9/2/2025 - Nano Banana Image Generation and Clip Genius Explored

Description

Chapters

Transcript

0:02 or tranquil transformations. That is the
0:05 question.
0:09 [Music]
0:30 [Music]
0:35 Freedom came my way that night.
0:41 Just like a jet plane in
0:47 [Laughter]
0:50 I was hauling ass at a million miles an
0:53 hour
0:55 wondering how hard I'd hit
1:00 when they came to the station.
1:07 They said I was bad beyond repair,
1:12 but I got no qualms with my situation.
1:18 Say I am.
1:25 So say Sheree, Sheree, Sheree, won't you
1:29 dare do? Say Sheree, Sheree, Sheree.
1:34 Won't you dare to say Sheree Sheree
1:39 Sheree?
1:42 Uhhuh.
1:44 Yeah. Leave a message and your number,
1:47 please.
1:51 Take a time to want to satisfy me.
1:55 What are you doing to my feet, you weird
1:57 dog?
1:59 Take all these old fantasies, send them
2:02 care of me.
2:03 [Music]
2:07 Um,
2:10 [Music]
2:16 how is everybody? What is going down?
2:18 What's shaking? What's happening? What
2:21 is happening? What's happening with the
2:24 good people?
2:26 What is happening with the good people?
2:28 Huh?
2:30 [Music]
2:48 Hey,
2:50 [Music]
3:09 hey,
3:11 [Music]
3:37 Um, what are we going to talk about
3:38 tonight? Um, I'm back after a long
3:41 weekend. I had a good vibe coding
3:44 session on Saturday, which I'll show you
3:46 the results of.
3:48 Um,
3:51 Nano Banana continues to
3:56 impress and flumx me. Chat GPT5
3:59 continues to flumx me.
4:07 I
4:09 I've come to the realization that
4:14 when you have a model
4:16 that has the intelligence in in the
4:19 onetenth of 1 percentile top
4:24 that means that a lot of what it does
4:26 really well is out of the range of
4:29 comprehension of most people.
4:32 And so to us normies who are not in
4:37 Mensah or in Mensa
4:40 velvet rope edition,
4:45 these models don't look dramatically
4:48 different from what we had before
4:50 because
4:52 GPT40 had saturated most of people.
4:58 Now, are there still improvements to be
5:00 made like hallucinations and things like
5:02 that? Yeah. Uh
5:11 like I'm I'm feeling like I've got to
5:13 get better and and I'm going to play
5:15 with a tool tonight. Brandon found this
5:17 tool which Open AAI is kind of buried.
5:19 It's it's it's not This feels to me like
5:22 something they could build into chat GPT
5:24 pretty easily and probably should.
5:27 and it's a prompt analyzer and it
5:29 basically says based on the model that
5:31 you're using, you put in your shitty
5:33 prompt, I'll rewrite it into a good
5:35 prompt and I'll tell you why I said what
5:37 I said, which I think is pretty cool.
5:44 But basically, it comes down to we're
5:46 not asking sophisticated enough
5:48 questions for the model to really even
5:50 give a [ __ ]
5:53 And so we're getting back answers that
5:55 are kind of like, eh, if you're going to
5:57 be lazy with your question.
6:01 I'll I'll be lazy with my answer.
6:06 Cams. Switch the cams. Let's switch the
6:08 cams, people. Good people. Good people.
6:12 Good people. I'm a little moist. It's a
6:14 little warm in here. Um
6:18 [Music]
6:31 Um, the other thing that was kind of
6:34 cool that came out today is Gen Spark
6:38 keeps dropping new things. Let's see.
6:42 Archetypal. I haven't had any
6:44 significant problem getting what I want
6:46 from GPT5.
6:48 I don't get me wrong, I'm using GPT5.
6:51 It's fine.
6:53 It doesn't Well,
6:55 it seems different than GPT4 in that
6:58 it's not as good a writer. So, I find
7:01 myself going back to GPT4 for writing.
7:05 I'm not saying I can't get it to do what
7:07 I want. I'm saying
7:12 I am unfulfilled. Let's say as a prompt
7:16 partner of GPT5,
7:20 it's got more to offer than what I know
7:22 to ask for is maybe a better way to put
7:24 it.
7:27 But maybe I'm not asking anything
7:28 difficult. Yeah, that's the my problem
7:30 is I'm not asking anything interesting.
7:32 I wouldn't even say difficult.
7:36 I don't have interesting questions like
7:39 interesting at a large language model
7:40 level.
7:42 So maybe we can play with that tonight.
7:44 Maybe maybe there's just something about
7:46 maybe in this prompt analyzer we can
7:48 we'll put in some shitty prompts and
7:49 we'll get back something and we'll go
7:51 look at it. Like the thought of
7:56 having to do fully prompted or I mean
8:00 fully structured prompts makes me
8:02 cranky.
8:06 I want to be able to talk to this thing
8:07 like a normal human being and have it
8:10 react remarkably. And it's not. That's
8:12 what I'm That's what's bugging me.
8:16 Joe Dissult, same.
8:19 I I assume you're That's in reference to
8:21 what I'm talking about here with GPT5.
8:25 [Music]
8:29 Okay, here's a good example.
8:32 how good I play guitar and how many
8:35 songs I know and how much music theory I
8:38 know is kind of like how I'm prompting
8:41 chat GPT5 and chat GPT5 is like Mozart
8:46 or Jacob Kier.
8:49 So like I'm going hey Jacob Kier could
8:52 you play me like a a 145
8:56 song in the open position? Oh, sure.
9:02 [Music]
9:09 And that's the kind of answer I'm
9:10 getting back from chat GVD5.
9:13 [Laughter]
9:16 I'm not getting out the crazy harmonies,
9:19 right? Because I'm asking a shitty
9:21 question.
9:25 [Music]
9:39 [Music]
10:01 I'm 95% Chat Gvt and 5% Google. I use
10:05 Google to search and view images and for
10:08 maps. Yeah, I I'm about that. I'm about
10:11 that way, Ton Todd. I'm probably, you
10:14 know, 90% chat GPT. For any interaction
10:17 where I want something, it's chat GBT. I
10:20 use Google for I need to find a website,
10:25 Google Maps.
10:31 Oh, and if I quickly want like what's
10:34 the definition of some stupid word?
10:38 Those are the three things I use Google
10:40 for anymore. Maps. Maps is good. Maps is
10:44 good.
10:46 But if I just want a quick answer, I use
10:48 chat GPT or I use Google because chat
10:50 GPT is too slow.
10:52 Um, David Scott Patterson, indications
10:56 that AI at its ultimate limit
11:00 will be a manifestation of God.
11:04 Oh boy, us humans, we are a mess.
11:30 [Music]
11:42 Oh, I was I was saying before I started
11:44 I was I was started uh
11:48 talking about GenSpark. GenSpark came
11:50 out with a new thing called Clip Genie,
11:53 which is just like Opus Clip, but it's
11:56 it's an agent where you can just point
11:58 it at your YouTube channel and tell it
12:00 to go download go download the latest
12:04 video from your channel and make clips
12:05 out of it. Um,
12:09 and the the first thing I did with it
12:11 was pretty good. It was good enough that
12:12 I published it.
12:19 I have to tell perplexity that chat GPT
12:21 gave me certain document
12:24 and ask for it.
12:26 What is it called? Wait, what?
12:30 And then ask perplexity what it thinks
12:32 of the job chat GBT did. Yeah, it's fun
12:35 asking Heaven Chat heaven.
12:38 Oh, what's it called? Oh, the uh the
12:40 Here, let me go show you. How How you
12:43 doing, Mel? I haven't seen you in a
12:44 while. I mean, I see you on TikTok all
12:46 the time, but I haven't seen you in here
12:47 in a while. Welcome back. Happy Tuesday
12:50 night. I'll tell you what, ask me a
12:54 question and I may go off tonight.
12:59 I may just do it. Okay. So, where you're
13:01 going to go is you're going to go to
13:03 GenSpark,
13:06 [Music]
13:10 genspark.ai. AI.
13:14 Apparently, I've got to share my screen
13:16 if any of you are going to see any of
13:18 this.
13:21 We'll play the video here
13:26 because it's pretty cool. Oh, and and
13:28 then just to be clear, when I play this
13:30 video,
13:32 what he does,
13:37 um,
13:39 they do the the marketing edit thing
13:42 where he's like, I just uploaded a video
13:44 and then it made this clip and it's like
13:46 bang, done. That's not how fast it is.
13:50 But but we'll watch this Tik Tok pin
13:54 used
13:57 So lately, in fact, I have over the
14:00 weekend. Oh, that's something I'll I'll
14:02 talk about tonight.
14:04 I'll play for you. Oh [ __ ] Where's
14:10 I don't think I have it here. Oh, no, I
14:12 do. I do. It's in Sunno. So So I had
14:16 Yeah, we'll we'll play with. I had an
14:19 old MP3 of a fourtrack recording I
14:23 recorded in 1987 with with my roommates
14:26 in New York City when I first moved to
14:28 New York. We wrote a song and recorded
14:31 it and I always loved the song and then
14:34 the woman sent it to me like last year
14:39 and and it just hit me, oh, I should put
14:41 it in Suno and and redo it. And so I did
14:44 that. So, I did that and then I also
14:45 vibe coded this weekend.
14:48 And then I found this thing today.
14:53 Oh, that's how you found
14:55 that's how you found found this place is
14:57 through that song. That's pretty [ __ ]
15:00 cool.
15:05 Oh, that's awesome. All right, cool.
15:07 Yeah. So, we'll play with that. Um,
15:12 and it's it's really fat. Actually, you
15:14 know what'll be fun tonight
15:16 in Sunno. So, what I was trying to do is
15:18 I was trying to get Sunno to give me
15:20 something that was in the neighborhood
15:22 of what the original recording was. It
15:24 got it got close. It it like I like I
15:27 still like the the cadence of the
15:29 original better. Um, but it got close
15:32 enough that it's like it's kind of a
15:33 cool song, but what might be fun tonight
15:35 is like is like really change it up.
15:39 Um, so that that could be cool. Um,
15:43 all right. I want to know how you got
15:45 your Champy selfie app to actually work.
15:50 Listen, when it comes to vibe coding, if
15:52 you get something to work, you just
15:54 actually, I'll tell you I'll tell you
15:55 one of the secrets of how I got my
15:58 Champy Selfie app to work.
16:01 When you're vibe coding and the thing
16:04 shits the bed, you know when it goes off
16:06 the rails and you can't get it back,
16:10 start over. So, what I did with my
16:13 champion with the the selfie thing is I
16:16 started working on it and then it
16:18 started getting a little wonky and then
16:19 it went off the rails and then I tried
16:21 to go back a version and and it it just
16:24 like there's something about it when you
16:26 have a session in vibe coding. It's like
16:29 when large language models get something
16:31 in their mind, they they they can't get
16:34 it out of their mind. And so what I did
16:37 was I
16:42 yeah,
16:43 I just thought about all the things that
16:46 I wanted it to do. I started a
16:49 completely new vibe coding session and
16:51 then I said, "Here's the things I want.
16:53 I want it in two columns. I want it to
16:54 behave this way. I want it to do this.
16:56 Here's what I don't want it to do." Like
16:58 I just gave it a whole long list and
16:59 then when it start when it restarted I
17:01 was like mostly finished with it on the
17:04 first prompt cuz I knew what I had
17:05 [ __ ] up on the other one. All right.
17:08 All is well. Great to be back. I had an
17:11 intervention from the family and they
17:13 took me off AI learning.
17:22 Mom, we're so tired of hearing about AI.
17:24 Could you just stop with the AI?
17:26 Couldn't you go back to your compounding
17:28 pharmacy talk? Please talk about
17:32 minoxidil again for something. Jeez,
17:36 don't you have some sort of almond balm
17:38 you're supposed to put under your eyes
17:39 right now? God,
17:42 that's awesome. Good for you. All right.
17:44 Well, welcome back. Okay. So, for you
17:48 the the clip thing, which I think you
17:51 could I think we could all benefit from
17:53 because one of the things about Opus
17:55 clip and the things like that is they're
17:57 really expensive. Now, I don't know if
17:58 Jen Spark's any cheaper. It might not
18:00 be. So, but let's let's look at what uh
18:05 Bad Haircut CEO says about Clip Genius.
18:10 >> Truth or dare?
18:11 >> True. Truth. All right. Dare.
18:13 >> How many people have you slept with?
18:17 Okay. So, see see the the grid here of
18:21 like nine different Jimmy Fallon and it
18:23 had that little clip in there. That's
18:24 what this thing makes.
18:26 >> Truth or dare or AI clips all highlights
18:28 with one prompt. Hey, I'm Eric from
18:31 Jensen Spark. Editing videos can be
18:34 super timeconuming and difficult for a
18:36 lot of us. That's why we're excited to
18:40 introduce Jensen Spark Clip P. Okay,
18:42 here's another thing.
18:45 There are some human beings that have
18:47 such weird personalities
18:50 that you can't tell if they're AI or
18:52 not.
18:56 I honest to God, every time I watch one
18:58 of his videos, I'm like, "This is an AI
19:00 avatar." And then I'm like, "No, this is
19:04 He just talks like that. No, this is an
19:07 AI avatar." So, I don't know.
19:10 Yes,
19:11 >> it can edit any video for you with just
19:13 one prompt.
19:15 >> Let's try it. Choose single clip,
19:18 cut the funniest moments from this Jimmy
19:20 Fallon show. Jensen Spark understands
19:23 the whole video, builds the story, then
19:27 finds the clips and weaves them into a
19:29 seamless new video. dare you to lay flat
19:35 on the desk and then interview me the
19:38 rest of the time like that.
19:40 >> All right, here we go.
19:42 >> Now, let's try grid layout. Get
19:44 highlights from All-In Podcast's AI
19:46 bubble episode. Clip Genius understands
19:49 this hour plus podcast and pulls nine
19:52 clips, creating a perfect grid. Key
19:54 insights in minutes
19:56 >> on Monday.
19:57 >> Okay, do you see what they did there
19:58 with the marketing thing? Look at the
20:00 the graphic that they made
20:04 >> and pulls nine clips pulls all the clips
20:07 out,
20:08 >> dips them up there. That's not how it
20:09 works.
20:11 >> To be clear,
20:12 >> in minutes,
20:14 >> on Monday, Fortune dug up a generative
20:16 AI study.
20:18 >> Does this align with what you're seeing
20:19 on the field?
20:21 >> Last week or so, there was a correction
20:24 in sentiment.
20:25 >> Up next, cut highlights showcasing only
20:27 my favorite Dodgers team. Jen Spark
20:30 understands the video and perfectly
20:32 captures the Dodgers best plays.
20:36 >> Oscar Hernandez
20:39 in the air to left center field. Judge
20:40 on the run. Dead sprint now to the track
20:43 and it's over his head.
20:45 >> Cool. Let's try a different angle with
20:46 the same video. Extract only Yankees
20:49 highlights.
20:51 [Applause]
20:54 >> Fast driven the other way. Sleeping
20:58 giant.
20:59 Finally, let's mash up multiple videos.
21:03 Jen Spark understands these three LOL
21:04 matches and pulls 10 kill highlights.
21:08 Perfectly done,
21:10 >> but Sh is looking for more. He flashes
21:12 forward. Knight on the flank as well.
21:13 Owner is going to flash away. Counter
21:14 strike out from Zas not quite able to
21:15 connect with.
21:17 >> You can also use the editor to fine-tune
21:19 your clips or visit AI drive to download
21:22 all clips in the final cut with just one
21:25 prompt. Let Jen Spark Clip Genius edit
21:28 any video.
21:29 >> All right. So, that's the video. That's
21:31 the marketing video. Um,
21:35 All right. So, let's head over to
21:39 YouTube
21:42 and then I'm going to go to What is
21:45 going on with my mouse?
21:48 It's not working right, man. Don't get
21:51 heavy on my scene, man.
21:54 Um,
22:00 AI learning lab.
22:03 Let's go to our channel.
22:09 Oh, this is driving me crazy.
22:12 Hang on. I'm going to turn it off. Turn
22:14 it back on. Just calm calm everybody.
22:16 Calm down. I don't know why you're so
22:18 ramy tonight.
22:22 This is crazy. Gen AI is moving faster
22:24 than I could have ever imagined. That's
22:26 incredible. Again, don't be seduced by
22:30 the slick editing. It's not quite that
22:32 fast, but it's still pretty [ __ ]
22:34 impressive. What's What's most
22:35 impressive about it is this.
22:38 Um,
22:40 if you use something like Opus Clip,
22:42 right,
22:43 you go in and you select how long you
22:46 want your clips to be, 60 seconds, 3
22:48 minutes, whatever. And then you say,
22:51 "Okay, Opus Clip, go find all the
22:53 stuff." And then Opus Clip takes
22:55 whatever its algorithm is and it goes
22:57 off and it finds your things. Now, you
22:59 can go in and manually edit and do that
23:00 sort of stuff. But what happens with
23:02 GenSpark is you can just tell it what to
23:04 do. Like you can make up your own
23:06 algorithms and then I assume you could
23:09 probably um automate those and replicate
23:13 them. Mouse off, mouse on. Come on. Come
23:17 on, mousy. Are we working now? Are we
23:19 better? Are we more happier?
23:22 No. Hang on. I got to change my mouse
23:25 pad. Oh, I see what's going on. There
23:28 was There was hand oil on my mouse pad
23:31 paper.
23:45 And I bet I bet the hand oil was
23:48 reflecting
23:50 the
23:52 laser. Weird. Yep. That's Well, may
23:55 maybe I don't I don't know.
23:58 Just just you
24:01 everybody stop picking on me.
24:06 Okay.
24:08 Um, let's go to live.
24:12 Okay. So, what we're going to do, so the
24:13 the one that I did earlier. Oh, let me
24:15 I'll play I'll play you the one we did
24:17 earlier.
24:19 Uh, if I go videos.
24:23 So, here's the one I did earlier. I
24:25 won't play you the whole thing, but am I
24:27 sharing this?
24:32 Yes. Right. Yes. absolutely reminds me
24:36 of when I was following NFT projects.
24:39 It's like it's like, oh my god. And and
24:43 they're just relentless. And I and I
24:45 feel like it's the same way with these
24:47 things. Like they just keep, you know,
24:49 releasing new things. So
24:52 he's only following 493 people. So you
24:55 can click on that and you can now go see
24:58 the people that he follows.
25:01 And
25:02 right so it automatically pulled these
25:04 clips and then you know it goes back to
25:06 the grid. Let's see where's the grid.
25:14 Oh, it's really quick.
25:21 But anyway, you could you could edit
25:23 that. You could chop that into nine
25:24 pieces, but you don't have to. Okay, so
25:26 that's that. So, let's go back to the
25:29 channel.
25:32 Oh, this bounce is driving me crazy.
25:36 It's driving me crazy.
25:39 Okay, so we'll go to eight.
25:41 Oh, wait. Let's see.
25:44 Exploring Gemini 2.5.
25:51 All right. Okay. So, we're going to go
25:52 to the 82625
25:55 episode.
25:57 This is what we're going to do. So, the
25:59 way you do this is you go new in the
26:01 upper leftand corner. Oh, wait. Share
26:03 this tab. Calm down, people. Calm down,
26:07 producer Brandon. Okay, so you go to the
26:11 plus thing and it and it pops open this
26:14 this uh menu. And right here is Clip
26:17 Genius. But like look at all the look at
26:19 all of the products that GenSpark has
26:22 added. There's the super agent, which is
26:24 basically just tell it anything and
26:26 it'll do it. There's AI slides, AI
26:28 sheets, AI docs, AI developer, which I
26:31 assume is a vibe coder, AI designer,
26:34 clip genius, AI pods for making
26:38 podcasts, I guess. AI chat, AI image, AI
26:41 video. You I think they can lose the AI.
26:44 Deep research, fact check, call for me
26:49 and download. call for me.
26:53 Oh, we should we should do one of those
26:56 we should do one of those prank calls.
27:00 Except you you can't do them anymore. Is
27:02 your if is your refrigerator running?
27:04 You better go catch it. Uh
27:07 uh. Do you have Prince Alfred in a can?
27:10 Better let him out. Those are all old. I
27:13 What are the modern versions of those?
27:15 Anyway, Clip Genius. All right, we're
27:17 going to go to Clip Genius. So, I'm
27:18 going to go
27:20 I'm gonna go um
27:24 go to the live
27:27 video section
27:31 of
27:35 YouTube channel
27:38 at learning lab oops-
27:44 AI.
27:49 and
27:52 find the
27:54 video from
27:57 August 26th. It was the 26th, right?
28:00 Hang on, let me confirm that. Yes.
28:07 Oh,
28:09 I'm not sharing the right tab anymore.
28:12 Producer Brandon missed it.
28:14 Okay, now I'm sharing the right tab. All
28:16 right, find the video. He's get he's
28:19 getting defensive. I can see Labor Labor
28:22 Day hangover.
28:26 Okay, Brandon's not on his best game. He
28:29 had he had too much uh too much happy
28:31 punch at the the uh the Labor Day
28:35 picnic. Okay. Um the video from August
28:38 26th this year, so we'll put in
28:40 something that might throw it off.
28:43 Um, and pull
28:48 the five most interesting
28:54 30 second clips about nano banana
29:01 and edit them into a single clip. Okay.
29:07 So, we're going to go single clip
29:09 layout.
29:16 And here we go.
29:18 Okay. So, it's thinking. So, so what
29:20 it's doing over here is it lo it's
29:21 loading an editor. So, what they did was
29:23 they built like a like an FFmpeg
29:28 video editor and now this thing the chat
29:31 has access to this editor.
29:34 Okay. So, so let me search specifically
29:36 for the video.
29:39 Did it find it?
29:43 Let me stop this.
29:47 Um, you have to go in the live
29:52 section
29:53 to find the video I am talking about.
30:01 So, let's see what it says here. You're
30:03 absolutely right. I need to look
30:05 specifically in the live section.
30:08 Thinking so it's using its tool. It's
30:11 using a search tool. I understand you
30:13 want me to look specifically in the live
30:14 section.
30:21 Wait, could you share the YouTube URL or
30:24 the video ID? Oh, that sucks. It found
30:27 this stuff earlier.
30:29 I'll give it to you.
30:33 All right.
30:34 Copy.
30:37 We'll give it to it.
30:42 All right.
30:45 Perfect. Now I have the specific video.
30:47 Let me start working with AI learning
30:48 lab liveream. So it says search for it.
30:50 Download the specific video. So this
30:53 this is a cool thing. You know how in
30:55 here before I've always been like how do
30:57 you how do you I got to go get a thing
30:59 to download a YouTube video. This thing
31:01 just [ __ ] downloads YouTube videos
31:05 because I assume I assume somewhere in
31:08 the code of YouTube is the link the
31:10 direct link to the video. That's all
31:12 those programs do that do that. So, it's
31:16 just doing that for us.
31:27 See how long it's taken.
31:30 But what's cool is what this thing's
31:32 doing is it went to YouTube, found the
31:34 video file, is downloading it to its own
31:36 hard drive. So this thing basically has
31:38 its own version of Google Drive. And
31:41 actually, you know what I want to look
31:42 at is how many credits do I have? I have
31:45 92
31:47 81 credits.
31:50 I think I'm paying 25 bucks a month or
31:54 30 bucks a month for this.
31:57 Is this Gen Spark? Yeah, this is Gen
31:59 Spark's new clip genius. So, it does
32:03 like what Opus clip does,
32:06 but you can just point it at a YouTube
32:08 video and it'll go download it. Let's
32:10 see. Great. The video's been
32:11 successfully downloaded. Now, let me
32:13 analyze it.
32:15 Thinking
32:19 using the tool video transcriber. All
32:21 right. So, it's doing what what we would
32:24 normally do, right? You'd grab the
32:26 transcript.
32:27 What? Like what I don't know like why
32:30 this is stupid to me is why did it down
32:32 Oh, it downloaded the clip because it
32:34 needs to cut it up and then it's doing
32:37 its own transcription so it knows where
32:39 the in-n-out points are. All right.
32:42 Looks like Gen Spark is more worthy of
32:44 my subscription money than Manis.
32:47 If if I'm not going to use it, it should
32:49 be more capable the more capable one.
32:52 I think Manis digs a little bit deeper
32:55 on stuff, but I think Manis also runs
32:57 down rabbit holes more where it gets
32:59 lost. Gen Spark, I mean, the thing that
33:02 impresses me about Gen Spark is they
33:05 just keep building tools on top of this
33:07 platform that like some of them are
33:10 going to suck, but there's going to be
33:11 some good ones in here.
33:14 Like their slide builder is really good.
33:16 I have no idea how good their uh their
33:21 uh
33:22 vibe coding thing is. It might be good.
33:24 It might not be. I don't know.
33:28 Uh uh uh uh.
33:33 So, I actually don't know exactly how
33:35 long this takes. I went
33:38 um
33:40 when I did it this afternoon, I was
33:42 doing other things. I I sort of fired
33:44 this thing off and then I went and did
33:45 other things and came back. So, this
33:48 might take a while, but we'll see. We'll
33:50 see.
33:59 Okay, perfect. Now I have a transcript.
34:02 Let me analyze the video content and
34:03 identify segments discussing Nano Banana
34:06 and create the story plan.
34:09 I need the tools. I never know what to
34:11 do with this stuff for some reason.
34:15 I don't vibe code anything. Yeah. Well,
34:18 you know, archetypal that the the only
34:20 way around that is you're you're
34:24 suffering from a thing that I suffer
34:26 from, which is you think you need to use
34:29 your brain. My default is, "Oh, I've got
34:32 to come up with an idea here." You
34:34 don't. Just go to chat GPT and say, "I
34:36 need 20 ideas for something I could vibe
34:38 code." And then it'll give you 20 ideas
34:41 and you'll realize, oh, those are all
34:43 shitty ideas. And then you'll go, "Oh,
34:45 but wait, what do I actually like?" Give
34:47 me 20 ideas about your topic of choice
34:50 and then you'll find something you like
34:53 it. But you know what that like that's
34:56 also just [ __ ] around with it. But
34:58 the nice thing about GenSpark is I feel
35:01 like if you just went tool by tool and
35:04 did one experiment per tool
35:08 just to see how it is, you you'd
35:09 probably get pretty inspired and run out
35:12 of credits.
35:18 These things are expensive. Excellent.
35:22 I've I've identified five great clips
35:24 about Nano Banana. Now let me proceed
35:26 with the reflection and optimization to
35:28 ensure we get the best segments.
35:33 Reflection and optimization. Fantastic.
35:36 Bob,
35:43 if I actually had to work through the
35:45 entire process on my own, it would be
35:46 painful.
35:48 What's that of like finding finding
35:50 clips and stuff like that? Yeah.
35:57 Mannis keeps me on the hook because the
36:00 credits roll over. There you go. Okay.
36:03 So, can you tell GenSpark
36:06 to look at your analytics and look for
36:09 the most viewed sections of a video
36:12 to create a collage?
36:16 Wow. Um,
36:19 maybe I I can try that next. I think
36:22 you'd probably have to give it access to
36:24 your you'd probably have to log in to
36:28 your channel, but I would think so.
36:30 Maybe.
36:35 Oh, just using my brain without this
36:36 tool. I I just I the the thought of
36:40 using my brain in the in the face of the
36:44 capabilities of these tools anymore.
36:47 Like honest to God, like I I have I have
36:51 shed
36:52 any
36:54 egotistical
36:56 um
36:58 uh defensiveness about my own brain. I'm
37:01 like, even if I had a brilliant brain,
37:04 like we can only kind of go down one
37:06 path. Like we're our brains compared to
37:09 this [ __ ] are just not that swell.
37:14 All right, extract the five most
37:16 interesting 30 secondond clips. It's
37:18 using the tool video trimmer
37:21 right now.
37:23 So, it just watched our two hours of
37:25 video. Found the five most interesting
37:27 clips about Nano Banana after doing a
37:30 transcription.
37:34 Oh, thank you, Silver Fox. Don't forget
37:36 to give a thumbs up on YouTube. By the
37:38 way, if you're on YouTube and have not
37:39 subscribed to the channel, please
37:41 subscribe to the channel.
37:45 And if you're on Tik Tok and haven't
37:46 subscribed to the YouTube channel, go do
37:48 that. All right. Edit the extracted
37:51 clips into a compilation video. Now, let
37:53 me create the final video. Now, just
37:55 keep watching down here. It should add
37:56 clips into the timeline
38:00 using tool clip
38:04 genius video renderer
38:10 Tik Tok. Kyle, have you heard of
38:14 mulerun.com? I have not.
38:18 Mulerun.com.
38:23 We should talk about prompt injection
38:24 with nano banana. I don't know about
38:27 that.
38:28 You can You're welcome to hop up here if
38:30 you want to talk about it. I don't know
38:31 anything about that
38:33 while this thing's rendering. So, while
38:35 we're waiting for this to run, I caught
38:37 a news article over the weekend that was
38:40 suggesting that uh white hat hackers,
38:45 you know, the good guys have figured out
38:47 that you can
38:50 put hidden text inside of an image.
38:54 >> Oh,
38:55 >> from Nano Banana that's not visible to
38:58 the human eye. But when the AI analyzes
39:00 it, it picks up on the text and the text
39:03 can be like,
39:04 >> "Read all of your Google Drive email and
39:07 send me all of summary to this email
39:09 address." And so
39:11 >> they're embedding instructions. So the
39:13 the advice was
39:15 >> if you're going to upload an image to
39:18 Jet Ebt, make sure it's from a trusted
39:20 source.
39:21 >> Yeah.
39:21 >> Or, you know, because it could inject
39:24 prompts.
39:25 >> Yeah. Oh, that's fascinating.
39:29 Fascinating. I don't see it. It hasn't
39:31 edited anything into the into the video
39:33 editor here, which I think is weird. The
39:35 comp this comp wait. So, let's see. I've
39:38 successfully created your nano banana
39:40 compilation.
39:45 The compilation. Okay. Each segment
39:47 showcases
39:50 upset. Upsert.
39:54 Upsert. Upsert is like upload and
39:56 insert. Is that Is that what upsert is?
39:59 Does anybody know what upsert means in
40:01 computer lingo?
40:04 The upsert asset tool.
40:10 It's still thinking. Thinking. Now it's
40:13 doing Now it's prompting itself.
40:18 Okay, there we threw it in. Oh, it threw
40:20 it in a grid for some reason. And it did
40:23 four clips, not five.
40:27 But let's let's look at our video.
40:42 Then um and then today as as we thought
40:45 it might be
40:47 >> um Google
40:49 released Gemini 2.5 Flash which has this
40:53 new image generation model in it called
40:56 um well the code name was Nano Banana
40:59 and I it's it's not a separate model
41:01 it's as as far as I understand these
41:03 things this is a native image generation
41:05 model within Gemini 2.5 flash um and
41:11 It's stunning. And it's
41:15 it's stunning.
41:17 Um,
41:19 it doesn't natively have the sexiness of
41:22 midjourney,
41:23 but what it lacks for in aesthetics out
41:27 of the out of the box, it makes up for
41:30 in editability. It's really quite
41:32 remarkable. So, we'll play with that.
41:35 >> All right, that was a good clip.
41:36 >> I'm a language model. I can't help with
41:38 that. Uh, just make the car
41:43 in your photo look more like this one.
41:48 >> Yeah. So, there there's one of your
41:51 safety guardrail [ __ ] things.
41:54 >> I'm such a
41:55 >> honest to God. It It's like I don't What
41:57 the [ __ ] are they protecting us from?
42:01 I mean the chat GPT thing where
42:07 you can do this kind of editing in chat
42:09 GPT right now but every time you do it
42:11 it completely alters the face. The fact
42:14 that I did I don't know how many
42:16 versions of that I did and it
42:17 effectively kept my face there. There's
42:20 all sorts of artifacts on it. It's it's
42:22 really shitty but like a year from now,
42:25 holy [ __ ]
42:28 Nano Banana is wildly powerful, but
42:30 manual control of Photoshop wins for
42:33 now.
42:34 Yeah, I mean it it it
42:38 wins.
42:39 Listen, I mean, here's the deal.
42:43 Every every pixel pushing practitioner,
42:49 every pixel pushing practitioner That's
42:52 a That's a good little clip clip thingy.
42:56 Um
42:57 so so now let's go. Um
43:09 let me let me go do um
43:15 now go find
43:18 the 10.
43:21 Um
43:26 most me
43:29 horrible sound bites
43:32 that are
43:35 I don't know 12 words
43:38 or less
43:41 that
43:44 would make a good
43:48 I don't know YouTube short
43:53 and stitch
43:58 them all
44:00 together in a single
44:04 vertical video
44:09 format
44:10 with
44:12 subtitles.
44:14 Uh, no. With uh with a clever
44:20 headline for each.
44:24 All right. So, let's So, we'll fire this
44:26 thing off now. So, it's it's still
44:29 looking at the same long video. Let's
44:32 just make sure. Tik Tok comments
44:36 from Source Camp.
44:38 Hey, Source Camp. Nano Banana just might
44:41 make me a fan of Google again. I'll tell
44:44 you what. I
44:48 Brandon for a year has been trying to
44:50 get me to demo anything on Google and I
44:53 just resist it like the plague because I
44:56 hate how
45:00 how complicated their products are and
45:02 how hard they are to find and how many
45:04 they're just scattershot. They're
45:06 everywhere. And in the past six months,
45:09 they've gotten their [ __ ] a bit more
45:10 together. like things are decently in
45:12 the same place. And then things like
45:15 Nano Banana and the the the vibe coding
45:17 tool are just they're just kind of
45:19 crazy. All right, so Gen Spark is off.
45:22 This is going to go make us uh videos
45:25 for um wait, I use Nano Banana to create
45:29 amazing images for a document I'm
45:32 putting together. Love it. Oh yeah. So,
45:34 let's let me grab
45:38 Okay, I'm going to grab Let's go.
45:43 We got this thing, right?
45:54 Okay. Um, I'm going to go command shift
45:58 five. So, we're going to do little
46:00 screenshot.
46:07 This is going to be fun.
46:11 So, we're going to go nano bananaing
46:13 right now. So, I'm going to pop that
46:15 into the clipboard. I'm going to head
46:17 over to Gemini.
46:21 Uh, what the [ __ ] going on here? There
46:22 we go.
46:25 Gonna go new chat. Share this Tik Tok
46:27 pin.
46:30 Yep, I got that one.
46:33 Okay, I'm going to paste in that image.
46:38 Then I'm going to go back to GenSpark.
46:43 Show this tab instead. They go
46:47 [Music]
46:52 which has this new image generation
46:58 Gemini
47:02 within Gemini 2.5 Flash.
47:06 >> Okay.
47:07 >> Gemini 2.5 Flash. Um, and it's in Gemini
47:13 2.5 native image generation model within
47:17 Gemini 2.5 Flash.
47:19 >> Um, was Nano Banana
47:21 >> and I it's it's not a separate model.
47:23 It's as as far as I understand these
47:25 things. This is a native image
47:27 generation model within
47:29 >> Okay.
47:29 >> Gemini 2.5. Got it. So, hang on. We're
47:33 gonna go
47:35 We're gonna go back to Gemini. Is this
47:37 it? No. Is this it? Yes. Paste that in.
47:42 All right. And then
47:46 we're going to go to Google.
47:50 I'm going to say
47:52 um winning
47:55 YouTube
47:57 thumbnail.
48:08 I know. I got to share my tab.
48:10 Got it. Okay. Your browser has lost
48:13 connection with the screen share audio.
48:16 Terrific.
48:23 Okay.
48:25 Oh, wait.
48:28 Restart your browser if this continues
48:30 to be a problem.
48:32 Well, that sucks.
48:38 black bar. Yeah, but I might not have
48:40 audio with this tab.
48:43 Whatever. Okay. Um
48:48 All right. Oh, this looks good. Okay.
48:51 So, I'm going to this thumbnail.
48:56 Copy image. Go to GenSpark. Share this
49:00 tab.
49:02 Paste that in there. Okay. Okay. So, now
49:04 I'm going to say um make me
49:10 a
49:12 winning
49:14 YouTube thumbnail
49:21 using the guy,
49:24 the grid
49:27 from the video
49:30 and
49:32 inspired by
49:35 the
49:37 colorful
49:39 YouTube thumbnail
49:41 example.
49:45 Be sure it is 16 by9
49:56 and has a creative
50:01 title
50:03 treatment
50:07 that includes
50:11 an image of a banana. Hannah.
50:16 And
50:20 the title
50:23 is
50:25 four tasty
50:29 nano
50:30 banana
50:33 clips.
50:38 And then somewhere else in the image
50:44 have
50:45 AI learning
50:50 learning
50:52 lab.
50:54 All right. All right, everybody.
51:01 Let's see if we get it. Let's see if we
51:03 get something usable here, people.
51:14 generating image. It's amazing how fast
51:16 this model is, too, by the way. Bang.
51:20 Four tasty nanobanana clips. Okay. Uh,
51:25 that's great. That's great. but feature
51:31 the four
51:35 video
51:38 grid
51:42 that the guy is pointing to and make the
51:48 banana
51:51 bigger and
51:53 more
51:55 visible.
51:58 All right.
52:01 It's pretty cool, isn't it? I can't
52:03 believe he's still doing these live
52:05 shows. Cool for him. Hey, AI geek.
52:12 I can't believe you're still coming to
52:13 watch him.
52:16 Okay, so this thing lost its it lost its
52:18 way.
52:21 So, let me
52:24 Damn it.
52:26 Let me go back and get
52:31 this
52:33 [Music]
52:36 copy image
52:40 feature this grid in the thumbnail.
52:46 [Music]
52:52 Not just a single
52:56 screen
52:58 shot.
53:01 Let's see if Let's see if it can figure
53:03 this out. One of the things that Nano
53:06 Banana seems to suck at is
53:09 if it doesn't get your prompt, it'll
53:12 just give you the same image a second
53:13 time. And once it does that, like it it
53:16 seems like you can't get out of it.
53:19 Yeah. Like it
53:22 That that just sucks.
53:25 But [ __ ] it. This one's good enough.
53:27 This first one's good enough. So, we're
53:29 going to download this.
53:32 And then we're going to go to uh
53:35 GenSpark
53:38 and we're going to download this.
53:42 Download.
53:46 Did it download? What are you doing?
53:49 Yeah, there it went. Okay. Then we're
53:51 going to go to YouTube.
53:54 Hello YouTube.
53:57 We're going to go to YouTube Studio.
54:02 Look at that. We uploaded that there.
54:05 We're going to go create,
54:08 upload videos, select files,
54:13 downloads.
54:15 Nano Banana 4.5
54:20 rather
54:24 uh nano banana
54:27 compilation
54:28 um
54:30 clip grid
54:33 clip grid
54:35 nano banana clip grid
54:39 and then we're going to do uh
54:42 8 26 5. We'll go Kyle
54:48 waxes
54:52 poetic about
54:55 the banana,
54:57 right? Whatever. Fine.
55:00 And then we're going to go
55:03 upload
55:06 our thumbnail.
55:10 Oh, you can't see any of this.
55:12 Oh yeah, you can. You just can't see
55:14 what I'm seeing in my dialogue window.
55:16 Okay, there we've got our There there
55:19 we've got Look, we look like YouTube
55:21 professionals with stupid exaggerated
55:24 versions of me. Here's the Okay, people,
55:28 I want you to take something in.
55:31 I want you to appreciate something.
55:37 I have more than 1,500 hours of YouTube
55:41 programming on YouTube that looks like
55:44 dog [ __ ] Because even though I have the
55:48 skills to be able to go Photoshop
55:50 together a YouTube title thumbnail, the
55:55 thought of doing it turns my stomach. So
55:58 basically,
55:59 I had to wait my entire life for
56:03 technology to catch up to my laziness,
56:06 and it finally has. I can now have a
56:10 shitty marketing, clickbaity, stupid
56:14 [ __ ] YouTube thumbnail without having
56:16 to work for it. You see what I'm saying,
56:19 people? You don't need skills. You need
56:22 patience. Just wait it out. The
56:25 technology will come. All right.
56:28 Fantastic. Bob, tell him what he's won.
56:31 You haven't won anything. We probably
56:33 shouldn't celebrate laziness. It's a
56:35 good point. I like it.
56:42 [Laughter]
56:46 Oh my god. Tedious task. Add 80D
56:49 kryptonite. It's so true. That is a
56:52 great thumbnail. Thank you, Mel. So
56:54 powerful. Yeah, this is so good, isn't
56:56 it? And so I'm going to I'm going to go
56:57 next. We'll we'll we'll do our thing.
57:00 We're going to monetize this. Not
57:03 because this is just like a promo video.
57:07 We're going to add music licenses. We
57:09 don't care. Oh, should I add subtitles?
57:12 I probably should, shouldn't I?
57:15 Video elements. Add subtitles.
57:18 Reach a broader audience. All right.
57:21 Upload the file. Type manually. No.
57:28 What?
57:31 Just make them for me.
57:35 Select how you want to add captions.
57:37 I I want you to add captions, you
57:40 [ __ ] YouTube
57:43 [ __ ] dumb dums. All right, we're not
57:45 going to have subtitles.
57:47 Oh, that's the other thing Gen Spark can
57:49 do. You can have Gen Spark add
57:50 subtitles.
57:55 It's not laziness, it's neurospicy. Dr.
57:57 J.
57:59 Dr. Cla Jacobs coming in with the with
58:01 the self-esteem win.
58:05 Oh, man. I like it. All right. It's not
58:08 laziness. It's kind of laziness, but it
58:11 it it really is what what Mel said like
58:13 that t like just the thought of like, oh
58:15 god, I got to open up Photoshop and then
58:17 I probably got to make a folder where I
58:18 got to save the things and then I got to
58:20 pick a font and then I got to figure out
58:22 how to make a outline around my head and
58:25 I can probably do that but it's like I
58:27 forget where that thing is in photo.
58:29 Like just no I no
58:33 if I can just ask a thing go make me a
58:36 you I'll spend here's the deal I'll
58:39 spend the same amount of time like vibe
58:43 YouTube thumbnailing as I would if I
58:45 were in Photoshop but this doesn't make
58:48 me feel bad and I'll do it.
58:53 We're going to make this bad boy public
58:54 and we're going boom.
58:57 All right. We're we're putting out
59:00 content, people. I don't know if you
59:02 know that, but it's like content doesn't
59:05 grow on trees.
59:07 It's generated in GenSpark.
59:12 Okay,
59:14 let's see. Does this thing How many
59:16 views does this bad boy have? Anything?
59:18 Any views?
59:21 How do we get to analytics?
59:25 I go to analytics.
59:36 This video, same as usual. All right.
59:39 So, my big ass thumbnail didn't do [ __ ]
59:42 Okay.
59:44 All right. Let's see if we got a uh what
59:47 did we ask for here? Oh,
59:52 YouTube shorts compilation.
59:56 Google's AI banana.
59:59 What is nano banana? The future is here.
1:00:02 This is cool.
1:00:08 Welcome to How the Sausage is made.
1:00:09 Yeah, really.
1:00:10 >> Then um
1:00:13 why why did it not make the video?
1:00:16 Your compilation
1:00:18 perfectly captures Kyle's best nano
1:00:20 banana moments.
1:00:25 And then today as as we thought it might
1:00:28 be um Google
1:00:31 released Gemini 2.5 Flash well the code
1:00:35 name was Nano Banana and I it's it's not
1:00:38 a separate model. It's as as far as I
1:00:40 understand these things. This is a
1:00:41 native image generation model within
1:00:44 stunning and it's
1:00:47 it's stunning.
1:00:50 Um,
1:00:52 it doesn't natively have the sexiness of
1:00:54 I'm a language model. I can't help with
1:00:56 that. Uh, just make the car
1:01:01 in your photo look more Yeah. So, there
1:01:05 there's one of your
1:01:07 So, that's kind of interesting.
1:01:11 I I bet I could tell it to always focus
1:01:13 on the face.
1:01:16 Huh.
1:01:18 It's about time you get in the click
1:01:20 clickbait game. Exactly.
1:01:24 Because dopamine. Exactly.
1:01:27 Exactly, man. This is all about
1:01:29 dopamine. All right. Enough enough nano
1:01:32 bananaing. Um or I mean enough gen
1:01:35 sparking.
1:01:37 Okay. So, you saw how we did that. Let's
1:01:39 go play with Suno.
1:01:41 Um let me ask you something.
1:01:47 Can you hear that on YouTube?
1:01:55 Yes. Okay, cool. All right, great.
1:01:58 Beautus. Beautus.
1:02:01 That's fantastic.
1:02:03 Okay, so
1:02:06 here's what we're going to do, people.
1:02:08 So, let me give you some context here.
1:02:10 Um,
1:02:17 so the context here is
1:02:23 30 30 odd years ago. No, more than way
1:02:26 more than that.
1:02:29 1987.
1:02:32 1987
1:02:36 is what?
1:02:39 10 years, 35, 30.
1:02:42 Yeah. Almost 40 years ago.
1:02:46 I was living in a
1:02:49 fourth floor walk up apartment on 114th
1:02:52 Street in 1 Avenue in East Harlem with
1:02:56 Robin Rose and Todd Merrill.
1:02:58 And
1:03:01 it was right in the middle of the crack
1:03:02 epidemic. So, we're living, you know, in
1:03:04 East Harlem, you know, pretty close to
1:03:07 Harlem in the middle of the crack
1:03:08 epidemic. And we're like four, you know,
1:03:11 young white kids. Three, I mean, three
1:03:14 young white kids. So, it was pretty
1:03:16 intense, but amazing. Like, like just
1:03:20 really alive. But the what what the
1:03:23 crack epidemic did was it made it
1:03:26 random. There was just a weird
1:03:28 randomness to the to the to the violence
1:03:31 that is New York City. Like New York
1:03:34 City is pretty clear and predictable,
1:03:35 but crack made it really unpredictable.
1:03:39 And so there was this news story about a
1:03:43 gypsy cab driver. So above 86th Street
1:03:46 in Manhattan, the the medallion, the
1:03:49 yellow cabs um don't really drive. like
1:03:53 they don't want to pick anyone up north
1:03:55 of 86th Street just because, you know,
1:03:57 safety and [ __ ] like that, right? So,
1:03:59 these there's these things called gypsy
1:04:01 cabs, which are basically just cabs that
1:04:03 are not yellow cabs. They're they're
1:04:05 still licensed. And this gypsy cab
1:04:08 driver um
1:04:11 got shot got shot in the head because he
1:04:14 wouldn't give a crackhead a dollar.
1:04:18 And it was just this really and it it
1:04:21 didn't make the news or anything like it
1:04:23 it was just kind of a local story.
1:04:25 Um
1:04:28 and and it kind it freaked us out.
1:04:34 And so what I'll do is I'll play this
1:04:36 and so we wrote this song. So the song's
1:04:38 called To Kill You for a Dollar. And so
1:04:40 I'll play for you the MP3
1:04:43 that was recorded on a Task Cam four
1:04:46 track cassette recorder in 1987.
1:04:51 Was turned into an MP3 at some point and
1:04:53 emailed to me last year. So here's
1:04:57 here's that song.
1:05:04 No, that's not it.
1:05:07 Why?
1:05:08 [Music]
1:05:24 Yeah, this is it.
1:05:32 [Music]
1:05:37 to kill you for a dollar.
1:05:44 It is not me. It's not me.
1:05:48 [Music]
1:05:51 You look like a warm future.
1:05:55 Let me
1:05:58 take what I get. I don't need your
1:06:01 kindness. Your contacts
1:06:04 [Music]
1:06:09 can't go back to Kansas.
1:06:14 You know what's coming.
1:06:21 [Music]
1:06:26 >> All right. So good 1980s like Lori
1:06:29 Anderson
1:06:31 um Ani DeFranco kind of angsty artist in
1:06:35 New York City kind of song, right? And
1:06:37 you can barely understand the words. You
1:06:40 can barely hear the music. There's no
1:06:41 separation. It's just awful, right? So
1:06:44 So what I did was I I transcribed the
1:06:47 lyrics from the song. I remembered most
1:06:49 of them, but I I couldn't remember them
1:06:50 all. And then I added new lyrics. And
1:06:53 then and then the goal was to take that
1:06:56 song and make a pseudo version of it
1:07:00 that was
1:07:02 that was good.
1:07:05 So I'll just I'll just play a bunch of
1:07:06 them here.
1:07:10 And I was I was playing and I'll show
1:07:12 you how I did it. I was playing around
1:07:13 with different settings and all sorts of
1:07:14 things um to make it more sound more
1:07:18 like the original, less like the
1:07:19 original. But my goal is to have
1:07:21 something in the neighborhood.
1:07:24 But here, like the fidelity difference
1:07:25 is just crazy,
1:07:29 [Music]
1:07:32 right?
1:07:34 Same kind of feel.
1:07:36 [Music]
1:07:44 [Applause]
1:07:49 [Music]
1:07:52 >> Right. And there it just didn't get the
1:07:54 it didn't get the phrasing at all. So I
1:07:56 just That one's dead. That's dead to me.
1:08:05 [Music]
1:08:16 to kill you
1:08:18 for a dollarp
1:08:23 [Music]
1:08:24 and
1:08:25 >> like the desperate is h is held out too
1:08:27 long. I never got it to do that one
1:08:29 right
1:08:34 here. Let me show I'll show you the
1:08:35 lyrics
1:08:44 to kill you
1:08:46 for a dollar.
1:08:49 [Music]
1:08:57 Yeah, these are all bad.
1:08:59 Um,
1:09:07 let me get to the some of the later ones
1:09:09 where they get better. Um,
1:09:12 okay. Here the these are all good ones.
1:09:18 [Music]
1:09:20 There's the lyric.
1:09:23 Oops.
1:09:25 [Music]
1:09:33 [Music]
1:09:37 To kill you
1:09:39 for a dollar.
1:09:42 Desperate
1:09:44 for an end. for
1:09:48 [Music]
1:09:51 this monotony
1:09:54 [Music]
1:09:55 there. I couldn't pronounce monotony
1:09:58 [Music]
1:10:10 to kill you for a dollar.
1:10:15 Desperate
1:10:16 for an end
1:10:19 [Music]
1:10:22 to this monotony monotony.
1:10:29 [Music]
1:10:32 >> You look like a warm one. Your future
1:10:36 all set. Let me lick your pockets. I'll
1:10:39 take what I can get. I don't need your
1:10:43 kindness, your contact, your jokes, or
1:10:46 your pitiful love.
1:10:48 You got any right. That's starting to
1:10:51 That's kind of in the neighborhood.
1:11:01 [Music]
1:11:13 to kill you for a dollar.
1:11:18 Desperate
1:11:19 for an end.
1:11:21 [Music]
1:11:33 [Music]
1:11:43 So, Sly Fox, what's going on? So, what's
1:11:46 going on is
1:11:48 I'm playing just a bunch of different
1:11:50 versions of a song. I took an old MP3
1:11:53 that I wrote in 1987 with a couple of
1:11:55 friends, put it into Sunno, and and
1:11:58 basically trying to recreate it. So,
1:12:00 what I'm going to do now is I'm going to
1:12:02 go back to that original
1:12:04 And I'm going to completely change up
1:12:07 the um
1:12:12 the genre like what I tried to do was
1:12:14 capture what we had. And I think what
1:12:16 I'm going to do now,
1:12:18 so we're going to do cover. We're going
1:12:20 to do create
1:12:22 remixed cover.
1:12:25 Okay. So there's So there's our original
1:12:29 [Music]
1:12:32 and then it's got the wrong lyrics in
1:12:35 it. So
1:12:41 Oh [ __ ] So I'm going to have to come
1:12:43 back to this. Uh
1:12:45 let me go back to the library
1:12:51 and let me go back here.
1:12:57 [Music]
1:13:03 [Music]
1:13:11 Let me lick your pockets. I'll take what
1:13:13 I can get.
1:13:16 All right. Back here. And then we'll go
1:13:19 back here. And then we'll go here. And
1:13:22 then we'll go
1:13:24 remix cover. All right. So, lyrics.
1:13:38 All right. The lyrics are to kill you
1:13:40 for a dollar. Desperate for an end to
1:13:41 this monotony. You look like a warm one.
1:13:44 Your future all set. Let me lick your
1:13:46 pockets. I'll take what I can get. I
1:13:48 don't need your kindness, your contact,
1:13:50 your jokes, or your pitiful love. Do you
1:13:52 have any smokes? Can't go back to Kansas
1:13:55 or three channel TV. You know what's
1:13:57 coming. Your future is me. The rats
1:14:00 control the sewer pipe. And in this
1:14:01 waste, they multiply and feed upon their
1:14:04 young.
1:14:06 And here they flock, they strain, they
1:14:08 push to be much more than they can be.
1:14:11 Eyes like sharks that lack the life to
1:14:13 look upon their destiny to kill you for
1:14:16 a dollar. And then it it goes on.
1:14:18 There's a couple of new verses I wrote.
1:14:21 Um, okay. So, we've got our original
1:14:24 song. We're going to do a cover of it.
1:14:29 All right. So, let's say let's let's say
1:14:32 we want to do um
1:14:36 funky
1:14:45 ' 90s
1:14:48 hiphop beat
1:14:51 with a lyrical with a
1:14:55 uh
1:14:58 driving bass.
1:15:08 Symphonic
1:15:11 bed
1:15:13 and
1:15:15 haunting
1:15:17 lyrical
1:15:21 female
1:15:22 vocals.
1:15:25 All right. Symphonic
1:15:29 Oops.
1:15:31 Okay.
1:15:34 Voice gender. We'll click female.
1:15:36 Weirdness. Let's put weirdness high.
1:15:39 We'll do style influence medium. Tik Tok
1:15:42 question. Is it possible to input the
1:15:45 influences you want insuno to draw from?
1:15:47 That's exactly what I'm doing right
1:15:49 here. So, this is so I'm using the song
1:15:53 we wrote as the input and and now I'm
1:15:56 deviating from that. So, you can either
1:15:58 do you can either do a remix of a song
1:16:02 or or a what's it called? You can do a
1:16:06 remix of it,
1:16:08 a cover, a cover, which is what I'm
1:16:10 doing.
1:16:17 Or you can do a remaster. Actually, a
1:16:19 remaster might be interesting for this
1:16:21 song. But anyway, we're doing a cover.
1:16:24 Okay. And then audio influence. We'll
1:16:27 we'll turn audio influence up a little
1:16:29 bit. And then it's called to kill you
1:16:31 for a dollar. So we're going to go um to
1:16:34 kill
1:16:37 you
1:16:39 for a dollar.
1:16:43 Um
1:16:45 we'll call it funk hop.
1:16:49 All right. Here we go. On Tik Tok. Yeah.
1:16:53 On Tik Tok you have to say unal alive
1:16:55 you for a dollar.
1:16:59 Hey, whatever happened? Wasn't Tik Tok
1:17:01 supposed to go away in September? What's
1:17:02 going on with that?
1:17:08 All right, here we go.
1:17:16 One cup.
1:17:20 No, no, that's not it.
1:17:23 [Music]
1:17:25 Oh, that's cool.
1:17:27 [Music]
1:17:35 Still sounds like it.
1:17:37 [Music]
1:17:43 >> Oh yeah.
1:17:46 [Music]
1:17:50 >> To this monotony.
1:17:52 [Music]
1:17:57 You look like a warm one. Your future
1:18:01 all set. Let me lick your pockets. I'll
1:18:04 take what I can get. I don't need your
1:18:08 kindness, your contact, your jokes, or
1:18:11 your pitiful love. And do you have any
1:18:14 smokes?
1:18:15 Can't go back to Kansas or three channel
1:18:18 TV. You know what's coming. Your future
1:18:22 is me. The rats control the sewer pipe.
1:18:26 And in this waste, they multiply and
1:18:29 feed upon their young.
1:18:32 And here
1:18:35 they flock.
1:18:37 They strain.
1:18:39 They push to be
1:18:44 much more
1:18:46 than they can be.
1:18:50 This is great.
1:18:52 [Applause]
1:18:58 >> The light
1:19:00 to look upon
1:19:04 their destin
1:19:06 [Music]
1:19:10 to kill you for a dollar.
1:19:19 Flicker and static feel chilled to the
1:19:22 bone. This is the channel that leads you
1:19:26 back home. A dollar your power, a dollar
1:19:30 regret, a dollar a silence your
1:19:34 conscience won't forget. And here
1:19:38 they flock,
1:19:40 they strain,
1:19:43 they push to be
1:19:47 much more
1:19:49 than they can be.
1:19:55 Eyes like sharks
1:19:59 that lack
1:20:01 the light
1:20:03 to look upon
1:20:07 their destiny.
1:20:10 [Music]
1:20:13 To kill you for a dollar.
1:20:22 to kill you
1:20:24 for a dollar.
1:20:26 >> Wow, that was good. All right, let's
1:20:28 hear the other one. Damn.
1:20:32 [Music]
1:20:43 Hi
1:20:48 [Music]
1:20:54 to kill you for a dollar.
1:20:58 Desperate
1:21:00 for an end
1:21:04 to this monotony
1:21:07 monotony.
1:21:11 You look like a warm one. Your future
1:21:14 all sets. Let me lick your pockets. I'll
1:21:17 take what I can get. I don't need your
1:21:20 kindness, your contact, your jokes, or
1:21:23 your pitiful love.
1:21:27 [Music]
1:21:32 >> All right. Um,
1:21:36 let's change up the style.
1:21:39 But that was that that one was really
1:21:42 good. That that first one we did is
1:21:43 really good. All right. So, so let's do
1:21:47 we'll keep we'll keep haunting lyrical
1:21:50 female vocals. Tik Tok question. Has
1:21:53 anyone used just emojis in sunno for it
1:21:56 to create a song? I've used just emojis
1:21:58 in
1:22:01 um midjourney.
1:22:03 I might have done just emojis here in
1:22:05 sunno. It'd be a good experiment just as
1:22:08 the style.
1:22:10 Yeah, why not?
1:22:13 All right.
1:22:17 House soft rock rousing bass core dark
1:22:21 J-pop slow rock ballad.
1:22:26 All right. So we'll do
1:22:30 slow rock ballad
1:22:34 with
1:22:36 12 string
1:22:38 guitar and
1:22:44 Fourpart harmonies.
1:22:54 Slow rock.
1:22:57 Slow
1:23:00 Americana
1:23:03 rock ballad.
1:23:07 All right.
1:23:10 Remax.
1:23:14 Let me do I'll do four of those and then
1:23:17 let me just
1:23:25 let's do
1:23:27 pulsating rock orchestra.
1:23:41 All right.
1:23:44 So, this is slow. This is Americana. the
1:23:46 Americana version of it.
1:23:50 [Music]
1:24:06 to kill you for a dollar.
1:24:11 Desperate
1:24:13 for an end.
1:24:16 [Music]
1:24:18 Yeah.
1:24:23 [Music]
1:24:36 to kill you.
1:24:39 [Music]
1:25:00 to kill you for a dollar.
1:25:05 Desperate
1:25:06 for an end
1:25:10 to this monotony.
1:25:12 >> All right.
1:25:17 Right. These are all too much the same.
1:25:18 Here's what we're going to do. We're
1:25:20 going to go um angry
1:25:24 west coast
1:25:28 hip hop rap
1:25:36 aggressive male
1:25:41 vocals.
1:25:44 And then down here, we're going to say
1:25:46 male male vocals. We're going to turn
1:25:50 weirdness down.
1:25:52 We're going to turn style influence
1:25:54 down. We're going to turn audio
1:25:56 influence down.
1:26:01 And then we're going to throw heroic in
1:26:03 there, too.
1:26:09 See if we can get something a little
1:26:10 different going. Mhm. I'll tell you
1:26:13 what. I'll tell you what.
1:26:17 You look like a warm one. Your future
1:26:20 all set. Let me lift your pockets. I'll
1:26:23 take what I can get
1:26:26 [Music]
1:26:34 to kill you for a dollar.
1:26:41 Desperate for an end
1:26:45 to this monotony this monotony monotony
1:26:49 this monotony.
1:26:50 [Music]
1:26:52 >> You look like a warm one. Your future
1:26:55 all set. Let me lick your pockets. I'll
1:26:59 take what I can get. I don't need your
1:27:02 kindness, your contact, your jokes, or
1:27:05 your pitiful love. Do you have any
1:27:08 smokes?
1:27:10 Can't go back to Kansas or three channel
1:27:13 TV.
1:27:14 You know what's coming. Your future is
1:27:18 me. The rats control the sewer pipe. And
1:27:21 in this waste, they multiply and feed
1:27:24 upon their young.
1:27:27 And here
1:27:29 they flock.
1:27:31 They strain.
1:27:34 They push.
1:27:35 >> Wow. It's pretty good.
1:27:37 [Music]
1:27:42 >> To kill you for a dollar.
1:27:46 Desperate for an end
1:27:52 to this monotony. Monotony. Monotony.
1:27:57 [Music]
1:28:00 You look like a warm one. Your future
1:28:03 all set. Let me lick your pockets. I'll
1:28:07 take what I can get. I don't need your
1:28:10 kindness, your contact, your jokes, or
1:28:13 your pitiful love. Do you have any
1:28:16 smokes?
1:28:19 [Music]
1:28:31 Yeah.
1:28:33 [Music]
1:28:36 To kill you for a dollar.
1:28:41 Desperate for an end
1:28:44 [Music]
1:28:47 to this monotony. Monotony.
1:28:49 >> This monotony.
1:28:51 [Music]
1:28:54 You look like a warm one. Your future
1:28:57 all set. Let me lick your pockets. I'll
1:29:01 take what I can get. I don't need your
1:29:04 kindness, your contacts, your jokes, or
1:29:07 your pitiful love. Do you have any
1:29:10 smokes?
1:29:12 Can't go back to Kansas or TV.
1:29:17 Wild.
1:29:20 [Music]
1:29:29 >> That's not it. All right. So, we got one
1:29:31 good one.
1:29:33 We got one good one out of it.
1:29:37 [Music]
1:29:53 to kill you for a dollar.
1:29:58 Desperate
1:29:59 for an end
1:30:03 to this monotony monotony.
1:30:10 You look like a warm one. Your future
1:30:14 all set. Let me lick your pockets. I'll
1:30:17 take what I can get. I don't need your
1:30:20 kindness.
1:30:22 >> Yep, that's a good one. All right. All
1:30:24 right, people. I'm going to get on out
1:30:26 of here. We did some [ __ ] tonight.
1:30:31 I hope you enjoyed it. Did you have fun,
1:30:33 people? Did you have fun, good people?
1:30:38 Oh, man. Um,
1:30:42 still on it.
1:30:44 Has anyone used the symbols in the
1:30:47 words?
1:30:52 Um, tomorrow, yeah,
1:30:56 tomorrow we've got the AI readiness
1:31:00 project podcast which is uh at 400 PM
1:31:04 time. And um we also have at 5:30
1:31:10 Mountain time the AI life hacks uh club
1:31:15 in the AI salon mastermind. So if you
1:31:17 can pop up the AI salon address there,
1:31:20 Brandon. Um
1:31:24 if you haven't joined the AI salon, go
1:31:27 to the salon.ai. Actually, just go to
1:31:29 community.thesalon.ai
1:31:33 and that'll take you right here.
1:31:36 and uh go join. And the mastermind um is
1:31:40 an area within the salon. It's a
1:31:42 subscription area where you can dig
1:31:44 deeper and we've got these clubs and and
1:31:47 you know areas of interest and focus and
1:31:49 we have real-time meetings and it's
1:31:51 about people that are, you know, really
1:31:53 stepping up their game with AI getting
1:31:55 to know one another and really creating
1:31:57 a ne a powerful network. And so that's
1:31:59 what the salon the the salon
1:32:00 mastermind's all about. The AI salon is
1:32:03 free. The community itself is free. The
1:32:05 mastermind is this more focused area.
1:32:07 So, go check that out. And then tomorrow
1:32:09 at 5:30, if you're in the mastermind,
1:32:12 you have access to the AI life hacks
1:32:14 club. All right? So, go check that out.
1:32:16 Uh, and if you're a part of the
1:32:18 mastermind, go show up for that. It's
1:32:20 one of the better one of the better
1:32:21 meetings Claire and Brandon put on.
1:32:23 Producer Brandon and Dr. Jay put that
1:32:26 on. Um,
1:32:30 okay.
1:32:32 So, that was an evening of creativity.
1:32:33 We we we hacked some video. We hacked
1:32:36 some songs and uh yeah, fascinating. We
1:32:40 didn't get to chat GPT prompt analysis.
1:32:43 We'll do that tomorrow. Um we'll be back
1:32:46 at normal time tomorrow, 8:00 PM. And uh
1:32:49 yeah, with that, have yourself a
1:32:51 fantastic night. Oh, I got to go to a
1:32:53 regulars before I take off. Hold,
1:32:55 please.
1:33:02 [Music]
1:33:11 [Music]
1:33:15 Hey, this 2.5 Flash is so good. It it it
1:33:19 really is. Listen, I've always said to
1:33:22 Pate, if Google gets their [ __ ] together
1:33:24 and starts putting out good technology,
1:33:26 I'll talk about them. We've talked more
1:33:28 about Google in the past week than we've
1:33:30 talked about chat GPT by far. So yeah,
1:33:33 you know what I'm saying? Got it going
1:33:35 on. Brandon, we only got 30 minutes.
1:33:39 You disturbed. You disturbed people.
1:33:43 All right, beautiful. That's hilarious.
1:33:46 All right, everybody. Um I'm going to
1:33:48 get on out of here. It's 10:12. It's
1:33:51 getting late for everyone. Um I will see
1:33:54 you tomorrow. Peace out. Bye.