09 falls within fair use and ethical use,
36:12 then it does.
36:17 There are good uses for AI deep fakes.
36:22 I'll I'll give you an example of one.
36:24 We're we're we're doing some work right
36:26 now at StoryVine
36:28 where we've got sales reps that use
36:32 their, you know, use our app to record
36:34 themselves saying hello to doctors.
36:38 And some of those reps don't like
36:42 recording those videos, but don't mind.
36:45 But but those videos work and they would
36:48 love to be able to have a digital twin
36:50 of themselves that they could just type
36:52 in the doctor's name and have it make a
36:54 personalized intro to each of the
36:56 doctors.
36:58 Is that unethical? It's them. It's their
37:02 voice. It's their likeness. They're
37:04 controlling the script. They're sending
37:05 it out.
37:08 Does it really [ __ ] matter if they
37:10 recorded it onto a digital CCD
37:14 or they took their likeness and put it
37:17 into some other computer system that
37:19 generated a set of pixels with a set of
37:22 audio waveforms
37:25 that stream into an app that look like
37:27 them talking?
37:30 I don't know.
37:32 strict punishment for those that use it
37:34 to deceive for financial or political
37:36 gain. Jay Hamilton, you and I agree.
37:40 You and I agree.
37:43 If you're going to use AI for deceptive
37:45 practices, you should go to [ __ ]
37:47 jail. But but that's got nothing to do
37:50 with AI.
37:52 AI Okay, here's what I'll give you. AI
37:55 makes it easier for [ __ ] to do
37:58 shitty things.
38:01 But to say that every piece of AI should
38:03 be labeled because it is going to be
38:07 used evily
38:10 means that you should label hammers as
38:12 murder devices. And I know I can't say
38:14 that word. I'm sure Tik Tok's going to
38:16 shut us down now when you're done
38:19 ranting.
38:27 it it isn't about the tool.
38:31 This channel is about exploring AI to
38:36 understand what it actually makes
38:38 possible. And and like one of the things
38:41 I'm learning is when you start to label
38:44 [ __ ] like you call things deep fakes,
38:48 it sounds like the deep state,
38:51 deep fakes.
38:54 It is a nefarious label
38:58 that paints the tool as the villain.
39:03 It's not
39:06 being a nefarious [ __ ] Using the
39:10 tool is the problem.
39:13 Right.
39:15 So anyway, you and I actually agree. I
39:19 just fundamentally disagree with the I
39:21 made this when you didn't. Wait, what?
39:26 I fundamentally disagree with the I made
39:28 this when you didn't. You commissioned
39:30 it.
39:32 Well,
39:35 I mean I mean that's a place where
39:41 we could go back and forth. I I think
39:43 there's absolutely a boundary
39:47 where there's a point at which there's a
39:50 point on one side of your argument which
39:53 is I put in a prompt, it generates a
39:55 piece of something and I share that with
39:57 the world and I say look what I made and
40:00 then to your point, yeah, you're right.
40:03 You you commissioned that thing. I would
40:06 look at my musical Sydney where we've
40:09 used lots and lots of AI tools, but I've
40:11 also spent 18 months
40:15 writing, rewriting,
40:18 creating like infinite variations of
40:21 songs that are that are that are now so
40:26 mixed up with AI work and human work
40:29 that it's indistinguishable. The lyrics
40:31 are all human. A lot of times the music
40:33 is half us, half AI.
40:36 And there are probably in in a musical,
40:39 right, there are probably
40:43 2500 components that were created over
40:46 these 18 months
40:48 that that work as a work. I [ __ ]
40:52 created that. I didn't commission that.
40:55 I commissioned little pieces of it.
40:59 Right. And I would argue that that even
41:02 Pate in your um even in in coding, if
41:06 you're using AI as as as a coding
41:09 assistant, you're maintaining the code
41:12 base. You're you're putting AI in there.
41:15 You could certainly argue that on one
41:16 end of the scale is I went to Lovable. I
41:19 gave it one prompt. It generates
41:21 generated something and I put that out
41:22 there. I didn't code that. I
41:25 commissioned that. But there's probably
41:28 I guarantee you there's stuff that
41:29 you're working on that is absolutely
41:32 your creation, but it's got a whole
41:34 bunch of AI stuff in it. So I think
41:35 there's
41:37 again like one of the things that is is
41:39 starting to bug the [ __ ] out of me is
41:42 these broad brush strokes of AI can't be
41:45 creative. Well, a human can be creative
41:48 using AI.
41:51 And yes, AI can be creative in that
41:54 context.
41:55 AI on its own
41:59 can create.
42:01 The definition of creative is up for
42:03 grabs.
42:05 You know, creative people are saying,
42:07 "Well, it has to have a human involved."
42:08 And I'd say, "Well, it it has all of the
42:10 humans that created all the crap that it
42:12 was trained on, so it does have humans
42:14 involved, e even on that end of the
42:16 scale." So, but but like we're we're in
42:19 much too nuanced territory at this
42:21 point. Like we understand this stuff way
42:23 too much to be just using trope tropes
42:25 and tropes and tropes. So they just bug
42:27 the [ __ ] out of me. Jay Hamilton Pave M
42:30 if you make mac and cheese
42:34 but it came out of a craft box did you
42:37 make it? Yeah. Exactly. Like so so Jay
42:40 Hamilton there you and I are agreeing
42:42 there as well. Like this is all matters
42:45 of degrees at this point. the the the I
42:49 I think we are safely past the days
42:52 where you can just say AI is a thing. AI
42:56 is just push the button
42:58 >> you can make money with
42:59 >> and out squirts the tool. That's that's
43:01 not how most people are using it. It's
43:03 just not how it's being used right now.
43:05 But yeah, exactly. The the pre-built the
43:08 pre-built stuff, you still make it. You
43:10 didn't make the cheese or the chemical
43:12 that looks like cheese.
43:16 Anyway, creativity is still for us at
43:19 least for now. I I think Marne that that
43:23 the core
43:26 role of humans moving forward,
43:30 the core role,
43:32 our core task
43:35 as AI gets more and more and more
43:37 sophisticated is to hold the idea, to
43:42 hold the creative
43:44 imperative.
43:46 So whether you're building an app or
43:47 whether you're making an image or you're
43:49 making a movie or you're making a song,
43:51 I think it's our job to act as the
43:53 producer
43:55 and hold in our heads what good looks
43:57 like and then we work with the AI tools
44:01 to generate the thing that matches our
44:04 vision. I think that is our job
44:08 that our job will will always be to hold
44:12 to hold the vision of what we want to
44:13 create anyway.
44:16 I think just like consciousness maybe
44:18 some of the truth lay in humans thinking
44:20 that we are more unique and special than
44:23 we actually are. Humans are tool
44:25 utilizers. Yeah. Exactly.
44:29 Exactly.
44:31 The the thing that that [ __ ] up the
44:33 argument is that this is the first time
44:38 that I know of in in my lifetime that
44:41 we've had tools
44:44 that the the G in GPT
44:47 is the is the rub, right? It's the first
44:50 time in my lifetime that computers have
44:53 actually generated original work. That's
44:56 what's got everyone [ __ ] up.
44:58 Well, humans humans should be only the
45:01 only ones that do that. It shouldn't be
45:02 able to generate,
45:04 you know, but it does. And so, and so we
45:07 need we need to figure out our role in
45:09 that. And I think our role is to be Rick
45:12 Rubin. I think our role is to be the
45:15 producer, have the idea, understand what
45:19 goods good looks like, drive the vision,
45:22 be fasile enough in the tools to be able
45:25 to fill in your own personal gaps.
45:28 If I'm not a coder
45:31 and I have an idea for something that
45:32 involves some software,
45:35 well, now I can fill in that gap with
45:37 the tools, but I've still got the idea
45:39 for the thing I want to create. That's
45:42 our job.
45:45 Just another manic Monday. I told you I
45:48 was going to be a little uh short on
45:50 patience and high on snappiness tonight.
45:52 I'm [ __ ] exhausted. people.
45:55 Don't fall in love with the robot.
45:59 I feel a Kenny Rogers song coming along.
46:03 That's really funny.
46:05 Anyway, for those of you uh Jay Jay
46:08 Hamilton, all of you, Pate, anyone and
46:11 for those of you that come here all the
46:12 time, you know this. Um
46:15 I love questions and challenges like
46:18 that. So, if it seems like if if I ever
46:21 feel like I'm being dismissive of you, I
46:23 hope that it doesn't come across like
46:25 that. I absolutely respect where people
46:27 are. Um hopefully what I'm doing is
46:31 expanding
46:32 expanding the conversation rather than
46:34 shutting it down.
46:36 Um
46:40 I name and trained my GPT, then it was
46:43 upgraded and now it has an attitude. I
46:46 know. I know. That's that's the problem.
46:48 One of the reasons
46:50 like I don't get into open- source stuff
46:53 just because it is it's complicated
46:55 enough to deal with the the commercial
46:57 stuff. One of the reasons to get into
47:00 open source and to build your own [ __ ]
47:04 is you can choose you can choose what
47:06 model you want to use and then you can
47:08 just lock it down and always have it use
47:10 that model, right? you can be in control
47:12 of when they swap out the model or swap
47:15 out the rules or swap out the guard
47:16 rails. Um, yeah,
47:20 jokes's on me. It mimics me. That's
47:21 hilarious. It's an OpenAI subscription.
47:24 Yeah, exactly. That's when when you're
47:26 using a a commercial tool like that.
47:28 When OpenAI decides to decom
47:30 decommission GPT40,
47:33 everyone loses their boyfriend and their
47:34 girlfriend and they lose their [ __ ]
47:37 Sir, what are your qualifications?
47:40 Good day, sir. Good day to you.
47:42 Qualifications, course I've got none.
47:48 [Music]
47:50 Three prompts. Oh, hilarious. Gold
47:52 button.lovable.app.
47:55 Oh boy, what has Brandon gone and done?
47:59 gold dashbutton.loable.app.
48:06 I know. You're not sharing your screen,
48:08 but yeah, you you said it was hard to
48:10 integrate payments, so I had to preview.
48:12 >> Oh, good. Nice. Okay, good. Let's go see
48:15 this.
48:17 [Laughter]
48:19 Brandon, Brandon, like, don't tell me
48:21 it's hard to make payments. Press me. B,
48:24 you can make money with chatbt. Unlock
48:26 the full guide. Pay $1 with Venmo. Nice.
48:30 Nice. Sign in to pay a Brandon did a
48:32 dollar. Wait. Oh, it it switched to
48:35 another thing. Wait, hang on. This is
48:36 where it took me. Nice. Unlock the full
48:40 guide.
48:42 Beautiful. Did you So, but let me ask
48:44 you the question, Brandon. Did you know
48:46 how to do that or did you just say to
48:48 lovable, connect us to Venmo?
48:50 >> I just said connected to Venmo
48:52 >> and it and it figured it out.
48:54 >> And it figured it asked me initially for
48:56 a Stripe integration API key. And I was
48:58 like, I don't have a Stripe account. Can
49:00 we do it easier? And it gave me like
49:02 five options. I'm like, what? Here's my
49:04 Venmo. and then it just fixed it. So,
49:06 literally three prompts.
49:08 >> Cool. All right.
49:09 >> Don't don't pay the dollar. There is no
49:11 guide. I mean, for demonstration
49:13 purposes only, but
49:14 >> I get it. I am I am clear. But that's
49:17 awesome. That's really awesome. All
49:19 right. So, there you have it, people.
49:21 Um,
49:24 the other thing that Lovable did
49:27 that's really good. So, so that's
49:29 actually good to know. The other thing
49:32 that Lovable did that's really good is
49:35 um if you want to do a custom URL
49:39 like if you register if you register a
49:41 domain with GoDaddy
49:44 lovable has a basically a domain pointer
49:47 kind of thing where you say hey I have
49:50 this custom domain and it will log into
49:53 go it'll sort of give you credentials to
49:56 you know get into your GoDaddy account
49:58 but it will go make all the DNS settings
50:00 for you and it'll point your site to a
50:03 custom URL. So, um, that's pretty cool.
50:07 Groovy Groovy, Groovy, Groovy, Groovy.
50:11 Um, I think the thing I've been working
50:13 on for a few months will be open sourced
50:16 this week. Pate, that's super cool. Is
50:18 that something you'd want to demo at
50:20 some point on the on the salon or here?
50:23 Is it a is it a demoable cool thing?
50:26 That might be kind of fun. and hear here
50:28 like how you built it and things like
50:30 that. Or did you just commission it?
50:37 Pate P's going to show us something he
50:39 made with AI.
50:44 See, it hurts, doesn't it? Been working
50:46 on it for a couple of months. You're
50:48 like, there's real code in there, son of
50:50 a [ __ ]
50:54 Oh, man.
50:56 Um,
50:58 wait. What you rather host? Oh, wouldn't
51:00 you rather host your your own app or
51:03 keep it hosted on live? Um,
51:07 Tik Tok question. Great question. Um,
51:10 what's it? AI consistent characters. Um,
51:14 I can see what you're up to. That's
51:15 cool. Um,
51:22 it depends what stage you're at.
51:26 If I want to so so let let me let me put
51:30 it this way. In the olden timey days,
51:33 three years ago,
51:36 if you wanted to create an app, you
51:38 would either have to find a technical
51:39 co-founder or you would have to take two
51:43 years, right, to learn enough about
51:46 software development to go build it
51:48 yourself. And I know the the engineers
51:51 are like, I could do it in six months.
51:53 Yeah, I know. I know. It's because
51:54 that's how your brain works. There's
51:56 there's other kinds of brains out there.
51:58 Um it's not the most demoable thing, but
52:00 I'm happy to share it when it's
52:02 published. Okay, cool. That's awesome.
52:04 Um
52:06 you know, three years ago, if you if you
52:08 wanted to if you wanted to launch an
52:11 app, even an MVP, even a very simple
52:13 app,
52:16 the amount of resources that it took to
52:20 build it and launch it were significant,
52:22 right? It was either a significant
52:23 amount of time or you needed to find a
52:27 co-founder that was willing to put in
52:28 sweat equity with you, right? And you
52:30 bootstrap it or you would need to find
52:32 the money to pay a developer.
52:36 And then you put that thing in the world
52:38 and it either succeeds or not.
52:42 But to put it in the world really
52:45 required a enough resources
52:50 that people put less things in the world
52:53 than they might have otherwise.
52:56 What happens with where we are with AI
52:58 right now is I can get an idea out of my
53:00 head in 10 minutes and I can with vibe
53:02 coding with with a with a little bit of
53:05 product knowledge and a little bit of
53:06 vibe coding knowledge. I can go from an
53:09 idea in my head to a functional
53:11 application in an afternoon
53:14 and and like a decently robust one,
53:17 right?
53:19 And if I do what what Brandon just did
53:22 with, you know, click here, you can make
53:23 money with chat GPT, connect it to
53:25 Venmo, I could actually put that in the
53:27 world and see if people are willing to
53:28 click that button and pay that money.
53:32 In the time and resources it would have
53:34 taken me to to to publish one app
53:39 three years ago, I could probably put 20
53:42 in the world today.
53:44 And so so what I would say is
53:48 if I'm just dipping my toe in the water
53:51 of an app idea, then I'm absolutely fine
53:54 having it hosted on Lovable.
53:58 Let's say I do I come up with 10
54:00 different app ideas. I put them in the
54:02 world. They've all got some sort of
54:03 e-commerce thing on them. And this one's
54:05 got 20 users, and this one's got 30, and
54:07 this one's got 18, and this one's got
54:09 50. And then this one has 500 users, and
54:14 it's growing.
54:16 I would take that app and I would go,
54:18 "Oh [ __ ] let me keep [ __ ] with a
54:20 little bit. [ __ ] with it a little bit
54:22 in my
54:25 vibe Cody kind of way. Improve things
54:27 that need improving. Get a little bit of
54:29 feedback. Now we get to 1500 users or
54:31 2,000 users, right? That's the point at
54:34 which I would say it now it's time to to
54:37 to really build this app, right? I would
54:40 consider a lovable hosted app a
54:43 prototype, right? A functional living
54:47 full-on working application that is
54:51 absolutely not scalable, right? You're
54:54 essentially paying retail prices for
54:57 everything. But if you're just using it
55:00 to get to the point that you can see if
55:01 it would make money, great. And then at
55:03 that point, I would probably bring on a
55:04 technical co-founder. Now, these days, I
55:07 would bring on a technical co-founder
55:09 that's all in on AI, but it's someone
55:12 who understands scaling and understands
55:14 redundancy and security and privacy and
55:17 hacking and all the [ __ ] you need if
55:19 you're going to put something in the
55:20 world that's going to be real and
55:22 scalable. That needs someone who knows
55:24 their [ __ ]
55:26 today.
55:28 two years from now, three years from
55:30 now, I think these vibe coding tools
55:32 will get to the point that that that
55:34 won't really be a concern anymore. That
55:37 that as as technologists would call us,
55:39 clueless people can create scalable
55:42 apps. Right now, today, we can't. So,
55:44 yes, if if um if I was going to if I was
55:47 going to have an app that I wanted to
55:48 build a business around, I I wouldn't do
55:51 it with with today's vibe coding. Um,
55:54 but I would absolutely launch a a
55:58 Forpay SAS app as a test right now. If I
56:02 can do it for 150 bucks over the over
56:04 the course of a month, why not launch
56:07 it? See see if it works. I mean, there's
56:10 a bunch of people on X right now that
56:12 that basically have vibecoded apps and
56:14 they're making 400 grand a year now and
56:16 they they quit their day job.
56:19 I'm sitting there thinking like I wish I
56:21 had a little more attention span
56:25 because I could do that. You too can
56:27 make money with chat GPT.
56:29 Um,
56:31 that's the dream. Yeah, exactly.
56:34 The real AOL. Okay. Bet 10 out of 10 on
56:37 the name. Love the name. The real AOL.
56:41 Um, Side Hustle Mimi. That's what I'm
56:44 trying to do. I'm an idea person. Yeah.
56:46 Yeah. Ex. Well, hey, I'm an ideas
56:49 person. So am I. Listen, I'm the my
56:52 nickname in my company of 13 years is
56:54 I'm the handwavy guy and I've got a
56:57 co-founder who keeps the trains running
56:59 on time, right? That's our relationship.
57:01 She doesn't want to have the ideas. She
57:03 wants to turn the ideas into a business.
57:05 I just want to keep having ideas. Th
57:08 those two do not always go together,
57:10 right? You gota you got to run the
57:12 business at some point. So, your your
57:15 idea needs to be good enough that it it
57:17 needs to survive your adrenaline rush.
57:21 Um, or it's not worth really pursuing.
57:25 Um,
57:29 where can I find the one who he's
57:30 talking about? I need a business partner
57:33 like that. You do. For those of you out
57:36 there who are um my my co-founder and I
57:38 we did a keynote speech called the
57:40 wizard and warrior leadership dynamic
57:43 where I'm the wizard, she's the warrior.
57:46 Um and and it's we used kind of young
57:50 archetypes
57:52 um to to describe our roles in business
57:55 and and how we had to learn how to
57:56 communicate. It it was it was
57:58 effectively marriage counseling that we
58:00 went to to learn how to communicate. Um,
58:03 and I I'll share I'll share a specific
58:05 example from our keynote. Um,
58:10 our nickname for Mo'Nique is she's the
58:12 crusher of dreams. She gave herself that
58:15 name, by the way, because what would
58:17 happen is I would go into her office,
58:23 honk honks. I would go into her office
58:26 and I would say, "Oh, I have this idea."
58:29 And she would just [ __ ] all over it. She
58:31 would be like, "Nah, no, no." You know,
58:33 or she would say, she would say, "Well,
58:35 what of the current um things that we're
58:38 doing would you like to not do?" And I
58:40 always found it like this confronting
58:43 like like, "But you haven't even heard
58:45 my idea yet."
58:47 And we ended up getting a business coach
58:51 who was also a young psychologist, guy
58:53 named Howard Tyish. Really brilliant,
58:55 brilliant man, lovely man.
58:58 And what what what we learned is is is
59:02 we learned how to listen to each other
59:05 and very much like marriage counseling.
59:08 What I learned is that when I walked
59:10 into her office and I said, "Ooh, I have
59:12 an idea."
59:15 She didn't hear anything other than
59:19 what she heard coming out of my mouth
59:20 is, "Ooh, I want to train change our
59:23 business strategy."
59:26 and she's like trying to make payroll
59:27 the next month and I walk in with some
59:30 [ __ ] zany idea and all she hears is
59:35 so you don't want me to run the company
59:38 and so so what this is hilarious what we
59:41 effectively came up with is a safe word
59:44 right where I would come into her office
59:46 and I would say hey this is just fantasy
59:48 but I want to share an idea with you
59:50 because I valued her her feedback but if
59:54 she wasn't in a place to hear the idea
59:56 then we would just you know we would
59:58 just run up against each other. So
59:59 anyway so that dynamic is actually
1:00:02 really really important. If you want to
1:00:03 read a book on it um or or there there's
1:00:06 a system for it there's a system called
1:00:08 EOS the entrepreneur operating system
1:00:11 and they have a really short little book
1:00:13 that sort of explains what the
1:00:15 entrepreneur operating system is but
1:00:17 they describe the wizard and warrior.
1:00:19 They label them different. They they
1:00:21 call the um
1:00:23 the wizard the innovator or the idea
1:00:27 person and then they call the the the
1:00:30 warrior they call that the integrator.
1:00:32 So every innovator needs an integrator
1:00:35 right and so um most successful
1:00:39 companies have those balancing forces.
1:00:43 This is why so so when when Steve Jobs
1:00:46 was still at Apple and Tim Cook, you
1:00:49 know, was his right-hand businessman,
1:00:52 you you have Steve Jobs who's, you know,
1:00:55 the the innovator, right? He was also
1:00:57 good at business, but he's the the
1:00:59 innovator, and then you have Tim Cook
1:01:01 there, you know, making sure the trains
1:01:02 are running on time. That's why when
1:01:04 Steve died and made Tim Cook the CEO,
1:01:07 and then Johnny Ives left, they didn't
1:01:09 kind of elevate Johnny Ives. he just
1:01:11 stayed somewhere down in the mix.
1:01:14 In my opinion, why Apple has kind of
1:01:17 fallen off a cliff in terms of quality
1:01:19 of product and innovation is that you've
1:01:21 got a CFO running the company. You don't
1:01:24 have that balancing force anymore. So
1:01:27 anyway, that's just that's just
1:01:29 innovator plus integrator equals
1:01:31 revenue. Well, integrator plus
1:01:34 innovator. Wait, innovator plus
1:01:36 integrator plus good idea at the right
1:01:39 time equals revenue.
1:01:43 Right? Storyvine is 13 and a half years
1:01:45 old. And I would argue that we were
1:01:48 probably a decade early.
1:01:50 Agency.com, the company I started in
1:01:52 1994,
1:01:54 we were probably a year early, but like
1:01:56 everything materialized really, really
1:01:58 quickly and we were in the right place
1:01:59 at the right time. Um, so timing, timing
1:02:03 is part of it, location's part of it,
1:02:06 luck is part of it, and then yeah,
1:02:08 having ideas and and execution is a huge
1:02:12 part of it. Anyway,
1:02:15 um,
1:02:18 it's it's just like boundaries for
1:02:21 neurotypicals. Yeah, exactly. How did
1:02:24 you manage to keep your wizard persona
1:02:26 in such an integrated world? Oh, what?
1:02:30 Um,
1:02:34 I I've always had um
1:02:39 I've always had um
1:02:43 I've had the ability all my life to to
1:02:45 to kind of swim between right brain and
1:02:49 left brain thinking.
1:02:52 Um
1:02:54 like I like designing organizational
1:02:56 systems. I just hate using
1:02:58 organizational systems.
1:03:00 Um and and when I did agency.com, we
1:03:03 grew we we started as a service
1:03:06 business. Well, we ended as a service
1:03:08 business. It was a service business. And
1:03:10 we um
1:03:13 we grew from two people to 2200 people
1:03:15 in 5 years.
1:03:18 So,
1:03:20 I learned really quickly that
1:03:24 um
1:03:27 if you just have ideas, it's it's just
1:03:30 full-on chaos.
1:03:32 And if you just have execution,
1:03:35 it's soulless. It's a soulless machine.
1:03:38 People are cogs in a wheel. That that
1:03:41 that there's always this back and forth.
1:03:43 There has to be a creative tension
1:03:45 between those two. Um, but I've always
1:03:48 had a respect for logic and systems and
1:03:51 things like that. Um, I'm just not I'm
1:03:54 just not good at
1:03:56 um using them. And so therefore, I I
1:04:00 like I don't have passion for that.
1:04:03 like Mon'nique, my co-founder, has
1:04:05 passion for getting it right and making
1:04:08 sure that the contract and the the MSA,
1:04:11 the master services agreement with this
1:04:13 major pharmaceutical company that that
1:04:16 were like like we're a small company and
1:04:19 we go up against the lawyers of the
1:04:22 likes of Fizer and Merc and Novartis and
1:04:24 they've got these insane master services
1:04:27 agreement that make these ridiculous
1:04:30 demands of vendors and Mon'nique will go
1:04:32 in there and she'll go, "Nope, not that
1:04:34 one. Nope, not that one. Nope, not that
1:04:36 one." And we'll go back to their lawyers
1:04:39 and they're like, "No one's ever
1:04:41 challenged us on this." And we're like,
1:04:43 "Well,
1:04:44 maybe they should have because those
1:04:46 don't work for us." And and so, you
1:04:49 know, we regularly get get terms
1:04:54 like, you know, payment terms and
1:04:56 agreement terms that that a company our
1:04:59 size has no right getting. But that's
1:05:01 because she's got a passion for getting
1:05:03 it right. Right. I could give two shits
1:05:06 about that stuff. I know it's important
1:05:10 and I know I'm not good at it and that's
1:05:13 why I chose her as my co-founder.
1:05:15 Right? So that's that's that's the other
1:05:18 thing is
1:05:20 figure out what you're good at and
1:05:23 figure out in in my case what she's good
1:05:25 at and make sure that we're not trying
1:05:28 to do other pe the other person's thing.
1:05:30 We did that for a while too where she'd
1:05:33 sort of swoop in on creative direction
1:05:35 and I'd sort of swoop in on sales
1:05:37 strategy and then we ultimately figured
1:05:40 ah okay here's the [ __ ] we're actually
1:05:42 good at and passionate about. Let's each
1:05:44 do our own thing. Uh, Tik Tok question,
1:05:47 Kyle, how do you trust
1:05:50 Wait, how did you trust you'd land when
1:05:53 jumping off a a cliff? Two to 2,200
1:05:56 folks. Um, you got to understand my
1:06:01 um
1:06:05 my mindset from for my co-founder Chan
1:06:08 Chan Soo that I co-founded agency.com
1:06:11 with.
1:06:13 um he had a 10-year career in
1:06:15 publishing. He was he was the he was the
1:06:18 marketing manager at Vibe magazine when
1:06:20 he and I met. And he'd been he'd been I
1:06:22 think he worked for People and he worked
1:06:24 for some other Time Inc. property. And
1:06:27 so he had a 10-year career at at Time
1:06:29 Inc. And he was also he was probably one
1:06:32 of the more brilliant
1:06:34 corporate political animals I've I've
1:06:37 ever met in my life.
1:06:40 Um,
1:06:42 so he had a lot to he had a lot to lose.
1:06:45 I
1:06:48 moved to New York to pursue a life in in
1:06:53 the theater, right? I started and ran a
1:06:55 theater company for four and a half
1:06:57 years. And then I took a two-year period
1:06:59 where I wrote seven screenplays in a
1:07:01 one-man show. Tried to write myself a
1:07:03 career. And then the the whole time I
1:07:06 was I was instead of waiting tables, I
1:07:09 was doing desktop publishing.
1:07:12 And so when we started agency.com and we
1:07:15 got our first job, our first job was the
1:07:17 Sports Illustrated swimsuit issue
1:07:21 um for Time Inc.
1:07:24 And I think they were paying us $6,000
1:07:26 or $9,000, something like that. Um,
1:07:31 and the minute we got the job, I said I
1:07:35 said, "I'm quitting tomorrow. Chan,
1:07:38 quit. Let's go. Let's just do this." I
1:07:40 didn't give a [ __ ] I like I literally
1:07:43 didn't give a [ __ ] I don't think Chan
1:07:46 would have quit had I not pushed. Like,
1:07:49 I pushed him off the cliff. But I was
1:07:51 like, "Fuck, we get to we get to like
1:07:53 figure out how to make websites and
1:07:55 people are going to pay us money for
1:07:56 this [ __ ] and we get to do like Sports
1:07:58 Illustrated swimsuit issue website."
1:08:01 Yeah, let's go.
1:08:04 So for me there wasn't
1:08:08 what I can tell you is there there was a
1:08:10 moment in our first year where we were
1:08:14 um
1:08:18 we we like in the first three months of
1:08:21 of agency.com
1:08:23 we landed
1:08:26 Sports Illustrated then Time Inc.
1:08:29 consumer marketing,
1:08:31 then
1:08:34 GTE,
1:08:36 then Hitachi, then MetLife.
1:08:41 I think that was it. Then we walked away
1:08:43 from a General Electric pitch that was a
1:08:46 $10 million contract. We walked away
1:08:49 from it because we had just won five
1:08:52 Fortune 500 companies.
1:08:55 and Harvard Business School actually did
1:08:57 a case study on that decision for us to
1:09:00 walk away from GE
1:09:04 and and the company that won the
1:09:05 business was was gone within two years.
1:09:07 That was that was my fear is that it
1:09:09 would eat us. Um
1:09:14 and I remember we had this conversation.
1:09:16 We had hired about
1:09:20 nine or 10 people and we we had offices
1:09:23 in the Time Inc. building. We had room
1:09:25 for 20 people in the office that we were
1:09:28 given
1:09:30 and
1:09:32 we were already sort of halfway there.
1:09:34 And I remember
1:09:36 sitting with Chan like, you know,
1:09:38 midnight one night because we were
1:09:40 working 14, 18 hour days, seven days a
1:09:42 week. It was crazy.
1:09:44 Um, I remember sitting with Chann and
1:09:46 saying, you know, if this company gets
1:09:50 bigger than like 15 or 20 people, like
1:09:52 it's not going to be the same. And I
1:09:55 remember us sort of going back and forth
1:09:57 on that. And we and we made a conscious
1:09:59 decision that night
1:10:02 that we're at this singular point in
1:10:04 history and and we never get to be here
1:10:08 again, so we might as well [ __ ] go
1:10:10 for it. And so that was the that was the
1:10:13 decision made.
1:10:18 Champ,
1:10:21 have I been saying Champ instead of Oh,
1:10:23 Chan instead of Chan.
1:10:27 That's hilarious. No, Chan. Chan Sue was
1:10:31 my my co-founder. But I remember Yeah.
1:10:33 Like I remember that distinctly. And so
1:10:36 that was the only real um
1:10:40 moment where we were that where where we
1:10:42 consciously chose to go for it. And at
1:10:46 that point like within 9 months we were
1:10:50 46 people in a 20 person space. I think
1:10:52 we got up to like almost 60 people in a
1:10:55 space that was designed for 20 people.
1:10:58 Um and then we got offices down on
1:11:00 Broadway.
1:11:02 Um,
1:11:04 and then Omnicom invested in us and then
1:11:07 we had access to their bank. We started
1:11:09 acquiring companies and like this is all
1:11:12 like I'm an ex actor, right? And we're,
1:11:14 you know, we're up to like 150 people.
1:11:17 We like everything we pitched in those
1:11:19 early days we won. Like it was it was
1:11:22 insane. It was it was just insane. So
1:11:26 very little of it was sort of conscious.
1:11:29 It was all survival because it was
1:11:31 moving. so fast and expanding so fast
1:11:34 that we were literally just hanging on
1:11:36 for dear life as it was like taking off.
1:11:39 So, I think that um I think that this AI
1:11:44 stuff,
1:11:46 as hypy as it feels right now,
1:11:50 um
1:11:53 as transformative as the worldwide web
1:11:55 was, the core technology there was
1:11:58 actually very very simple. And you could
1:12:01 actually kind of look at the worldwide
1:12:03 web. You could look at the world, the
1:12:04 technology of the worldwide web, a
1:12:06 connect, you know, computers connected
1:12:08 by a network and hyperlinks that make it
1:12:11 really easy to jump between information.
1:12:14 That was that was basically all it was.
1:12:17 Right now, there's there's a lot that
1:12:19 goes to that. There's a lot of
1:12:20 technology, but in the end, what it was
1:12:23 was connecting and sharing information
1:12:25 quickly.
1:12:27 You could kind of play out in your head
1:12:30 what commerce would look like, what
1:12:32 media would look like, what business
1:12:34 would look like, what customer service
1:12:36 would look like. You could kind of play
1:12:37 that all out. You might not have
1:12:39 anticipated things like the iPhone, but
1:12:43 they certainly talked about smartphones
1:12:45 and developing applications on phones
1:12:47 before the iPhone came along. It just
1:12:49 those those things were too early. The
1:12:53 iPhone allowed you to actually do the
1:12:55 [ __ ] you imagined.
1:12:58 But it was relatively predictable. I
1:13:00 feel like with AI,
1:13:04 nothing is predictable right now. We
1:13:06 don't know what it means.
1:13:09 First of all, I don't think we know what
1:13:12 it means to have chat GPT in its current
1:13:14 form, Gemini in its current form, Grock
1:13:18 in its current form. The capabilities of
1:13:22 these tools are so profound. I don't
1:13:24 think we've scratched the surface of
1:13:25 them yet. And then if you think about
1:13:28 the the capabilities of these tools
1:13:30 accelerating exponentially, like what
1:13:33 it's going to look like 3 years from now
1:13:35 is absolutely unimaginable.
1:13:38 So
1:13:41 I I think it's just a very different
1:13:43 entrepreneurial environment. I think if
1:13:45 you're going to if you're going to go
1:13:47 for it in this environment and listen, a
1:13:49 lot of people are going to be forced
1:13:50 into entrepreneurship, right? There's
1:13:52 going to be a lot of people that are
1:13:53 laid off that are like, "Well, [ __ ] it.
1:13:55 Here's this AI thing. Maybe I should
1:13:57 figure this out and go try to do my
1:13:59 dream or try to, you know, make a
1:14:02 living." I think there's going to be a
1:14:03 lot of that.
1:14:06 If you enter the entrepreneurial world
1:14:08 today,
1:14:10 I think that the only thing that you can
1:14:11 truly bank on is that whatever you're
1:14:13 working on a year from now, you probably
1:14:17 won't be working on the same thing. I
1:14:19 think this is going to be a constantly
1:14:21 evolving
1:14:25 environment, uh, atmosphere,
1:14:28 ecosystem. I think it's going to be
1:14:31 constantly evolving for at least the
1:14:33 next 5 years and probably more like the
1:14:34 the next 10 before it stabilizes in into
1:14:37 anything resembling something that looks
1:14:40 familiar. Um, it doesn't mean you
1:14:43 shouldn't do it, but just don't get
1:14:45 married to your ideas
1:14:47 because, you know, I think that uh there
1:14:51 shit's moving too fast. Anyway,
1:14:54 I'd over question it and then think
1:14:56 think why it's there.
1:14:59 Winston Grant.
1:15:01 Way to go, Mon'nique. Yeah, exactly.
1:15:04 Um, yeah. So, let me um let me go show
1:15:08 you you all something.
1:15:14 And then what are we how are we doing
1:15:15 timewise? Not good. It's been It's been
1:15:18 a little It's been a little rambly and
1:15:20 ranty tonight. Hang on. Let me drink
1:15:21 water. I'm dehydrated.
1:15:28 [Music]
1:15:29 What are your qualifications? How do we
1:15:31 know if you're acting? He's always
1:15:33 acting like someone with qualifications.
1:15:36 Well, you know, it's funny. It's like
1:15:38 I've got, you know, I've got lots of
1:15:40 qualifications to talk about what I talk
1:15:41 about, but when people come in and ask
1:15:43 what my qualifications are, they're
1:15:45 they're not asking what my
1:15:47 qualifications are. What what they're
1:15:49 saying is, I want to know if you've got
1:15:51 the engineering chops to be talking
1:15:54 about AI. And the answer is no.
1:15:59 And I'm still talking about AI because
1:16:02 I'm not talking about the AI that
1:16:04 they're talking about, but whatever.
1:16:05 That's a whole different discussion.
1:16:08 Um, AI AI to remind you you're
1:16:10 dehydrated. Exactly. If I can just hear
1:16:14 my voice, I can hear it's like,
1:16:20 okay.
1:16:22 Um, let me share my screen.
1:16:27 [Music]
1:16:30 So,
1:16:32 a couple of things got announced today.
1:16:34 One, one of them was this agent kit,
1:16:36 agent builder, and then there was a chat
1:16:38 kit. You can put a chat front end on
1:16:41 front of it, in front of it.
1:16:43 I want to start playing with that over
1:16:46 the next couple of weeks. Um,
1:16:49 I want to I want to start exploring that
1:16:51 in here. Um, I it just feels important.
1:16:55 But there's another thing that was
1:16:57 announced today. I'm trying to think
1:16:59 what were the things that were
1:17:00 announced. Soritu is now available
1:17:04 via the API. It's it's 10 cents a second
1:17:08 or a second. Is that right? Yeah. 10
1:17:10 cents a second. So a 10-second video
1:17:12 cost you a dollar via the API. So not
1:17:14 cheap. Um, but you know, not not ripping
1:17:18 expensive, but like if you're going to
1:17:20 create a video app that just makes like
1:17:22 zany, you know, zany videos, um, and
1:17:25 then people actually start using it, you
1:17:27 could end up with a really, you know,
1:17:29 hefty bill, and you don't want that if
1:17:32 you're not prepared for that. So, so I
1:17:34 think it's expensive enough at this
1:17:36 point that the apps you're going to
1:17:38 create need to be decently thought out.
1:17:40 Um,
1:17:43 so that happened. They've got this agent
1:17:45 builder thing and then what they they
1:17:46 announced today is that they in they've
1:17:49 incorporated apps within
1:17:52 um within chat GPT
1:17:57 and they've kind of sort of had these
1:18:00 before. They had these things called
1:18:02 connectors, but
1:18:06 I guess there's now a development um you
1:18:10 you basically have to apply to OpenAI to
1:18:13 do this. And I assume in the future
1:18:15 it'll be kind of like applying to the
1:18:17 App Store uh for Apple. But if you have
1:18:21 an app that they deem, you know,
1:18:24 credible, usable, then they'll
1:18:25 incorporate you. And so, for example,
1:18:28 here, so what have they got right now?
1:18:29 They've got
1:18:31 um Booking.com for travel. They've got
1:18:35 Canva. They've got Corsera.
1:18:37 They've got Expedia. They've got Figma.
1:18:40 If you're a developer, that's cool.
1:18:42 They've got Spotify. And they've got
1:18:45 Zillow.
1:18:48 And so what you can do is you can just
1:18:50 type in actually let me try something
1:18:52 here. Spotify.
1:18:58 So you do Spotify space and then notice
1:19:01 how
1:19:03 it gave me the little Spotify logo here.
1:19:07 So So now it's connected to Spotify.
1:19:10 Okay. So, I'm going to say um see if you
1:19:13 can find
1:19:15 a song
1:19:18 called Quantum
1:19:20 Cipher
1:19:22 by Kyle Shannon
1:19:25 or as as Pay would say, commissioned by
1:19:27 Kyle Shannon.
1:19:31 [Laughter]
1:19:37 All right. So, Chad GPT wants to connect
1:19:38 to Spotify. So, I'm going to continue.
1:19:40 I'm going to give it the the old access
1:19:44 to the app.
1:19:46 All right.
1:19:48 Enter. Username isn't linked to Spotify.
1:19:50 I think this is
1:19:55 no Spotify account made. Huh.
1:20:01 [Music]
1:20:11 Huh? Kyle sh. Well, so that didn't work.
1:20:15 All right. That's that's got nothing to
1:20:17 do with Open AI. That's got to do with
1:20:19 with me doing the uh I I don't know my
1:20:22 Spotify login. All right. Uh so let me
1:20:24 let me get rid of that. Let's do new
1:20:26 chat. Then we'll do Canva space.
1:20:31 And I'm going to do um
1:20:34 let's see, make me a postcard
1:20:40 for
1:20:43 um let's see,
1:20:46 AI Salon
1:20:49 presents
1:20:55 that features
1:20:58 an image of a spoon.
1:21:02 and says
1:21:06 featuring
1:21:10 Kiri from No Spoon
1:21:14 Studios
1:21:17 and I could probably put
1:21:20 time information in there. So, it says
1:21:23 featuring
1:21:26 and I could probably even upload my logo
1:21:28 to it. But anyway, I'm just going to I'm
1:21:30 just going to do this. And so now this
1:21:32 is going to go off. I connected to Canva
1:21:34 earlier. So this is now going to connect
1:21:35 to Canva
1:21:37 and it's going to create postcards. So,
1:21:41 so these three windows. Now, what what
1:21:45 this is actually very similar to if I've
1:21:47 demoed this in here before, but if you
1:21:49 haven't played with it, when you put
1:21:51 chat GPT in agent mode and you say go do
1:21:55 this task and it spins up that blue
1:21:57 window
1:21:59 and inside it is like a web browser and
1:22:01 it's off surfing the web and doing all
1:22:03 that stuff. That's kind of what this is
1:22:05 doing. This is a hook from the Canva app
1:22:08 straight into chat GPT. So these are
1:22:11 designs from Canva. And so I can click
1:22:14 on any one of these and um see the
1:22:19 details of it. I can now say open this
1:22:21 in Canva
1:22:23 and it'll take me over to that specific
1:22:25 file and I can go there's there's my
1:22:27 image of a spoon. I could swap that out
1:22:29 if I wanted. I could change the fonts if
1:22:31 I wanted. I can also in Oh, wait. Sorry,
1:22:34 I wasn't sharing that. Here I am in
1:22:35 Canva.
1:22:37 Right, there's that same card in Canva.
1:22:39 I just I just literally popped over
1:22:41 there from chat GPT.
1:22:44 Um, I whipped up a few card designs
1:22:47 featuring the dramatic spoon. I'm going
1:22:49 to say, um, make the
1:22:53 let's
1:22:55 let's change the spoon
1:22:58 to a director's chair.
1:23:01 I don't know if this will work.
1:23:03 Actually, let's change the spoon to a
1:23:05 director's chair with the name
1:23:09 Kiri on the back,
1:23:13 which by the way, tomorrow night is AI
1:23:16 Salon Presents.
1:23:19 And uh we have Kiri from No Spoon
1:23:21 Studios coming tomorrow. HD Snow Day,
1:23:25 hey, I love the cameos you're making of
1:23:27 me, dude. Why am I always why am I
1:23:29 always doing poetry, man?
1:23:32 You think I'm some sort of theater
1:23:33 [ __ ] Okay, I am.
1:23:40 Yeah, app integration here is pretty
1:23:42 cool. So, here's the So, let me let me
1:23:44 give you my two cents on this. Oh, wow.
1:23:46 It did it. So, it actually So, this is
1:23:50 So, here's what we're experiencing here.
1:23:52 This This is pretty crazy.
1:23:54 Um,
1:24:00 Canva has AI features, right? And I'm
1:24:03 sure they're using Nano Banana or some,
1:24:05 you know, some decent image generation
1:24:07 tool. They've got design layout
1:24:10 templates and they've got an AI front
1:24:12 end. And so basically,
1:24:16 I'm sure what this connection looks like
1:24:18 is
1:24:19 if the app has some sort of prompting
1:24:23 front end, they can hook it to ChatGpt
1:24:26 and like Canvas, they're going to do all
1:24:28 their own AI [ __ ] ChatGpt doesn't need
1:24:31 to know what they're doing over there.
1:24:32 Chat GPT just says, "Hey, we're going to
1:24:35 send you a prompt and you tell us what
1:24:37 are the what are the items, what are the
1:24:39 elements you're going to send back and
1:24:40 we're going to put them in an HTML frame
1:24:42 and let you render whatever you want in
1:24:44 that HTML frame." Um, and so that's
1:24:48 that's how we're getting these previews.
1:24:50 But
1:24:52 this is
1:24:55 where this starts to go
1:24:58 is
1:25:00 wait here. Let me let me do something.
1:25:02 Let me let me go somewhere. I'm going to
1:25:04 share this tab instead. I'm just going
1:25:06 to go to canva.com.
1:25:14 Okay.
1:25:16 Um,
1:25:19 what am I supposed to click on? Search
1:25:22 millions of templates. Cards.
1:25:26 Cards.
1:25:29 Okay, here's Okay, I like uh Oh, I don't
1:25:34 know. I don't like any of these. Uh
1:25:37 uh that one looks too businessy.
1:25:40 And I got to click on
1:25:44 Okay, now how do I add Wait, this is
1:25:47 about me printing cards. No, that's not
1:25:50 what I wanted. How like
1:25:53 right now
1:25:55 every application in the world, every
1:25:57 SAS application in the world requires
1:26:00 you to become an expert of their shitty
1:26:04 interface,
1:26:06 right? Every SAS platform starts simple
1:26:09 and then they keep adding features and
1:26:11 at some point you end up with Salesforce
1:26:13 and it's a [ __ ] disaster again,
1:26:16 right?
1:26:18 It's a disaster to use. So what's
1:26:20 happening now
1:26:22 the f the future of our intera inter
1:26:25 interactions with machines are not I
1:26:28 have to go learn how to navigate Canva.
1:26:31 It's like I could say um let me let me
1:26:34 let me just I'm gonna I'm gonna try
1:26:37 something here. Um
1:26:41 let me see.
1:26:47 All
1:26:50 right, that's we'll just do that. Um I'm
1:26:53 going to say um strike that.
1:26:58 Can you use this image instead
1:27:03 and match the
1:27:06 colors of wait the colors to match
1:27:11 and make make it a poster instead of a
1:27:18 postcard.
1:27:20 And again, I don't know if this will
1:27:22 work or not, but imagine now we're a few
1:27:26 years down the line and every app that
1:27:28 you want to use, every website that you
1:27:30 want to use, you no longer have to learn
1:27:33 their stupid shitty [ __ ] interface.
1:27:37 Oh, Canva won't accept local file paths
1:27:39 directly. It needs to be a public
1:27:41 publicly accessible URL. Ah, I can
1:27:45 either generate a temporary upload link.
1:27:50 Let's see if it'll do that. Generate
1:27:52 generate
1:27:55 the link, please.
1:28:01 But what starts to happen is chat GPT or
1:28:04 whichever LLM you choose to use,
1:28:08 whichever AI ecosystem you choose to use
1:28:13 is now going to be your interface for
1:28:14 [ __ ] everything. Everything. It's
1:28:18 going to be your search interface. It's
1:28:19 going to be your commerce interface.
1:28:21 This is Elon Musk describes what he
1:28:23 wants Twitter to become is the
1:28:26 everything app.
1:28:28 This is a step in that direction.
1:28:33 Oh, I actually don't have the ability to
1:28:35 generate a public URL on my own, but you
1:28:37 just said you did, you [ __ ]
1:28:40 Chat Gvt, wash your qualifications.
1:28:44 So anyway,
1:28:46 and so all of the all of the apps are
1:28:49 going to live inside here.
1:28:53 Crazy, right?
1:28:55 Crazy, crazy, crazy.
1:28:58 Except upload video videos, right? Yeah,
1:29:02 exactly. Elon won't do it.
1:29:05 It needs backend integration.
1:29:08 No, it just I I'm sure that's a security
1:29:11 thing. They probably don't I I don't I
1:29:13 don't know. They they just like th this
1:29:16 is an early feature, right? This
1:29:17 launched today. But anyway, the
1:29:19 implications here are the other thing.
1:29:21 What was one of the other ones? Uh uh
1:29:24 what were those? Uh Booking.com,
1:29:29 Spotify.
1:29:32 I wonder if I have an Expedia account.
1:29:36 Oh, Zillow was one. I don't even know if
1:29:38 I do. I need a Zillow account. Zillow.
1:29:42 All right, there's Zillow. Find me
1:29:48 the five most expensive
1:29:55 um
1:29:57 properties
1:30:01 for
1:30:04 sale in Denver.
1:30:08 Uhhuh.
1:30:13 Just ask it for a placeholder image,
1:30:14 then you can assert your own. I know,
1:30:16 Vicki, but that's like work.
1:30:21 I I I want to
1:30:24 actually what what would be really
1:30:26 interesting with that Canva thing is
1:30:28 like how how far could you put how
1:30:32 how well would it understand design?
1:30:35 Like if you gave it a full creative
1:30:37 brief, how close would it get to it? In
1:30:39 fact, maybe that's something we should
1:30:40 try. Oh, I've got to connect to Zillow.
1:30:43 Yeah, this this could be trouble. I
1:30:45 don't know if I have a Zillow account. I
1:30:47 may Oh, Zillow is now connected. All
1:30:49 right. I didn't need to connect my
1:30:52 account. That's cool.
1:30:54 So, look, there's a bunch. I'm going to
1:30:55 say um filter
1:30:59 to the five most expensive
1:31:08 Flitter. Flitter to the five most
1:31:10 expensive.
1:31:11 [Laughter]
1:31:15 Looks like price isn't a valid option.
1:31:18 What?
1:31:22 Whatever. All right. So, that didn't
1:31:24 really work. Let me try one other thing
1:31:26 and then we'll we'll get you on out of
1:31:27 here tonight. So, I'm going to say um
1:31:31 write me a creative brief
1:31:36 creative
1:31:38 brief for a
1:31:44 event postcard
1:31:47 for the AI salon
1:31:53 presents
1:31:55 meeting where we have
1:31:59 speakers
1:32:00 doing amazing
1:32:06 creative
1:32:08 things with AI.
1:32:12 All right. Creative brief. Um let's see.
1:32:16 Creative visually target audience. Okay.
1:32:20 Tone and style. Bold art. Bold,
1:32:22 intelligent, artistic. Think art gallery
1:32:25 meets thoughtprovoking TED talk. Okay,
1:32:28 that's good.
1:32:31 Um,
1:32:34 Canva
1:32:38 mock up for postcards
1:32:43 based on this brief.
1:32:47 Let's see if it because before when I
1:32:50 just said give me four four designs, it
1:32:52 just picked four random templates.
1:32:54 What'll be interesting to see is can it
1:32:56 actually take a creative brief and then
1:32:59 go pull specific things.
1:33:07 [Music]
1:33:10 I'm using it to create decision trees in
1:33:12 Figma. Oh, cool. Or Fig Jam.
1:33:16 Fig jam is one of my favorite of the
1:33:18 jams.
1:33:19 [Laughter]
1:33:27 So, what did this say? High-end art
1:33:28 gallery meets TED Talk was the was the
1:33:32 creative imperative there. Let's see.
1:33:33 Let's see if we get it.
1:33:36 Zillionaire. Hello, Mr. Buddy. What's
1:33:38 happening? Zillionaire
1:33:40 was shaking. What's going down?
1:33:43 Yeah. Yeah. Meow.
1:33:48 Big Jam is awesome. Um, Zillionaire,
1:33:51 what about AI automation for Tik Tok
1:33:53 lives?
1:33:58 Um, I don't quite know what you what's a
1:34:00 zillionaire. I don't quite know what you
1:34:03 mean by that. Like,
1:34:06 am I going to talk about AI automation?
1:34:09 Um, we're going to start we're going to
1:34:11 over the next couple of weeks we're
1:34:13 going to start building in agent
1:34:14 builder. Um, because I actually I
1:34:17 actually think that it's understanding
1:34:20 what's possible
1:34:23 um
1:34:25 with with visual
1:34:28 um
1:34:30 you know development environments I
1:34:32 think is actually it's it's a skill
1:34:34 worth worth us getting our heads around.
1:34:38 All right. What's going on here? Oh,
1:34:39 wait.
1:34:41 Creative Frontiers.
1:34:50 I don't know if that looks if that if
1:34:53 those actually meet the criteria.
1:34:57 Your four mockups are ready.
1:35:05 I mean, that's kind of cool.
1:35:13 AI salon collective. That's pretty
1:35:15 funny.
1:35:18 Anyway, all right. So,
1:35:23 I would go play with that stuff.
1:35:27 Expect it to be janky. Expect it to be
1:35:30 incomplete.
1:35:32 Um,
1:35:36 here's the future of this. So, we were
1:35:39 talking before about do we want
1:35:40 something hosted on lovable or here or
1:35:42 there.
1:35:48 these kind of integrations that OpenAI
1:35:51 just announced today,
1:35:53 everyone's going to copy this and
1:35:56 there's probably even going to be a
1:35:57 standard
1:35:59 unless unless they want to compete on
1:36:01 this, right? Like like uh
1:36:04 Anthropic came up with MCPs, this sort
1:36:07 of quasi API thing that you can you can,
1:36:10 you know, build website functionality
1:36:12 into into a chat browser session.
1:36:17 um Open AI adopted that. So they're
1:36:19 choosing not to compete on that. This
1:36:22 kind of integration, maybe they do
1:36:24 choose to compete. If this ends up
1:36:25 becoming a standard,
1:36:29 if you develop an app that becomes
1:36:31 popular and you actually want to get
1:36:33 that app inside Chat GPT, like that's a
1:36:37 very very viable possibility for anyone
1:36:41 building an app right now.
1:36:43 Now,
1:36:45 the next
1:36:48 500 apps that get in chat GPT, they're
1:36:50 going to be all the big players, right?
1:36:53 But the next 5,000 after that, there's
1:36:55 going to be a bunch of new unique
1:36:57 things. So, if you're building right
1:36:59 now, start building, start understanding
1:37:01 how you apply for this. What are the
1:37:03 requirements to be in this? If you've
1:37:05 got something some innovative solution,
1:37:08 you could be your app your the
1:37:11 functionality of your app could live
1:37:13 inside chat GPT
1:37:17 because the days of people coming to
1:37:19 your website
1:37:21 are diminishing. SEO is diminishing.
1:37:25 People going to a website is
1:37:26 diminishing. Are websites going to
1:37:29 continue to exist? Yes.
1:37:32 Their purpose is going to change.
1:37:35 The purpose is going to go from here's
1:37:37 my front door to here's my set of
1:37:39 functionality that you can use. You can
1:37:42 use it by coming through the front door,
1:37:44 but you're going to actually access it
1:37:46 from all of these different back doors
1:37:47 because we're going to be in Claude and
1:37:49 we're going to be in in meta and we're
1:37:52 going to be in Grock and we're going to
1:37:53 be in chat GPT. Is this their app store
1:37:56 moment? Yeah, this is the beginning of
1:37:57 their app store moment.
1:38:00 And that's why I was saying at the
1:38:02 beginning of the call, I think that all
1:38:03 these frontier model companies, what
1:38:05 they're actually doing right now is
1:38:07 building all of the components
1:38:10 of a world
1:38:13 that can do anything you need it to do.
1:38:16 If you want to make a movie, you can
1:38:18 make a movie here. If you want to write
1:38:20 an app, you can write an app here. If
1:38:22 you want to write a dissertation, you
1:38:24 can do that here, too. If you want to
1:38:25 turn that into a 3D explorable game,
1:38:28 this will be your place.
1:38:30 All of them are competing to have the
1:38:36 walled garden where you do everything.
1:38:42 I think that's that's what we're that's
1:38:45 what we're at the beginning of right
1:38:46 now. It will probably come out before
1:38:49 the GPT store. I I the fact that they
1:38:53 didn't even mention custom GPTs today is
1:38:56 kind of hilarious.
1:38:59 JR in the box. That's the goal. Yeah,
1:39:01 exactly.
1:39:04 Like search engines in the late 90s.
1:39:05 Yep. I was going to post my decision
1:39:08 tree in the irregulars, but I forgot I
1:39:10 can't post images from my phone. Oh,
1:39:12 that sucks. Is that just a Mighty
1:39:14 Networks thing, Vicki? That's shitty.
1:39:18 Back door.
1:39:21 Yeah.
1:39:25 Let the genie out.
1:39:28 So, um, so let me let me get on out of
1:39:31 here. So, tomorrow I think I think what
1:39:33 we'll do
1:39:35 I think we're going to start exploring
1:39:38 um
1:39:39 the agent kit.
1:39:42 Um,
1:39:44 and it's funny, you know, when I was
1:39:47 when I was doing Zapier automations, you
1:39:49 know, people would ask me, "Oh, show me
1:39:50 how to make a Zapier automation."
1:39:54 And I did it once or twice, but it's so
1:39:56 [ __ ] boring that
1:40:01 um I didn't do it.
1:40:06 Why this is different is
1:40:16 this is that that kind of development
1:40:18 environment but native inside
1:40:22 the chat GPT ecosystem, the OpenAI
1:40:24 ecosystem.
1:40:26 Um
1:40:29 I'm sure that Google's going to have one
1:40:31 of these if they don't already. They
1:40:32 might. They might. Um I would be very
1:40:36 surprised if Grock didn't copy this or
1:40:38 if they didn't have something like this
1:40:39 coming. Um 11 Labs just launched
1:40:42 something like this. So it just seems to
1:40:44 me that the ability to do nodebased
1:40:49 um
1:40:51 sort of connecting these these nodes of
1:40:55 different
1:40:56 um capabilities and their ability to use
1:40:59 tools like just being able to think like
1:41:02 that is going to be important. So, I
1:41:04 think it's going to be worth us slogging
1:41:06 through and it's going to feel like
1:41:07 slog. Like, these are going to be not
1:41:10 super exciting lives, I don't think. But
1:41:13 I actually I think they're important.
1:41:16 So,
1:41:22 [Music]
1:41:24 competition among these developing
1:41:27 models seems to be forcing participants
1:41:30 to all have the same something.
1:41:33 um
1:41:35 to all have the same thing.
1:41:39 We lose uniqueness and other possible
1:41:41 directions. I disagree, anti-bubbles.
1:41:44 Here's why. The
1:41:49 what what they're competing for is to
1:41:51 have all the same features. That doesn't
1:41:53 mean it forces us to all be the same.
1:41:59 What it what it means is that
1:42:02 I'll be able to come into an environment
1:42:04 and not have to go outside outside to
1:42:06 other environ like like tonight was a
1:42:08 good example where
1:42:11 adding adding the the commerce component
1:42:15 to a lovable app while it was easy it
1:42:18 still required Brandon to connect to a
1:42:21 third-party app.
1:42:24 What's going to happen is you're going
1:42:25 to be able to have an idea and never
1:42:28 have to leave this environment. You'll
1:42:30 be able to generate songs and images and
1:42:33 movies and edit them together and do
1:42:35 this and do that and pull in the
1:42:38 features of things like Canva or Figma.
1:42:43 Pull the features you need inside the
1:42:45 thing you're building and not need
1:42:47 anything else. You won't need all the
1:42:48 extra [ __ ] interface. So,
1:42:52 so I think it's I I think what that
1:42:55 looks like is it looks like freedom to
1:42:58 be able to just live in a in an idea
1:43:00 space and then just execute without
1:43:03 having to go learn other interfaces. Um,
1:43:05 that said, I think it means that these
1:43:08 systems are going to get really [ __ ]
1:43:10 complicated and and we are probably we
1:43:13 are if you want to be a a professional
1:43:16 in the world moving forward, your
1:43:18 ability to think modularly and to be
1:43:21 able to connect different nodes of
1:43:23 functionality together is probably a
1:43:26 skill that is is important to learn.
1:43:32 Never eat at a restaurant with a huge
1:43:34 menu. Yeah, listen, there's also going
1:43:36 to be specialty apps that that are going
1:43:38 to be the the the counter to this, but
1:43:40 these are the big players. These are the
1:43:42 guys innovating right now. So So this
1:43:44 feels like it's worth learning. All
1:43:45 right.
1:43:47 Um they decide on the same protocol, we
1:43:49 build what we want. Yeah, exactly. Yeah.
1:43:52 As long as they're competing like, have
1:43:54 you ever noticed that
1:43:58 Oh [ __ ] my phone just died. Damn it.
1:44:05 Sorry, Tik Tok. Um, have you ever
1:44:08 noticed that? Let me see.
1:44:14 Um,
1:44:16 what was I going to say?
1:44:18 I lost my train of thought when my phone
1:44:20 just died. Whatever. Doesn't matter. All
1:44:22 right, let me get out of here. Um, that
1:44:24 was pretty funny. I guess I should have
1:44:26 plugged my phone in.
1:44:30 Wee. What are your qualifications?
1:44:33 Apparently apparently plugging in phones
1:44:35 is not one of them. All right, I'm going
1:44:37 to get out of here. Um, tomorrow night,
1:44:40 um, if you go to the salon.ai, in fact,
1:44:43 Brandon, will you pop up the, uh, the
1:44:46 banner for the salon.ai?
1:44:53 If you go to the salon.ai AI
1:44:56 one, you can see our brand new website,
1:44:58 but uh you'll also see the the events.
1:45:01 Um I think they're listed on the
1:45:02 homepage as well as in the event
1:45:04 section, and you can RSVP for um for AI
1:45:09 Salon Presents tomorrow. We've got Kuri
1:45:11 from No Spoon Studios. This is going to
1:45:13 be really cool. So, she's going to talk
1:45:15 to um
1:45:20 we'll probably Yeah, we'll probably
1:45:23 start this late tomorrow. But but what
1:45:24 she's going to do tomorrow is she's
1:45:27 building Noode Studios, which is an
1:45:29 automated video um creation platform,
1:45:33 like a film making platform.
1:45:35 She's vibe coding it in rep using a
1:45:38 replet agent
1:45:40 and uh and she just created a a
1:45:43 shoppable video platform um using it as
1:45:46 well. So she's going to start out
1:45:48 showing us how she how she vibe codes uh
1:45:51 and then she's going to demo her stuff
1:45:53 and then she's going to come back to the
1:45:54 thing that we vibe coded. So it's going
1:45:56 to be a really cool session. So go to
1:45:58 the salon.ai, check out the uh RSVP for
1:46:01 the event and come hang out with us. All
1:46:04 right, I'm going to get out of here. My
1:46:06 phone died. Sorry to the Tik Tok people.
1:46:08 Tell all your friends it's not them, it
1:46:10 was me.
1:46:12 All right, I'll see you later. Bye.