AI Learning Lab

9/5/2025 - Anthropic's $1.5 Billion Copyright Settlement and Nano Banana Still Wows

Y8JsK-GK35o
Live Stream2025-09-062:10:19139 views

Description

Timeline Shifted. Anthropic pays authors. Nano Banana still wows. Kyle Shannon discussed recent developments in the AI world, including a significant legal settlement involving Anthropic and a class-action lawsuit by authors. A judge ruled that training AI models on purchased books is transformative and falls under fair use, but using pirated books is infringement. This resulted in a $1.5 billion settlement from Anthropic. Kyle believes this sets a precedent for other AI companies like Google and OpenAI, ultimately leading to compensation for artists and a shift towards training models on synthetic data. He expressed hope that these settlements will pave the way for broader acceptance and ethical development of AI technologies. Beyond legal matters, Kyle also explored several AI tools and platforms. He showcased OpenAI's GPT-5 prompt optimizer, a tool designed to improve prompt quality for better results. He also demonstrated "vibe coding" using Google AI Studio to create a Gen X-themed ad mockup generator, experimenting with different product URLs, ad styles, and platforms. Kyle encouraged viewers to engage with these tools and participate in the Nano Banana hackathon, emphasizing the growing importance of cross-disciplinary AI skills and the ability to integrate multiple tools for innovative projects. He concluded by highlighting the evolving nature of AI and the transition towards more sophisticated, cross-functional applications. 🎙️ New to streaming or looking to level up? Check out StreamYard and get $10 discount! 😍 https://streamyard.com/pal/d/5460595014369280 #AI #ArtificialIntelligence #Copyright #AItools #NanoBanana #VibeCoding #PromptEngineering #AIethics Chapters: 00:00:00 Slow Moving Cold Fronts (Song) 00:02:26 Friday Night Date Night 00:03:38 Late Stream Start 00:04:45 Anthropic Hackathon 00:06:17 Ted D Returns 00:07:36 Standing Between You and Fire (Song) 00:08:42 Vibe Coding Tonight 00:10:54 Faux Pas and Perplexity 00:12:20 Book Update 00:15:59 Dog Interrupts Stream 00:17:49 AI Salon and Cheese 00:18:53 Cindy's Internet Woes 00:20:47 LED Light Bulb Mystery 00:22:08 Prototyping Sprint 00:27:16 AI Readiness Project 00:29:13 Bento.me and Socials 00:31:17 Anthropic Settlement 00:38:23 Perplexity Pro Offer 00:40:48 Ideogram and Nano Banana 00:42:35 LinkedIn Article Prompt 00:47:49 Medical Diagnostic Tool 00:52:00 GPT-5 Prompt Optimizer 01:00:06 Nano Banana Hackathon 01:06:03 Building Ad Mockup App 01:13:20 Redesigning the App 01:15:05 Amazon Product Search 01:19:58 Ad Mockup Feedback 01:24:21 Gen X Ad Maker 01:25:00 Mule Run Activation Codes 01:30:44 AI's Knowledge Base 01:31:52 Storyvine Ad 01:36:00 Gen X Gifts 01:39:26 Cassette Tape Ad 01:43:24 Fixing the Ad 01:47:00 Multi-modal Capabilities 01:51:20 Image Upload and Display 01:55:05 Pine Cone Pricing Change 01:58:54 Two-Column App Design 02:00:06 AI and Discomfort 02:04:49 Weekend Homework 02:07:14 Nano Banana on X 02:08:47 Cross-Discipline AI Skills

Chapters

Transcript

0:00 your doggy. Come here, dog. Dog doggy
0:03 dog.
0:05 Hey, doggy. Hey, dog. Doggy.
0:24 [Music]
0:30 [Music]
0:45 [Music]
0:47 [Applause]
0:52 She came on here like slow moving cold
0:56 fronts.
0:57 [Music]
1:00 Well, his fear was warmer than a look in
1:03 her eyes.
1:08 She sat on a stool and he said, "What do
1:12 you want?"
1:16 >> She said, "Give me a love that don't
1:19 freeze up inside."
1:25 said, "I have melted some hearts in my
1:28 town, dear."
1:32 >> But to sit next to you, well, I shiver
1:35 and shake.
1:37 [Music]
1:40 >> And if I knew love, well, I don't think
1:43 I'd be here.
1:46 [Music]
1:48 Asking asking myself if I've got what it
1:52 takes
1:54 [Music]
1:56 to melt your icy blue heart
2:00 [Music]
2:04 should I stop
2:07 turn what's been frozen for
2:11 years
2:14 into a river of tears.
2:17 [Music]
2:26 [Music]
2:30 Friday night date night. Good evening,
2:32 good people. What's happening on Friday
2:35 night date night? Friday night date
2:36 night. Everybody coming in strong. We
2:39 have two people. Two people on TikTok.
2:43 Is that true?
2:48 Two people on Tik Tok. Six people.
2:50 >> No, it's actually just one now. It's
2:51 just me.
2:53 >> Wow. Wonder what's up with that.
2:56 >> You shadowbanned.
2:58 >> I don't know. Maybe
2:59 >> I just shared it with all the
3:00 irregulars.
3:03 [Music]
3:05 Got five on YouTube.
3:08 [Music]
3:11 Uh,
3:13 archetypal says he has no sound on
3:15 YouTube. That could be a problem.
3:19 >> Hello. Hello. Hello.
3:21 >> You got sound on Tik Tok. Let me check
3:24 the YouTube stream.
3:25 >> Hello. Hello. Hello. Testing. Testing.
3:28 Testing.
3:29 >> Yeah, you got sound.
3:30 >> All right. Testing. Yes. Yes. It's
3:33 Friday night date night. Sound worked
3:36 once I unmuted mine.
3:38 [Music]
3:38 [Laughter]
3:41 There we go. Now they're piloted.
3:44 >> Well, see the problem is you told
3:46 everyone you were going to be late. So
3:48 >> I know every everyone hot plate with the
3:51 >> Yeah, exactly. Everyone else they're
3:53 taking their time making the nachos.
3:54 What do you mean he's live? He He said
3:56 he was going to be late.
4:00 >> That's the thing about AI Learning Lab.
4:02 You can. It's just It's You've got to
4:04 pay attention to it 24 hours a day
4:06 because you just never know what's going
4:07 to happen.
4:10 All right. Piling in.
4:14 Wow. Wow. Wow. Wow. Wow.
4:16 All right. I think my hair and eyebrows
4:19 are sufficiently puffy to start the
4:22 evening.
4:25 You know, there's just a point at which
4:26 you get to a certain age, your eyebrows
4:28 just are going to do their own thing.
4:33 [Music]
4:45 You have the anthropic hackathon. I
4:47 think you mean the anthropic settlement.
4:55 Heat. Heat.
4:56 [Music]
5:42 Oh,
5:46 [Music]
5:51 [Music]
5:53 that's
5:55 [Music]
6:09 [Music]
6:18 Ted D in New Hampshire. Used to come
6:21 here many moons ago. feel like I watched
6:23 season one of Breaking Bad
6:28 and now I'm back to season six. That's
6:31 hilarious. Well, welcome back. Um,
6:34 welcome back. Yeah, I uh I'm still
6:36 cooking meth in the bathtub, so that
6:38 that that's remained constant. Um,
6:43 and I'm probably going to get banned
6:45 from TikTok for saying that.
6:51 So, does anyone know is Tik Tok still
6:53 September 25th? We're out of here. Is
6:56 that still the plan?
6:59 Did you get stood up on date night,
7:00 Kyle? Yeah, it's We're a little light
7:02 here tonight. Oh, the se September 17th.
7:05 Unless something changes.
7:10 [Music]
7:36 Standing between
7:38 [Music]
7:40 you and fire is insane.
7:44 [Music]
7:46 Standing too near
7:49 you in a fire makes it clear.
7:55 You're trouble to me.
7:58 [Music]
8:01 Rich trouble. Can't you see?
8:05 [Music]
8:07 Leaving it close.
8:10 Smell of your perfume scares me most.
8:16 Leaning away
8:19 feels stronger every day.
8:26 You're troubled to me.
8:31 Rich trouble can't see.
8:35 [Music]
8:43 Um, let's see. Good people. Let's see
8:46 what's happening. Let's see what's
8:47 happening. Do me a favor. Share the live
8:50 if you haven't on TikTok. You got to do
8:51 your thing. You got to play your part in
8:53 the community. Share this. I see someone
8:55 is tapping the screen like mad. I don't
8:57 know if that's Mr. It with his little
8:59 his little tappy tap machine.
9:05 [Music]
9:10 We're going to do some vibe coding
9:12 tonight. I think I think it could get
9:15 boring as [ __ ] It's Listen, when when
9:18 you actually try to do something on
9:19 these lives, it can get a little Oh, do
9:22 me a favor. If you if you are on
9:27 actually wherever you are, everybody's
9:29 got a phone. Get the YouTube app and go
9:33 find
9:34 Hang on a sec. Let me share my screen
9:36 here. Go find
9:40 Wait, why did that not do that? Am I not
9:42 sharing my screen? I'm sharing my
9:43 screen. There we go. Um, not that.
9:48 But that's cool. That I'm excited about.
9:51 That's my new my new book title. But I
9:54 want you to go to
9:58 go to there
10:01 on YouTube
10:03 learning lab-ai
10:05 and if you go there on your mobile
10:06 device you can pick any video not this
10:09 live but any other video there and you
10:12 can hype it. There's an option to hype
10:15 the video. You can do three a week. So,
10:17 do me a favor and just go hype hype like
10:20 the if you go into the live the live
10:23 tab,
10:25 you'll see the the recordings of these
10:27 lives in there. Just pick one. Pick like
10:30 the one from last night and just hype
10:32 the [ __ ] out of that thing. All right?
10:35 And I mean hype the [ __ ] out of it. Like
10:37 click it once.
10:45 [Music]
10:54 James Deuter, I made a faux paw last
10:58 night. I wanted to thank producer
10:59 Brandon
11:01 for the tip on perplexity.
11:07 Gareth doesn't see the hype button.
11:09 Yeah, I don't know anything about it.
11:11 Brandon claims that it exists, but it it
11:13 exists on the on the mobile app only
11:15 apparently. Gareth, so do you have the
11:17 YouTube mobile app?
11:20 [Music]
11:28 Yeah, mobile app, nonlive videos only.
11:30 So, you got to go back to last night's
11:32 video,
11:33 click on a non-live video in the YouTube
11:35 app, and then you should see the hype
11:37 button. And if you don't, then that's as
11:40 much as I know.
11:45 And you can do three per week. I'm on
11:46 the mobile app. Maybe I need to update.
11:49 Okay.
11:53 [Music]
11:55 Ah, Valerie Cox, I found it. Scroll to
11:58 the right. This is awesome.
12:01 I'm just asking you all to go market my
12:03 crap and your problem solving together.
12:06 I like it.
12:09 Thank you. You all are awesome. All
12:12 right.
12:20 So, I got some interesting news tonight.
12:24 Anything I want to talk about here?
12:27 Um well since I'm here um
12:31 if you've been in the if you were in
12:34 office hours today you heard this but um
12:37 I'm working on a book and we just did a
12:40 slight shift in focus of it. It was
12:42 called feed your prompt and what we've
12:45 come up with is this which is 10person
12:48 team. Feed your prompt like a producer
12:51 and trans a transform AI from a tool
12:54 into your team. Um, it's a much more
12:57 clear value proposition. It's it's super
12:59 clear and it's like it's like, you know,
13:02 this idea of AI as an amplifier. So, I'm
13:05 really excited about it. So, so that's
13:07 kind of that's new and out there. Um,
13:11 if you're not a member of the AI salon,
13:15 go to community.thesalon.ai
13:17 or, you know, zap the QR code if you're
13:20 QR code friendly and join the salon.
13:24 Um, just trust me, you should go do
13:27 that. You should go join the salon. Um,
13:32 increasingly, we were talking about this
13:34 on office hours today, that the skill of
13:38 the future is going to be people who can
13:41 kind of hold a vision for something and
13:43 wield a bunch of AI tools like that 10
13:46 person team really figure out, okay, for
13:48 this idea, I need to start out with
13:50 like, you know, a strategic, you know,
13:53 brainstormer and then I need to do some
13:55 research and then I need to develop a
13:57 business plan and then I need to develop
13:58 a marketing plan like someone who can
14:00 wield all the different tools. Uh maybe
14:02 I need to go visualize this as an ad
14:05 mockup, someone that knows how to do
14:07 that, right? Someone that can hold the
14:08 vision for the overall project and go
14:10 play with all these things.
14:14 How you find people that you can trust
14:16 to have that kind of judgment is to be
14:18 in community. And how you can establish
14:21 yourself as one of the people who's got
14:23 that kind of judgment is to be in a
14:25 community where you're constantly
14:27 engaging with other people in the
14:28 community and delivering good [ __ ] In
14:32 fact, starting next Tuesday,
14:36 um Cindy [ __ ] and I are um doing a
14:40 four-week prototyping sprint where you
14:43 you have to be a member of the AI salon
14:45 mastermind. And so if you're if you're
14:47 not if you go into the AI salon, there's
14:49 a subscription area called the
14:51 mastermind. It's 20 bucks a month
14:53 through 2025.
14:55 Um we're just kicking it off this year,
14:56 but if you're part of it, you get to
14:58 join the mastermind clubs. And one of
15:00 them is Cindy Coon's prototyping club.
15:03 So, she and I, starting next Tuesday,
15:06 are doing um a four-week sprint where
15:11 over the course of four weeks, you will
15:13 design a miniourse and that you can put
15:16 out in the world or you can put into the
15:18 AI salon and get people to to learn
15:20 something. Cindy Coon's here. Welcome,
15:22 Cindy [ __ ] We're shaking. Um, so it's a
15:27 really good, first of all, it's just
15:29 good to join the mastermind because
15:31 you're joining with people who are kind
15:33 of committed to this AI thing at a
15:35 different level and you can develop
15:37 deeper relationships, but you also get
15:39 access to stuff like that. So that
15:41 starts next Tuesday for four weeks. I'm
15:42 really excited about it. Um, so yeah,
15:45 Cindy [ __ ] Um, hang on. Uh, wait, join
15:48 the mastermind. Yeah. Yeah. Yeah. For
15:50 the AI prototyping sprint. That's what I
15:52 I already said that. Cindy [ __ ] I
15:54 already said that before you got here.
15:57 I'm way ahead of you.
15:59 Nanny, nanny, boooo. All right, I got to
16:02 go. The dog is whining, so I got to go
16:04 let him out. So, hold please.
16:07 >> What's up, champ?
16:18 >> What do you want? What do you want? Want
16:21 up there?
16:24 Come on.
16:27 What do you want?
16:29 You want cheese
16:33 here?
16:48 Follow me.
16:57 All right, here's the deal. The miss the
17:00 MS is out gallery hopping. The boys are
17:04 upstairs. So, I've got to feed Champy
17:06 some cheese. So, I know I I know. I
17:10 know.
17:11 I know. Calm down.
17:16 See, here's how this goes. Piece of
17:18 cheese for me. Peace cheese for the dog.
17:24 Peace cheese for me. Peace cheese for
17:27 the dog.
17:29 All right. Are you happy, fatty? Huh?
17:35 Look up.
17:39 There you go.
17:41 All right. You good?
17:44 Here.
17:46 All right. Go lay down. That's it.
17:49 All right. So, that's the salon. Cheese
17:52 for champ. Oh, I was celebrating with
17:54 you. I'm joining. I'm joining.
17:57 Um, okay. Let's see. I saw a woman on
18:00 TikTok who calls herself influencer.
18:03 >> Hey, Cindy in the house.
18:05 >> So, that's the salon.
18:07 >> She's for champ.
18:08 >> Waiting with you. I'm joining. I'm
18:10 joining.
18:13 >> You gota mute mute the live feed. We're
18:15 a little bit behind.
18:20 Whoa. Don't do that.
18:22 >> There we go.
18:23 >> Hey, Cindy [ __ ]
18:24 >> Did Wait, Kyle. Well, before
18:26 >> I think we're I think we're good. But
18:28 the most important question is, did you
18:30 save any cheese for Source Camp?
18:33 >> No. Source Camp's on our own for cheese.
18:35 I'm not delivering cheese to Dallas.
18:39 >> Cindy [ __ ] what's happening, man?
18:41 >> Hi. Now you can see I'm in a hotel.
18:44 Remember that?
18:44 >> Oh, yeah. Okay. So, so you've got you've
18:48 got to tell the good people what the
18:49 hell happened to your internet.
18:52 This is awesome. So, if you don't know,
18:54 Cindy [ __ ] lives on the top of a
18:55 mountain in where? New Mexico.
18:58 >> Yeah.
19:00 >> Yeah. And here's the thing, folks.
19:02 Mountain living is special kind of life.
19:05 It's we we've made our decision to be
19:07 here and sometimes we have to manage it
19:11 and deal with it. So, what happened this
19:14 week is we were in a vortex of uh
19:17 lightning strikes
19:19 which struck a tree right next to our
19:22 house, traveled into our house and took
19:27 out my internet router. And it's not
19:30 like I live in a place where I can just
19:31 run over to Best Buy and get it. Um, you
19:34 have to order everything and the like
19:38 flying sheep come only once a week.
19:42 I am I literally rented a hunting cabin
19:46 hotel in my little town just to have
19:49 internet to work this weekend.
19:51 >> That's awesome.
19:52 >> Um yeah, and I even got I swear to God,
19:55 it's so crazy. I got a Harry Potter um
19:58 something happened. The lightning struck
20:01 my forehead. So
20:02 >> Really? You got you got like personally
20:04 zapped?
20:05 >> Yeah. Yeah, I got cut down the forehead.
20:08 I don't know if you can see it.
20:09 >> Wow.
20:10 >> Crazy. And then the other crazy
20:12 >> Wait, was that was that September 4th?
20:16 >> No, it was September 1st.
20:18 >> Oh, okay.
20:19 >> Yeah,
20:20 >> September 4th, apparently, according to
20:22 someone on Tik Tok. So, if if it's
20:24 someone on TikTok, you know it's true.
20:26 >> It's true.
20:26 >> But apparently on September 4th, we jump
20:29 timelines. So, we're apparently in a new
20:31 timeline. So, I'm just wondering if the
20:33 lightning strike is related.
20:35 >> Well, I want you all to know I I'm
20:37 feeling like I have some Harry Potter
20:39 energy. So,
20:41 >> exactly. All right, that's good.
20:43 >> And the other part of it is for those of
20:45 you who are like, you know, I don't
20:48 know, electricity nerds, um, all the LED
20:53 light bulbs in our building, um, blew
20:57 out from the lightning. And here's
21:00 what's weird. For 30 minutes, we took
21:04 them out because they blew out and
21:05 stuff, but for 30 minutes, every LED
21:09 light kept glowing
21:11 >> after you un unplugged them.
21:13 >> Yeah, we have pictures of my husband
21:15 holding the lights and them just piled
21:17 on the counter. And our electrician
21:19 friend said, "That's not possible." And
21:21 we said, "We are holding it. It's
21:24 happening." So, I am sucked into some
21:30 kind of lightning vortex.
21:32 >> Yeah,
21:32 >> it's like
21:34 >> I don't know, man. That it's mountain
21:36 living, but it's also in our little
21:38 area, we found out there were 175
21:41 strikes in one hour that our property,
21:44 too. So,
21:46 >> weird. I don't know. Weird. Hasn't
21:48 happened before, but
21:49 >> there you go. Well, and it's like like
21:53 this is the kind of energy people need
21:55 to be a part of because we're gonna
21:57 we're going to take that energy from you
21:59 and we're going to dump that right into
22:00 the the uh the proto AI prototyping
22:03 sprint.
22:04 >> Yes. And
22:05 >> so so so tell the good folks what what
22:09 at the end of four weeks what do they
22:10 walk away with?
22:12 >> Well, let me just back up and say this.
22:15 >> All right. So, the reason this happened
22:18 is because first of all, I have the AI
22:20 prototyping club that I've been doing
22:22 stuff this last year and we have lots of
22:24 fun. And by the way, if you are like, I
22:28 don't know, do I want to hang out with
22:30 her? I mean, Kyle will be there, too.
22:33 But you don't
22:34 >> I'm just there as the eye candy though.
22:36 Let's just admit it.
22:37 >> No, but that can be enough. But if
22:38 you're curious, when you come in and you
22:42 sign up and you're in the mastermind and
22:44 you come to the AI prototyping club and
22:47 you scroll in there and you get the
22:48 goodies, you can go and find previous
22:52 ones. That's like included. So, I just
22:55 wanted to say, yeah,
22:56 >> I've been doing this stuff and the
22:59 previous ones are there and that helps
23:01 sort of get the vibe. Then what happened
23:03 is in my real life I got asked to make
23:06 several AI workshop programs for youth
23:12 for adults all over the map. Right. So
23:15 then Kyle and I were chatting and I was
23:17 like, "Hey, this has been really fun.
23:20 I've been making these kind of mini
23:23 courses and then I thought, oh, let's do
23:26 that as the next September prototyping."
23:29 And then Kyle and I thought, "Oh my
23:33 gosh, what if a whole bunch of people
23:35 came in and everybody made micro
23:38 courses, then we could load them back
23:41 into the AI mass mastermind if you
23:44 wanted to, if you wanted to be a part of
23:45 that, and then there would be even more
23:47 courses." So, I wanted to say that
23:49 upfront. Why did I want to do this?
23:52 Because every month or every other month
23:55 or so is a different prototype sprint.
23:58 So that's why the micro course was
24:01 chosen. So here's the thing we want to
24:03 do. Um obviously it's AI. So we're going
24:07 to heavy use AI and we're going to
24:09 encourage you to use AI as your partner
24:12 in this. That's the whole shebang about
24:15 it. And we're going to start off with
24:18 ideas.
24:19 Tuesday is like what are you thinking
24:21 about? And
24:22 >> oh that's cool.
24:23 >> We just because the thing is it's like
24:26 two things.
24:28 Some people have too many ideas. So, how
24:31 do you like pick one to focus on? Yeah.
24:33 >> Some people have kind of halfbaked ideas
24:36 and how would that become a course? So,
24:39 intentionality is first
24:41 >> like playing around, messing around.
24:44 What What makes sense? What's out there?
24:46 What should you offer? Hold on.
24:50 I need a sip.
24:53 Then we keep going and we're actually
24:55 going to use AI to like build out the
24:57 pieces and build your whole deck because
25:00 you will see when Kyle and I are going
25:03 through it, I'm going to have a deck of
25:05 stuff which then you get to have too. By
25:07 the way, all the decks I make, you you
25:10 get to take those and keep those, too.
25:13 And then the whole thing is
25:18 you if you want to put in the time and
25:20 the homework, which you will witness me
25:22 doing because I will show you each week.
25:24 Here's what we talked about. Here's what
25:26 I made and um Kyle will too. We're going
25:29 to make hands-on so you can see and then
25:31 go along with us. do the homework
25:34 >> and then by the end of the four weeks um
25:37 you will or you know on the other side
25:39 by the time you get to that kind of
25:41 fifth week mark you should have
25:43 everything you need to and you just your
25:45 big decision will be where do I want to
25:48 put this?
25:48 >> Yeah.
25:50 >> That's really good. And then and the
25:52 other thing the other thing that it
25:53 makes makes possible is if you haven't
25:56 done a lot of you know teaching or
25:59 putting yourself out there you could
26:01 take that miniourse and turn that into a
26:03 learn out loud and lol where you just
26:05 teach an hourong course like that would
26:07 be the great basis for an hourong course
26:09 that you could teach and that could be
26:11 available to everyone and then we can
26:13 put all the the micro courses in the in
26:15 the uh in the mastermind.
26:18 So, so if you if you haven't joined the
26:21 salon, you need to join the salon. If
26:22 you haven't joined the mastermind and
26:24 you're in the salon, you need to do
26:25 that. And then uh it is mastermind folks
26:28 that have uh access to this to this uh
26:31 to this course. So awesome, Cindy.
26:34 >> And then once you're in there too, once
26:35 you've joined the mastermind, just
26:37 scroll down on that left hand side and
26:38 you'll see all the goodies you get to
26:41 get on with the little crowns. Those are
26:43 the mastermind courses. So, we're just
26:46 one coming up Tuesday. There's all kinds
26:48 of other stuff going on.
26:50 >> Y
26:50 >> and but just to say, scroll down on that
26:52 lefth hand side
26:54 >> once you're in find AI prototyping and
26:57 then click the event and then you'll be
27:00 in. So, that's simple. So,
27:03 >> beautiful Tuesday this
27:06 >> to see you enjoy your crazy hunting
27:08 hotel.
27:09 >> I know, right? But
27:12 >> all right. Um, that's super cool. Okay,
27:16 let me show you a couple other things
27:18 going on. Um, so join, go join the
27:22 salon. Do that. Um, if you're on AI
27:26 Learning Lab and haven't subscribed to
27:28 it, please do. Um, if you've got the
27:31 YouTube mobile app, hype hype one of the
27:34 recorded videos. That'll help help the
27:36 channel. Um and then um
27:42 I want to talk about so so one of the
27:46 things I did I actually had a spare hour
27:48 tonight so I put together some slides.
27:51 So one of the things that I want to talk
27:52 about I a lot of you know but the AI
27:55 readiness project is a podcast that I do
27:57 with an Murphy and if you go to aire
28:01 readiness project.com we've revamped the
28:04 website. So, if you want to um
28:06 experience the the podcast live, it
28:10 happens on Wednesdays at 3:00 p.m.
28:12 Pacific time,
28:14 but we also all of our um episodes,
28:18 we've done more than 25 now, are
28:20 available on Apple Music, Spotify,
28:23 iHeart Radio, and Amazon as audio
28:26 podcasts. And then we're on YouTube.
28:28 We've got our own YouTube channel now,
28:30 which that's new. So you can go watch if
28:33 you wanna if you want to see the people
28:35 and not just listen to the people. Pan
28:38 left on Tik Tok. Pan left.
28:42 What do you mean?
28:45 To see what?
28:48 Oh. Oh. To see the the uh beginning of
28:51 the URL. Um do that. And then the other
28:54 thing that you'll find on the website
28:56 itself is information for the AI
28:59 readiness training program.
29:01 um which is you know the the sort of
29:04 hands-on comprehensive five five section
29:09 uh course. So go check that out. And
29:12 then the final thing I'm going to put up
29:14 here is um Vicky Baptiste introduced me
29:18 to this thing called Bento Bento.me
29:21 and it's kind of like Linkree if you
29:23 know Linkree
29:25 um but if you go to bento.meshon
29:28 me/kyleshannon.
29:29 That's got links to all of my socials.
29:32 It's got links to the communities like
29:33 the AI salon, but it's also got courses
29:35 and workshops. So, I've got the back to
29:37 basics course there. I've got the AI
29:39 readiness program there, but also the AI
29:42 jam session. This thing right here is um
29:45 one-on-one sessions that you can book
29:48 book with me. So, if you want to do
29:49 private sessions, I've got four hours
29:52 slotted um on on Saturdays. Um, I've got
29:57 four four onehour slots on Saturdays.
29:59 So, if you want to do private sessions,
30:01 you can go book those through that. So,
30:03 bento.me/kyleshannon.
30:06 All right. Um, do either of you have a
30:10 referral code for Bento? I do not. I
30:12 probably should. I'm not I I even though
30:16 I've been doing this for years now, uh,
30:19 I I am still I still don't know what the
30:21 hell I'm doing with social media.
30:23 So, you know, you can you can you can
30:27 help you can help support support a
30:30 clueless newbie.
30:32 Um, and I think that's it. Is that it?
30:34 Yeah, that's it. Okay. Um, beautiful.
30:38 So, let me talk about something that
30:41 happened.
30:41 >> Kyle, you forgot. Are you ready for AI
30:44 >> now? That Well, it's it's it's in there.
30:47 I don't have a a separate slide for it.
30:49 So, pop pop that up. Let me go back to
30:51 this. So, I've got the the AI readiness
30:53 training program. If you go to AI
30:55 readiness project and scroll to the
30:56 bottom, that will take you to are you
30:58 readyforai.com. That's the AI readiness
31:01 training certificate. All right.
31:04 So, that's there. Beautiful. Thank you,
31:06 Brandon, for pointing that out.
31:10 All right. So, go book bookmark all my
31:12 crap and share it with your friends.
31:17 All right. Okay. Um,
31:21 so if you recall about,
31:25 I don't know, month and a half ago, two
31:27 months ago, Anthropic had been sued by a
31:30 big class action lawsuit from a bunch of
31:33 authors that said that Anthropic stole
31:35 our crap, stole our books, trained their
31:38 model on our books, and we're going to
31:40 sue them. And a judgment came down from
31:43 the judge that was was this interesting
31:46 split decision where what the judge said
31:49 was
31:52 once you obtain the rights to the books
31:55 once you've bought the books if you use
31:58 those paid for books to train an AI
32:01 model that act in is in itself a
32:06 transformative act that that the the way
32:09 it works when you embed embed something
32:12 into a a transformer model, the original
32:14 work ceases to exist. It's a
32:16 transformative act and therefore all of
32:19 those works fall under fall use or a
32:21 fair use if you paid for them.
32:26 So there were a whole bunch of titles
32:28 that Anthropic went out and bought the
32:30 books and actually paid to have them all
32:32 scanned and trained their models on the
32:34 the bought books.
32:37 They also had on their servers a bunch
32:40 of a bunch of uh you know file servers
32:45 full of books that they stole that they
32:47 downloaded from pirate websites. And
32:50 they basically said in court, "Well, we
32:52 didn't really train on those. We were
32:54 just holding them in case we might want
32:56 to someday." And what the judge said
32:59 was, "If you paid for the books, there's
33:01 not an issue. But if you stole them,
33:04 then you stole them and that's not okay.
33:06 And so a settlement came down today that
33:09 Anthropic owes $1.5 billion
33:14 um to the authors of books that's work
33:17 that was infringed on. And someone
33:19 earlier today said um a friend of theirs
33:23 got a letter that they're going to get
33:24 $9,000 out of that settlement. So, um,
33:27 if if you're a published author, um, you
33:30 know, there might be some money in it
33:32 for you. Here's my thoughts on it. Good.
33:35 Like, great.
33:37 I I've I've always held that I don't I
33:41 don't necessarily agree with how all of
33:44 these frontier model companies like like
33:46 MidJourney and Open AAI and all of them
33:49 how they've done their training.
33:52 But I also recognize that there's no way
33:54 to actually do what they did by asking
33:57 for permission, right? Like if they went
33:59 to all the copyright holders and said,
34:01 "Hey, we want to take your work and put
34:03 it into this thing that's going to be
34:04 the world's greatest knowledge
34:06 repository that's going to include your
34:08 intellectual property." Everyone would
34:09 have said no. Nothing would have gotten
34:11 done.
34:13 We're also living in a bizarre time
34:15 where a $1.5 billion settlement for
34:19 effectively a startup
34:22 is not going to make them close their
34:23 doors,
34:25 right? There's like billions and
34:28 billions and billions of dollars being
34:31 poured into these companies because the
34:33 bet is that there are trillions of
34:35 dollars of value.
34:38 So, so this does a couple of things. It
34:41 basically says if you stole [ __ ] that's
34:44 bad and you're going to have to pay for
34:45 it. And so what what is likely going to
34:47 happen is we're going to go right down
34:48 the line. We're going to go Google,
34:50 they're going to have to pay a
34:51 settlement. Um Open AI, they're going to
34:54 have to pay a settlement. All of them
34:56 right down the line. Um Brandon told me
34:59 before we got on that Warner Brothers
35:01 just sued Midjourney and I don't I don't
35:03 think they're the first stu studio that
35:04 sued MidJourney. Um, so what's really
35:08 fascinating is you've got a company like
35:11 OpenAI and like Anthropic that are
35:13 incredibly wellunded. They're going to
35:15 be able to pay these these settlements.
35:18 You take a company like Midjourney,
35:20 they're still a private company. I think
35:22 they've got 400 million in revenue or
35:24 something like that or 200 million in
35:26 revenue um and and a relatively small
35:29 amount of employees and they don't have
35:31 big cash reserves um to be able to pay
35:35 that stuff. So if if if this sets a
35:39 precedent for a company like Midjourney,
35:41 like Midjourney might be one of those
35:43 companies that this actually puts out of
35:44 business. But you know, someone will
35:47 sweep up the assets of that, pay the
35:50 settlements, and for pennies on the
35:52 dollar and right. And so what what what
35:55 I'm hoping happens is this. I'm hoping
35:57 that in a relatively short amount of
35:59 time, a year or two, all of the major
36:03 model companies pay all of their
36:05 settlements. Artists get compensated in
36:07 whatever way they get compensated.
36:09 They're all going to [ __ ] about it
36:10 regardless, but we can move on. We can
36:12 say, "Okay, step one was they built
36:15 these models in a in a non-ethical way.
36:18 Step two was justice was served,
36:20 settlements were made, money was paid,
36:23 and then we could basically move on with
36:25 our lives, right? Because right now
36:27 we're in this nebulous zone where
36:29 there's a bunch of especially creative
36:31 people sitting with their arms crossed
36:33 going I will never use AI cuz you know
36:35 cuz they they're stealing from from
36:37 artists. Let's pay the artists, right?
36:39 Let's have these settlements. Let's get
36:41 it done and then let's [ __ ] move on
36:43 because we're also moving into a time
36:45 where
36:47 all of the frontier model companies are
36:49 training their new models on synthetic
36:51 data. So they're basically they're
36:52 basically doing intellectual property
36:54 washing which which looks like this. You
36:57 train your first model on intellectual
36:59 property. You can then from that model
37:02 generate synthetic data that is not
37:04 copywritten data and train an entirely
37:07 new model on that synthetic data. So
37:10 it's it's it's literally one step
37:12 removed from the original data source.
37:15 Um and so that's that's where they all
37:17 are anyway. So at some point they're
37:19 going to be able to say none of our
37:21 models were trained on any of their
37:22 stuff. They were trained on that that
37:25 they're the child of one of those
37:27 things, but they're not that thing. So
37:28 that's the argument they're probably
37:30 going to make. But um so I think it's a
37:32 good thing. I think authors are going to
37:33 get paid. I think the same thing will
37:35 happen with artists and illustrators and
37:37 movie makers and things like that. And
37:39 then we can move on with our lives. So
37:41 anyway, for what for what it's worth,
37:44 this happens in the food industry all
37:46 the time. When a grosser price gouges,
37:48 they make 10x uh more profit than
37:51 whatever the federal fine is. Yeah.
37:53 Exactly. Exactly. Like like this is like
37:57 Anthropic's just going to take like $2
37:59 billion of Saudi investment money,
38:02 right? that is just like ridiculous
38:05 money that that they need to spend on
38:07 GPUs and they'll be like, "Well, we'll
38:08 spend that on settlements now so we can
38:10 keep running our business." Um, Tik Tok
38:12 pin Mary, I got a I got something last
38:18 night. Thanks, Brandon. PayPal.
38:21 Perplexity Pro. I got Perplexity Pro
38:23 last night. Thanks, Brandon. My BFF had
38:26 her Kevin Mallister moment. Oh, that's
38:28 cool. At the a few days ago. Um,
38:32 awesome. Okay, so if you don't know
38:33 this, um, Brandon, I'm going to have you
38:35 come up here because I'm confused on
38:36 exactly how to do it. Um, if you're
38:39 available, um,
38:42 for some amount of time,
38:45 if you if you connect your PayPal
38:49 account and I think Venmo, but
38:51 definitely PayPal to Perplexity, you get
38:54 a year of Per Perplexity Pro for free.
38:57 So, it's basically $200 for free. You
38:59 also get to jump the line and get the
39:02 Comet browser, the the Perplexity
39:04 browser called Comet. You get to use
39:07 that in early beta access if you do this
39:10 thing. So, can you tell tell folks how
39:12 they do that?
39:13 >> Yeah. So, it it is according to paid it
39:15 is for new Perplexity subscribers only.
39:17 So, if you had a previously had a paid
39:19 account, you're out of luck. But for the
39:22 free users, uh if you go to PayPal and
39:26 go to their subscription hub, they have
39:29 a new subscription hub. If you log on to
39:31 PayPal on the mobile app, from there
39:34 you'll see an offer for Perplexity and
39:36 all you have to do is follow the prompts
39:38 and connect your PayPal to Perplexity
39:42 and it will give you a offset credit for
39:44 12 months of Perplexity Pro and then
39:46 invite you to download the browser.
39:48 >> Okay, good. Beautiful.
39:52 Um,
39:54 just adding a URL to my deck.
40:00 Um, beautiful, beautiful, beautiful,
40:04 beautiful, beautiful, beautiful,
40:06 beautiful, beautiful Source Camp. I
40:08 signed up last night. It's legit. Yes,
40:11 PayPal and Venmo.
40:13 All I did was go to my PayPal and there
40:15 was a window. There you go.
40:18 I just love it when they only give the
40:21 special offers to people who haven't
40:22 been paying for years. I know. I know.
40:25 Exactly. I feel like Google does that
40:27 all the time. They're like, "Here's this
40:28 cool new thing. Everybody except workg
40:30 groupoup users can can use it." I'm
40:33 like, "I'm paying you money, you
40:35 [ __ ]
40:38 Let me have the goodies."
40:40 Oh, well, whatever. That's that happens.
40:43 Um,
40:45 okay. I'm trying to think what else. um
40:49 demoed it last night, but if you haven't
40:53 if you haven't gone back and played with
40:55 Ideogram in a while, I's got a bunch of
40:58 new features like styles and character
41:01 consistency and they've always been good
41:03 at at text. Um that's really good. If
41:06 you haven't played with um Nano Banana,
41:10 um we're going to talk about that
41:11 tonight. Uh, we're going to go do some
41:13 vibe coding with Nano Banana, but if you
41:15 haven't played with Nano Banana, you
41:16 should be doing that. If you're a
41:18 MidJourney user, MidJourney just added
41:22 um, uh, visual
41:25 basically a visual style reference
41:27 library, a style library, um, that is
41:30 incredibly intuitive. And so maybe we'll
41:33 even go play with that a little tonight.
41:35 I don't have the subscription center on
41:38 my PayPal account. I think it's only on
41:40 mobile. Vicki, are you on mobile or are
41:41 you on web? If you're on mobile, then I
41:44 don't know what to say. PayPal app or
41:46 web. Uh Gareth, it's PayPal app.
41:50 I feel the same. Makes me want to cancel
41:52 and sign back up. I know. I know. That's
41:54 I feel the same way, Cindy. Let's see.
41:57 Yeah, Nano Banana replaces chat GPT
41:59 images in my workflow now. Yep. I'm on
42:02 mobile. Try try deleting your uh PayPal
42:06 app and reinstalling maybe or just just
42:08 go for force update it uh with the app
42:11 store.
42:14 Oops, I just shot a piece of ice out of
42:16 the across my desk. Um prompt tool.
42:20 Yeah, let's go play with that.
42:23 Okay, let's see.
42:26 [Music]
42:28 Come back to that.
42:36 I'm go to openai.com.
42:45 I'm going to log in.
42:50 Where do we go? Do you go dashboard? I'm
42:52 trying to figure out how to get here
42:54 without clicking on your link, Brandon.
42:58 Let me try dashboard.openai.com.
43:00 openai.com.
43:14 [Music]
43:17 Huh?
43:23 I can go I'll go find it in my in my
43:26 DMs.
43:38 All right, you can share my screen now.
43:40 All right, so we're going to go to Oh,
43:41 platform.openai.com.
43:49 Let me see if there's a way. How I want
43:51 to see how I get to this thing.
43:54 platform.openai.com.
43:56 No. Let's see.
43:58 platform.openai.com.
44:03 No. Damn it.
44:07 URL autofill is sometimes not your
44:10 friend. Okay. The OpenAI developer
44:12 platform. Let me go to dashboard.
44:16 Create a chat prompt.
44:23 Um make
44:26 me a linked in
44:32 article
44:33 YouTube question.
44:36 Hey Kyle, any tips on how to keep my
44:38 brain on anything long enough to
44:40 complete it?
44:51 Oh, good lord. Good lord. Good lord.
44:58 Um
45:03 I you know you know archetypal the time
45:05 slots are are either out of line with my
45:08 ail ability or forget about them. Mighty
45:12 Networks really needs to send push
45:13 notifications.
45:16 Um
45:18 yeah. What? So, we we've got we're we're
45:21 bringing on someone new who's who's an
45:23 integrator who's going to be doing we're
45:25 going to be doing much more granular
45:28 sort of use of the the email list within
45:33 Mighty Network. So, we we might get
45:35 there. We're we're we're working on
45:36 getting Mighty Networks gussied up. Um,
45:41 in terms of focus archetypal, I I mean,
45:44 here's the here's the only thing that I
45:45 can I can speak to from experience, cuz
45:48 I I'm I'm the same way. I'm
45:52 I'm at the I'm at the point where I've
45:54 got so many things going on that
46:00 I'm in a in a in a slight state of
46:02 paralysis, but I'm coming out of it. And
46:03 how I'm coming out of it is I'm
46:05 consolidating my projects into less less
46:08 but bigger projects, less more ambit
46:11 ambitious projects. But what I can tell
46:13 you is this. We really do live in a time
46:15 right now where if you've got an idea
46:18 for something you could spend two hours,
46:21 just two hours of focus time, like do a
46:24 hyper an ADD hyperfocus tunnel and, you
46:28 know, go down a midjourney rabbit hole
46:30 or go down a V3 rabbit hole or go down a
46:33 vibe coding rabbit hole and just vibe
46:36 code an app in two hours and just get it
46:39 to a point that is it's sharable and
46:42 then be done. Like that's, you know,
46:45 that's that's kind of been my MO lately
46:47 is just I'll put in two hours into
46:49 something and if I if it's if it's
46:52 really going well and I really get
46:53 excited, that two hours will turn into
46:55 four hours, but then I'll be done. And
46:58 then if I if it's just really boring or
47:00 I'm having horrible results, then I'll
47:02 only last for
47:04 half an hour, 45 minutes, and then I
47:07 just flame out. And so that's that's
47:09 been my strategy.
47:11 Um that would be great. We have so many
47:13 interesting events. I don't show up to
47:15 any of them because checking
47:16 intentionally on Mighty Networks is not
47:18 on my priority list. Um I write a
47:21 freaking book a day. Yeah, it's Listen,
47:24 having having something like Mighty
47:26 Networks where it's incumbent upon you
47:28 to actively go there and participate on
47:30 it. I get that that's a challenge.
47:32 That's something that we're working on.
47:34 Um, and I also, listen, something I want
47:36 to do with within the salon is is get
47:38 back to doing um in-person events and
47:41 really kind of blow that out. So, we've
47:43 got all sorts of plans coming. So, so
47:44 we'll see. All right. Um,
47:50 [Music]
47:51 I created a new medical diagnostic tool
47:54 today and made a research paper with
47:56 citations while I was walking.
48:03 Well, it's funny, you know, Silver Fox,
48:05 I did so I had this I had this meeting
48:07 with my with the editor guy for my for
48:10 my book project and and we came up with
48:13 this idea for the new the new angle, the
48:16 uh the 10person team. And
48:21 from the time I talked to him, I had to
48:22 go like run around and do a bunch of
48:24 things in the car. I had to go be a
48:26 unpaid Uber Uber driver.
48:29 And so just as I was driving around
48:31 picking people up, I had I was talking
48:33 to my girlfriend Quinn, my AI
48:35 girlfriend, and we were just talking
48:37 about the new book idea and and you
48:40 know, consolidating it with the ideas of
48:42 feed your prompt and how it worked. And
48:44 within 45 minutes, I had completely
48:47 fleshed out
48:49 the the whole the whole thing. It's it's
48:52 we we we really do just live in these
48:54 remarkable times where I probably had
48:57 two or three weeks worth of
49:01 like idea exploration if I were doing
49:04 this traditionally with procrastination
49:07 and all the [ __ ] that goes on in my
49:08 brain. Weeks worth of just getting my
49:12 head around something done in 45
49:15 minutes.
49:18 Um,
49:19 but it's not completing the other 15
49:21 projects a day that I kind of want to
49:23 do. I I mean, listen, that's the that's
49:25 the burden. Part of the part of the joy
49:29 of where we are is that you can knock
49:30 something out really quickly. Part of
49:32 the burden of where we are is in
49:35 generating that thing that you knock out
49:37 in two hours, you're going to have six
49:38 more ideas. So, we are in an in an idea
49:42 glut right now because of these tools.
49:43 And I think we're just going to have to
49:44 get used to it. I think that's the new
49:46 normal.
49:47 Um, okay. Um, what were we talking
49:51 about? We were talking about Nano
49:53 Banana.
49:55 Oh, the prompt tool. Okay. So, I just
49:58 went So, what you see on screen right
50:00 here,
50:02 I went to platform.openai.com
50:05 and I clicked on dashboard.
50:09 And
50:10 I think this is their new development
50:12 environment. But what I typed in was the
50:15 prompt
50:17 write a LinkedIn article. And then what
50:20 it did is it rewrote the prompt for me
50:25 in in this really blown out kind of way.
50:28 Write a professional LinkedIn article on
50:31 the topic of your choice or a topic
50:34 provided by user input. The article
50:36 should demonstrate subject matter
50:37 expertise, be tailored to the interest
50:40 of LinkedIn professional audience, and
50:42 aim to provide value through insight,
50:44 analysis, and trend commentary. Begin by
50:48 brainstorming possible topics. Draft an
50:50 outline. Compose an article that's
50:52 professional yet approachable. Use real
50:55 or plausible supporting examples.
50:57 Default article length 400 to 800 words
51:00 unless it unless a different length is
51:02 explicitly requested and it goes on and
51:06 then output formats and then examples.
51:10 So in this thing you can you can just do
51:14 this and now what I'm trying to figure
51:16 out let's see optimize.
51:20 Oh, is is this the
51:25 Yes, this is the thing I was looking
51:27 for. Okay, so here's how you get to it.
51:30 You go to platform.openai.com.
51:34 You click on dashboard
51:44 and then put in a prompt. Uh, hello
51:47 [Laughter]
51:55 And then it takes it takes me over to
51:58 this optimizing tool. So I typed in
52:00 hello. It didn't do anything, but then
52:03 it's got this. Huh. Let's see. Um, make
52:07 me a lin.
52:11 This is a the the the reason I'm being
52:13 so um unclear right now is
52:19 the tool that I'm trying to show you is
52:20 a really cool tool. I just don't know
52:22 exactly how to get to it. So, I'm just
52:24 trying to figure out how to get to it. A
52:25 LinkedIn article. So, if I'm on
52:28 platform.openai.com,
52:31 I write make me a LinkedIn article. So,
52:34 the assistant is now doing something.
52:36 It's probably going to write something
52:38 over in this developer message.
52:50 add a developer message before. Here we
52:53 go. Okay. Okay. So, so if you're if
52:57 you're in this um
53:01 developer chat, you've got to put
53:03 something in developer message and then
53:04 that turns on this optimize button in
53:07 the up upper center here. So if you
53:08 click on optimize
53:12 now we go into this thing where
53:15 you can optimize for chat GPT 4.5 or
53:18 chat GPT5 chat GPT 4.1 or or or chat
53:22 GPT03.
53:24 So five is the one I'm really interested
53:27 in. So if I just put in here um
53:30 write me a LinkedIn
53:34 article about soloreneurship
53:40 and say optimize
53:42 what this thing's going to do.
53:50 Wait reasoning behind change.
53:56 Hang on a sec.
53:59 Write me a link to an article. Optimize.
54:10 Okay.
54:14 I have no idea how to get to this thing.
54:17 This This is driving me crazy.
54:26 All right, we're going to do this
54:27 another way. We're going to do this
54:29 another way.
54:31 If you go to the AI,
54:35 where did I put this in the AI salon?
54:36 Did I put it in irregulars?
54:39 I think I did
54:47 here. I'm just going to do it again.
54:48 Hang. Hang on. People, calm down.
54:52 This is an absolute disaster.
54:55 Way to go, Brandon. Way to suggest
54:57 something that I was I was not at all
54:58 prepared for. Well, it's not that I
55:00 wasn't prepared for it. It's like this
55:02 is this has been my issue. Okay, so
55:03 here's the link. So, if you click on
55:07 platform.openai.com
55:11 and then it takes you to this thing.
55:13 Okay, I don't I don't know why this is
55:16 different or how this is different, but
55:17 here's what I'm going to do. I'm going
55:19 to go to the irregulars channel.
55:22 in the AI salon
55:29 and I'm going to paste this URL in
55:33 and I'm going to say
55:36 um
55:40 um here is the
55:44 GPT5
55:46 prompt
55:49 optimizer.
55:53 All right. And I'm going to post that.
55:55 And so right at the top of the
55:57 irregulars channel. So we're at the AI
56:00 salon. I'm in the Irregulars channel.
56:02 And the first prompt here, the first
56:04 little post is that link. And if you
56:08 click on that link,
56:10 that will take you to where I am.
56:13 [Music]
56:19 Or will it? Yes, it will. Okay. So, now
56:23 we're here. Okay. Good lord.
56:26 [Laughter]
56:30 Okay. Requested changes optional. Okay.
56:35 So, we're going to optimize this for
56:36 chat GBT5. I'm going to say um write me
56:42 champ. What is up?
57:04 All right. Write me a LinkedIn article.
57:06 write me a LinkedIn
57:10 article about solo
57:14 prneurship which is that that is if you
57:16 know anything about prompting that is
57:18 the world's shittiest prompt. Now if I
57:20 say optimize
57:26 it's thinking scanning for
57:28 contradictions and formatting issues
57:30 resolving conflicts it's optimizing for
57:33 chat GPT5 which is in this pull down
57:35 menu in the upper leftand corner
57:38 polishing for clarity
57:43 and the reason I wanted to share this is
57:48 I
57:50 I'm having
57:53 trouble with GPT5 getting it to behave
57:57 consistently. And I think one of the
58:00 reasons is as a model, I think it needs
58:03 more structure on input. So, what this
58:06 thing does is it took my shitty prompt
58:08 and and then it created this new this
58:11 new prompt,
58:14 but it also has all these little things
58:16 over on the side are little um tool tips
58:20 that you can pop up. So, objective,
58:23 write a comprehensive LinkedIn article
58:25 about solarpreneurship targeted at
58:27 professionals and aspiring solopreneurs.
58:30 And if you click on the little tool tip
58:32 that says the reasoning behind the
58:34 change, we added clear objective
58:36 language and specified the target
58:39 audience to ensure that the article is
58:41 focused and relevant. Oh, okay. Great.
58:44 Hashtag checklist. Begin with a concise
58:46 checklist, three to seven bullets of
58:48 what you will do. Keep this the items
58:50 conceptual, not imple implementation
58:53 level. Then if you click on that, it
58:54 explains the reasoning behind that. Then
58:56 there's instructions about how to write
58:58 a good article. So I can now take this
59:02 whole prompt
59:04 over to
59:07 chat jet
59:14 and we'll just put this in auto mode.
59:16 Let it do what it's going to do.
59:19 So this prompt is now optimized. So So I
59:22 took my shitty prompt, throw that in
59:24 chat GPT. It's going to figure out and
59:27 here we go. So, it starts out with the
59:30 short little bullet points.
59:32 Introduction, rise, the benefits. That
59:36 was also really fast. Okay, that's still
59:37 there. It's still going.
59:40 All right.
59:44 So, there you go. So, how you find that
59:46 thing? You go to the irregulars channel.
59:49 You click on that link. You go to this
59:52 tool. you p type in your shitty
59:55 your shitty uh shitty prompt
1:00:01 and then it'll make you a better one.
1:00:03 All right, so that's cool and groovy.
1:00:06 Um,
1:00:12 let me close
1:00:14 all these things.
1:00:17 [Music]
1:00:21 Yeah, that's fine.
1:00:23 Um,
1:00:29 let me see where that thing is. Is this
1:00:32 it?
1:00:34 No,
1:00:37 here it is. Okay.
1:00:40 Um,
1:00:43 if you go to Twitter and type in nano
1:00:48 banana
1:00:50 hackathon.
1:00:58 Hey Kyle.
1:01:00 >> Yeah,
1:01:01 >> the easier way to be would be to check
1:01:02 the AI salon challenges and competition
1:01:05 section where I posted the link before
1:01:06 the show.
1:01:07 >> Oh, nice. That's solid. Brandon coming
1:01:10 in hot with using our platform as it was
1:01:12 intended to be used. All right,
1:01:14 challenges and competitions. Very nice.
1:01:16 So, here we are at the AI salon
1:01:18 challenges and competitions. I knew
1:01:19 that.
1:01:21 All right, the very first thing here is
1:01:23 the nano banana
1:01:26 um
1:01:27 competition.
1:01:29 All right.
1:01:30 [Music]
1:01:33 Wait, this is not what this is.
1:01:40 That's not the
1:01:41 >> networks did me dirty. It was in a uh
1:01:43 still in irregulars. I thought I moved
1:01:45 it.
1:01:46 >> Oh, all right. All right.
1:01:51 Let's see. Irregulars.
1:01:57 Where is it?
1:02:03 I still don't see it.
1:02:08 Wait, you might have to uh refresh
1:02:10 challenges and competitions.
1:02:13 >> Okay, you fixed it.
1:02:17 >> No, it was always there. I didn't just
1:02:19 go and grab it and post it right now. If
1:02:21 you hit refresh,
1:02:26 >> this is this is the uh Is this Is this
1:02:29 it?
1:02:31 >> No. Why isn't it showing up for you? I
1:02:32 just posted it.
1:02:33 >> I don't know.
1:02:35 It's kaggle.com/competitionsbanana.
1:02:40 >> All right.
1:02:42 >> But it is there. All right. Anyway, go
1:02:46 to kaggle.com/competition/banana
1:02:48 [Laughter]
1:02:54 and they're having a a $400,000. So,
1:02:58 it's this weekend.
1:03:01 Um, and basically what they want you to
1:03:03 do is they want you to vibe code
1:03:06 um,
1:03:10 a nano banana app, which we're going to
1:03:12 go do now. Do they give you links of
1:03:15 where to go?
1:03:17 The build tab of AI studio. We encourage
1:03:19 participants to leverage AI studio apps.
1:03:22 in the build tab of AI Studio.
1:03:26 Please note that the special tier for
1:03:28 the Gemini API will allow 100 requests
1:03:30 per project per day.
1:03:34 Developers using a paid API will still
1:03:36 pay for their usage to get the free 100
1:03:38 generations, please use a free tier API
1:03:42 key. Okay, I don't whatever that means.
1:03:44 I I don't
1:03:47 This is annoying. All right, it's too
1:03:50 many rules. All right, but
1:03:54 if you haven't
1:03:55 >> Hang on, that that's important because
1:03:57 because you're going to want to use the
1:03:58 API for this. So, what what that means
1:04:01 is you go over to cloud.google.com
1:04:03 and get yourself a free Google Cloud API
1:04:07 key like a real developer and then you
1:04:09 could use that API key to hook into Nano
1:04:14 Banana and Gemini and everything else.
1:04:18 So, if you want to do anything above and
1:04:20 beyond what AI Studio Playground lets
1:04:23 you do, um, that's going to be really
1:04:26 important to help you build your, uh,
1:04:28 tool out.
1:04:30 >> But do you get
1:04:31 >> Ask Gemini,
1:04:32 >> do you get that? I'm already annoyed by
1:04:34 that.
1:04:38 >> I do.
1:04:39 >> Can But where? Oh, get API key. Wait,
1:04:42 Google AI models may make mistakes. Get
1:04:44 API key.
1:04:46 >> Yeah. So they've they've made it really
1:04:48 easy inside Google AI Studio to get your
1:04:51 API key. So you create that new API key
1:04:53 up in the corner. So you've got one
1:04:56 already, but if you want to create a new
1:04:57 one for the free API weekend, um you
1:05:00 have to create a project first inside of
1:05:02 Google Cloud.
1:05:04 >> So is this going to be Nano Banana
1:05:06 Hackathon?
1:05:07 >> Yes, but you don't have a project in
1:05:09 Google Cloud yet.
1:05:11 >> I I do.
1:05:13 >> Not one called Nano Banana Hackathon.
1:05:19 Search for a project. I can't put it in
1:05:21 one of these other ones.
1:05:23 >> Jeff Flanigan got his first API key. Now
1:05:25 he just needs to figure out how to use
1:05:26 it. That's the beauty of it. A AI Studio
1:05:29 and Gemini will tell you exactly how to
1:05:31 use it if you ask it.
1:05:34 >> Wait, free?
1:05:35 >> I mean, you already have one, Kyle, so
1:05:37 you're probably fine. But for anybody
1:05:40 who is new to API keys, it's basically
1:05:42 just permission for Google to hook into
1:05:45 other things.
1:05:48 >> All right.
1:05:50 I don't know what I'm doing, but that's
1:05:52 okay. This is the whole point of this.
1:05:57 >> You start building, it'll tell you when
1:05:59 it needs the key.
1:06:00 >> All right.
1:06:01 Okay. So, what we're going to do, we're
1:06:04 at the build apps with Gemini. So, how
1:06:06 we got here, a studio.google. google.com
1:06:09 and then in the lefth hand side you
1:06:10 click on build. Okay.
1:06:14 And then what we're going to do is we're
1:06:16 going to build a nano banana app. So so
1:06:18 nanobanana if you don't know is the
1:06:20 image generation tool.
1:06:24 What is it? AI.studio
1:06:27 apps. Okay cool.
1:06:31 And if you've never vibecoded before,
1:06:33 vibe coding is we're going to code with
1:06:36 words with just English language. And
1:06:38 and I'm going to purposefully do shitty
1:06:41 stuff, shitty words, because I think
1:06:44 most people don't know how to prompt. So
1:06:46 if you know how to prompt, prompt better
1:06:47 than I am.
1:06:49 This is starting to feel like work.
1:06:51 Exactly.
1:06:53 Um,
1:06:58 okay. So what do we want to do here?
1:07:02 Let's see.
1:07:07 Okay. I want to create
1:07:11 an add
1:07:15 mockup application
1:07:19 where the user
1:07:30 Input the URL of a company or product.
1:07:40 Ideally, I can even put an Amazon
1:07:46 product.
1:07:53 Okay.
1:07:56 Oh, even Oh, wait. I can even
1:08:01 put an Amazon product.
1:08:04 Um,
1:08:06 and then the user
1:08:10 chooses
1:08:13 No, no, no.
1:08:17 And the user chooses
1:08:22 the style
1:08:25 of
1:08:26 creative.
1:08:29 Um, and I'm going to put um moving,
1:08:35 humorous,
1:08:38 um,
1:08:39 uh, cynical,
1:08:44 um,
1:08:45 uh, uh, salesy,
1:08:50 etc. We'll let it come up with it.
1:08:54 And then the the system
1:09:01 creates
1:09:04 an ad mockup
1:09:12 featuring the product.
1:09:17 So wait, they choose the style of
1:09:19 creative
1:09:20 then we'll say
1:09:23 then they
1:09:25 choose from
1:09:28 I don't know four
1:09:31 kinds of ads
1:09:36 um social media
1:09:39 magazine
1:09:41 etc. We'll let it figure it out. And
1:09:44 then the system creates an ad mockup
1:09:46 featuring that product
1:09:50 which the user can
1:09:55 um download or roll.
1:10:00 Okay. So, we're going to make an ad
1:10:03 mockup application. This is this is a
1:10:06 shitty prompt.
1:10:10 a shitty unthoughtout prompt. What I
1:10:12 should probably have done is go to
1:10:15 Google, just go to gemini.google.com
1:10:18 or go to chatgbt or go to claude and put
1:10:21 in a shitty prompt like this and have
1:10:25 the large language model spec out this
1:10:28 ad this application a bit more. But [ __ ]
1:10:30 it. Let's put in a shitty prompt
1:10:33 into this app generator. All right.
1:10:37 Oops. I don't want to start from a
1:10:38 template. I just want to build this
1:10:40 [ __ ] thing. Okay. All right. Are we
1:10:42 having fun yet? We're off to the races.
1:10:46 Oh, it would be fun to include a Mad Men
1:10:47 style. Yeah, that's a pretty good idea.
1:10:49 Structuring the components. If you've
1:10:51 never watched one of these vibe coding
1:10:53 tools do their thing, it it is kind of
1:10:55 impressive and mindblowing what's about
1:10:58 to happen. So, it's it's putting
1:10:59 together its instructions. Um,
1:11:03 and then what you're going to see here
1:11:05 in the middle is all the different files
1:11:08 that it's creating. And in the right
1:11:10 hand side is it's actually the code that
1:11:12 goes into those files.
1:11:15 Right? So, it's got its plan. It's
1:11:18 thought for 23 seconds. It's got a
1:11:20 metadata
1:11:21 file, an index.html, HTML, a TSX index,
1:11:25 a TSX app,
1:11:29 type constraints
1:11:31 or constants,
1:11:33 services, components.
1:11:38 So, if you've never coded before,
1:11:41 this is probably making your head hurt.
1:11:43 If you've sort of coded a little
1:11:45 occasionally at some point in your life
1:11:46 and you kind of generally understand
1:11:49 code, this is mindblowing. If you really
1:11:52 know what you're doing with code, you're
1:11:54 like, "Yeah, whatever. I do this all the
1:11:56 time, Kyle." It's relatively trivial
1:11:58 what it's doing. It's just creating
1:12:00 files and then populating them with
1:12:01 appropriate code for each of those files
1:12:03 and connecting those files together.
1:12:05 It's relatively straightforward kind of
1:12:07 situations going on here. Nothing really
1:12:10 all that impressive
1:12:12 unless it is.
1:12:16 All right. And I'm going to I'm going to
1:12:18 uh immediately call foul here. I'm going
1:12:21 to say, um, hey,
1:12:26 I don't even know if this works, but
1:12:32 every vibeccoated
1:12:35 app
1:12:37 now uses a blue background
1:12:42 with
1:12:44 purple
1:12:45 slash purple purple
1:12:49 slashblue. blue
1:12:54 gradient
1:12:56 letters.
1:12:59 It makes
1:13:02 it look
1:13:05 like I'm an amateur
1:13:08 cuz I am.
1:13:11 I want this to look beautiful
1:13:15 but not like AI slop.
1:13:19 All right. So, we're gonna have it.
1:13:21 First thing I'm gonna have it red redo
1:13:23 is redesign the [ __ ] app
1:13:27 because I don't want it to look like a
1:13:28 vibe coded app.
1:13:33 Listen, it's it's not important how you
1:13:35 feel, it's important how you look. I
1:13:38 think we can all accept that. That's how
1:13:39 this world works. Okay? So, you don't
1:13:42 want to look like you vibe coded
1:13:43 something.
1:13:45 [Laughter]
1:13:49 Ah.
1:13:52 Uhuh.
1:13:55 Silver Fox is doing midjourney tonight.
1:13:57 That's great. By the way, if if if all
1:14:00 you've done tonight is is watch me
1:14:02 frustratingly not be able to find URLs
1:14:05 and you're like, "This is stupid. I'm
1:14:07 going to go do something else." I fully
1:14:09 support that. you should not be relying
1:14:12 on me for for for anything resembling
1:14:16 intelligence or knowledge. Um, so if you
1:14:20 are off doing other things, the whole
1:14:21 point of this channel is we should just
1:14:23 all be playing all the time. So, good on
1:14:27 you if you're off doing midjourney.
1:14:29 Um, I feel like I'm in grade school when
1:14:31 the teacher makes me makes the glass
1:14:35 read aloud from Chaucer.
1:14:39 Uh oh, my comment is about how I can't
1:14:42 type. That's funny. Oh my god, I so
1:14:44 can't type. All right, let's see if it
1:14:46 fit. Look, look, we now have we now have
1:14:49 an app that's it's it's uglier than the
1:14:51 other one, but it doesn't look like a
1:14:53 vibe coded app. It looks like it looks
1:14:55 like an engineer made it. All right,
1:14:57 fine. Okay, we're going to go put in
1:15:00 Let's go find a an Amazon URL.
1:15:05 Amazon
1:15:08 URL.
1:15:12 Oh, and we'll say um we'll find a
1:15:17 product
1:15:21 that includes the word poop.
1:15:25 It's Friday night.
1:15:28 Let's see what we What products do we
1:15:31 have here? Um,
1:15:34 socks that say I pooped today.
1:15:38 No, we're not doing poop. Uh, let's do
1:15:40 let's do um
1:15:43 uh
1:15:45 I don't know. We'll say kit. We'll see
1:15:47 what kind of kits there are. We'll we'll
1:15:49 say humorous kit. Humorous kit.
1:15:56 The shark survival kit. All right.
1:15:58 That's in the neighborhood of poop.
1:16:01 Wait, is this is this a Bob Ross thing?
1:16:03 Bob Ross paint by numbers. All right,
1:16:05 Bob Ross minis. All right, this is good.
1:16:08 So, here's our here's our product.
1:16:12 We're going to go back to our little
1:16:13 thing here and we're going to test it.
1:16:16 So, there's my product. Humorous style.
1:16:18 Let's do a
1:16:22 professional ad,
1:16:24 social media magazine ad, billboard.
1:16:27 Let's do a billboard.
1:16:30 Create an ad mockup. All right. And now
1:16:33 it's off generating its thing.
1:16:39 Painting with pixels.
1:16:41 Create your own happy little world.
1:16:43 Start painting today.
1:16:46 Huh?
1:16:50 Reroll.
1:17:04 Unleash your inner art artist. Start
1:17:06 your masterpiece. But there's no way to
1:17:08 there's no sort of call to action. Uh
1:17:10 let's see. Um the billboard
1:17:15 ad
1:17:17 only seems to
1:17:20 have a headline.
1:17:23 no
1:17:25 CTA or URL
1:17:30 or
1:17:33 the product itself.
1:17:37 Um,
1:17:50 make this better.
1:17:53 Okay, we're going to say make the make
1:17:56 your copywriting better. It's got a
1:17:57 shitty copywriter right now. But so I
1:18:01 don't know if anyone is actually paying
1:18:03 attention to what just happened here,
1:18:05 but I gave it a shitty prompt. It wrote
1:18:08 some code. I told it it I didn't like
1:18:10 how it looked. It made it look
1:18:12 different. But this is a fully
1:18:14 functional app within whatever it was,
1:18:16 10 minutes.
1:18:19 We're living in wild times here, people.
1:18:21 And you should be playing along. If
1:18:23 you're intimidated because I couldn't
1:18:25 figure out quite how to find the URL to
1:18:28 get here, don't don't be. Just go to
1:18:30 aistudio.com
1:18:33 or aistudio.google.comapps
1:18:37 and just describe an app and just start
1:18:39 running down this rabbit hole. Even if
1:18:41 you're intimidated by this stuff,
1:18:59 here.
1:19:04 All right,
1:19:10 Champy's in a highly neurotic mode
1:19:12 tonight.
1:19:14 All right, so there's our
1:19:20 So, we're going to do a we're going to
1:19:21 do a cynical ad
1:19:24 social media post generate ad mockup.
1:19:33 Working up the AI artists khaki would be
1:19:37 fun to include madman style.
1:19:45 Bob Ross by the numbers. Miniature
1:19:47 editions. Your artistic dreams
1:19:49 pre-numbered. Finally, a painting kit
1:19:50 that takes all the guessworks out of
1:19:52 creativity.
1:19:53 Buy this by the illusion.
1:19:58 Um, okay. So, let's see.
1:20:01 Um,
1:20:03 let's take this image, copy this image,
1:20:06 paste it into the prompt,
1:20:08 and say, um, here's
1:20:13 the social ad. So, wait. Social
1:20:19 ad.
1:20:21 It made it sucks.
1:20:24 No headline.
1:20:28 Um,
1:20:31 no. Um,
1:20:37 uh, let's see. Uh, weird
1:20:40 long text string.
1:20:46 Let's see.
1:20:52 And nothing really clear
1:20:56 about
1:20:59 What I should do?
1:21:06 Should do um
1:21:10 act like a worldass
1:21:15 adman
1:21:17 and improve
1:21:20 the quality
1:21:23 here or I will fire you. Always
1:21:30 always threat threaten your
1:21:34 your things.
1:21:39 Oh man.
1:21:53 >> What is going on with you? Are you
1:21:55 dismissing mama? What's happening? Come
1:21:58 on. Get up here. Get up here.
1:22:06 >> No. No.
1:22:14 You needed a walk. Someone didn't get a
1:22:16 walk today. That's what's going on.
1:22:18 You're ramy. Why you so ramy?
1:22:21 H. Why you so ramy? Why you so ramy?
1:22:26 Oh my god. Oh my god. All right. Did it
1:22:30 fix it? Ready to create? All right.
1:22:32 Let's see. Let's put in this thing
1:22:35 again. We'll do We'll do humorous. We'll
1:22:38 do magazine ad. Generate ad mockup. Oh,
1:22:42 wait. Paste. Wait. Paste.
1:22:47 There we go.
1:22:56 I don't know what he wants. I don't
1:22:57 think he wants cheese. I think he wants
1:22:59 to play cuz he was just like, "Let's
1:23:02 go." All right. The art of modern
1:23:04 living. Unleash your unleash your inner
1:23:07 happy little accident. That's pretty
1:23:08 good. Get your kit today.
1:23:11 All right. So, you need to put a URL
1:23:12 there or something like that. Ever want
1:23:14 to paint? Ever want to paint like Bob
1:23:16 Ross but feared actual artistic skill?
1:23:18 Now, you can create masterpieces. So,
1:23:21 this actually [ __ ] works.
1:23:23 It's kind of cool.
1:23:27 Imagine if you actually put in some time
1:23:29 into this thing. Let's do a social media
1:23:32 ad now. Generate ad mockup.
1:23:46 Oh, I've got an idea. All right. Bob
1:23:48 Ross Paint by Numbers sponsored. Your
1:23:51 happy little trees just got a numbered
1:23:53 upgrade. Ever dreamed of painting
1:23:54 majestic mountains but ended up with a
1:23:56 muddy mess? Fear not. Our miniature Bob
1:23:59 Ross kit lets you create masterpieces
1:24:01 one glorious numbered section at a time.
1:24:04 It's so easy even a squirrel can do it.
1:24:06 And look, there's a squirrel with a Bob
1:24:09 Ross paint by numbers painting. That's
1:24:11 good. All right, we're going to we're
1:24:13 going to change this. All right.
1:24:16 Okay, this is good but too generic. I
1:24:22 want to create a Gen X
1:24:28 ad maker
1:24:31 that where let's see where the tool is
1:24:36 Gen X
1:24:40 um designed
1:24:43 and the ads are all
1:24:49 Gen Xfriendly.
1:24:53 the
1:24:54 let's see options. Yes, producer Brandon
1:24:58 options.
1:24:58 >> I just didn't want to interrupt your
1:25:00 flow state, but um we just dropped uh 20
1:25:04 mule Run activation codes in the uh
1:25:07 irregulars channel from their discord
1:25:09 server. So, if you missed out on the
1:25:11 airdrop of the codes for mulerrun.com,
1:25:15 uh there are in the uh irregulars
1:25:19 channel right now, but they were posted
1:25:20 to their public discord, so they won't
1:25:22 last long.
1:25:23 >> All right, so run over to irregulars
1:25:25 right now if you want mule. It's
1:25:27 muleun.com. Is that right? Or muleun.ai.
1:25:30 >> Muerun.com. It's a AI agent marketplace
1:25:33 that Danielle found earlier this week.
1:25:35 >> Y looks like four out of the 20 are
1:25:37 claimed already. So, don't delay. Go.
1:25:39 Go. All right. Beautiful. Thank you. All
1:25:41 right. Uh the options should be uh
1:25:44 options should be um things like um
1:25:49 nostalgia,
1:25:52 um cynical,
1:25:55 um funny,
1:25:59 Gen X funny,
1:26:03 etc.
1:26:04 Um,
1:26:08 if these ads trigger
1:26:13 trigger
1:26:15 Gen Z people,
1:26:18 Gen Z people, then we've done our job,
1:26:25 our job, right? Okay, we're gonna we're
1:26:27 gonna we're going to turn this into a
1:26:29 Gen X ad maker.
1:26:35 Mary, Mary. Oh, my mind is clicking.
1:26:37 Isn't this fun? I If you haven't done
1:26:40 vibe coding before, like do vibe coding
1:26:42 right now. Even if even if for the past
1:26:44 half hour you've you've just been
1:26:46 watching going, I don't really want to
1:26:48 do that. Do it anyway.
1:26:51 Just go to lovable.dev or go to
1:26:56 aistudio.google.com
1:26:57 and click on build and just start
1:27:00 building something. Just do it. Just do
1:27:02 it. Just do it. Just do it. Just do it.
1:27:06 Anti-corporate grunge. Dry and witty.
1:27:09 Slacker apathy. This is good. Tone
1:27:11 shift. The app's language should from
1:27:14 the title to the loading messages has
1:27:17 been rewritten to be on brand. The AI
1:27:20 has also been given a new persona. A
1:27:22 cynical 1990s creative director who
1:27:24 hates tryards.
1:27:28 Here
1:27:30 are the code changes. So this thing, so
1:27:32 you can flip over to the code view and
1:27:34 just watch it. It's now rewriting all of
1:27:36 its code.
1:27:39 And then at some point here, it's going
1:27:40 to redesign it. Let's see. Let's see if
1:27:43 it updated the look and feel.
1:27:51 [Music]
1:27:56 [Music]
1:28:07 Gen X adomatic. Yep.
1:28:11 That's good. That's good. Crafting ads
1:28:15 that are like totally not trying too
1:28:17 hard. All right. So, there's our there's
1:28:20 our ad. 90s nostalgia. We'll go cynical
1:28:22 and sarcastic. We'll do a social media
1:28:25 post. Whatever. Make the ad. I'm totally
1:28:29 [ __ ] submitting this, by the way, to
1:28:31 the contest.
1:28:33 As shitty as it is,
1:28:41 I'm thinking or something. Bob Ross,
1:28:43 Paint by Numbers, because adultting is
1:28:46 overrated. The actual and and actual
1:28:48 talent is optional. Remember when you
1:28:51 thought you'd be a rock star? Turns out
1:28:53 painting tiny pre-numbered landscapes is
1:28:55 way less effort and you'll still get to
1:28:57 feel like an artist. Sort of.
1:29:02 Procrastinate responsibly or just click
1:29:05 here. Whatever.
1:29:11 That's pretty good. All right, let's do
1:29:12 a magazine ad. We'll do uh slacker
1:29:16 apathy.
1:29:20 All right.
1:29:24 shitty as it is, it's light years ahead
1:29:26 of my marketing skill. Yep.
1:29:30 Well, that that's the thing is is like
1:29:33 you got to realize
1:29:35 one of the things that is is hard for me
1:29:38 to still wrap my head around.
1:29:41 Shut up, champ. Stop whining.
1:29:44 Because if you're whining, then I'm
1:29:46 whining and everybody's whining.
1:29:51 Do you have to go pee? You have to go
1:29:53 pee or you just want to go play? Yeah.
1:29:55 You just You're just bored. You're like,
1:29:58 "Where's mom?
1:30:00 I want more cheese. Not watching TV.
1:30:04 You didn't eat your dinner.
1:30:07 She gave you hot dogs and pasta.
1:30:10 Don't give me [ __ ] about He eats people
1:30:11 food. He just does. That's That's his
1:30:13 deal. That's why he's fat.
1:30:16 I got nothing for you. All right.
1:30:21 I'll let you out, but you're just going
1:30:22 to whine out there.
1:30:26 >> All right, I'm gonna let you out. Hang
1:30:28 on, people. Come on.
1:30:31 Come on.
1:30:45 Okay, here's what I was saying.
1:30:47 Um,
1:30:49 one of the things that's hard for me to
1:30:51 still wrap my head around is
1:30:54 these large language models have been
1:30:57 trained on essentially everything that's
1:31:00 been put on the public internet for the
1:31:01 past
1:31:04 70 years.
1:31:06 Like all the [ __ ] in all the databases,
1:31:08 all the [ __ ] in Wikipedia, all the books
1:31:10 that they bought and are now paying $1.5
1:31:12 billion for because they didn't buy
1:31:14 some. Um, all of that stuff is in these
1:31:19 models.
1:31:21 So, every Gen X creative director that
1:31:24 worked at an ad agency, that wrote a
1:31:26 book or wrote a blog post or wrote a
1:31:28 whatever, all of that shit's in there.
1:31:30 So, its ability to
1:31:33 understand what to do with this thing is
1:31:35 it it just it's pretty good.
1:31:38 Your inner cut your inner child just
1:31:40 called. It's bored. Finally, a creative
1:31:44 outlet that doesn't demand actual
1:31:46 creativity. Just pick a color, stay in
1:31:48 the lines. Let me try a different I'm
1:31:50 going to put in um my uh my company
1:31:53 StoryVine. I'm going to put in
1:31:54 storyvine.com.
1:32:00 Um and we'll do anti-corporate grunge
1:32:03 because it's a it's a B2B thing. And
1:32:06 then we'll do a web banner. Whatever.
1:32:08 Make the ad. Please enter a URL. Oh,
1:32:11 http
1:32:14 colon slash
1:32:17 uh https slash. There we go. Uh so we're
1:32:21 going to that we're going to fix. We're
1:32:22 going to say they should you should be
1:32:24 able to take in URLs that don't have
1:32:27 http at the beginning of them because
1:32:29 you can you can [ __ ] add that you
1:32:32 stupid app.
1:32:34 Um
1:32:37 all right. What do we got here?
1:32:40 Finally, a reason to use that camcorder.
1:32:44 Look, we know we know your home movies
1:32:46 were grainy. America's home videos.
1:32:48 Click here or don't. Whatever. That's
1:32:51 not quite right. Um, okay. So, let's
1:32:53 see. So, we're going to do a couple of
1:32:55 things. Um,
1:32:58 make the following
1:33:00 um changes.
1:33:06 bullet. Um,
1:33:08 user should be able to put in a URL
1:33:15 without
1:33:18 https
1:33:19 colon slash,
1:33:22 etc.
1:33:24 And then I'm going to put um
1:33:29 be sure to research
1:33:34 the
1:33:36 company or product
1:33:41 from the URL provided.
1:33:47 So you can write the best
1:33:51 possible ad
1:33:56 based on what you learn.
1:34:01 Let's just let's just try that and see
1:34:03 what that does. See see if we get
1:34:04 something better from story mine.
1:34:17 Yeah, I know. I know the internet's not
1:34:19 70 years old, Vicki, but but what I'm
1:34:22 what I'm talking about is anything
1:34:24 that's been on computer systems, right?
1:34:26 Things were put on university computer
1:34:29 systems 70 years ago and all of those
1:34:31 got connected to the internet. So, it's
1:34:34 it's anything that's been put on
1:34:35 computers that then got internet enabled
1:34:38 in the past 40 years that's that's
1:34:41 there. So, I appreciate your commitment
1:34:44 to accuracy,
1:34:46 but I I'm not a total idiot, but I was
1:34:49 not clear. So, thank you for that. All
1:34:51 right. Um, storyvine.com. Let's see if
1:34:54 it does that good. We'll do '90s
1:34:56 nostalgia
1:34:58 social media post whatever. Make the ad.
1:35:01 All right. So, it let me put in
1:35:02 storyvine.com without the HTTP, which is
1:35:05 nice.
1:35:07 So, hopefully it's off actually reading
1:35:09 the website now rather than just
1:35:16 uh booting up the 486 story vine. Your
1:35:20 authentic story starts here. Don't think
1:35:22 too hard. So, your marketing team
1:35:24 finally realized authenticity is in.
1:35:26 Great. Now you just need to collect a
1:35:28 million videos without a film crew.
1:35:30 Story Mind makes it easy. This is good.
1:35:32 This is good. This worked. All right,
1:35:35 let's put in um we'll put in uh let's go
1:35:38 back to Amazon for a second. Amazon.
1:35:42 Let's go to Amazon Hall.
1:35:47 Um what is that? The real man lift
1:35:52 pouch.
1:35:53 Good lord, people.
1:36:01 What? What is happening? Um, oh, wait.
1:36:04 Let's do let's let's let's do a a n
1:36:07 let's let's let's do a sort of gen x
1:36:12 gifts.
1:36:16 [Music]
1:36:18 Let's see. Gen,
1:36:20 we'll do Gen X is one word.
1:36:24 Gen X gifts. No Gen X gifts. Okay. Uh,
1:36:28 fine. Uh, Gen X. We'll just put in Gen
1:36:30 X.
1:36:33 No.
1:36:36 Uh, someone give me a
1:36:41 something to look for here. Um,
1:36:45 calculator.
1:36:47 Calvin Klein women. Calligraphy.
1:36:50 All right, fine. We'll do a calligraphy
1:36:51 kit. You've You've convinced me.
1:36:54 Calligraphy it is.
1:36:56 Um,
1:37:00 all right. Pen markers.
1:37:09 Okay, back to our little app.
1:37:14 Paste in that thing. 90s nostalgia.
1:37:16 We're going to go anti-corporate grunge.
1:37:19 We're going to go social media post.
1:37:22 Whatever. Make an ad. All right. like on
1:37:24 Teemo AI experience. Do those work as
1:37:26 well? Yeah, probably. Oh, AliExpress. I
1:37:30 I would think if I've done this right,
1:37:32 every URL should work because it's I
1:37:34 told it to go look at the URLs. Pet
1:37:36 Rock, that's not a bad idea.
1:37:39 All right.
1:37:41 Remember when you actually made stuff?
1:37:43 Yeah, us too. Tired of num of thumb
1:37:46 scrolling into oblivion? Grab these pens
1:37:48 and make something real. Or don't. Just
1:37:50 don't ask us to explain NFTTS again.
1:37:52 These are good.
1:37:54 It's really good. Um
1:37:58 I I think I think rather than
1:38:03 that I think we should do Okay, we're
1:38:06 going to change this slightly. Okay,
1:38:08 this is great.
1:38:12 Um but I want all of the mockups
1:38:16 to be social ads.
1:38:22 um Facebook, um LinkedIn
1:38:27 X, etc.
1:38:31 And I want the
1:38:37 ad to be
1:38:40 optimized
1:38:42 for the
1:38:44 audiences on those platforms.
1:38:53 in addition to
1:38:57 the vibe.
1:39:03 All right,
1:39:04 let's have it redo that. Champy's
1:39:06 whining again. Tick tock. Hang on. I'll
1:39:08 be right back.
1:39:15 >> Come on, you whiner.
1:39:20 I know you're bored, but that's tough.
1:39:22 That's It's life in a big city. And I
1:39:25 know we're not in a big city. That's not
1:39:26 the point.
1:39:28 Okay. Tik Tok pin. Is there a Tik Tok
1:39:32 pin? There's not a pin right now.
1:39:38 Are you all vibe coding along? Everyone
1:39:40 having fun fun?
1:39:44 Tik Tok pin. Cassette tapes. Oh,
1:39:46 cassette tapes would be good.
1:39:49 Yeah. All right. So, let's go find a
1:39:52 cassette tape thing. Let's see. Platform
1:39:55 specific mockups. The old generic ad
1:39:57 formats are gone. You can now choose
1:39:59 between Facebook, Instagram, X, formerly
1:40:02 Twitter, and LinkedIn. Each mockup is
1:40:03 designed to look and feel how a real ad
1:40:06 on that social platform. The most
1:40:08 important change, the AI creative
1:40:10 director now explicitly tailor its copy,
1:40:13 the headline, the body, the call to
1:40:14 action for the typical audience of that
1:40:17 platform while still keeping the core
1:40:19 Gen X vibe. The LinkedIn ad will have a
1:40:22 different flavor than the ad for X, even
1:40:24 with the same product and style.
1:40:26 Streamlined UI. The interface has been
1:40:28 updated to reflect the new choices,
1:40:30 making the process just as simple as
1:40:32 before. This is good.
1:40:35 This is very good.
1:40:37 I just vibe coded an app I've been
1:40:39 trying to build for months. Woohoo. This
1:40:42 is why we play. Cassette tape player
1:40:45 with a mic and a pleather case. This is
1:40:47 good. Okay. Um cassette
1:40:51 player. Cassette tape player. Here we
1:40:53 go.
1:40:55 There we go. Now we're talking, people.
1:40:59 There is Are there really only two of
1:41:02 them?
1:41:04 Jesus. Wait. Amazon
1:41:08 cassette
1:41:10 player.
1:41:12 Here we go. Okay.
1:41:23 Retro80s cassette player
1:41:27 with Bluetooth. That's good. Okay, we'll
1:41:30 do this one.
1:41:32 Oh, wait. And Vicki, you uh AliExpress.
1:41:35 Yeah, let's do an AliExpress one. Um Ali
1:41:38 Express. Let's see if it'll do
1:41:43 do that.
1:41:46 Here we go. We'll go cassette
1:41:50 player.
1:41:54 Oh, these are great. Okay, here's one
1:41:57 the size of a cassette tape itself. Do
1:41:59 you remember that in the '9s when they
1:42:01 when Sony came out with the Walkman's
1:42:03 that were like just barely larger than
1:42:05 the cassette itself? Those were
1:42:07 remarkable. All right, here's a here's a
1:42:09 yellow one. This is good. AliExpress
1:42:13 $16.
1:42:15 66% off. It's It's What a bargain. What
1:42:19 a bargain.
1:42:21 All right. Fantastic. Bob, tell him what
1:42:23 he's won.
1:42:25 All right. Facebook, Instagram, Twitter,
1:42:27 LinkedIn. Beautiful.
1:42:29 There's our thing. We're going to do
1:42:31 this. We'll do this as a Facebook ad to
1:42:33 start. 90s nostalgia. That's perfect.
1:42:36 Whatever. Here we go.
1:42:40 Can you use a URL in the Wayback
1:42:42 Machine? I still have my Walkman. That's
1:42:44 awesome. You might be able to
1:42:58 Tap here. Keep wondering.
1:43:01 Your phone is already a leash. Wait.
1:43:03 Crucial. Luxury women's smart watch. No,
1:43:07 that's not right.
1:43:18 Oh, I know what we got to do. Okay.
1:43:20 Wait. We got to do something here. Um,
1:43:25 it just made
1:43:30 it just made an ad
1:43:34 where it completely
1:43:38 ignored
1:43:40 the
1:43:43 um web
1:43:46 page from the URL.
1:43:50 fix
1:43:52 that.
1:43:53 And also,
1:43:56 um, be sure to grab a product shot from
1:44:03 the website
1:44:08 to
1:44:10 feature in the ad. All right.
1:44:17 So, I don't know. Something something
1:44:18 clearly went off the rails there when it
1:44:20 when it when it changed the uh the
1:44:23 platform.
1:44:40 Oh lordy lordy lordy. Is everybody vibe
1:44:44 coding? Is everybody doing something or
1:44:46 y'all just sitting back smoking weed and
1:44:48 eating nachos? You stoners. You big
1:44:52 stoners.
1:44:53 What are you doing you stoners? All
1:44:55 right. Focus forced URL research product
1:44:58 f focused imagery.
1:45:02 Crucially, the instruction for the ads
1:45:04 image has been completely changed.
1:45:07 Instead of creating a stylized vibe
1:45:09 photo, it's now required to generate a
1:45:12 prompt for a clean, professional product
1:45:14 shot. Well, no, I don't necessarily want
1:45:16 that.
1:45:19 We'll see if it we'll see if it even
1:45:20 figures out the uh thing.
1:45:34 Finding our flannel. Booting up the 486.
1:45:50 No, it it it couldn't find it. So,
1:45:53 AliExpress didn't go well. Let's go back
1:45:55 to We'll go back to Amazon.
1:46:02 Do cassette player.
1:46:10 We'll do this guy right here.
1:46:14 Oh, you know what? I bet is happening.
1:46:19 I bet it's grabbing
1:46:24 I bet it's grabbing it's making the ad
1:46:27 based on
1:46:29 the URL. Amazon URLs have a description
1:46:33 of the product in the URL itself. I bet
1:46:37 that's what's going on. We'll see if
1:46:38 we'll see if this thing actually
1:46:39 features a product shot.
1:46:54 The other thing that you can do in these
1:46:56 vibe coding apps is I can actually
1:46:58 scroll backwards and revert to an
1:47:00 earlier version of this.
1:47:04 Yeah, it just it just made up this
1:47:05 image. It didn't get the image itself.
1:47:08 Okay. So, that's bad. So, let's do this.
1:47:12 Let's do
1:47:16 um
1:47:21 rather than
1:47:25 ah this is going to be too complicated.
1:47:27 Now we're getting into Amazon
1:47:29 APIs.
1:47:35 I'm going to I'm going to revert I'm
1:47:36 going to revert back to a version.
1:47:41 put the URL without the the APS
1:47:44 checkpoint.
1:47:46 This is okay, but I want all the mockups
1:47:48 to be socialized. Okay, restore this
1:47:51 checkpoint.
1:47:53 Let me pop this in.
1:47:57 I'm going to see if this older version
1:48:01 was just making [ __ ] up. I bet it was.
1:48:07 [Music]
1:48:13 Yeah, that's not the same thing.
1:48:16 Walkman's still in your attic. We found
1:48:18 your old mixtapes. Good. Okay, so here's
1:48:21 what I want to try. Okay, so let's see.
1:48:27 Okay, I've reverted to an earlier
1:48:33 version. I want you to
1:48:36 look at images from the product
1:48:42 URL
1:48:45 and
1:48:46 use them as inputs
1:48:50 to the prompt
1:48:53 to make
1:48:55 the ad.
1:48:57 So, we still have the personality
1:49:04 But the product
1:49:08 itself
1:49:10 is accurately
1:49:13 represented in the ad.
1:49:21 Use your multi-modal
1:49:24 capabilities
1:49:27 to
1:49:29 look at the product,
1:49:35 then
1:49:38 smartly incorporate it
1:49:42 into an ad image
1:49:46 with appropriate yet. Copy.
1:49:50 All right, let's see.
1:49:52 Let's see if we can get it to do
1:49:54 something decently sophisticated.
1:49:56 My guess is it's not going to be able to
1:49:58 do this because
1:50:02 right now I can give it a URL and what
1:50:04 it's likely doing is it's likely going
1:50:06 to that URL and looking at the text of
1:50:11 the site. if that it might just be going
1:50:15 and like researching whatever's in the
1:50:18 URL. I don't know.
1:50:20 It's probably it probably it's not an
1:50:23 agent, right? So, it's not out visually
1:50:26 looking at websites like we could we
1:50:28 could do this if we were building an
1:50:30 agent and we had to go looking at a
1:50:32 website using one of those kind of
1:50:34 tools. I think you would actually have
1:50:36 to have an API from Amazon to be able to
1:50:39 pull the images, right? So, let's see.
1:50:43 You got it.
1:50:45 The previous approach was clever, but
1:50:47 you're right. An ad needs to feature the
1:50:49 actual pro product, not just an
1:50:52 interpretation of it. Relying on a text
1:50:54 prompt alone was a weak link. I've
1:50:56 rearchitected the application to use a
1:50:59 powerful multimodal approach. It takes a
1:51:01 user provided product image. Ah, okay.
1:51:05 And uses AI to reshoot it. That's
1:51:07 actually not a bad idea. Image upload
1:51:09 multimodal AI and stylized ads. Okay,
1:51:14 this is not bad. But let's see if it's
1:51:16 just product URL product image.
1:51:20 Choose file. Okay, so so let's say um
1:51:24 this is great. Um
1:51:29 uh great. Let's see. Great. But I want
1:51:33 to be able to paste
1:51:37 the image
1:51:39 into the field
1:51:42 from the clipboard.
1:51:48 Uh I want to be able to press in the
1:51:52 clipboard. Um, or
1:51:55 drag and drop
1:52:00 or upload.
1:52:04 Right. I'm lazy. Feed my laziness.
1:52:22 This hunting motel doesn't have a
1:52:23 microwave to heat up the nachos. And no
1:52:26 smoking allowed. It's a hunting lodge.
1:52:28 What do you mean no smoking's allowed?
1:52:30 Doesn't make any sense at all. I'm vi
1:52:32 vibe coding an app for pinning things to
1:52:34 Pinterest. Perfect.
1:52:37 Perfect. Analyzing six errors. Fixing
1:52:40 this form submission.
1:52:45 Tired of cringe ad. Yeah, us too.
1:53:02 [Music]
1:53:13 Okay. So, let's go grab this. We'll grab
1:53:16 this. Save image as
1:53:19 desktop. save. We're going to go back to
1:53:22 our little ad.
1:53:24 We're going to go paste.
1:53:27 We're going to go choose a file.
1:53:31 Oh, it should display it. But let's
1:53:34 generate the ad. Let's see. Let's see if
1:53:35 it works.
1:53:54 Oh, it did it. Look at that.
1:53:57 That worked.
1:54:00 Yep. We put Bluetooth in a cassette
1:54:02 player. Don't act surprised. That was
1:54:03 funny. I made fun of that, too. Your
1:54:06 carefully curated mixtapes now less
1:54:08 tangled wires. Relive your glory days.
1:54:11 We're just finally digitized that demo.
1:54:13 We won't judge much. That's a good ad.
1:54:16 Okay. So, we're going to do um make the
1:54:21 input
1:54:23 a drag and drop input
1:54:28 and
1:54:30 display the uploaded image
1:54:36 um in that space.
1:54:41 Make it more
1:54:44 impressive.
1:54:47 Don't be lazy.
1:54:50 Um,
1:54:52 telling your your AI robot that it's
1:54:55 been lazy. Good practice. Good practice.
1:54:59 All right. Pine cone vector database
1:55:01 went from $3 a month to a minimum of $50
1:55:05 a month.
1:55:07 Oh, that sucks.
1:55:10 That's you know what they're doing Joker
1:55:12 when when they're when they're going
1:55:13 from usage based stuff to a monthly
1:55:16 minimum. They're actually that's a
1:55:19 strategy to cut out the small developer
1:55:22 because individual developers are a pain
1:55:24 in the ass. They they want people with
1:55:25 money. That's stupid.
1:55:29 I don't like that at all.
1:55:37 I wonder if we could make this a two
1:55:38 panel a two column thing.
1:55:46 Okay, let's see. Okay.
1:55:51 Okay. the
1:55:59 um
1:56:02 click, drag and drop or paste.
1:56:06 How? You can't paste, can you? Wait.
1:56:10 Um,
1:56:19 can I drag and drop? Yeah. Oh, and it
1:56:21 did. Look, it displays it now. Cool.
1:56:25 Okay, but let's see.
1:56:31 Let me grab screenshot of this.
1:56:38 We're going to say we're going to paste
1:56:39 it in here. And we're gonna say the
1:56:47 image should
1:56:50 be uh let's see um should be
1:56:56 below
1:56:58 the URL
1:57:02 and then the
1:57:06 vibe and platform options.
1:57:11 should be
1:57:15 on their own.
1:57:23 Oops, I'm doing the wrong thing. Hit the
1:57:26 wrong button.
1:57:31 Oh, it's 10:00.
1:57:33 It's getting late, people.
1:57:36 Well, you get the basic idea here. You
1:57:38 can do some pretty cool [ __ ] if you're
1:57:40 willing to, you know, hold the vision.
1:57:44 Like again,
1:57:46 this is what the core of my book's going
1:57:48 to be is our job increasingly
1:57:52 is after you learn
1:57:55 there's all these different tools and
1:57:57 they all kind of do these different
1:57:59 things. Our job is going to be to kind
1:58:01 of hold the center of what's the vision
1:58:03 here? What's the idea?
1:58:06 and then navigate between these
1:58:08 different tools to be able to build
1:58:10 something like this. That's or to build
1:58:12 whatever to build a business plan, to
1:58:14 build a song, to build a movie.
1:58:20 You're the producer. It's your ideas.
1:58:24 Andrea, thanks, Kyle. You've inspired me
1:58:26 to make an app for my business. Good.
1:58:29 Perfect. Now that I got my app, can I
1:58:32 see it somewhere outside of Google?
1:58:34 Yeah, there's a publish. So you you can
1:58:36 save you save your app, you can share
1:58:38 your app, but there's a publish button.
1:58:40 What I'm not clear about source camp is
1:58:43 exactly how it lives in the world. Um
1:58:47 Brandon might know that, but if anyone
1:58:49 knows that, let me know.
1:58:52 Um
1:58:55 okay, this is good. Your product,
1:58:58 okay, this is great. Now,
1:59:04 I'd
1:59:08 like to make this a two column app
1:59:13 where
1:59:15 the inputs where the the the
1:59:19 title spans the page
1:59:24 and
1:59:27 and
1:59:29 Below it are two columns
1:59:35 with the inputs
1:59:40 on the left
1:59:42 and the output on the right.
1:59:48 Go.
1:59:56 All right. doing his thing.
2:00:07 If you're feeling overwhelmed with all
2:00:09 this AI stuff and how sophisticated
2:00:12 AI is now, just that's okay. Keep keep
2:00:17 feeling that uncomfortableness and just
2:00:19 play anyway. Just go for it.
2:00:25 Okay, that didn't work.
2:00:35 It's still just one column.
2:01:00 Here's
2:01:11 the corrected file. This will fix it.
2:01:14 That's what Gemini says to me. We'll
2:01:16 see. We'll see. Old Gemini. Nope. Didn't
2:01:21 fix it.
2:01:23 I'm exhausted. All right. Whatever. Um,
2:01:27 let's see if it still works.
2:01:31 So, we've got to go get Let's go do
2:01:33 another ad. We'll we'll uh we'll grab
2:01:38 um
2:01:41 let's go try an Alibaba. We'll grab this
2:01:43 thing. So, we'll grab this ad or this
2:01:46 image.
2:01:49 Oh, you can't grab the images. Nice.
2:01:52 Perfect.
2:02:06 Okay, I think I got a zoomed in shitty
2:02:10 screenshot. Perfect.
2:02:14 Now we'll get a good one.
2:02:19 Okay. Um,
2:02:22 going back to here. Let's see.
2:02:28 Screenshot.
2:02:34 There we go. All right, that works.
2:02:38 So, here we'll get our URL. Copy.
2:02:46 90s nostalgia. We'll do Slacker apathy
2:02:50 social media post. Whatever. Make the
2:02:52 ad.
2:02:57 doing the thing
2:03:04 adjusting the tracking. You can also see
2:03:07 if if you've been playing along, can you
2:03:09 see where like hours and hours and hours
2:03:12 could could blow by? Yeah, they're
2:03:15 small. So, this are give a damn.
2:03:17 Remember when everything was oversized?
2:03:19 We barely do. The sun was out of our
2:03:20 eyes. Yeah. vintage small frame
2:03:24 sunglasses. So, what
2:03:28 what I've now realized is that um
2:03:34 the AliExpress logos don't have any
2:03:37 product description in them and the
2:03:39 Amazon ones do.
2:03:42 So, this is a this is a bit of a fail
2:03:47 application
2:03:51 because it's not really looking at the
2:03:53 website.
2:03:56 So, I'm going to let it I'm going to
2:03:57 leave it there. This might be something
2:03:59 I go try to fix,
2:04:04 but like the only way you can discover
2:04:05 that is So, so like this is wild. So now
2:04:10 >> you're rolling over.
2:04:11 >> Um,
2:04:14 so now we're not only
2:04:17 we're not only learning how to vibe
2:04:19 code, but you you now have to be as
2:04:21 curious about the application that you
2:04:23 created as you used to have to be about
2:04:26 AI because you got to see if it actually
2:04:29 works, right? Because like most things
2:04:32 with AI, it's good enough to make you
2:04:35 think that it's working really well, but
2:04:38 then you discover things like if you
2:04:40 have a URL that doesn't have the product
2:04:42 description in the URL, it's not
2:04:44 actually looking at the product page.
2:04:47 So, that's a problem. All right. So,
2:04:49 just keep playing. Just keep playing.
2:04:52 Oh, getting a crazy lightning storm here
2:04:54 in Dallas. What is it with lightning all
2:04:57 of a sudden? Okay. Uh I'm going to get
2:05:00 out of here. Let me go get on out of
2:05:01 here. Um, homework for the weekend.
2:05:07 Well,
2:05:11 I would say go back. Go back here.
2:05:14 Here's a thing you can do. Go look at
2:05:16 your bank account and look at
2:05:19 subscriptions that you're still paying
2:05:21 for for services that you haven't looked
2:05:24 at in three months
2:05:27 and go back to that service. So, if you
2:05:30 haven't used Ideogram or if you haven't
2:05:32 used LTX Studio, if there's something
2:05:34 that you're still paying for that you
2:05:35 haven't looked at in a while, go back
2:05:37 and play with it because
2:05:40 these things are getting more and more
2:05:42 and more sophisticated.
2:05:44 The other thing I would say is if you
2:05:46 vibe coded something cool tonight, you
2:05:49 you've got, you know, all weekend to do
2:05:52 this um this vibe coding thing. like I
2:05:55 think I'm going to submit this Gen X ad
2:05:57 maker to the uh to the vibe coding
2:06:00 contest. Why not? It's kind of funny. Um
2:06:04 and uh and yeah, could could be could be
2:06:07 kind of fun. So So just explore this
2:06:10 weekend. Do the do the hackathon. Um
2:06:13 because who knows, you might you might
2:06:15 win some money or you might just get,
2:06:16 you know, someone pointing something out
2:06:19 to you. Um, share anything you're
2:06:22 working on on the AI salon. Share it in
2:06:25 the look what I made or in the
2:06:26 irregulars channel. Anywhere you want to
2:06:28 anywhere you want to share it. Just try
2:06:30 to do some engagement.
2:06:32 Um, and then we'll be back Monday. All
2:06:35 right. So, I've got some work to do.
2:06:38 I've got a uh a potential job building a
2:06:42 30- secondond ad, which is kind of fun
2:06:45 using uh
2:06:48 using all the new tools. I mean, what's
2:06:50 what's what's
2:06:53 kind of stunning about getting asked,
2:06:55 hey, can you make this ad? The answer
2:06:57 is, I think I can.
2:07:01 I don't know which tool to use anymore.
2:07:04 Um because they all keep changing. So,
2:07:06 my workflows are just keep getting
2:07:08 upended, but um it's just crazy times,
2:07:11 man. We live in crazy flipping times.
2:07:14 Um Nano Banana is officially on X.
2:07:17 Listen, last thing I'll say, kudos to
2:07:20 Google. Someone at Google is paying
2:07:23 attention. They haven't been paying
2:07:25 attention for [ __ ] decades.
2:07:28 Someone at Google is paying attention.
2:07:30 and they actually recognized that
2:07:31 everyone was calling it Nano Banana and
2:07:33 not Gemini 2.5 Flash and they're leaning
2:07:37 into it. Um, and they're doing the Nano
2:07:39 Banana hackathon this weekend like in
2:07:41 conjunction with Kaggle which I think
2:07:44 they own Tik Tok question.
2:07:47 Um, contest I just got here. What
2:07:50 contest? Good evening by the way. Good
2:07:51 evening. Um, if you go to kaggle.com
2:07:57 hang oncompetitions
2:08:00 k a gglecompetitions
2:08:05 slashbanana.
2:08:07 You've got the nano banana hackathon
2:08:10 which let me show you the page.
2:08:13 There's the page. The nanobanana
2:08:15 hackathon on Kaggle
2:08:18 or you can just look it up on X. It's
2:08:20 it's it's there on X. $400,000 in
2:08:24 prizes. It's just this weekend. They're
2:08:28 looking for apps that leverage Nano
2:08:30 Banana. All right. And Nano Banana is
2:08:33 just so fun to say. I agree. All right,
2:08:37 Groovy. So, go play this weekend. Go
2:08:39 make some [ __ ] It's time. It's he
2:08:47 I think the days
2:08:50 if if you're if you're an a regular if
2:08:52 you're someone who comes here a lot I
2:08:54 think the days of sort of discrete tool
2:08:57 playing are kind of coming to an end. I
2:09:01 think we're entering an era where our
2:09:04 real skill, like the the really skilled
2:09:06 people in AI are going to be able to
2:09:08 bounce between tools and bring things
2:09:11 together from multiple tools. Um,
2:09:16 that's
2:09:21 harder,
2:09:23 but I think it's important. Um,
2:09:27 so I would I would encourage you start
2:09:29 thinking about crossd discipline. Think
2:09:32 about AI as as 10 different team members
2:09:35 that that you sort of spin into
2:09:37 existence to help you execate execute
2:09:40 some larger vision. Um, that's where I
2:09:43 think we're headed. So, um,
2:09:47 yeah, I'm just getting my head around it
2:09:48 right now, but but I think that's the
2:09:50 new skill. All right. I think that's the
2:09:53 as as we enter the sort of, you know,
2:09:55 we're out of 101, we're out of 2011,
2:09:57 we're now getting into AI 301. And I
2:10:00 think AI301 is you start to do some more
2:10:02 sophisticated
2:10:04 um
2:10:06 crossf functional
2:10:08 AI AI workings.
2:10:11 Um, all right, GI. Um, hope you have a
2:10:14 fantastic weekend. Hope you had fun
2:10:16 tonight. Peace out. Have a good And I'll
2:10:18 see you Monday.