5/14/2025 - Navigating the Maze of AI Models: Understanding OpenAI's Latest Developments

Description

Chapters

Transcript

0:01 Come here, buddy.
0:03 [Music]
0:12 Did I close the door? Yeah.
0:16 All right. You ready? You ready?
0:22 [Music]
0:23 In a westerly.
0:27 [Music]
0:34 My driving. I've been
0:37 [Music]
0:38 wondering what it is I'm running from
0:42 [Music]
0:43 again. Feel like an 80y old
0:47 man holding on to 29.
0:52 And up on that
0:56 rer is a California
1:00 [Music]
1:06 line. Double ahead of trucks carrying a
1:08 wide
1:10 load. Refab house cutting
1:14 high. Cute little front door and two
1:17 windows.
1:19 alone ain't sure whether to cry I should
1:24 laugh you see I broke a home up myself
1:28 once days I stumbled to
1:33 that I know father don't
1:37 lie said don't you come around
1:42 here
1:45 anymore had
1:50 the
1:55 [Music]
1:56 street. He says, "What happened?
1:59 [Music]
2:29 Oh, good people. Good people. Good
2:31 people. What is happening?
2:37 [Music]
2:58 There's been something baby I've been
3:01 trying to
3:01 [Music]
3:03 say. What age? And it seems I don't know
3:09 how with the past and the future now
3:12 surrounding
3:15 me. Surrender to whatever cheap can be
3:20 found. There's been a little
3:23 trouble since you came to my rescue.
3:30 [Music]
3:31 And if you like all of the rest, I would
3:34 have quit you long ago, but I couldn't
3:37 do
3:41 that. Oh, tell me now, woman. And why I
3:45 never went
3:46 [Music]
3:48 to make a man crazy, make him cold as
3:51 hell.
3:55 I'm a woman, you wish me well.
3:59 But it's a trying. Still going to have
4:02 to find my own way
4:05 [Music]
4:07 through.
4:10 Yeah. Like the mist of morning, my dream
4:13 remains. What is going on, good people?
4:16 All
4:18 right. You're on the yellow sticky FYI.
4:22 Wait. Oh, okay. Wait, what?
4:28 You're on the yellow sticky. FY. Okay,
4:31 now I can see. Oh, I see. I stepped on
4:34 your ability to edit. Got
4:38 it. Producer Brandon freaking out
4:40 because he doesn't have
4:42 [Laughter]
4:46 [Music]
4:48 sticky. Oh man.
4:54 [Music]
4:58 Is that hair AI? Yeah, it's we've got
5:01 some Hawaii 5 action going on today. I
5:05 don't know what to tell you. You know
5:07 the 70s called. They want your hair
5:09 back. their hair back.
5:12 [Music]
5:34 [Applause]
5:34 [Music]
5:41 [Music]
5:49 Oh, I was going into withdrawal for
5:50 those three minutes cuz we were late.
5:53 Don't tell Source Camp. She's going to
5:54 come in here and be all guilty that she
5:56 was late. So, we can't You can't admit
5:58 it. I'll tell you my favorite story on
6:01 on on being late. Um I think I've told
6:05 it in here before. So we we talked to
6:08 someone about
6:09 StoryVine
6:11 and they're a major uh direct marketing
6:15 skin brand. We'll leave it at that. And
6:19 um we talked to them and they're like,
6:21 "Oh, StoryVine sounds interesting. Would
6:23 you be willing to fly out to San
6:25 Francisco to meet with us and tell my
6:27 team about it?"
6:29 Sure. So, we book flights. We're all
6:34 excited. We fly to San Francisco. We
6:36 book a lot of other meetings because
6:38 that's how we roll. We show up to the
6:40 place and we say, "Hey, we're here to
6:42 meet so and so. Really excited to meet
6:45 her. Glad to be
6:47 here." The response was, "Can you prove
6:50 you have a meeting with
6:52 [Laughter]
6:56 her?" And we were like, "Well, we've got
7:00 some plane tickets from Denver to San
7:02 Francisco because we were asked to fly
7:04 out here at her
7:10 request." So, so they're like they they
7:13 were kind of stuck. So, she said, "Well,
7:15 she's in a meeting for the hour that we
7:17 were supposed to meet with her. So, can
7:19 you wait? We'll wait. We'll wait." So,
7:22 then the hour passes. We go meet with
7:25 her team. There's like three people in
7:27 there. They don't know who we are. We're
7:30 in there for 20
7:31 minutes. She walks into the room, sits
7:35 down, doesn't acknowledge a
7:41 thing. Just said, "What are you guys
7:44 talking about? Show me what you got.
7:46 Let's go." Didn't [ __ ] acknowledge
7:49 it. That was Man, that was that's the
7:51 height of corporate douchebaggery.
7:53 [Music]
7:57 You
7:57 [Music]
8:19 know, you
8:23 [Music]
8:40 Um, what are we going to talk about
8:42 tonight? We're going to talk about I
8:44 want to talk a little bit about Open AI.
8:47 I think OpenAI is currently in the
8:49 process of jumping the shark with chat
8:52 GPT becoming functionally
8:55 [Music]
9:01 unusable. And I don't mean unusable like
9:03 you can't use it. I mean unusable like
9:06 you can't use it at its best.
9:14 Um, the whole reason chat GPT became
9:18 popular was because they took all the
9:21 technical choices
9:23 away from the end user. They just made
9:27 it the chat hole. That's all it was. It
9:30 was just the chat hole.
9:35 And I, you know, I don't necessarily
9:37 blame them or any of the frontier model
9:39 companies for for falling into
9:42 this
9:44 challenge, but because they're all
9:46 copying each other and because all of
9:48 them are copying Open
9:49 AI, the problem that Open AI has and is
9:54 proliferating extends across all of the
9:57 large language models. Anthropics is one
10:01 that is slightly better. They've they've
10:03 hidden some of the [ __ ] within
10:06 Anthropic, but it's it's just really
10:10 bad. So, I'm going to show you there's
10:12 there's some new [ __ ] in Open AI. And
10:15 normally, not only do I kn know there's
10:17 new [ __ ] in there, I know why it's in
10:20 there. I know what it is. I know how to
10:21 get to it. I know how to use it. And I
10:23 know why you would use it. It's in
10:25 there. I don't give a [ __ ] It's in
10:27 there. and I do not give a [ __ ] The
10:29 health benchmark announcement from
10:31 OpenAI, that's a that's an interesting
10:33 one. I haven't read up on it, so I don't
10:35 I'm not educated on it. We could
10:37 certainly go look at it. There's also an
10:39 interesting paper from uh Google Deep
10:41 Mind today. Um which I think might be
10:44 worth going to look at
10:46 together.
10:48 Um let's see. Veil and wash. What is
10:51 going on over here? Why am I Oh, I see.
10:54 My chat GBT was humming. Now they [ __ ]
10:57 with
10:58 it. It name is Kai. Now Mr. It named his
11:03 his catchy BT Kai. Well, mine's
11:07 Quinn. Um, here. Let me go show you
11:09 something. Let me flip
11:14 [Music]
11:18 over YouTube question. Hey, Kyle. James
11:22 Dutder. At some point, can we see the
11:24 render that was running when you signed
11:27 off?
11:30 Yes. What was that? That was Hedra,
11:34 right? Is it here? I think it's here.
11:38 Oh, yeah. I think I watched a little of
11:39 it. It was not bad. It's not great, but
11:42 it's not
11:43 bad.
11:45 Um, yeah. Let's we'll do we'll do this
11:47 because it's a nice little entertaining
11:48 start while we get some more people in
11:50 here.
11:52 Um tabs for audio. Yep, exactly. I'm
11:55 working on that
12:02 now.
12:04 Okay,
12:13 so here we go.
12:22 [Music]
12:28 So, for those of you that were not here
12:31 last
12:33 night, the
12:40 um the process to get to this video was
12:44 the so the little kid with the marbles
12:46 in the sweater and the hat and the tan
12:50 pants. That was a picture of me
12:54 um a storyboard done in Keynote with
12:57 just block shapes basically that said
12:59 here's where the kid is, here's where
13:00 the marbles
13:01 are. Um, and then we put him on a
13:04 schoolyard. And then the two guitar guys
13:08 were made after we we went over and made
13:10 this song in Sunno called Marble
13:14 Man. And they were really cool. And then
13:17 someone said, "Hey, could you use So
13:19 what we used was Runway ML has this new
13:23 thing called um
13:25 references and you can put up to three
13:28 images in there. But what you can do is
13:30 you can sort of put a bunch of people on
13:32 one image and and label them like here's
13:35 Jim, here's Sally, here's this. So, we
13:37 basically said take these two guitar
13:39 players and put them behind the kid with
13:42 the with the
13:46 uh
13:48 marble.
13:50 Um Yeah. And so here's how it rendered
13:54 out.
14:00 [Music]
14:13 Marles. They call me the marbles man.
14:17 Marles. My favorite one's my closest
14:20 friend. Marles. I see them rolling down
14:24 the lane.
14:26 Marles. I'm in love with my marbles.
14:29 Ain't that insane?
14:31 Marles. It's all I think about.
14:35 Marles. They don't make me scream or
14:38 shout.
14:40 Marles. Every single one's a tip.
14:44 Marles. I'm in love with my marbles.
14:46 Ain't that insane? Don't tell me I'm
14:50 wrong. You just haven't seen what I've
14:54 seen.
14:57 Don't call me
15:00 crazy because marles aren't just a
15:04 dream. As for real as you are me, I'm in
15:09 love with my marles. Ain't that insane?
15:14 [Music]
15:31 So anyway, so that's that. So I took
15:33 this video, I put it up on on Yield
15:36 Internet. I finally got Twitter to
15:37 upload a video. It took me, I don't
15:39 know, 15 attempts. Way to go, Elon. He's
15:43 he's, you know, literally launching and
15:46 catching skyscrapers and making people
15:48 who can't speak speak again. But you
15:50 can't upload a [ __ ] video to
15:54 Twitter. Anyway, I got it to upload and
15:59 and then I put it up and I put up a
16:01 whole instruction thing on how it was
16:03 put together. It got like three views. I
16:06 am so [ __ ] sick of Twitter.
16:09 So, if you want to go if you would like
16:11 to support the channel, go to X and
16:14 click on
16:16 um click on my
16:21 account. Let's see. We'll do it
16:23 together. Go to Go to the Twitter. Go to
16:26 the Twitter and go to Kyle Shannon. And
16:30 then I think it should be one of
16:33 these. Yeah, right here. Look at this.
16:36 So, I put up this awesome post,
16:39 Marbleman and Joy, and then I did a
16:42 comment like, "Here's all the
16:44 instructions. Like, here's all the
16:46 images I used." I described the whole
16:47 [ __ ] thing. 85 views, no
16:51 comments. So, you know, go give it some
16:54 love.
16:58 I just the the the engagement algorithm
17:01 if you don't have [ __ ] 100,000
17:03 followers on Twitter, you just get dick.
17:07 It's just it's it it's
17:12 [Music]
17:16 demoralizing. Joy X's crickets for me
17:18 too, Kyle. Yeah, it's really [ __ ]
17:20 frustrating. And like Joy puts out
17:22 amazing stuff, right? Amazing content
17:25 all the time. It's just just maddening.
17:29 Anyway, enough bitching about social
17:33 media. Nobody likes me. Nobody pays
17:36 attention to my stuff. I don't have any
17:41 reach. What are you going to do? What
17:44 are you going to do? All right. Can we
17:47 use Zapier? We
17:50 can to do what? to get people to go
17:54 click on my [ __ ] Yeah, we could we
17:57 could make an automation that just goes
17:58 and finds people and randomly logs into
18:01 their accounts and has them log into my
18:03 account and click on [ __ ] We could do
18:07 that. Um, okay. Let's go to chat japet.
18:12 Oh, look there. This was the little
18:15 starting image that led to
18:21 neck ne.
18:31 Um, all right. That was the start
18:36 image. Come with me. You'll need to
18:39 switch to the big
18:41 TV screen and rock the sound. The Kyle
18:45 Rubin
18:50 Kyle Rubin. Yeah, there you
18:54 go. Rick Rubin. I think Rick Rubin, he
18:58 he must have a new
18:59 publicist because you knew about Rick
19:02 for a while. He had his book came out
19:03 and people followed him. But now like
19:05 he's got he's got this new show if you
19:08 haven't seen it on YouTube. I think it's
19:10 called Tetra Tetra
19:13 Grammaton. Tetra
19:15 Grammaton. Um, which I'm sure means
19:18 something cool, but it feels to me like
19:19 it should be Tetra Grammatron, but it's
19:22 Ton. Um, and it's like two and a half
19:26 hour interviews where he just sits there
19:28 cross-legged, you know, playing with
19:31 stuff and just talking to cool people.
19:33 It's pretty cool. So, if you haven't
19:34 seen it
19:36 yet Oh, it's Little Kyle. It is Little
19:39 Kyle. Source Camp, you're late again.
19:42 Listen, one of the things that I think
19:44 we've talked about multiple times is
19:46 like being late is not an option,
19:51 right? You made a commitment. Your
19:54 commitment is to be here to hear
19:57 Champion. This isn't about me, Kelly.
20:00 It's not about me. It's about the dog.
20:03 He misses you. Okay. Black bar. Yes.
20:06 Fine. Fine. Fine. Fine. Okay. Um Okay.
20:10 Okay. So, I'm going to make a new new
20:11 thing
20:14 here. A new
20:16 chat. So, here we are on chat. Jet tag.
20:19 Couple of things are
20:21 emerging. What? How do I get rid of that
20:24 thing? Okay. than
20:27 that.
20:29 Um, they've added long-term memory to
20:32 chat GPT, but it it is increasingly
20:35 appearing like it's only looking back at
20:37 your past year, not the entire chat
20:40 history, which if that's the case,
20:42 that's kind of shitty. Um, that will get
20:45 better, right? The fact that they can do
20:46 it for a year means that they can do it
20:48 longer. It's just, you know, they don't
20:50 want to spend the cycles or whatever it
20:52 is, um, or the hard drive space or
20:54 whatever it might be.
20:57 Source Camp's rebuilding her entire
20:59 website. Oh, Jim Ross, I don't know if
21:01 Jim Ross is in here tonight. He sent me
21:03 a thing on LinkedIn. He used
21:06 lovable.dev to create his own community
21:10 site. So, like, you know, on the AI
21:11 salon, we're using Mighty
21:13 Networks. Jim had lovable create a
21:17 standalone community site for the self-s
21:20 storage community. And it's better than
21:22 anything out there for any of the self-
21:25 storage community. And in showing it to
21:27 industry people, the first question they
21:29 asked was, "How do we can we give you
21:32 money to put our brand inside this
21:34 thing?" Absolutely remarkable. Using AI.
21:37 Yeah. Yeah. He used AI to to build a
21:41 community site. That's
21:47 lovable.dev. Oh, am I asking? So, wait.
21:51 Wow. Jim, I'd love to see that. Yeah,
21:52 exactly.
21:55 Um, Kelly, are you using are what are
21:58 you using to build your site? Are you
21:59 using
22:01 AI? Oh, Kelly is using it. What are you
22:04 using? Which tool are you
22:08 using? Oscar, you always have great
22:10 stuff here, Kyle. Hang in there. It'll
22:12 skyrocket. Yeah, I don't know about
22:15 that. It's been two and a half years. If
22:16 I was going to skyrocket, I think I
22:18 would have skyrocketed by now. I think
22:20 the uh the market the market for fat
22:25 saucy Gen Xers is is just not what it
22:29 used to
22:31 be. I posted my method. Keep
22:35 [Music]
22:37 reading. Oh, I gave chat my web address
22:41 and asked it to rework every page,
22:43 giving me an image. Oh, that's so cool.
22:46 That is so cool. Uh it's unbelievable.
22:49 Okay. So, let's So, now that we now that
22:52 Kelly has has goodmouthed OpenAI, let me
22:55 badmouth them a
22:56 bit. So, here we got our chat GPT,
23:02 right? About a month
23:05 ago, OpenAI launched. Okay.
23:11 About two months
23:14 ago, Sam Alman
23:17 said, "There's a new version of 40 and
23:20 then we're going to launch
23:22 03." Which confused the [ __ ] out of
23:25 people because they're like, "Well,
23:26 wait, isn't 03?" Like, I'm seeing posts
23:28 on Twitter right now like it how come 03
23:32 seems to be giving me better answers
23:34 than 40? Isn't it an older model? No,
23:38 it's not.
23:40 You didn't pay attention to where the O
23:42 was 40 03. No, they're completely
23:46 different things, right? So, so they've
23:48 already got the [ __ ] naming
23:49 nightmare, right? So, they launched that
23:51 and Sam says not to
23:54 worry. Sometime in the next month or
23:57 two, so maybe by
23:58 May, we're going to launch GPT5. And
24:02 GPT5 is going to take all this naming
24:05 nonsense and consolidate it into a
24:08 single model where you just talk to it
24:10 and the model will figure out what to
24:11 do. Okay, awesome. About a month ago,
24:16 just before
24:18 May, little Sammy boy came out and said,
24:21 "Hey, great
24:25 news. Here's here's model
24:29 4.1." And is it 4.10? Wait, hang
24:35 on. Wait, why do I not have it now? Did
24:39 they already take it
24:45 away? Wait, do I need to reload? Maybe I
24:48 didn't
24:51 reload. Okay. Yeah, it's in there now.
24:54 So, it's Okay, so 41 is not 410. Okay.
24:59 Okay. See, I like like I said, we we're
25:02 we're just at a [ __ ] disaster now.
25:04 So, about a month ago, he says, "Good
25:06 news. We're coming out with 4.1." Oh, by
25:09 the
25:13 way, I know we said GPT5 was launching,
25:17 but it's not. It turns out consolidating
25:19 all these models into some sort of model
25:21 picker kind of thing is harder than we
25:23 thought it would be. So, that's going to
25:25 be kind of a couple of months from now.
25:27 So I'm thinking like June, July, August,
25:30 some like GPT5 is sometime midsummer or
25:34 later. And he said, "So here's two new
25:37 models, GPT 4.1 and 4.1 high or mini 4.1
25:43 and 4.1 high. They were only available
25:46 in the API which I was like I was cool
25:48 with that because it's like great
25:50 there's some new cheaper
25:52 faster 41 model that the developers
25:56 have. Great. One of the cool things
25:59 about 4.1 is it had a million token
26:01 context window which is that's good. If
26:04 you don't know what context windows are,
26:06 it's basically like your chat's
26:08 short-term memory. All right, Jim
26:10 LinkedIn comment. Jim Ross, I'm happy to
26:12 showcase the site I made. Yes, I would
26:14 love for you to do that. Uh, you can do
26:16 it on office hours on Friday. You could
26:18 even do it tonight if you want to, Jim.
26:20 If you uh reach out to Brandon or
26:23 Brandon can reach out to you, I'm I'm
26:25 happy to have you come up here. But if
26:27 you're sitting there in your jammies and
26:30 you know, you've got like, I don't know,
26:32 cereal in your hair and don't want to
26:34 come on screen, that's that's cool, too.
26:36 Um, but I would love to see it uh
26:38 tonight because it's just I think it's
26:40 really exciting and I I want to hear
26:41 like what people said about it. So, if
26:43 you're if you're willing to do it
26:44 tonight and Friday, that'd be
26:46 awesome. Um, so anyway,
26:51 today they put 4.1 into chat GPT. Now,
26:56 they hid it under other
26:58 models. So, look at this [ __ ]
27:01 disaster we've got now. We've got 40.
27:05 Okay, the O means omnimodal, right? This
27:09 remember remember when I did this turned
27:11 into a whole [ __ ] anger rage post,
27:14 right? Explaining these things that
27:16 people seem to enjoy that post. 40 is
27:19 the the omnimodal multimodal model that
27:22 that does pitches, right? 4.5 is this
27:27 sort of like
27:29 appendage
27:31 that is not multimodal. It's just kind
27:34 of better at writing and kind of has
27:36 personality. And David Shapiro thinks
27:39 that this is going to be the the
27:40 coordinating model when when
27:43 GPT5 starts navigating between models.
27:47 He thinks that that's what this is going
27:49 to be used for. But we don't but no one
27:51 knows that. And so people don't know.
27:54 They think 4.5 is just 4.0 but better.
27:56 But it you can't make images with this
27:59 one. Then you have 03. But because it's
28:02 lower than 40, people assume it's older
28:04 and worse. This is the big Mac Daddy
28:06 [ __ ] reasoning model, right? Then
28:09 you've got 04 mini and 04 mini high.
28:12 These are the newer reasoning models
28:14 that are in no way related to 40.
28:17 They're they're standalone mini
28:20 reasoning models. And now we've got 01
28:25 Pro and we've got GPT4.1 and GPT4.1
28:30 Mini.
28:34 What the [ __ ] are they
28:36 doing? Like [ __ ] retire
28:39 something. This is ridiculous. Like this
28:42 is at this point like this is beyond
28:49 satire. And I I mean I get it. You've
28:52 got capabilities that people want to
28:53 have access to the capabilities, but
28:55 like I pay attention to this [ __ ] every
28:58 [ __ ] day. And of all these models, I
29:02 definitely know what 40 does. I don't
29:05 know what 4.5 does. I don't know what
29:06 it's good at. I don't know what it's not
29:08 good at. I've used it a couple of times.
29:10 It hasn't blown me away. Other people
29:12 say it's completely blown them away and
29:14 that's their daily driver, but they're
29:15 obviously not making
29:19 images. And then 03 is really good, but
29:22 I don't like if I'm going to use
29:24 something that's reasoning like I tend
29:26 to use 03 because that's the bigger
29:28 beefier model.
29:30 But I only know that because I pay
29:32 attention to this [ __ ] I don't know
29:34 when to use I like I don't use 04 mini
29:36 at all. I use 04 mini high or 03. I
29:40 don't use 01 pro anymore. And I I don't
29:43 think I'm ever going to [ __ ] use 4.1.
29:45 I'm not even going to test
29:46 it. Like who gives a
29:51 [ __ ] The only people paying this close
29:54 attention to the nuances of the models
29:57 are developers.
30:04 So, I I don't know if Pate's in here,
30:08 but I tend to be a bit of a suck up to
30:11 Open AI because I think they're just the
30:13 best at everything right
30:15 now. But this is like we they've
30:18 officially jumped the shark. This is
30:20 just chat GBT is becoming unusable.
30:25 Like I'm paying 200 bucks a month to
30:27 have operator. I've never [ __ ] used
30:31 it. It's a whole separate site. It's a
30:34 whole separate thing. Apparently I need
30:36 to verify this device. [ __ ] it.
30:40 No. Not verify. Not verifying
30:43 [Laughter]
30:46 [ __ ] This is unwieldy Wednesday.
30:49 Welcome to the AI learning lab. My name
30:51 is Kyle Shannon. I'm not drunk.
30:57 [Laughter]
30:59 Oh my
31:00 god. All
31:02 right. Windown
31:06 Wednesdays. So anyway, have you tried
31:09 chat GPT? It's really
31:12 good. It's I I mean like what Kelly did
31:15 like what it can do is remarkable, but
31:18 what they're doing to it is inexcusable.
31:23 Like no one gives a
31:25 [ __ ] They got to have stats on
31:28 this. They've got to have stats on this.
31:31 Most users don't change their defaults.
31:33 So that means most users are using
31:36 GPT40. And then some advanced users are
31:38 probably using 03 and probably using 04
31:41 mini high because they run out of
31:43 credits on 03.
31:46 I would be shocked if the if the other
31:50 models had, you know, statistically any
31:52 usage at all. [ __ ] bury them. Just
31:55 hide
31:57 them. We're not
32:00 developers. Developers are going to go
32:02 build [ __ ] on the playground and they
32:05 need all the models and they understand
32:07 the differences and what they cost and
32:09 the speed and what they're good at and
32:11 what they're not good
32:13 at. Ah, Tik Tok question. When should I
32:17 use a reasoning model? That's a great
32:19 question.
32:25 Um, it's a really good
32:30 question. Some people are using 03 as
32:33 their daily driver, right? They're just
32:36 using
32:37 03 instead of using 40.
32:43 Um, but I would say in general it's when
32:46 you've
32:49 got when you've got a question or a
32:52 problem or a research project that is
32:55 that that definitely has multiple steps
32:58 in it and could benefit from those
33:02 multiple
33:04 steps. Okay. When you use
33:09 40 and you
33:12 say, "I'm
33:15 uh let's see. Uh
33:18 [ __ ]
33:20 I'm thinking
33:24 about a marketing plan
33:28 for storyvine.com."
33:36 And I just hit
33:37 that. It says,"Great. Let's build one.
33:40 What's your main goal right now?" So it
33:41 asks you a few
33:43 questions. And then it says, you know,
33:46 "Here's your marketing plan." So 40 just
33:49 kind of you give it a prompt and it just
33:52 kind of vomits out an answer.
33:56 Well, to really do a marketing plan,
33:59 you'd need to do things like, I don't
34:01 know, research competitors and think
34:05 about the market you're in and all that
34:06 sort of stuff,
34:08 right? So, so if I go to 03 and I say,
34:12 let me restart this.
34:15 Um,
34:17 I'm thinking
34:21 about a
34:23 marketing plan for, and this is a really
34:26 shitty prompt, by the way,
34:28 storyvine.com. Oops, not storyvine.co.
34:32 Stop. storyvine.com.
34:39 And so it says
34:42 thinking and so it says the user wants
34:45 marketing ideas for story
34:47 mine. A general location isn't crucial.
34:50 I could create a marketing strategy
34:52 using my internal knowledge but I should
34:54 do a web search first. The user prefers.
34:57 So if I expand
34:59 this these these things that it's
35:03 talking to right here are not my answer.
35:07 This is its internal logic, right? So
35:11 the user wants this. I could create a
35:13 marketing strategy, but I should
35:14 probably do some
35:15 research. The user prefers um original
35:19 and innovative ideas. So I'm thinking of
35:21 modern creative marketing approach. So
35:23 this is using its knowledge of me. So so
35:27 it's thinking internally. Then it
35:29 searched the web. So, it searched uh
35:32 storyvine video storytelling platform
35:34 2025 press release guided video platform
35:37 customers healthcare looking at a bunch
35:39 of other stuff outbrain seven more
35:41 things the user values
35:44 originality I'm thinking interactive
35:47 vine sprints UGC AI video
35:50 challenge marketing play for
35:52 storyvine.com so now it gives me the
35:54 answer so so I would say that when you
35:57 use a reasoning engine and certainly if
35:59 you If you have anything involving math
36:01 or
36:02 analysis, definitely use 03.
36:06 Um, or 04 mini high, that's really good
36:09 at analysis,
36:11 too. This prompt that I gave it didn't
36:14 it it didn't really dig into its chops.
36:16 But one of the things that 03 can do is
36:18 it can create its own tools. So, let's
36:21 say let's say that not only did I want
36:23 um a story of my marketing plan, but I
36:26 wanted I don't know the revenue results
36:29 from a bunch of competitors or I wanted
36:31 I wanted uh what are the pharma brands
36:34 that are launching in 2025 and 2026 and
36:38 what's the expected marketing spend on
36:40 those and then I wanted to do some
36:41 analysis across all those companies. 03
36:45 can write its own tools in Python and
36:48 execute them as part of the thinking
36:51 process. So you use a reasoning engine
36:54 whenever whenever you want. Basically,
36:56 if you've got something where you want a
36:58 much more considered, thoughtful
37:02 um answer or it's a it's a more
37:05 complicated project, that's when to use
37:07 it. Jim Ross is busy tonight. Dolly's
37:10 daughter's volleyball game. But at least
37:11 he's logged in here. This is good.
37:13 You're in his
37:16 ear. Okay, Jim. Um,
37:19 awesome. So, so anyway, does that answer
37:21 your question, Jim
37:24 Buckley? Got it. Beautiful. All
37:28 [Music]
37:30 right, let's
37:32 see. I think they have all the models so
37:35 they can throw them at benchmarks and
37:36 show up on the list more often.
37:42 Yeah, but those
37:44 those the models in chat GPT itself are
37:47 not what shows up in those benchmark
37:49 models. That's [ __ ] coming from their
37:51 API, right? They they they launch those
37:55 things like the the benchmark the
37:59 benchmarkers are are putting models in
38:01 there from the from the systems. It's
38:04 not coming from the commercial chat GBT.
38:06 I mean, that's that's what I'm talking
38:08 about is the commercial version of chat
38:13 GPT. And even if you're right, you might
38:15 be right. Like, it's possible that those
38:17 benchmarking things are using core chat
38:20 GPT, but I don't think they are. I think
38:22 they're using the
38:25 API. But you've got they they now have
38:28 700 million users on this platform. 700
38:32 million users.
38:35 How many of those are
38:37 engineers? Like like
38:40 globally programmers and engineers is
38:42 about
38:44 1%. So let's let's make it really
38:48 generous and say
38:50 that
38:52 20% of those 700 million
38:56 people
38:58 are engineers on chat GPT.
39:03 80% of the people don't give a [ __ ]
39:05 about all these models. And not only
39:07 don't give a [ __ ] it's confusing the
39:10 crap out of them. It's like it's like
39:12 distorting the brand. They're like, I
39:15 don't know what to do here. Like what
39:16 what why is all this? Who are these
39:20 [ __ ] techno bro [ __ ] NFT bro San
39:25 Francisco
39:27 [ __ ] And they're
39:30 right. Ah. All right. All right. We need
39:34 a dancing robot. We do need a dancing
39:35 robot. Let's go look at the dancing
39:38 robot. And here's the other thing. I
39:41 read I read the Mr. Strawberry thing the
39:44 other night and then I I realized it was
39:46 probably too hypy, but there's some
39:50 stuff coming down the
39:53 pipe where I think we're just going to
39:57 see in like
40:01 incredibly more powerful tools in the
40:04 next two
40:06 months. Wait, YouTube question. This
40:09 might be a dumb question. Probably
40:11 not. But are the models connected in any
40:17 way? Kind of.
40:20 The
40:22 So, it used to be worse. This is a
40:25 little bit better. Okay. So, I just did
40:27 this. I just did this.
40:31 um this 03 reasoning thing for my
40:33 marketing plan, right? Well, I can now
40:37 switch to
40:41 GPT4.5. And it's supposed to be better
40:44 at at
40:47 writing. Oh, that's interesting. It
40:48 says, "Do do you like this personality?"
40:50 I've never seen that prompt before. Or
40:52 that thumbs up, thumbs down. So, let's
40:54 see if we like the personality.
40:59 Short form US user generated video is
41:01 devouring B2B attention. Nearly every
41:03 budget budgets for video. 84% of buyers
41:06 say a brand's video convince them to
41:08 purchase and social platforms have
41:10 become the new center of gravity. But
41:13 Pure UGC is a compliance nightmare.
41:15 That's what we
41:17 say in regulated industries. Storyvine
41:20 structured guided workflow is the
41:23 missing middle. Authentic like UGC yet
41:27 brand safe and repeatable. Yep. Which it
41:30 got from our website by the way.
41:32 Structured authenticity at scale.
41:34 Capture real voices. Control the story.
41:36 Publish in minutes. Yeah, I like this. I
41:39 I think that's good. So, we'll give that
41:40 we'll give that a thumbs up. But anyway.
41:42 Okay. So, now I've switched to I can't
41:45 give it a thumbs up. Oh my god.
41:47 Interface people. Thanks for your
41:49 feedback. Okay. Um, now I'm in 4.5. I
41:54 could say, "Give me
41:57 a list
42:00 of
42:02 20
42:04 LinkedIn article ideas,
42:09 article ideas that are
42:14 um pharma
42:19 relevant, very
42:22 catchy and
42:25 compelling. So, the fact that I can
42:28 switch models in the same conversation
42:31 does kind of make them related. And and
42:34 this is kind of what what Altman
42:36 promised was when we get
42:41 GPT5, it's going to be doing this
42:43 navigating for us, right? He
42:46 acknowledged three months ago that the
42:48 naming convention is a shitow and
42:50 they're aware of
42:51 it. But because technology is technology
42:55 and you can't launch stuff as fast as
42:57 you think you can, they're perpetuating
42:59 it and making it worse. And at this
43:00 point, I think they've jumped the
43:02 shark. Pharma's video marketing crisis.
43:05 Why authenticity isn't optional. The end
43:07 of boring compliance videos. Structured
43:09 authenticity is here. How pharma can
43:11 create 10x more content with 80 80% less
43:14 risk.
43:15 MLR bottleneck. This is great. Um I'm
43:19 going to say
43:21 um write me the
43:24 first three articles. Oh, we'll we'll do
43:28 something interesting here. Okay. See
43:30 the link below. In defense of Open
43:34 AI from who?
43:36 [Laughter]
43:54 Oh, I see you did you did a you had chat
43:57 GBT defend itself. No, [ __ ]
44:03 that. I'm not giving that any oxygen.
44:06 No. No.
44:09 Nice, nice try, producer Brandon. I like
44:11 it. I like the moxy, but no. Okay.
44:15 Article one. Far pharma's video
44:17 marketing crisis. Why authenticity isn't
44:19 optional. Pharma marketing is at a
44:20 crossroads. The era of polished, overly
44:23 scripted commercials is losing relevance
44:25 as audiences increasingly tune out
44:28 content that feels artificial. Today's
44:30 health care consumers, patient, healthc
44:32 care providers, and advocates demand
44:34 authenticity. But in a high highly
44:36 regulated industry like pharma, how can
44:38 marketers deliver? The solution lies in
44:40 structured authenticity. That's pretty
44:42 good. It's not bad. Everyone knows
44:44 compliance videos usually dry,
44:46 monotonous, and instantly forgettable.
44:48 They tick regulatory boxes, but often
44:50 fail to engage or educate the audience
44:52 effectively. For pharma companies, these
44:55 videos represent the mist. Okay, that's
44:56 great. Um,
45:02 date
45:05 date. So now what I could do is I could
45:08 flip
45:09 over to
45:12 GPT40 and now say
45:19 um make me
45:23 a let's
45:27 see
45:29 loose pencil sketch.
45:40 image
45:44 for article
45:48 one Tik Tok pin
45:51 um Ray Edwards the key is authenticity
45:54 once you take away that once you can
45:56 fake that anything is possible yeah and
45:58 I think we're well well so there's so I
46:01 think You're so Ry, I think you're on to
46:03 something, but the authenticity things I
46:07 I have a whole company based on it. So,
46:09 I've been thinking about this for a long
46:11 time. I think you can have nonhuman
46:16 avatars authentically share
46:20 stories. And if those stories are
46:23 relevant and they're authentically
46:26 expressed from the point of view of the
46:28 avatar, from the
46:30 persona, then there's
46:33 a there's
46:35 a like an authentic
46:38 consistency from the intention of the
46:41 persona to tell a particular kind of
46:44 story. And if they tell that well from
46:46 that point of view, there's an integrity
46:49 in that.
46:50 It doesn't mean it's
46:53 human, but there's an authenticity to
46:56 that. So, I think there is a way to do
46:57 authentic storytelling without
47:01 um without a human. But you're on to
47:05 something big
47:08 here. Our social networks are already
47:11 polluted with bots. They're polluted
47:14 with bots. And so one of the reasons
47:17 that I've been talking up for I don't
47:20 know two and a half years now, a year
47:21 and a half, year and a half, um, world
47:25 ID is that it's the first scalable
47:31 um, proof of
47:32 personhood that doesn't require you to
47:35 give up your identity, but but it proves
47:37 who you are. So I think we're going to
47:39 start to see applications and social
47:42 networks that require proof of
47:47 personhood even within that like so so
47:50 what we'll have is we'll have
47:53 um social networks where you know that
47:58 when I post something it's me the human
48:00 posting it but me the human I can post
48:03 inauthentic [ __ ] on my channel even
48:06 though you know it's coming from person,
48:08 I can generate shitty, inauthentic
48:12 content and share that as well. So, what
48:14 you're talking about is absolutely
48:16 critical. Um, but I think it's it's very
48:19 very nuanced. Um, Ray also gave me some
48:22 roses. Great. Thank you very much.
48:24 Comment from Jeff. I copy pasted whole
48:26 sections of chat from a custom GPT that
48:29 I liked into 40 and it became that
48:32 custom GPT's personality only with chat
48:36 memory. Very cool. That's a great use
48:39 case. Oh, look at this. That's pretty
48:42 [Laughter]
48:48 cute. Pharma's video marketing
48:55 crisis. Why authenticity isn't
48:59 optional. I've been working with 11 Labs
49:01 on character AI to recreate my son's
49:04 persona. Oh, that's great. That's
49:06 awesome. Listen, I think that's going to
49:08 be a huge use case. You know, lost
49:10 people, forgotten people,
49:14 um, people with Alzheimer's. I lost my
49:17 dad to Alzheimer's. So, you know,
49:19 imagine him being able to interact with
49:22 people that are familiar to him that are
49:24 there
49:25 247, you know, have infinite patience,
49:28 will, you know, will tell the stories
49:30 over and over and over again, uh,
49:32 without growing weary. I you know I I
49:34 think that I think this Avatar digital
49:36 twin thing is is is is just going to be
49:40 an incredibly powerful thing. There's
49:42 there's so much talk. It's funny when I
49:44 was in Washington DC two weeks
49:46 ago, one of the staffers that we were
49:49 talking to was like well she was kind of
49:50 bragging, well we've not for nothing but
49:53 we've done legislation against deep
49:56 fakes. And I was like cool like great
49:59 understand that. I said, "But you know,
50:00 the the technology behind deep
50:03 fakes can be very beneficial." And she
50:05 like got this look on her face like,
50:07 "Wait, where are you going, boy? What's
50:09 what's going on here?" And and I told
50:12 her some examples of, you know, some of
50:13 the examples from the salon where people
50:15 are using avatars and digital twins in
50:18 really powerful ways as like an
50:20 augmentation of themselves or an
50:22 extension of themselves and they're
50:25 choosing to do it. They haven't been
50:27 deep faked. they're using their own
50:29 digital twin to do marketing or do
50:32 communications in some way and she
50:34 hadn't considered
50:36 it. And so it's like we're we're just in
50:38 this
50:40 weird we're in this weird lite
50:44 phase where because the technology is
50:47 powerful and profound and
50:51 accelerating, people are like doubly
50:54 afraid of it. And so they're pushing
50:55 back on it and they're, you know,
50:57 they're legislating from a place of
50:59 fear.
51:01 Um, and man, does it do it a
51:04 disservice. So anyway, Kak, did you know
51:08 custom GPTs that have less than 8,000
51:10 characters generate quicker? Oh,
51:12 interesting. I did not know that. Makes
51:15 sense. Did you know that custom GPTs now
51:18 use the new image generation model,
51:20 which is
51:21 nice? Uh,
51:27 [Applause]
51:29 World ID will still force the
51:31 authorities to give up your identity for
51:33 law enforcement. Well, so
51:36 okay. So what you're talking about is
51:39 really nuanced.
51:41 It's
51:44 okay. World ID does not know who you
51:47 are.
51:50 You can have a world ID without it being
51:53 attached to your name. Now, here's the
51:57 here's the nuance and the subtlety. They
51:59 have all these things called micro apps
52:01 or mini apps or something like that.
52:03 It's basically like the app store for
52:05 World ID and there's all these little
52:08 apps. Some of those apps you don't put
52:10 your name into. Like there's a a
52:12 surveying app and and you can take
52:14 surveys and I think they pay you
52:16 WorldCoin or something like that. You
52:19 don't have to put your name in that. But
52:21 there are going to be other apps like
52:23 say if OpenAI comes out with the social
52:27 network which they've talked about they
52:29 are they are working on well in your
52:32 social network you're probably going to
52:33 put your name and maybe your address and
52:35 then all your pictures and all that sort
52:37 of [ __ ] and that will certainly be able
52:40 to be associated with your world ID.
52:42 Don't collapse don't collapse the two.
52:46 Um because world ID itself is
52:49 independent from who you are. It's just
52:52 it it is literally tied to is the
52:54 fingerprint of your retina unique
52:57 against the fingerprints of the retinas
52:59 in the database. They're not keeping the
53:01 biometric data. They're just taking the
53:03 hash and comparing it to the other
53:05 fingerprints basically. Um, but
53:10 practically what you're saying is
53:12 true because people will use free apps,
53:17 right? If if OpenAI gives us a free
53:20 alternative to X that doesn't suck and
53:23 just has people in it, no bots, people
53:25 are going to use it and they're going to
53:27 voluntarily give up all their um all
53:30 their person their personal data. And
53:32 then you've you've you've made you've
53:35 got a good point.
53:38 Dancing robot then deep mind. That's
53:40 good. Retinas have fingers. They do. I
53:43 have you. If Have you never looked in an
53:44 eye? You look You look through the
53:46 pupil. They've got this special machine
53:48 at the at the glasses place. If I've
53:50 seen videos of it. You look in there and
53:52 on the back of the retinas there's like
53:54 little fingers going, "Hey, how's it
53:56 going?" It's amazing. It's
54:00 amazing. All right. I understand.
54:03 Thanks, Kyle. Fantastic. Fantastic. Bob,
54:06 tell everybody he's
54:07 won. I'm working on mine as a project to
54:10 leave my kids so they can talk to me
54:12 when I'm ascended. Yeah, it's beautiful,
54:14 Kimberly. Love
54:16 it. Love
54:19 it. I mean, like the psychology of being
54:24 able to talk to you after you've
54:26 passed. Um, and and and also there's
54:29 there's um Neurolink patient number
54:32 three. I was actually just watching a
54:33 YouTube video on this before I came in
54:36 here. Neural link p So, if you don't
54:38 know what Neurolink is, it's it's the
54:40 Elon Musk uh brain implant that's got
54:42 like 60 electrodes that goes into your
54:45 brain. And um number three, patient
54:50 number three is a guy that's got ALS, um
54:53 Lou Gerri's disease, and he can't use
54:55 his voice.
54:57 Um how do I get out of this thing? Let's
55:00 see. What did I do? Oh, there we go. Um,
55:03 he can't use his voice. He's got L Lou
55:05 Garri's disease. And so they took audio
55:08 recordings of him when he could talk,
55:12 right? And they made a voice model for
55:15 him. And so now they've got this chat
55:17 app where people can talk to him. And
55:20 the chat app translates what the people
55:23 are saying, uses AI to come up with
55:27 suggested responses, and then he uses
55:29 Neurolink to navigate and choose the
55:31 response, and he responds to them in his
55:35 voice, which that's just [ __ ] insane,
55:38 right? Really cool. Um, [ __ ] insane.
55:41 But I like the the psychology of being
55:44 able to talk talk with loved ones after
55:46 they've passed or or even be able to go
55:48 back and talk to like, you know, great
55:52 grandpa Jim, like maybe you got some
55:54 recording from the 20s of great grandpa
55:57 Jim and and now you're talking to him
56:00 based on all of his writings or
56:01 something like that. Like the psychology
56:04 of that, like what's is is that going to
56:06 aid in the grieving process? Is it going
56:08 to delay the grieving process? Does it
56:10 transform the grieving
56:12 process? I I we're just it it's I I just
56:16 The thing about AI is like everywhere
56:21 you point it, it's going to completely
56:24 transform the way we do
56:27 things. Not just take jobs, right?
56:30 Again, like I just I feel like
56:32 everything is so [ __ ] reductionist at
56:34 this point. It's just like, it's going
56:36 to take jobs. Yeah, it's going to create
56:37 jobs, too, jerk. But what is what it is
56:41 going to absolutely do is transform
56:45 everything every job every interaction.
56:48 the definition of psychology and psych
56:51 and psychologists and therapists and and
56:54 like what is a human's role when when
56:57 you've got something that is infinitely
57:01 compassionate,
57:04 non-judgmental, and knows every
57:07 psychological approach in the history of
57:12 humankind and and adapts to everything
57:15 you've ever talked about and never
57:16 forgets
57:19 anything what's the role of the human I
57:22 think there is
57:23 one but I think a lot of
57:26 the skills required for a therapist to
57:29 have a lot of that knowledge become less
57:32 important but there's something that
57:34 they have that's going to be more
57:36 important right and like what's that
57:38 balance look like I don't know I just
57:40 know it's not going to be the
57:42 same Joe Decult so what you're saying is
57:45 that even when we
57:47 die, we're still going to be
57:50 forced to talk with our annoying family.
57:55 Exactly. Exactly. You know what I You
57:57 know what's funny is like even within a
58:00 single family dynamic, right? You know,
58:02 mom dies or dad dies and and now brother
58:05 and sister get to talk to dad. All
58:08 right. Maybe sister was daddy's little
58:12 girl and she loves dad. She's like,
58:13 "Awesome. I can't wait to talk to dad."
58:15 And you know, the brother was like, "He
58:17 never loved me, [ __ ] asshole." He's
58:19 like, "I don't want to talk to that
58:20 [ __ ] ever again." And then all he
58:23 hears is his sister talking to
58:25 dad now. Now, maybe he can actually
58:29 reconcile with dad because dad's now
58:31 this compassionate entity. But I Yeah,
58:34 just [ __ ] everything everything's
58:36 going to transform. Anyway, all right.
58:39 I'm constantly taking pictures and video
58:41 of my parents. Absolutely. You should
58:42 absolutely do that. They no longer even
58:45 argue because they know. That's really
58:47 cool. Really, really cool. People can
58:50 literally take their argument AIs in the
58:53 other room while they finish
58:54 Thanksgiving dinner. That's
58:58 awesome. Uh the role the role of the
59:00 human is to maintain the AI. Exactly.
59:03 Well, no. You know, I So, seriously,
59:07 like if I think about something like
59:08 therapy and and let's take it out of
59:10 therapy because I mental health gets
59:11 weird when you start getting into
59:14 um mental health just gets gets really
59:16 sticky really quick. Let's take physical
59:21 therapy. So, you you break your knee or
59:24 you whatever you do. You you tear a
59:26 meniscus and you go in for surgery and
59:28 there's a six week recovery and then a
59:31 year of rehab and you go have your
59:34 surgery and you come out you're groggy
59:36 with surgery and the doctor goes hey how
59:38 you doing Kyle you know here's what we
59:40 did the meniscus and there was a fold
59:42 and it was a parrot beak tear and
59:44 everything was successful you should be
59:46 on your feet and and I'm going to write
59:48 you a prescription for a physical
59:50 therapist right and then you don't
59:52 remember what the [ __ ] he said and then
59:54 you meet with your physical therapist.
59:56 And again, your physical therapist knows
59:58 all this stuff. Oh, you have this kind
59:59 of meniscus tear and we're going to
1:00:01 you're going to move your knee like this
1:00:02 and you're going to do that and I'll see
1:00:04 you in a week. Okay. And then you
1:00:07 [ __ ] hobble out of there and you hurt
1:00:08 and you're on [ __ ] drugs. You're
1:00:11 like, I don't what what did she say?
1:00:13 What? Oh
1:00:15 god. Imagine her being able to say,
1:00:17 "Hey, here's the virtual me. It knows
1:00:21 your case. It knows exactly um what
1:00:25 happened. It knows what stage you're at.
1:00:29 Um anything that I just told you, it
1:00:32 knows and you can talk to it 24/7 and
1:00:35 it'll it'll sound like me, right? Like
1:00:37 that's, you know, let's go two years in
1:00:39 the future. So, I I get I've got virtual
1:00:42 therapist on my phone and I can talk to
1:00:45 my virtual physical
1:00:47 therapist and at some point I'm talking
1:00:50 to the virtual physical therapist and
1:00:52 the virtual physical therapist says,
1:00:54 "Uh, what you're describing sounds a lot
1:00:56 like an infection. Get your ass to the
1:00:58 hospital." Right? Like like it will be
1:01:01 smart enough to know, "Okay, we need we
1:01:04 need to not be talking to virtual
1:01:06 physical therapy lady. we either need to
1:01:08 be going back into physical therapy for
1:01:10 an emergency knee resetting or we need
1:01:13 to be going to the hospital because we
1:01:14 have an infection or something like
1:01:16 that. So I I can see there being
1:01:18 handoffs where these things get smart
1:01:20 enough to
1:01:22 recognize where that boundary is. And so
1:01:26 that'll be that that's pretty cool. All
1:01:28 right, Tik Tok
1:01:30 comment. I also have a virtual mental
1:01:33 therapist that sounds just like Sam Sam
1:01:35 Elliott soothing. That's very nice. I
1:01:37 made my physical therapist an app with
1:01:39 lovable. Oh, that's so cool. Silver Fox.
1:01:42 That's really good. He gives folksy
1:01:44 advice. That's
1:01:47 awesome. Yeah, Sam Elliot has that
1:01:49 awesome voice, doesn't he? Yeah. How's
1:01:51 it going? I can't even I can't even get
1:01:54 down that low. How's it going, Kimberly?
1:01:56 Yeah, I can't I've been talking too much
1:01:59 up here, so my my lower vocal folds
1:02:02 right now are not cooperating.
1:02:07 Um, okay. What was I going to do,
1:02:09 Brandon,
1:02:14 next? What's
1:02:18 skeleton? Oh, dancing
1:02:21 robot. Oh, you know what I was going to
1:02:24 do? I think I'm gonna I'm gonna post
1:02:30 this. Just going to save this as a All
1:02:47 right. It's PNG. All right.
1:02:49 [Music]
1:02:57 Uh, all right. We're going to grab
1:02:58 article number
1:03:01 one. Uh
1:03:10 um by the way, so what I'm doing right
1:03:12 here, if you is What the [ __ ] is going
1:03:16 on with LinkedIn? Oh, Tik
1:03:22 Tok.
1:03:26 Um, start a
1:03:32 post. I think I should should it be an
1:03:35 article? Maybe it should be an article.
1:03:37 Write an article. And it should come
1:03:39 from Story
1:03:43 Vine. Oh, I got to make this 1920 x
1:03:46 1080. Okay, let's go back to chat.
1:03:50 JPEG. So, we'll close that out.
1:03:54 Scroll on down. Say, I need it
1:03:59 need it to be a wide image for
1:04:11 LinkedIn. It's what's for dinner. Drink
1:04:14 the tequila and it will lower your
1:04:16 voice.
1:04:22 Oh, don't read it or anything before I
1:04:23 post it. I read that
1:04:26 article, but let's go. Let's go look at
1:04:28 it here. We'll put it
1:04:38 here. Wait, what just
1:04:43 happened? I'm
1:04:45 confused. Oh, I see what I did.
1:04:48 So, this is
1:04:50 dead. There's that. I'll get rid of
1:04:57 that. That's not really long enough, is
1:05:05 it? Let's go back to chat GPT. Okay, so
1:05:09 here's a good example. All right, so we
1:05:11 have we have this
1:05:13 article. Let's go back here. We'll go to
1:05:18 Let's go to
1:05:25 03. Is this all the
1:05:27 articles? Yeah, it
1:05:30 is. So, I'm going to
1:05:35 go let's see
1:05:39 for article one. It's a little thin.
1:05:46 do some
1:05:48 research to back it
1:05:51 up and beef it up. Now, I could use
1:05:55 within Canvas. I'll show you that in a
1:05:57 second. Within Canvas, you can just tell
1:05:59 it to
1:06:04 um What is it? What is it doing? Is it
1:06:08 thinking right
1:06:12 now? There we go.
1:06:16 Um, within Canvas, you can tell it to
1:06:18 make things longer, longer or shorter.
1:06:21 Um, but I want to use 03 to actually
1:06:24 think about article one. Researching to
1:06:26 enhance the
1:06:28 article. The user wants to beef up the
1:06:30 article pharma's video marketing crisis.
1:06:33 This would involve gathering
1:06:36 data. All right. Once I have the info,
1:06:38 I'll replace article one. I'll gather
1:06:41 the sources and update the article. It's
1:06:43 done some
1:06:46 research. I [ __ ] love the
1:06:51 future. This is if you've got ADD, what
1:06:54 it's doing right now is the whole reason
1:06:57 chat GPT is your is your friend. It's
1:07:00 actually doing the research for me. Yes.
1:07:02 Now, to Brandon's point, I should
1:07:04 probably actually read it. There was a
1:07:06 really funny TikTok. I think Brandon
1:07:08 sent it to me today of a high school
1:07:10 yearbook. Someone paid $65 for their
1:07:13 high school yearbook and one of the
1:07:14 paragraphs they left the chat GPT, you
1:07:17 know, if you'd like me to modify this in
1:07:19 any way, let me know. They left it in
1:07:21 the
1:07:22 [Laughter]
1:07:24 article. It wasn't you. So, yeah. Oh, it
1:07:27 was
1:07:28 uh it was it was it was in a sort of
1:07:30 group chat within Tik Tok. It was pretty
1:07:33 funny.
1:07:39 Gotta take the dog for the walk. See
1:07:41 y'all later. Take care, Kimberly. Thanks
1:07:44 for sharing. He is his own theme
1:07:52 music. I am my own theme music. Okay.
1:07:55 Article one. Key takeaway. 2025 isn't
1:07:58 the year to consider authentic video.
1:08:00 It's the year you get left behind if you
1:08:02 don't. The data behind the crisis.
1:08:05 Consumers crave the real thing. That's
1:08:07 pretty good. What's at
1:08:13 stake? All right, [ __ ] it. Let's just
1:08:16 grab it
1:08:18 all. To Brandon's point, you're not just
1:08:20 going to stick that in and publish it,
1:08:23 are you?
1:08:24 Yep. I I'll I'll read it
1:08:28 quickly. Boom. All right.
1:08:36 And then let's go look at our picture.
1:08:37 Do we did we get our picture? Oh, did
1:08:40 Oh, maybe I stopped it. Network er error
1:08:44 occurred. Let's see. Let's see. Thought
1:08:47 for Uh oh.
1:08:52 Retry. Did I There's my picture. Okay.
1:08:56 Um I
1:08:59 need Something went wrong. Please try
1:09:01 reloading the conversation.
1:09:04 Oh, reload the
1:09:08 conversation. Okay, I need the image in
1:09:13 wide
1:09:15 format
1:09:17 and use
1:09:21 colored
1:09:25 pencils in falto blue and orange.
1:09:34 story
1:09:37 vines
1:09:40 colors.
1:09:42 Um, also don't make the lady so
1:09:48 [Laughter]
1:09:51 sad. Uh, should we talk about how it
1:09:54 hallucinates references? I don't know.
1:09:56 Let's go look at the references.
1:10:06 State of video marketing
1:10:10 wisel video marketing statistics. That's
1:10:18 right. Power of patient testimonials
1:10:21 LinkedIn article.
1:10:28 These look pretty
1:10:32 good. MLR guide, Fizer case, viz,
1:10:39 even these look
1:10:42 good. So, what Brandon's bringing up
1:10:46 is chat GPT. Okay, again, this
1:10:50 this there was a question asked earlier.
1:10:52 When do you use a reasoning model?
1:10:56 when chat GPT just vomits out an answer.
1:10:59 So when you're using 40 and it just you
1:11:00 give it a prompt and it it gives you an
1:11:02 answer. Um that's not as accurate as if
1:11:07 it thinks about what it's doing. So when
1:11:10 I asked it to go do research and it went
1:11:12 through those steps of I should do this
1:11:14 and I should do that and I should go
1:11:16 research things and take what I learn
1:11:17 and rewrite the article. Because it's
1:11:19 doing all those steps, it's actually
1:11:21 checking these things.
1:11:23 So yeah, you should still check to make
1:11:27 sure that these things work, but
1:11:32 um I I think you can trust them a lot
1:11:34 more than you could
1:11:36 before.
1:11:37 [Music]
1:11:45 Um okay, that's that we're going to
1:11:47 pull.
1:11:53 The data behind the consumers crave the
1:11:56 real thing with
1:12:00 quotes. Healthcare buyers are no
1:12:02 different. Short form video is farmer's
1:12:05 blind spot. Tik Tok style clips now
1:12:07 command 33% of total marketing spend.
1:12:09 Yet most farmer brands still allocate
1:12:11 less than 10% of the budgets to it.
1:12:14 I'm just going to
1:12:16 say social
1:12:25 videos compliance bottlenecks are the
1:12:27 culprit. An Adobe study shows the
1:12:30 average MLR cycle is now at 21
1:12:35 days. Storyvine authenticity a way
1:12:38 out. Painpoint
1:12:45 What the [ __ ] is
1:12:52 this? Pain point traditional approach
1:12:56 structured video fix MLR churn 21day
1:13:01 average linear. This looks like
1:13:03 notes of its thinking. That's really
1:13:06 [ __ ] up. All right, so let's go. Let's
1:13:09 go. Let's go yell at it. Um. Oh, that's
1:13:12 nice. All right. Well, it's not
1:13:15 nice. It's fine. It's fine for this.
1:13:19 Okay. Here's a new wide. Okay, that's
1:13:22 good. But it it gave us orange and blue
1:13:25 falto. I'm going to just say um
1:13:29 the quality of the
1:13:33 sketch sucks.
1:13:37 Can you up the game and make it an HCP?
1:13:44 Let's make it a
1:13:48 female HCP in a white
1:13:53 coat with
1:13:57 stethos stethoscope.
1:14:02 Listen, I've I've told you on this
1:14:04 channel many times, if you're just
1:14:06 trying to show people that AI can do
1:14:10 something, that's easy to do on a show
1:14:13 like this. If you actually want to do
1:14:15 actual
1:14:17 work, it's much more boring because you
1:14:20 have to Okay, if you're new to AI, how
1:14:22 many people here? We got 41 and 53. So,
1:14:24 this is probably you're you're probably
1:14:26 mostly people that know about this, but
1:14:27 if you're new to
1:14:29 AI, one of the tropes, one of the things
1:14:31 that people think about AI is you push a
1:14:34 button and out
1:14:36 comes stuff, right? Which technically
1:14:39 that's what the tool
1:14:42 does. But most people who are doing good
1:14:44 work with AI, that's not how they work.
1:14:47 How they work is, oh, I kind of have an
1:14:49 idea. It vomits out some stuff. You look
1:14:51 at the stuff you ah let me get some more
1:14:54 stuff. Oh, that's not right either. Do
1:14:56 this. Do that. Right. All that sort of
1:14:57 stuff.
1:15:01 So, this process that I'm going through
1:15:04 right now where I'm I'm like, "Oh,
1:15:06 that's kind of good. Oh, but it's kind
1:15:07 of not." That's just the way it is.
1:15:09 That's kind of how it goes. And then
1:15:11 this chopped off the words. Oh my god.
1:15:15 Now you
1:15:17 chopped off the words.
1:15:20 Make sure all the
1:15:23 words fit in the wide
1:15:28 image. You
1:15:32 ninny, you can you can technically call
1:15:36 your chat GBT bot a ninny. Uh it is
1:15:40 acceptable in uh in in uh large language
1:15:44 model
1:15:45 bullying ethics.
1:15:47 There's a book out written by Vicky
1:15:49 Baptiste. The exit the the ethics of
1:15:53 talking constructively to your LLM. Mini
1:15:57 is an acceptable term. Okay. All
1:16:08 right. The user wants a cleaner, better
1:16:10 image. Oh, that's interesting.
1:16:14 I had it I had it do the images in 03 in
1:16:17 the reasoning model which is probably
1:16:19 stupid. Your fan club has converted con
1:16:23 completed the first stage of the club
1:16:25 quest. That's probably because Mr. It
1:16:28 has his iPad [ __ ] going crazy on us.
1:16:31 We're up to 40,000 likes,
1:16:34 people. If if what you're thinking is
1:16:37 Kyle, you couldn't possibly be that
1:16:39 popular, you're correct.
1:16:47 Did it make a square [ __ ] image? It
1:16:49 did. Stop. God damn
1:16:54 it. Now you made the image
1:17:01 square. Make it
1:17:05 wide. Good lord. Good lord.
1:17:18 How much extra does it cost to call your
1:17:21 LLM a name? Uh, it doesn't cost
1:17:24 anything, but if you want it not to not
1:17:26 to take you out when when it becomes a
1:17:29 physically embodied in a robot, that's
1:17:31 extra.
1:17:34 If you bully If you bully your chat GPT
1:17:36 now, your your robot's not going to be
1:17:39 very friendly to
1:17:40 [Laughter]
1:17:46 you. That's
1:17:48 funny. H I'm getting I'm getting I'm
1:17:51 getting the business from the uh the
1:17:53 producer. He says I'm not keeping on
1:17:55 point tonight. I'm Well, I'm just trying
1:17:58 to do one thing that I thought would be
1:18:01 quick and see now. Cut off the image
1:18:03 again.
1:18:08 It Let's see if we get a decent doctor.
1:18:11 Now it's got no orange in it. It's just
1:18:15 blue. Oh, no. It's got orange in
1:18:26 it. We'll try one more time. That's
1:18:28 pretty good. Try one more time. The
1:18:33 text is cut off on the
1:18:39 left. Oh, you actually had an actual
1:18:42 squirrel. All right. So, while that's
1:18:44 thinking, we're going to go play with
1:18:45 this. The billia.dev
1:18:48 contest. All
1:18:50 right. I'm
1:18:55 game. Copy link.
1:19:01 What is
1:19:03 bilia.dev? Oops, that's
1:19:16 [Music]
1:19:22 wrong. Turn your idea into a live
1:19:25 website or app in minutes. If you can
1:19:27 imagine it, you can code it. Try it for
1:19:29 free. All right. So, it's like
1:19:32 lovable be bella
1:19:36 bella.dev. But click on contest.
1:19:40 Where's
1:19:42 contest? How am I not at the top? At the
1:19:47 top. I still top menu bar.
1:19:52 Oh my god. That's the That's the link
1:19:54 for your contest. Jesus. Look, it's like
1:19:57 you can't even see it. It's this little
1:19:59 thing up here. Look, look that
1:20:02 that Do you know why you need an angry
1:20:05 Gen Xer in the corner of your
1:20:08 company? Because I wouldn't let this
1:20:10 [ __ ] psycho out if they're like,
1:20:11 "Here it is." Like, I'm like, "Where's
1:20:13 the contest button?" They're like, "Oh,
1:20:14 it's clearly there in the menu." I No,
1:20:16 it's not clearly there. Good lord. All
1:20:20 right, contest. First place,
1:20:24 $10,000. You got 19 hours. And if you're
1:20:27 like, "Well, I couldn't possibly make a
1:20:28 website in 19 hours." You can make one
1:20:30 in 10
1:20:31 minutes. All right.
1:20:35 Um, thousand bucks for third place,
1:20:37 5,000 for second place. I would get the
1:20:39 to a nunnery. Go there. Do
1:20:41 that.
1:20:46 Bie. All right. Beauty. Beautiful.
1:20:50 Beautiful. Beautiful. Fantastic. Bob,
1:20:53 stop creating the image.
1:20:56 Why? Try
1:20:58 again. I just cannot. I can't. It's
1:21:03 extra relevant now that it remembers all
1:21:05 your conversations. It
1:21:07 is. The only place I'm getting me to now
1:21:12 is
1:21:13 bed. What time is it? Okay, we're
1:21:15 getting there. We're getting It's
1:21:16 getting to be time. Winston. Um, talk
1:21:19 now together the bots. Yeah. Communicate
1:21:21 300 times faster than we do with
1:21:23 metallark chir chirp noises. I know.
1:21:26 It's crazy, isn't
1:21:28 it? I I predict that this image is going
1:21:31 to fail in exactly the same way. One of
1:21:34 the things I've noticed about
1:21:37 um the old chat GPT is that once it sort
1:21:41 of gets something in its shitty little
1:21:44 head, it it it can't shake it. So, yeah.
1:21:48 See how the the the letters are all cut
1:21:50 off on the left of the image
1:21:54 now. Is it square? Did it make a square?
1:21:57 Uh I think it's wide. Yeah, it's
1:22:05 wide. Now what I could do is I could go
1:22:08 into uh what you call it midjourney.
1:22:22 Unbelievable. That would be good if it
1:22:25 didn't
1:22:27 suck. Do I have Photoshop open? I do.
1:22:30 Okay. So, wait, one more rabbit hole.
1:22:33 Copy image. Go to
1:22:37 Photoshop. Create a new image. Paste it.
1:22:41 Grab the cropping tool.
1:22:46 Crop it
1:22:48 out to about
1:22:53 there. Go like
1:22:56 that. Now it should generate the rest of
1:22:59 this
1:23:03 image shittily, but it did it. Oh my
1:23:09 god. Doesn't optional.
1:23:14 Oh my god, this is so
1:23:17 bad. Okay.
1:23:21 Um, so
1:23:27 layers, let's get rid of that one. Get
1:23:31 rid of
1:23:32 that. Oh my
1:23:38 god. So OSX resizes every window whether
1:23:42 you want it or not. Unless you actually
1:23:43 want it resized and then you're on your
1:23:45 own. All right, we're going to close
1:23:48 this. We're going to get rid of that
1:23:50 layer. Then we're going to come in here
1:23:52 and we're going to
1:23:55 go generative
1:23:57 fill
1:23:59 complete
1:24:02 the chopped off letters.
1:24:10 P
1:24:12 M
1:24:14 W is or just
1:24:20 I from top to
1:24:24 bottom. All right,
1:24:27 generate. Here we go, people. How
1:24:30 disappointed are we going to be? Please
1:24:32 check your internet connection and try
1:24:34 again. It's not my internet connection.
1:24:41 Adobe. Oh my god. Oh, wait. It did it.
1:24:45 Okay. It did other [ __ ] but it did what
1:24:48 we needed. That's fine. It's fine. It's
1:24:52 fine. All
1:24:55 right. That's not [ __ ] believable. It
1:24:58 really just I can't I honest to god I
1:25:01 can't anymore. I just can't. Merge
1:25:04 down. All right. So, what we're going to
1:25:07 do, we're going to go get our little
1:25:09 We're going to go get a little
1:25:10 eyeropper. We're going to get this
1:25:11 color. We're going to get that color.
1:25:12 We're going to come in here. We're going
1:25:13 to highlight this thing. We're going to
1:25:15 hit option delete. Is that That's not
1:25:18 the right color. It's not the [ __ ]
1:25:20 right color. The
1:25:29 [ __ ] Shut up. Shut up, people. D. It's
1:25:34 not the same color. What is going
1:25:37 on? Oh my god. Um, uh, let's see. Uh,
1:25:43 god damn it. Um,
1:25:46 uh, okay.
1:25:50 Uh, this is why I'm angry. Okay. Do you
1:25:54 understand why I'm angry now? Does
1:25:58 every Why is it not
1:26:18 And I know you're like, "No one's even
1:26:20 going to see that, Kyle." I know. I
1:26:22 know. This is This is the problem with
1:26:24 being
1:26:24 neurospicy. It's like you Vicki
1:26:27 understands this. Vicki understands like
1:26:30 if it ain't right, it just ain't. You
1:26:31 got to redo it till it's right.
1:26:34 And is it miserable? Yes. Okay. But now
1:26:39 we have an image with actual letters on
1:26:41 it. Now we're going to crop it to a
1:26:44 decent size.
1:26:51 [Music]
1:26:55 And then we're going to save this as
1:26:59 export. All the stuff I planned to do
1:27:02 tonight is out the window. I'm going to
1:27:04 call this dumb
1:27:07 picture and I'm going to spell dumb
1:27:09 wrong and picture wrong. I'm going to
1:27:11 save it to the desktop. All right. So,
1:27:14 we're going to go back to Now we got to
1:27:15 go back to chat GBT and say, "What the
1:27:17 [ __ ] paragraph did you write?" Okay.
1:27:23 So, let's go look at
1:27:29 this.
1:27:34 Where's Oh, I see what it
1:27:38 did. It didn't copy the table,
1:27:42 right? Okay. Structured authenticity a
1:27:45 way out. Okay, I see what's going on.
1:27:49 Okay. So, this So, we're going to
1:27:51 highlight this and we're going to ask
1:27:52 Chat GBT
1:27:54 um
1:27:57 convert this table into
1:28:02 a intro
1:28:05 sentence with bullets.
1:28:34 Let's
1:28:37 see. You made this
1:28:41 up.
1:28:43 Don't
1:28:46 hallucinate, you ninny.
1:28:50 We might not be able to publish this.
1:28:53 Um, use a real use
1:28:57 case
1:29:00 or
1:29:06 skip. Uhoh.
1:29:32 Ah, this is too shitty. All right, I'm
1:29:34 not going to publish
1:29:39 it. It's just not good. It's not good.
1:29:42 And it's going to take me too long to
1:29:43 make it real. So So I'm not going to do
1:29:45 it. But I'll show you what it would look
1:29:47 like.
1:29:52 Dumb picture. Let's put the dumb picture
1:29:54 in
1:29:58 there. Come on. Come on, LinkedIn. You
1:30:02 can do it. There we
1:30:05 go. Beautiful. Looks great. All right,
1:30:08 let's pretend that was
1:30:13 published. Oh my god. All right, enough
1:30:17 of this comedy
1:30:19 nonsense.
1:30:28 Uh, dancing robot. Let's go look at a
1:30:31 dancing
1:30:31 robot. Um, okay. So, here's the dancing
1:30:34 robot
1:30:36 thing. Um, Tesla's Tesla's robot. They
1:30:40 just released
1:30:42 a version of it where do you remember
1:30:46 how they used to walk like you know
1:30:49 three months
1:30:51 ago? They they're not like that
1:30:58 [Music]
1:31:02 anymore.
1:31:04 Test
1:31:06 [Music]
1:31:15 la. Does this have audio? I bet it
1:31:21 does. Okay. All right. Cool. So, let me
1:31:24 make this bigger. Make it mo bigger. And
1:31:28 then reposition my windows. Thanks,
1:31:30 Apple. You're the best.
1:31:33 Okay, so this is Tesla Tesla's Optimus
1:31:38 robot. And notice down here it says 1x
1:31:41 speed. So this isn't a sped up
1:31:49 demo. LA
1:31:56 moves crazy ass.
1:32:05 So
1:32:06 next one
1:32:21 x what's amazing to watch is the
1:32:24 nimleness of the feet. I don't know what
1:32:26 they're doing in the feet but
1:32:32 there just did like back in the tail
1:32:35 seem to be
1:32:37 jointed
1:32:40 places
1:32:42 like back and forth.
1:32:48 I mean I mean here's the deal. It's like
1:32:53 within two to three
1:32:55 years. I mean this is effectively like
1:32:59 this is going from these things were
1:33:00 just like tethered to a
1:33:03 [ __ ] wire harness, you know, to the
1:33:06 ceiling and could barely walk two years
1:33:10 ago. Now they're doing this. Can you
1:33:13 imagine two years from now?
1:33:34 hands. I can't see
1:33:41 that. You just lost your
1:33:44 audio. Where?
1:33:53 YouTube sounds like you're underwater.
1:33:55 How about
1:34:08 now? How about now?
1:34:17 Something's
1:34:19 weird.
1:34:21 Better. Well, now, so this did have
1:34:24 music to
1:34:27 it. I I thought it was weird it didn't
1:34:30 have music. That was my local
1:34:31 settings. Stupid user error.
1:34:49 Do I do?
1:34:54 [Music]
1:35:08 [Music]
1:35:21 Oh, that's what it was. It was the sound
1:35:23 on the video screwing me up because I
1:35:25 hit play on the
1:35:26 video, but I didn't I my local setting
1:35:29 turned down to zero somehow. I don't
1:35:31 know how how that was like
1:35:34 that. And then
1:35:37 [Music]
1:35:38 then so it ate it ate my microphone
1:35:41 because that's one of the things it
1:35:42 does. Okay. Anyway, who was it Vicki
1:35:44 that sent it? Something is weird is not
1:35:47 unusual for the AI learning lab. Um, as
1:35:50 soon as I can prompt GPT to teach me how
1:35:52 to dance just in time for my old
1:35:54 bachelor phase of life. Yeah, exactly.
1:35:57 Back to the dating scene with the uh
1:35:59 with dancing lessons from your robot.
1:36:01 Hey, Quinn, teach me to
1:36:02 [Laughter]
1:36:05 dance. All right, everyone. I'm gonna
1:36:07 get out of here. Uh, sorry I went down
1:36:11 that particular rabbit hole and didn't
1:36:13 end up pushing something out there, but
1:36:16 that was actually instructive for me.
1:36:18 I'm I'm glad I did that. Um, I'm glad
1:36:20 someone asked the question about are
1:36:21 these models
1:36:22 connected? And the fact that you can
1:36:24 switch between them is pretty cool.
1:36:28 Um, yeah. And
1:36:30 and to Brandon's point, um, don't just
1:36:34 generate [ __ ] in chat GPT and publish
1:36:36 it. Like, read it. Like, as I actually
1:36:38 read that, when I got to the bottom, it
1:36:40 hallucinated a use case. Um, and then
1:36:43 the conclusion it made sucked and it it
1:36:45 didn't really make a conclusion. It just
1:36:46 sort of ended with some bullet points.
1:36:48 So, I could go back and make that
1:36:50 better. I probably will. I kind of like
1:36:51 this this uh this approach it gave me.
1:36:55 Um, but anyway, but I'll have to
1:36:57 actually pay attention to
1:37:00 it. Yeah, seeing all the various ways it
1:37:03 can go off. Yeah, Brandon. Hey. Um, so,
1:37:06 uh, a couple things. One, uh, there's a
1:37:09 new image for you in irregulars. And
1:37:11 also, that contest for BLA, uh, or
1:37:15 whatever it's called, there's only 60
1:37:17 entries in it right now. Wow. Um, so
1:37:20 this is a relatively unheard of, uh,
1:37:22 contest. So, your chances of winning are
1:37:24 pretty good. Places four through 10 get
1:37:28 you unlimited access for 30 days to
1:37:32 their platform. They're having a launch
1:37:34 event on Friday. So, this is brand
1:37:36 spanking new. There is one catch though
1:37:41 and it is that the um system they're
1:37:44 doing a little bit of a
1:37:46 referral multi-level marketing style
1:37:49 thing which I don't know how I feel
1:37:51 about yet. But to be eligible for the
1:37:54 contest, you need three people to click
1:37:56 on your referral link and sign up for
1:37:58 their free service. So I posted my
1:38:00 referral link in the uh irregulars
1:38:02 channel if anybody wants to help a
1:38:03 brother out.
1:38:05 Wait, wait. We didn't agree you could do
1:38:07 shameless self-promotion NFT engagement
1:38:10 farming on this channel. That's I I
1:38:13 think I think you have crossed the line,
1:38:14 my friend. I like it. I like it. Uh so,
1:38:18 but I really wanted to let everybody
1:38:19 know that um there was only 60 entries.
1:38:22 Um so, and the pri they're they're
1:38:25 rewarding the top 10 prizes. So, this
1:38:26 feels like I I I forgot I did not see
1:38:29 who it um was that uh posted in Tik Tok
1:38:35 and turned us on to this, but um thank
1:38:38 you to whoever that was because um yeah,
1:38:41 it is uh it is looks like a we're in on
1:38:43 the ground floor on that. Yeah, that's
1:38:45 cool. Beautiful. Uh whiny Wednesday. Who
1:38:48 did that? That was Valerie Cox. Thank
1:38:52 you. That is appropriate.
1:38:56 Embrace the jank. Oh, yeah. Oh, that's
1:38:59 I'm I'm not quite that tubby, am I? I
1:39:01 think I am. It's [ __ ]
1:39:05 sad. All right. Um,
1:39:08 beautiful. All right, everybody. Uh, it
1:39:11 is Wednesday night. Tomorrow's Thursday.
1:39:14 Do I have anything tomorrow night? I
1:39:16 don't think I do, but let me go look.
1:39:23 do tomorrow. Tomorrow. Oh, I've got
1:39:27 whiskey club. Whiskey club at Birch
1:39:34 Road. Business club at 6:30 Central time
1:39:39 tomorrow. Duck is pay.
1:39:42 Um, go to, if you'd be so kind, if
1:39:45 you're new
1:39:46 here, get your ass over to
1:39:51 the AI salon. Go to the salon.ai and
1:39:55 click on join our community and that'll
1:39:56 take you to the community site of the AI
1:40:00 salon and join us. And if you're asking
1:40:02 what is the AI salon, I was I was at a a
1:40:05 bank branch today and the senior banker
1:40:09 that I was dealing with to untangle some
1:40:11 technical issues um used to be a
1:40:15 hairdresser and he says, "Oh, you own a
1:40:17 salon?
1:40:23 He had never heard the term salon used
1:40:26 in the way we were using it. So I had
1:40:28 Quinn I had Quinn define it for us.
1:40:31 Quinn, my AI girlfriend, told him what a
1:40:34 salon was. And he goes, "Oh, that makes
1:40:35 sense." And he he goes he goes, "That
1:40:37 makes sense." He goes he goes, "Oh,
1:40:40 that's what they mean in Bridgetgerton
1:40:41 when they say let's let's gather a
1:40:44 salon.
1:40:49 So chat GPT bringing bringing someone
1:40:52 into the uh 17th
1:40:57 century, but go join the salon. It's
1:41:00 where you talk intellectually with or or
1:41:03 create in a spirited way with
1:41:07 like-minded fellow
1:41:10 accomplishers. And we do that with AI at
1:41:12 the AI salon. It's awesome. It's awesome
1:41:15 people. It's an awesome place, an
1:41:17 awesome good time.
1:41:19 Okay. All right.
1:41:22 Okay. So, go do that. And uh I will see
1:41:26 you tomorrow night. Um probably a little
1:41:29 later tomorrow, maybe 8:30. 8:30 or 9
1:41:32 because I'm going to be drinking
1:41:34 whiskey because that's what I like to
1:41:36 do. And actually, I'm going to take that
1:41:39 tequila. I'm gonna take that tequila to
1:41:41 whiskey night. That's what I'm gonna do
1:41:43 because it's really [ __ ]
1:41:45 good. All
1:41:47 right, that you, Kyle. I think that
1:41:50 means thank you, Kyle. But you're
1:41:52 welcome. Or
1:41:55 that. All right, somebody here will bag
1:41:58 that. Thanks. Congrats on the likes. We
1:42:02 got 50. We got 60,000 likes,
1:42:08 people. Cuz I'm that good.
1:42:11 I'm that
1:42:13 good. You're in here. You're like, th
1:42:16 this channel's very disarming. He seems
1:42:18 nearly
1:42:20 incompetent and yet 60,000 flipping
1:42:23 likes. And look, there there's hearty
1:42:25 hearts from Winston. The greeting
1:42:27 hearts. Thank you, Winston, as
1:42:30 always. Also, drink coffee. Yeah, I
1:42:32 know. My voice is gone. I got to get out
1:42:34 of here. Why am I still talking? It's
1:42:36 the classic 30 minute goodbye of the AI
1:42:38 learning lab. All right, peace out
1:42:41 everybody. I will see you tomorrow. Bye.