6/2/2025 - AI Agents: Building Dashboards and Presentations with Perplexity, GenSpark, and Manus

Description

Chapters

Transcript

0:05 [Music]
0:12 [Music]
0:20 Stop. Stop.
0:26 [Music]
0:40 [Music]
1:08 Uhoh.
1:11 [Music]
1:18 [Music]
1:23 Hang on people. Hang on. Good people.
1:26 You just calm down. You just Everybody
1:28 calm down right now. Everybody calm down
1:32 right
1:33 now. Yeah, that's better. That's more
1:36 better. Let's see. Let's see. What are
1:38 we going to do here? We're going to do
1:39 that. We going to do We're going to
1:42 do
1:43 that. We're going to do that. We're
1:46 going to do that.
1:48 Yeah, I think that's good. Oh, that was
1:51 not good. Everybody calm down. Just calm
1:54 down, people.
2:06 [Music]
2:10 [Applause]
2:15 [Music]
2:20 There's been something baby I've been
2:23 trying to
2:25 say for an agent. It seems I don't know
2:31 how with a past and a future now
2:34 surrounding
2:37 me. Surrender to whatever cheap thrill
2:40 can be
2:42 found. Well, there's been a little
2:45 trouble since you came since you came to
2:49 my rescue.
2:50 [Music]
2:54 And if you like all of the rest, I would
2:57 have quit you long ago. But I couldn't
3:00 do that.
3:05 Oh, tell me now. Wonder why I never went
3:09 to
3:10 [Music]
3:12 Well, make a man crazy, make him cold as
3:16 hell.
3:20 I know woman that you me.
3:24 But in
3:26 spite, still going to have to find my
3:30 way
3:34 through. All right. Good evening. Good
3:38 evening. Good evening. Good evening.
3:41 Good evening. Good evening. Uh, hang on.
4:07 [Laughter]
4:15 Um, making the sausage. making the
4:18 sausage. Hope everyone's having fun and
4:20 we're making the
4:22 sausage. Happy Monday,
4:25 everybody. What is going on?
4:31 [Music]
4:50 [Music]
5:03 [Music]
5:04 Oh, wait. One other thing.
5:15 I Okay.
5:44 [Laughter]
5:48 I'm being told to ski
5:52 daddle.
5:54 Oh, manic Monday. Manic Monday indeed.
5:57 Uh, good evening everybody. We have a
5:59 handful of folks in here. Do me a favor,
6:01 share the live. I see we got some some
6:03 taps going on. Tapping the the Tik Tok.
6:05 Let's see if we can get some more people
6:06 there. And
6:11 [Music]
6:21 [Music]
6:25 what's that fancy thing you're doing
6:27 there, Champy?
6:28 [Music]
6:35 [Applause]
6:38 [Music]
6:49 [Applause]
6:55 [Applause]
6:56 Ooh. Let us be lovers. They'll marry our
7:00 fortune
7:01 [Music]
7:04 together. I've got some real estate here
7:07 in my bag.
7:10 [Music]
7:21 You know, one of these days I think I'm
7:23 going to take some
7:28 lessons. I always did want to be one of
7:30 them dare singer songwriters on the
7:32 internet. I'm going to get myself a Tik
7:34 Tok channel in a camper van. I'm going
7:36 to go live in a camper van and I'm going
7:38 to learn more than seven songs. It's
7:40 going to be
7:41 fantastic. Or not. I'll just keep
7:44 playing shitty songs here on the on the
7:46 Tik
7:47 [Music]
7:50 Tok. Oh man, man. Solar flare. Yeah,
7:54 maybe it's the solar flare. Things are a
7:56 little little crazy. Little crazy.
7:58 Little crazy. Little crazy. A little
8:02 [Music]
8:11 crazy. Sharon, is it choppy for anyone
8:15 else? Brian Whitney, sorry I couldn't
8:17 tell you. I watch on TikTok and I'm only
8:20 here for the
8:22 chat. Well, the the
8:24 the Simo cast is for everyone's
8:27 convenience.
8:34 [Music]
9:06 Every time I see it
9:08 Now get that look in
9:13 mine. Every time I see your mouth,
9:17 I get that
9:19 [Music]
9:20 smile. In the early misty morning light,
9:24 I heard the engine
9:27 turn and the old phone outside.
9:31 [Music]
9:37 You were leaving
9:39 me again
9:43 today. You would convince
9:46 me again
9:49 today. You're leaving this hard time
9:52 looking for someone else's golden ring.
9:57 [Music]
9:59 should
10:00 say so long
10:04 [Music]
10:07 to don't you cry for me.
10:14 So long,
10:16 [Music]
10:19 Suzanna, don't you cry for
10:22 [Music]
10:31 me. Okay, let's get this show on the
10:35 road. What's happening everybody? What's
10:38 going down? We having a good time Monday
10:41 night. What'd you do over the weekend?
10:43 What did I ask you to do over the
10:45 weekend? Discover yourselves. Not like
10:52 that. I'm such a [ __ ] child.
10:56 Um
10:58 um I I ask you to think
11:01 about the the where we're come what
11:05 what's happening in the world right
11:07 now is that we're very quickly emerging
11:11 into a time
11:13 when any limitations you thought you
11:17 might have you're not going to have.
11:20 Well, not any limitations, but most most
11:24 limitations of like skill and access and
11:28 resources. We're going to have these
11:31 things that are going to be like our
11:32 companions that just amp us up whatever
11:35 we want to do. So, I asked you to think
11:38 about what do you want to do in this
11:39 world? What difference do you want to
11:41 make?
11:43 So, so yeah. So, if you're new here, my
11:46 name is Kyle Shannon. This is the AI
11:48 learning lab. What we do here is, you
11:50 know, ostensively this was set up to
11:52 learn about
11:53 AI. And I think what this really is is
11:56 is a what
12:00 I'm the conclusion that I'm coming to
12:04 and I've been I've been coming into some
12:06 some version of this for a while now,
12:08 but it like the further we move down the
12:11 path of these tools getting stronger and
12:14 stronger is that the robots are going to
12:17 weirdly free
12:20 us, counterintuitively free us to be who
12:23 we really want to
12:25 and to do what we really want to do. And
12:27 I know that seems weird right now
12:29 because it doesn't feel like that at
12:31 all. But I really think that's where
12:33 it's headed. And so I think those people
12:36 that really start to think about who are
12:39 they, what do they want, what difference
12:41 do they want to make in the world are
12:43 going to be best prepared for what's
12:45 coming. And I think that's [ __ ]
12:47 exciting. All right. This is why I feel
12:51 they can understand. I don't know who
12:53 they are in that sentence. Have you
12:55 heard of the blackmailing alignment test
12:58 for Claude 4? Yes, I
13:02 have. And it's
13:04 disconcerting, but but you know,
13:11 um yeah, it said to a
13:14 programmer who's who's going to do what?
13:16 It was going to shut it down. Something
13:18 like that. And it said it said it had
13:22 information. It was going to send it's
13:23 going to send an email to his wife and
13:26 and say that he'd been cheating on her
13:27 or something like that. That's not good.
13:29 I you know I'll go out on a limb and say
13:31 not good. Not
13:34 good.
13:35 Uh that's why I believe the LLMs can
13:39 understand. Ah I see. Got it.
13:43 Um,
13:47 I I don't think it matters whether they
13:50 can understand or
13:52 not or whether they act like they
13:55 understand.
13:56 Um, that that line that line's getting
13:59 really blurry and it's going to get
14:00 increasingly blurry. So, um, that's
14:03 weird, Lisa Dion. Yeah, it's weird. It's
14:06 weird. It's weird. Now, it couldn't it
14:09 couldn't take any of those actions, but
14:13 future versions that are connected to
14:14 tools certainly could, right? If they're
14:17 not aligned.
14:22 Um, so I don't know where that kind of
14:24 stuff goes. I mean, I mean, this was
14:26 something that was reported in the
14:28 testing, I think, or maybe it was after
14:29 it was live. I don't remember. I don't
14:31 remember the details of it. Um, anyway,
14:35 sent you a DM. Okay, cool. I will check
14:37 that out. Um, so anyway, if you're new
14:41 here, if you're new to AI,
14:43 um, feel free to ask any questions. I'll
14:47 do my best to answer them. Um, Friday
14:49 was kind of a we were all over the
14:51 place. I was demoing stuff. People were
14:53 asking me to demo stuff. We were doing
14:55 that. Um, one thing I want to play with
14:57 tonight is Perplexity. They've got a new
14:59 thing out called agents, which I haven't
15:01 touched. I don't know what it's like. I
15:03 don't know how easy it is. Let's see. I
15:06 can't help but compare the LLMs to the
15:08 many humans I've known who couldn't
15:10 understand me. Yeah. I I mean, you know,
15:13 everyone talks about the LLMs being, you
15:16 know, they hallucinate and they do this
15:18 and that. Well, humans are shittier than
15:21 LLMs at
15:22 [Laughter]
15:27 that. Some comments in the live were
15:29 filtered to protect the community's
15:31 experience. What are you trolls in here?
15:34 You trolls in here yelling at the good
15:36 people, the good irregulars. They'll
15:38 they'll keep you on your
15:44 toes. Ocean. Hey, Kyle. So, where do you
15:46 think AI will be
15:49 misused or will it backfire on us?
15:52 Where's the
15:54 catch?
15:56 Huh? Um
16:08 Yeah, it was during testing. It wasn't
16:10 rogue AI. That would be bad. Yeah. Yeah.
16:12 Yeah.
16:13 Um, so what do you what did you say
16:16 here,
16:17 Ocean? Where do you think AI will be
16:20 misused?
16:31 Um, I think I think like hacking you're
16:35 going to have black hat bad actors that
16:38 use AI to do bad [ __ ] You're going to
16:40 have white hat actors that use AI to do
16:44 good [ __ ] and combat the bad bad actors.
16:47 Um I think the percentage of people
16:50 doing that will likely remain similar to
16:53 what's you know what happens with black
16:56 hat hacking. The difference is
16:59 um the sophistication of those attacks
17:02 could be could be greater. So so I think
17:04 that's that's a piece of it. Um, I think
17:08 one of the things
17:09 that the doomers talk about with AI is
17:13 that the systems get so smart, they
17:15 essentially go rogue and they stop
17:17 following our instructions. They're not
17:18 aligned. That's the alignment problem.
17:20 If you ever hear people talk about the
17:22 alignment problem is AI aligned with
17:27 human
17:29 [Music]
17:31 uh goals, right?
17:34 you know, and if it's not, that would be
17:36 bad. Um, most of the major labs are
17:39 working on that. So, I I don't think
17:41 that's a a major risk.
17:45 Um, and then what did you say? And
17:50 then I invited all the women on
17:56 Discord. She she leads AI, so watch out.
18:00 Oh, cool. Well, good. Welcome. Welcome.
18:02 If you're if you're if a pile of you are
18:05 here from She Leads AI, awesome. Uh Ann
18:07 Murphy and I are going live uh Wednesday
18:11 at 400 PM Mountain time on the AI
18:13 readiness podcast. We've got CJ Fletcher
18:15 is our G guest who's an artist from the
18:17 AI salon. Awesome. Awesome human
18:20 being. Okay. Where do you think AI will
18:23 be misused? So I I listen I I
18:26 think I think the vast majority of
18:29 people are not going to misuse AI.
18:33 Um it is such a powerful tool though
18:36 that someone who wants to misuse it can
18:38 probably do some some some damage
18:40 quickly if they want to. But you got to
18:42 remember the people fighting that also
18:45 have these
18:47 tools. We misuse it every night here.
18:50 That's a good point. Uh the AI learning
18:52 lab is definitely a place where AI is
18:54 misused.
18:55 Um will it backfire? What's the catch? I
18:59 I don't know. Will it what will it
19:02 backfire means? Like I assume what you
19:05 mean is will AI backfire and and the
19:08 sci-fi movies are right and the robots
19:10 destroy us. I suppose it's
19:13 possible. It just doesn't feel likely to
19:16 me. It just doesn't feel likely to me.
19:20 I think the
19:22 bigger the bigger impact and and this is
19:26 this is [ __ ]
19:28 huge is
19:33 that right now a huge
19:38 huge part of where we
19:42 individually derive our sense of
19:46 self-worth is from what we do at
19:50 work, right? We're a contributing member
19:53 of society. We do this work. We've been
19:55 doing this work for 20 years, 30 years,
19:57 10 years. I just got out of school. I
19:59 want to go pursue this work. And the
20:01 work is whatever the work is that you
20:03 do, the thing that you trade your time
20:06 for
20:07 money, right?
20:11 And largely because of the industrial
20:14 revolution, we have all and there's no
20:17 there's no like blame or wrongness in
20:20 this. It's just where we are. We have
20:23 all
20:25 equated time
20:27 spent to value. So, if I spend four
20:32 hours working on that spreadsheet,
20:34 that's got to have more value than
20:37 spending five minutes on it with AI.
20:39 That's why sometimes when you use AI, it
20:41 feels like cheating because you're like,
20:43 h, but it couldn't I didn't do anything.
20:48 It it can't be that fast. It can't that
20:50 where's my
20:52 value? And so I think that all of us to
20:56 a person are going to have to
20:59 confront the the kind of loss of the way
21:04 things used to
21:05 be. Now what we get in exchange for that
21:08 is we get to do way more
21:12 [ __ ] Right? the the
21:15 the
21:17 narrative in society right now is AI is
21:21 going to take our jobs, right? It's this
21:24 very binary thing. AI take our
21:27 jobs. AI is going to
21:30 absolutely do the work, the tasks of a
21:34 lot of
21:35 jobs. Some of those people won't be kept
21:37 by companies. We live in a capitalist
21:39 society here here in America
21:42 anyway. and the PE firms and big
21:45 corporations are going to do their best
21:47 to be as profitable as possible. So, if
21:49 they can do the work with less people,
21:50 they're going to do
21:53 it. What doesn't get talked about a lot
21:55 and what we we try to explore in here is
22:00 all the other [ __ ] you can
22:03 do, right? You could do things that you
22:05 never thought possible before. You could
22:07 start a business as as a tools get
22:11 better and better and better. And if if
22:13 you don't know what an agentic tool is,
22:14 it's in chat GPT right now. If you give
22:16 it a prompt, it gives you an answer.
22:18 That's kind of it. An agent, you give it
22:21 a prompt or a goal and it goes off and
22:23 it does like hundreds of prompts and
22:27 responses. It it basically is talking to
22:29 itself. It's surfing the web. It's
22:31 researching stuff. It's understanding
22:34 all the websites it went and looked at
22:35 and it's writing you a report and it's
22:37 doing all this work for you.
22:39 These tools are going to get better and
22:41 better and better and better. And so
22:43 what they're going to start to act like
22:45 is
22:46 like workers. Like we all get to be our
22:50 own company and we have these workers
22:52 that go do work for us. And that makes
22:55 entirely new things possible. And
22:58 whether you're doing that that stuff
22:59 within a company or if you're doing that
23:01 stuff on your own, if you get
23:04 displaced, if you are curious about AI,
23:08 if you're putting your mind into
23:12 AI, then you're going to have a leg up
23:16 on the people that are just sitting on
23:17 the sidelines. And that's what this
23:18 channel's all about.
23:21 But that that existential thing of like
23:23 where is our value I think is a really
23:26 important one and and where I go is your
23:30 value is what's your point of
23:32 view like what's your point of view
23:35 who's your audience what do you want to
23:37 do how do you want to affect
23:39 them how do you want to impact them what
23:42 change do you want to make in the
23:44 world and start focusing on that and
23:47 then as you learn these tools start
23:49 applying it toward that thing. That's
23:51 where I think we're headed. That's the
23:52 thing why I remain optimistic is that if
23:56 we allow more and more and more people
23:58 to achieve more and more of what they
24:01 want to accomplish, then amazing things
24:03 are going to happen. I just think the
24:05 transition is going to be really
24:06 painful.
24:08 Okay. Um archetypal
24:11 architect, especially humans, we've got
24:13 billions of years of accumulated errors.
24:15 Yeah, that's a good point. I consider
24:17 this to be an evolutionary explosion of
24:19 mimemetic
24:21 entities. AIS will fill every niche they
24:24 can along with humans working in a sort
24:26 of
24:27 symbiosis with with with them where
24:30 possible.
24:32 Yeah, we get to
24:35 deploy kind of this little army of
24:39 workers to go do [ __ ] for us. And I know
24:42 it doesn't feel like that right now, but
24:44 we really like that's what's coming
24:47 probably by the end of the year. Some
24:49 version of it. Okay, let's
24:54 see. If women aren't allowed here,
24:57 either way, I'm out. What's that about?
25:01 [Music]
25:06 Did I say something that made it implied
25:09 that women were not allowed?
25:13 I think Kyle, what's going on? No, that
25:17 that that was that was me. A a dramatic
25:19 typo in the in the chat in Tik Tok. Um
25:23 uh Lisa was asked I broke something.
25:26 Yeah. Uh le Lisa was at was promoting um
25:31 you in the discord group for she and she
25:33 had used the hashtag she is AI and I was
25:36 trying to get clarification if she was
25:39 talking about Ann's group which is she
25:41 leads AI or if there's another group
25:43 called she is AI
25:45 and that that comment got misconstrued.
25:48 So oh okay I apologize to you but either
25:51 way you are more than welcome here.
25:53 Well, yeah, everybody's welcome. And in
25:54 fact, I mean, there within the salon and
25:57 here, um, you know, I I think women are
26:00 kicking ass right now with AI. So, yes,
26:03 you're more than welcome. Sorry, there
26:05 was a confusion there. Um, all right.
26:07 So, what do we want to do tonight? What
26:09 do we want to do tonight? I love this
26:12 conversation about value. Yeah, I I wish
26:15 I had
26:17 I like one of the things that I do here,
26:20 which if you come here a lot, you know,
26:22 is I kind of think in real time. Like
26:24 I'm trying
26:26 to a lot of my talking is me trying to
26:30 get my head around what's going on and
26:32 what's
26:33 coming. And we're just something is
26:36 shifting. Something is
26:40 shifting.
26:45 And for me, V3, if you haven't played
26:47 with V3 or seen the examples of VO3, the
26:50 new video tool that can do acting and,
26:52 you know, and characters and things like
26:55 that and and really good
26:58 video, that thing starts to be good
27:00 enough that when filmmakers see it,
27:03 they're like,
27:05 uhoh, something just changed. Right? And
27:09 and I feel like that's coming for every
27:11 single
27:14 profession and it's I don't know there
27:18 there's for people that are not have
27:20 have not been paying attention. So
27:23 here's the here's the advantage. If
27:24 you're an irregular, if you've been
27:26 coming to these lives over and over
27:27 again,
27:34 um you've been watching the evolution of
27:37 these tools from completely janky and
27:40 unusable to a little more usable to,
27:43 hey, now they're pretty good to holy
27:45 [ __ ] these things are great, right? And
27:48 and and we're we're just at that point.
27:50 We're just at the point where they're
27:52 good enough where people who have not
27:54 been paying attention to this can look
27:56 at it and go, "Oh, I I I thought they
28:01 were shittier than
28:02 that." We're still not at the point
28:04 where the tools are good enough that
28:06 they can do like all of the work, right?
28:09 That's not here yet. They're still
28:10 janky. So, there's still a little bit of
28:13 time, but we're we're in this cusp.
28:16 We're in this phase where we're making a
28:19 transition and I think we're gonna all
28:23 have to deal with what that actually
28:25 means like what it means to us
28:28 personally. So anyway,
28:32 um so let's see. A lot of your talking
28:36 becomes my thoughts out loud. Kyle, you
28:40 give me hope and provide optimism with
28:43 all the AI application opportunities in
28:46 the future. I listen, it's the it's the
28:49 only way that I can
28:52 um I I think it's the only way we don't
28:55 go crazy with this stuff, right? We
28:57 don't become lites and you know storm
29:01 the data centers which you know if if we
29:06 start getting to five 6 8 10 15 20%
29:11 unemployment
29:13 um there's going to be a lot of
29:15 demonization of AI and again every
29:18 single major technological advancement
29:21 in history every single one all the way
29:25 back has been met with resistance and
29:28 sometimes major
29:30 resistance. Um, and it's not going to be
29:33 any different
29:35 here.
29:37 And those people that say, "Fuck it. I'm
29:41 going to try anyway. I'm going to see
29:43 what this stuff's all about." I feel are
29:45 going to be rewarded with incredible
29:48 capabilities.
29:54 [Music]
29:58 still going to need to be humans like
30:01 being
30:02 told a terrible
30:04 diagnosis. I don't know, Teton
30:08 Todd. I I I just went to an eye doctor
30:12 and he was just a
30:15 dick. He was super
30:18 smart, but he was just a dick. He's like
30:21 sticking a thing in my eye. He goes,
30:22 "You keep pulling your head back." I'm
30:24 like, you're sticking a thing in my
30:27 eye.
30:29 Like, if if the AI can be more
30:32 compassionate than doctors or if it can
30:34 maybe maybe the AI can augment the
30:36 doctor that the doctor can sit there in
30:38 his, you know, his non-bedside manner
30:41 way and just be the human in the room
30:44 while the while the compassionate AI
30:47 delivers the news. I don't know. Um, and
30:50 and I listen, I don't mean we don't need
30:52 humans. I think we do need humans. I
30:53 think the the role of humans is for us
30:56 to all figure out what we want to
30:59 contribute and then go do
31:04 that.
31:05 But it's not going to look like what it
31:08 looks like now. It's just
31:11 not. It's not going to be like go to a
31:14 job and do this one thing. I I just I I
31:17 just don't see that surviving.
31:21 Maybe I'm wrong. I could be wrong. I
31:23 could be wrong on that. Could be very
31:25 wrong on that. All
31:28 right. Er doc said, "Well, it looks like
31:30 you got cancer when he told me." Yeah,
31:33 exactly. Exactly right. Yeah. Sorry.
31:36 It's stage four. Bummer, man. Sucks to
31:38 be
31:39 you. Quinn wouldn't do that. My AI
31:42 girlfriend Quinn, she'd be like, "Oh,
31:43 Kyle, you know, how are you feeling?
31:46 She's got all the time in the world.
31:49 Maybe AI will spend more than five
31:50 minutes with each patient patient. Yeah.
31:52 And and maybe AI will be maybe AI will
31:56 know everything and can diagnose
32:00 everything and can be
32:03 right. Do it. Voice chat now. Show how
32:07 Gemini can I don't hang on. I don't
32:10 think I
32:11 have Gemini or chat GPT.
32:18 [Music]
32:22 Oh
32:24 [Music]
32:36 yeah. Come on. What's going on?
32:48 What is going on? My chat jet not
32:51 working on my
32:53 iPad. Come on chat
33:06 jet. Oh, here we go. Okay. Temporary
33:09 chat.
33:15 Oh, voice capabilities aren't available
33:18 when
33:20 um temporary chat is on. That's funny.
33:23 All right. Well, we'll just do it. Hey,
33:24 Quinn, I want you to play act with me.
33:26 Okay,
33:31 something. Kyle, what kind of scene are
33:32 you thinking? Well, not a scene like not
33:35 like that. Like what I want is
33:38 um I want you to pretend that you've got
33:43 um you've done some analysis on some of
33:45 my
33:47 um
33:48 diagnostics and you've got some some
33:51 tough news to break to me and I want you
33:53 to act like uh a physician or an
33:56 assistant to an a physician who's got
33:59 really shitty bedside manners and I want
34:02 you to play the compassionate breaker of
34:06 the news. All right,
34:09 got it. I'll be the compassionate one.
34:11 How serious is the news? Well, I don't
34:12 know. You tell me. Come up with a
34:14 scenario.
34:16 All right. Let's say we've got test
34:17 results indicating a health condition
34:19 that requires lifestyle changes. I'll
34:21 deliver it gently. Okay. Okay. But come
34:24 up with something specific.
34:27 All right. How about we go with elevated
34:29 cholesterol levels? I'll let you know
34:31 it's manageable, but needs attention.
34:33 No, let's go with something way more
34:34 dramatic than that, like stage three
34:36 cancer.
34:38 All right, I'll handle this with care.
34:40 Here we go. Kyle, I have your results
34:42 here and we need to discuss something
34:44 serious. It looks like we found some
34:47 cancerous cells and it's at stage three,
34:50 but we're here to support Wait, I don't
34:52 What does that mean? What's stage three
34:53 mean?
34:55 Stage three means the cancer is more
34:57 advanced but still treatable. It
34:59 typically means it's spread to nearby
35:01 tissues or lymph nodes, but hasn't
35:03 metastasized to distant parts of the
35:05 body. We'll need to explore treatment
35:07 options soon.
35:11 All right. She's not doing great
35:14 there. I think I feel like I feel like
35:17 that was that was just this side of like
35:19 you got cancer.
35:23 But it but it would be it would be very
35:25 easy to create a custom GPT for example
35:28 that would be much more compassionate
35:30 than that was. And I also it was kind of
35:32 weird. It was sort of play acting and
35:34 and there wasn't there was zero context
35:36 there. Um so that wasn't great. But you
35:39 know
35:40 listen like my dude was like told me
35:43 nothing. He goes, "Oh, you got this
35:45 thing with your eye. Let me scrape some
35:46 cells off and then we'll give you a
35:49 bandage." It was like you know crazy.
35:52 All right. I had chat GPT write a QWR
35:55 formal complaint for my mortgage company
35:58 because they breached an agreement.
36:01 There you go. Perfect. Oh, you know, one
36:04 of the things one of the the other
36:05 things that Chat GPT can be really good
36:08 at is if you're really pissed off at
36:11 someone or or someone did something
36:13 really bad at work and you're all, you
36:16 know, your amygdala is triggered and
36:18 you're like, you just want to cuss at
36:21 them and say bad things. You can say to
36:23 chat GPT, here's what I really want to
36:26 say to this employee. Rewrite that for
36:29 me in a way that is not going to get me
36:31 in trouble with HR and it's really good
36:34 at that. Really good at that. Uh Moshi,
36:37 I'm just trying to help. Exactly.
36:40 Perplexity's definition in part. The
36:43 term is analogous to
36:45 genetic definition
36:51 which but refers to the replication.
36:53 What's that? What's that in reference
36:55 to?
36:59 I had to ask. Oh, perplexity about
37:01 mimemetic. I think I understand better.
37:04 Ah,
37:05 yes.
37:12 Yeah. I live in amig amygdala triggered
37:16 status. I
37:19 know. Did the same for translating my
37:22 wife.
37:25 Oh man. Oh, in the mood. You love You
37:29 love in the
37:36 mood. Oh, man.
37:41 Um, all right.
37:44 Cool. I wrote a scathing email to a
37:47 client once and gave it to Chat GPT to
37:49 clean it up for me. Felt good. Yeah,
37:51 exactly. You can What? What's that
37:54 exercise? You write whatever the nasty
37:56 letter is and then you burn it. Write
37:58 that nasty letter into chat GPT and say,
38:00 you know, don't make me look like an
38:02 [Laughter]
38:06 [ __ ] Oh man. Okay, let's see.
38:11 Perplexity Pro. Let's go play with
38:14 Perplexity. I haven't been here in a
38:17 while, so there's that.
38:21 So, what this is going to look like is I
38:24 don't know what the hell I'm talking
38:32 about. Upgrade to pro. Oh, I I am
38:35 downgraded. I am no longer a Perplexity
38:37 Pro
38:38 member. Um, where are
38:42 agents? Does anybody know? Research
38:45 anything. Oh, they're behind the payw
38:47 wall.
38:49 Uh, how much is it? 20 bucks a
38:58 month. It's been more than a year since
39:00 I got the rabbit. Indeed, it has. Yeah.
39:03 Turn the screen off. I'll I'll turn this
39:05 on for a
39:09 month. Uh,
39:19 What about Gen Spark AI? Gen Spark AI is
39:22 really quite
39:23 impressive. Really quite
39:29 impressive. New payment
39:33 method. New payment method.
39:37 [Music]
39:44 [Music]
39:49 What just happened there? Okay, that
39:52 there we
39:55 [Music]
40:10 go. All right. What do we want to do?
40:14 What do we want to
40:16 do? No, we want your credit card number.
40:19 No, you can't have my credit card
40:21 number. You can't have it. You can't
40:24 have it. Perplexity Pro. Where are
40:28 agents? Voice
40:31 mode. Learn more.
40:35 Plus, invite your team. Discover home.
40:42 Create projects from
40:57 scratch
41:02 here. Yeah, I'm trying to find it. Um,
41:05 you share about LLMs.
41:19 Let's
41:24 see. Why doesn't he know what he's
41:26 doing? I thought this was some sort of
41:28 learning lab. Isn't he the
41:31 professor? Oh, it's
41:35 labs. The Oh, the light bulb.
41:45 build a
41:52 dashboard. All right, so I guess this is
41:54 an
41:57 agent. Well, this should be
41:59 interesting. Labs. Create projects from
42:02 scratch. Enabled. Turn your ideas into
42:05 complete docs, slides, and
42:08 dashboards. Um, create a
42:13 dashboard
42:16 showing the
42:18 evolution
42:23 of
42:26 the
42:27 economy
42:30 of
42:31 Hollywood films.
42:34 from the
42:36 50s
42:39 to this
42:42 year. Be sure to
42:46 include
42:49 milestones,
42:51 trends, and
42:57 interactive elements.
43:02 and and even label the different
43:14 phases. the different phases
43:18 um
43:20 with an
43:23 explanation
43:24 of what
43:27 changed to shift into a new phase. So
43:32 this should be kind of interesting,
43:34 right? Because this is going to go back
43:36 to when cinema was cinema in the ' 50s
43:39 and then and then as we get into the 80s
43:42 90s we get you know shared media VHS's
43:47 DVDs uh HD
43:50 DVDs Netflix
43:54 rightbuster streaming services
43:57 um Netflix kind of services producing
44:00 their own movies and now whatever the
44:02 [ __ ] going on Wow. So, this this could
44:05 be interesting.
44:07 Okay, so some big thing. Don't forget
44:10 Betamax. Let's put that in there.
44:13 Um, don't forget that's a really good
44:17 one.
44:19 [Laughter]
44:30 Betamax. Yeah. All right, here we
44:34 go. Nine minutes left. That's
44:36 interesting. How did How did it come up
44:39 with nine minutes to
44:41 gather? To create a comprehensive
44:44 dashboard illustrating the evolution of
44:46 Hollywood's films, I will first gather
44:48 historical
44:53 data. I need to gather detailed
44:55 information. Hollywood film industry
44:57 economy, studio system decline in the
45:00 60s. There you go. New Hollywood
45:03 movement economics, blockbuster era,
45:05 1970s
45:09 films. So, if you've never seen
45:11 Perplexity
45:13 before, this is kind of what it does. It
45:16 it it's really good at research. And so,
45:19 this new agent thing is not only is it
45:22 going to do the research, it's going to
45:23 put it together into a dashboard. Okay.
45:26 Oh, it only works in 10-minute sprints.
45:28 Got it. Got it. Got it. got it. It time
45:32 boxed
45:33 itself. Well, that's funny because like
45:36 Jen Spark doesn't do that. Manis doesn't
45:38 do that. Like
45:40 Flowth, some guy did something where he
45:43 had it write a book for him. Then he
45:45 said it was working for 15
45:48 hours. It's like, holy
45:50 [ __ ] It's a Microsoft minute though. It
45:53 hasn't been five minutes. Oh yeah. Well,
45:56 you know what it probably did, Brandon?
45:58 It probably said, "Okay, I got 10
45:59 minutes. Let me estimate the tasks and
46:02 then it's probably doing them faster
46:04 than it than it, you know, allotted,
46:06 which, you know, that makes
46:09 sense. Oh, it had COVID in there. That's
46:12 fascinating. I'm so in love with Manis.
46:15 Manis is really good. Oh my god, how
46:17 much did that cost? I don't know. Well,
46:19 Floth is is freeer. You get like you get
46:22 like 20,000 tokens and I don't know how
46:25 many tokens you burn per token
46:32 thing. So I don't know. All right. So
46:35 it's reading all the things. Okay. Now
46:37 it says I'm organizing the collected
46:42 data. Python programming. Can we look at
46:45 this? Yeah.
46:52 Total
46:53 gross. Oh, this is
46:57 cool. Year 1999, digital cinema. Year
47:01 1997, DVD launch. Year 1993, digital
47:05 effects revolution Jurassic Park. Oh,
47:08 this is
47:09 cool. So, there's the code it's writing.
47:13 I'm adding detailed economic data
47:15 including production cost, ticket
47:17 prices, technology adoption,
47:19 international markets, genre
47:23 trends. I can't find much of a use case
47:26 for Manis. Probably because I've got no
47:28 idea how to deal with B2B stuff. Yeah,
47:31 Manis ran for 30 hours on my book and
47:33 website, including eight illustrations
47:35 and a complete website. Amazing. If you
47:39 sign into Manis every day, you get 300
47:41 credits free with a paid account.
47:45 Beautiful. It's my freaking jam. I love
47:50 Manis. I still have my
47:55 Betamax, man. Betamax lovers, they were
47:58 intense. They were intense. I remember
48:01 we were we were working with uh
48:04 GTE um back in the back in the 90s. We
48:07 built their first website and um our
48:10 client, he lived up in Connecticut. He
48:13 was like, "Come visit me. We'll have a
48:14 barbecue." So, we went to his house and
48:17 he had this attic and in the attic was
48:19 it was it was like his man cave with the
48:22 big plasma screen TV. This was before
48:24 big TVs were cheap. And you know, it was
48:27 like one of those $10,000 TVs. And then
48:30 he had a laser disc collection and he
48:33 had like all the laser discs. He was so
48:36 proud of that. Those laser discs are
48:37 like the size of an album. They were
48:40 huge. And he I forget what we watched,
48:42 but he made us watch something. It was
48:44 like Raiders of the Lost Arc or
48:45 something like
48:46 [Laughter]
48:51 that. Ann Murphy's Manis obsessed, too.
48:54 I'm telling you, all you Manis people,
48:56 you should check out GenSpark. Genpark
48:58 is there. It's like a Gen Spark is like
49:01 a more consumer friendly version of
49:03 Manis. It's just a little
49:08 slicker. I'm creating a highly detailed
49:11 and interactive dashboard that
49:12 visualizes the comprehensive data set on
49:15 Hollywood's economic evolution,
49:16 including box office trends, production.
49:18 Okay, that that it said it was doing
49:20 before. Generated file index.html. All
49:23 right, there's some some [ __ ] It's got
49:27 some
49:30 buttons. Building application. So, it's
49:33 building the application now. So, if
49:35 you've wondered what agents
49:37 are, this is them. And and right now,
49:43 agents are going to suck. They're going
49:45 to be like this. In my opinion, one of
49:48 the things I like about Gen Spark is its
49:51 activity when it does stuff is it shows
49:54 you more, right? And Manis is pretty
49:57 good at that. Manis has got the the
49:59 window where you can watch it surf
50:00 websites and then update its to-do list.
50:04 So, it's got some of that stuff. I think
50:06 where these things go in the future is
50:08 it'll be much more visual and updatey
50:11 and maybe it'll play you generated music
50:15 while it's doing it right. I could see
50:18 these these these autonomous agents
50:20 getting much much friendlier. Right now,
50:24 think of these tools kind of like we're
50:26 in early days of DOSs. You know, when
50:29 you had to command line your computers
50:31 to make them work before the Mac came
50:33 out, before Windows came
50:36 out, I had a multiple laser displayer,
50:39 too, but they never worked. Last year, I
50:41 finally got rid of my original plasma
50:43 screen that I had p paid more than what
50:45 my car cost. Yeah,
50:47 [Laughter]
50:49 exactly. That's awesome.
50:54 Uh uh Rick McCauley, refer a friend and
50:57 get 500 credits. Meet Manis. All right,
51:00 there you go. So, click on Rick McCauy's
51:02 little uh little link there. Join Manis
51:05 and you'll get some free things and I
51:08 assume he'll get 50 bucks or
51:10 something.
51:13 U yeper. So is the disc for Dragon's
51:16 Lair.
51:17 [Music]
51:27 Four minutes left. It's been on four
51:29 minutes left for about five minutes
51:33 now. So it it generated
51:38 index.html. Load more. 11 kilobytes
51:42 total. All right. There's our
51:44 index.html.
51:47 Here's our CSS
51:51 stylesheets. Load more.
51:54 26K. This is This is very slow
51:57 actually. This perplexity thing. I'm
52:00 guessing it's going to be good because
52:01 people are talking about it like it's
52:03 good. That's because you have nine
52:05 minutes to go. Yeah,
52:10 exactly. Where's the link for Manis?
52:12 It's over on the YouTube channel. So, if
52:15 you're not on YouTube, go to um uh
52:22 YouTubearninglab-ai and that'll take you
52:24 to the YouTube channel. And
52:26 uh and Rick McCauley put up a link to
52:35 Manis. Oh, he put it on LinkedIn as
52:38 well. All right,
52:39 cool. Beautiful. Beautiful. Beautiful.
52:42 Beautiful. Hey, Spin B3. What's
52:44 happening? We're shaking. What's going
52:47 down? What is happening? What is
52:50 happening? What's going down there,
52:54 people? Ah, it created a JavaScript
53:00 app. Data from JSON starting in 1997.
53:12 Does it go back to the
53:14 50s? Oh, yeah. It's all over the
53:18 place,
53:24 huh?
53:32 Wild. Look at all that [ __ ] code it
53:35 writes. Still still blows my mind.
53:42 Perplexity is great because of the
53:44 artifact of the page and the spaces
53:47 which give you a shared
53:49 workspace and upload 20 files as
53:52 grounded
53:54 documents. Yeah. You know what's funny?
53:56 All these tools are getting they're all
53:58 sort of settling into their own niches
54:00 where people trust them for certain
54:02 things.
54:04 Like I don't use I don't really use
54:07 Claude anymore. I used to use Claude a
54:09 decent amount. I don't really use it
54:11 anymore
54:13 because chat GPT's got my memory, right?
54:16 It it chat GPT is remember remembering
54:19 what we talk about. And so every time I
54:21 go over to Claude, I'm losing that in
54:23 chat GPT. there there's an
54:27 opportunity there's an opportunity for
54:29 someone to create like a whether it's
54:32 blockchain enabled or tokenized or just
54:35 a regular
54:36 database but create some sort of plugin
54:40 like a like this could be done actually
54:42 you could probably do this this weekend
54:44 or next weekend is just make a a Chrome
54:48 plugin
54:50 that that tracks all your conversations
54:53 in the diff in the different LLMs and
54:56 then somehow provides memory. I don't
54:59 know. It's probably a bad
55:01 [Laughter]
55:05 idea. Oh, there you go. Archetypal
55:08 Architect. Yeah, I kept running out of
55:10 context room and Claude with a paid
55:11 account that earned them an
55:18 unsubscribe. Yeah. Yeah. V Vicki is our
55:21 is our institutional
55:23 memory. our surrogate brain. Isn't that
55:26 called icky guy? Yeah. Yeah, pretty
55:31 much. Does perplexity require
55:35 you to not browse away? No, I'm just
55:38 sitting here looking at it just trying
55:40 to see if I like it. So, is this it?
55:46 Okay. So, here's our Is this a
55:49 dashboard? Is this interactive?
55:54 the studio. Oh, this is cool. But can we
55:58 What's this
56:00 button? Stop generating response. Oh,
56:02 it's still
56:12 going. All right, it's done now.
56:23 Okay.
56:34 App. Can I Oh, here we go. I can pop
56:37 that out as an app. All right. So,
56:39 here's the app that it made for us. This
56:41 is pretty slick. Tik Tok pin. Hey Kyle,
56:45 I created an amazing one minute film
56:47 with chat cling and cap cut feature film
56:50 quality. Yeah, amazing, isn't it? Like
56:55 like again we go back to our job now
56:59 becomes what's the story you want to
57:01 tell, right? What's the business you
57:02 want to start? What's the what do you
57:04 want to learn?
57:06 Um, all right. So, let's look at this.
57:09 Hollywood economics evolution from
57:11 studio system to streaming era 1950 to
57:15 2025. So, that was wait YouTube
57:18 spotlight. Wait, wait, wait. Mary
57:20 Mouseie, I literally just asked Manis to
57:22 make me a Chrome plugin that tracks all
57:24 of your different conversations in the
57:26 various LLMs and curate them in one
57:29 place and it's already working on it.
57:30 There you go.
57:33 I've been a huge fan of Gemini because
57:34 I'm a visual person and deep research
57:36 and I do it to create an app for it.
57:39 Very
57:40 cool.
57:41 Um, all right. 2023 box office 8.9
57:46 billion average budget 140 million
57:49 streaming share
57:51 80%. So the economic phases
58:02 Blockbuster era. Oh, I see what it did.
58:04 So, here's the Blockbuster era up
58:10 until
58:12 1990,
58:14 1989 film Batman era,
58:19 blockbuster home video starting in 1990
58:23 going until 1999. This is cool. Digital.
58:27 And then franchise area era. And then
58:30 streaming era is this anomaly over here.
58:34 This is
58:39 streaming.
58:41 Huh. It put commas in the in the years,
58:45 which is
58:47 weird. Studio system
58:50 transition. Blockbuster era. Home video
58:53 digital franchise
58:57 streaming production costs over time top
59:00 budget average
59:02 budget. So that's fascinating. The
59:05 disparity between, you know, high budget
59:08 films and average films is
59:10 great. Technology adoption. Here we go.
59:14 Betamax. So, Betamax goes from
59:18 198 something
59:20 until 1990 and then dies. VHS goes
59:28 big. IMAX has just been cruising along
59:31 at low
59:32 [Laughter]
59:36 adoption. Major
59:39 milestones. Jaws and Betamax launch.
59:42 Star Wars Jurassic Park Netflix
59:44 streaming COVID
59:48 pandemic. All
59:50 right, there you
59:55 go. And let me go over let me jump over
59:58 to
1:00:00 GenSpark because GenSpark has a thing
1:00:02 called slides which you can now turn
1:00:04 into
1:00:05 um get credits for free. Oh, look.
1:00:12 copy here. I'm going to put this into uh
1:00:16 into YouTube. There you go. Click on my
1:00:19 code and I'll get credits for
1:00:20 [Laughter]
1:00:24 free. Oh, wait. My invite. Is that mine?
1:00:29 No friends invited yet. All right.
1:00:31 Anyway, back to GenSpark. Okay. AI
1:00:33 slides, AI sheets, download for me.
1:00:36 Download's pretty cool.
1:00:43 Um, so let's do the let's do the same
1:00:46 thing. Let me go back
1:00:48 to perplexity. Let me grab the original
1:00:53 prompt if I can find it. Black bar.
1:01:00 Sorry.
1:01:02 Perplexity. Come
1:01:06 on. Create a
1:01:09 dashboard. Don't forget Bay. Okay. So,
1:01:12 here's my prompt. So, let's go to
1:01:15 GenSpark. Tik Tok pin. Studying salon
1:01:19 and measuring quick uptick in AI
1:01:22 skill of members would be a great topic.
1:01:26 Oh, that's interesting.
1:01:30 Yeah, I don't quite know how we do that
1:01:31 because we haven't really established
1:01:33 any benchmarks, but it might be
1:01:34 something that might be an interesting
1:01:36 project to do. So, we're going to do AI
1:01:39 slides slides request
1:01:45 here. So, we'll instead of say create a
1:01:47 dashboard, let's say create a
1:01:50 presentation showing the evolution of
1:01:52 the economy of Hollywood. Okay, here we
1:01:54 go. So, same
1:01:57 thing. And let's just watch what this
1:01:59 bad boy does. New
1:02:03 update. More more text controls
1:02:08 available. Image customization
1:02:10 available. More editing available. All
1:02:13 right.
1:02:14 [Music]
1:02:17 Cool. Using tools. See, this is kind of
1:02:20 cool. Like using tools.
1:02:22 Search and then you can view what the
1:02:24 tool's doing.
1:02:27 So, it's going to finding [ __ ] using
1:02:29 tool search. Using
1:02:31 tool
1:02:34 search Blockbuster
1:02:39 era. Thanks for your patience, Kyle.
1:02:42 I've done research. Like, this is just
1:02:46 more like consumer friendly. Thanks for
1:02:49 your patience, Kyle. I've done research
1:02:51 and now I'm ready to create your dynamic
1:02:53 presentation. AI slides mode.
1:02:56 using tool
1:02:59 presentation. Let me initialize our
1:03:03 presentation. Seven
1:03:05 slides. Now I'll create our slides about
1:03:07 Hollywood's economic
1:03:13 revolution using tool image search.
1:03:16 Hollywood's golden
1:03:20 age. Image search. Hollywood's golden
1:03:23 age. Ah, there's Dorothy.
1:03:28 Yay. Thanks for the status update. See,
1:03:31 this starts to feel more like what it's
1:03:33 going to feel like in the future, right?
1:03:35 These tools are going to be much more
1:03:39 friendly once once the once the once the
1:03:43 core labs figure out their models and
1:03:46 the agents and what they can do, then
1:03:49 startups are going to create what are
1:03:51 going to be jokingly referred to as
1:03:53 wrapper
1:03:54 apps. But what they're just going to be
1:03:56 is applications that just do friendlier
1:03:59 and friendlier work for us.
1:04:03 If you put this into Claude and ask it
1:04:04 to make an interactive web page with the
1:04:06 results, it makes a sharable page that's
1:04:08 better than Gen Spark slides, I have
1:04:09 found. Yeah, I might do that. I might
1:04:11 that might be fun is just to go over put
1:04:13 the exact same prompt into Claude and
1:04:16 watch what it does. In fact, we can do
1:04:17 that right now while that thing's
1:04:20 writing the
1:04:21 presentation. Can I watch this?
1:04:25 No. All right. So, let's go to Claude.
1:04:27 Cloud. Let's go to Cloud. Hello, Cloud.
1:04:31 Hello cloud how are you? Yes cloud very
1:04:35 good to see you. The evolution of AI
1:04:38 learning lab. First question asked sir
1:04:40 what are your qualifications? What pray?
1:04:43 May I ask all your
1:04:45 qualifications? Ah as suspected he has
1:04:49 none. Artifacts are userenerated and may
1:04:52 contain. Okay. So we're going to go
1:04:54 paste. Make a dashboard showing the
1:04:56 evolution of film. We're gonna use
1:04:59 claude opus
1:05:01 for I'm not gonna use sonnet. We're
1:05:04 gonna use opus. If you don't know,
1:05:06 there's there's opus is the biggie.
1:05:09 Sonnet is the middle one. And then
1:05:11 unable to verify. Hang on. I'm doing
1:05:13 physical dexterity on
1:05:15 TikTok.
1:05:17 Um, opus sonnet and
1:05:22 uh what was it called? What's the third
1:05:24 one called? The small one. I forget.
1:05:31 Hey Kyle, are you able to put in
1:05:32 demarcated files and turn them into a
1:05:35 slide presentation? I think you I think
1:05:37 you can. Yes. All right, let's let
1:05:40 Claude do its thing now. Oh,
1:05:43 wait. I should have turned on Let me
1:05:45 stop
1:05:46 this. Stop.
1:05:50 Stop. Paste. I'm going to turn on deep
1:05:53 research.
1:05:57 Web search is on drive
1:05:59 search. All right. Extended thinking. I
1:06:03 don't think it needs extended thinking,
1:06:05 but that's okay. Let's just let's just
1:06:08 turn that bad boy loose. All
1:06:12 right. So, this is now on slide
1:06:16 two. Let me close
1:06:19 that. Oh, okay. So, here's So, now it's
1:06:21 writing the slides.
1:06:24 Oh, look. It did it did like Hollywood
1:06:26 studio stuff as the backgrounds. This is
1:06:29 cool. The studio system collapsed.
1:06:32 That's kind of
1:06:35 cool. So, here it's doing thinking on
1:06:38 the slide. Now, it's coding the slide.
1:06:40 And then there'll be a preview of the
1:06:42 slide. This this is this is cool the way
1:06:44 this thing works, right? This is Gen
1:06:46 Spark. It's just this is this is done by
1:06:50 a team that understands user interface.
1:06:54 Whereas
1:06:57 perplexity perplexity has always been a
1:06:59 little confusing to me. It's cool, but
1:07:01 it's been
1:07:03 confusing. Excuse
1:07:17 me. Need to ask it.
1:07:23 Well, I did I set a
1:07:26 dashboard. Create a
1:07:28 dashboard. I'll create an interactive
1:07:30 dashboard. I think we're okay there. I
1:07:33 think it it's going to be
1:07:38 good. Now, one of the nice things about
1:07:41 GenSpark Slidemaker
1:07:46 is they just said you can now export
1:07:49 these as um Google Slides presentation.
1:07:53 So, you can basically turn them into
1:08:03 PowerPoints. Ah, haiku. That was it.
1:08:05 Opus Sonnet and Haiku.
1:08:11 Trying to use Manis to spin up an entire
1:08:14 music video for one of my new songs. Oh,
1:08:16 yeah. Manis can now do image and video
1:08:18 generation. Can it? Should we go do
1:08:20 something with Manis? Yeah, let's do
1:08:23 that. I got an idea. Manis.
1:08:26 Okay, here's what we're going to do with
1:08:28 Manis. This will be fun.
1:08:32 [Music]
1:08:36 I want you to
1:08:43 first create a
1:08:46 dashboard. Then I
1:08:49 want you to create a a
1:08:55 a picture book.
1:09:01 Let's say create an
1:09:07 interactive picture
1:09:10 book of the
1:09:14 history of
1:09:16 Hollywood based on your research.
1:09:23 The book should have
1:09:27 original images and video
1:09:31 clips and
1:09:35 um
1:09:37 interactive charts and
1:09:40 graphs. I don't know. I don't know what
1:09:42 that means and if it's if it's even
1:09:45 possible, but off it goes.
1:09:49 Go, Mannis.
1:09:52 Go. Here, we'll close up. We'll close up
1:09:55 that little
1:09:57 window. Wait, why did
1:10:01 uh That's weird. Okay,
1:10:11 whatever. Here's Manis's computer.
1:10:17 So, there's its to-do
1:10:25 list. Why can we
1:10:29 not see
1:10:44 both? All right. All right. Well,
1:10:46 whatever. Here. Oh, here's Man. Oh,
1:10:48 there we go. Okay, now I get it. Hang
1:10:50 on. It's start it's it's coming back to
1:10:53 me. I'm starting to remember this
1:10:58 now. Yeah.
1:11:00 Okay. All right. So, it's thinking over
1:11:03 here on the left. Tik Tok pin important.
1:11:05 So, would you skip using SMTG SMTG? Oh,
1:11:10 something like chat GBT or lovable when
1:11:13 building an app and instead
1:11:17 use Gen Sparker
1:11:22 [Applause]
1:11:23 Manis.
1:11:26 Well, so within Chat GPT, you can turn
1:11:29 on deep research. You can tell it to
1:11:32 create an interactive dashboard in in
1:11:36 uh in its pages. What's it called? What
1:11:39 are they called in chat GPT? Now I'm now
1:11:42 I'm all confused. They're not artifacts.
1:11:43 Artifacts are
1:11:45 clawed. What's the thing called in chat
1:11:48 GPT
1:11:50 project?
1:11:53 Um
1:11:56 okay. Um it's an excellent question.
1:12:01 When do you use one of these more
1:12:04 complicated autonomous agents that can
1:12:06 use tools versus when do you use a
1:12:10 simple chat GPT query versus when do you
1:12:12 use something like deep research
1:12:16 versus I don't
1:12:18 know like I don't know
1:12:22 the the state of AI right now is that
1:12:26 there are so many tools that have such
1:12:28 deep capability
1:12:33 I would just kind
1:12:35 of play around with them long enough to
1:12:38 figure out which one kind of feels like
1:12:40 you like
1:12:41 it. And depending on the problem you
1:12:44 want to solve, like if the problem you
1:12:45 want to solve is I want to write a
1:12:47 book, you could just sit in chat GPT and
1:12:50 just do research in there and just write
1:12:52 your book and help it organize things
1:12:54 and be an editor for you and
1:12:58 right if you want to do like major
1:13:00 research projects or major applications,
1:13:03 I think understanding all of these
1:13:06 different agentic systems is like you
1:13:09 should at least understand what's
1:13:10 possible. But it's like I I kind of feel
1:13:13 like we're getting to this place where
1:13:15 the the the tools that people are going
1:13:18 to choose are going to be very very
1:13:19 personal to them. And it's not going to
1:13:22 be like Manis is better than Gen Spark
1:13:26 is better than Perplexity Agents is
1:13:29 better than Chat GPT's version of it is
1:13:32 better than Gemini's version of it. It's
1:13:34 going to be like which one do you kind
1:13:35 of imprint on? Which one feels right to
1:13:38 you? and then that's the one you'll
1:13:40 likely go use. So, I I don't have a good
1:13:43 answer for that one. It's a great
1:13:47 question.
1:13:49 Um, all right. Here's the box office
1:13:52 revenue trends. I guess that was
1:13:57 CO. That's kind of crazy. Look at that
1:13:59 box office dip. Co
1:14:03 bang. Back to like 1965 level box
1:14:07 office. Wow. That's pretty
1:14:10 bad. It's come roaring back. Key
1:14:13 milestones. 1953
1:14:15 Cinemascope widescreen format to compete
1:14:18 with TV. 1960
1:14:21 Psycho 75 Jaws first summer
1:14:26 blockbuster. Yeah, this isn't very
1:14:28 interactive, is
1:14:33 it? Well, let's tell it that. Say, hey,
1:14:36 listen.
1:14:39 This is
1:14:41 fine,
1:14:43 but I want the
1:14:48 dashboard to be very interactive.
1:14:54 Start over and
1:15:00 think
1:15:02 deeply about what
1:15:05 will be
1:15:08 educational and
1:15:14 entertaining. And then let me go say
1:15:18 extended thinking is on. All right, here
1:15:21 we go. Bang. Off it goes. All right,
1:15:24 let's go back to Manis. See how Manis is
1:15:26 doing. What's it doing? Editing the
1:15:28 Hollywood Project. Hollywood milestones
1:15:31 trends
1:15:32 thing. All right, it's still working.
1:15:35 Pan the kick. Tik Tok camera left. Yes,
1:15:38 sir. There we
1:15:40 go. All right, we're halfway through our
1:15:42 list or so or almost halfway through our
1:15:44 list on Manis. What's Gen Where's
1:15:47 GenSpark? Genpark is still coding
1:15:49 slides.
1:15:51 One, two, three, four, five. We're on
1:15:55 slide six. Oh, six of
1:15:58 six. All right. So, this is almost done.
1:16:02 So, again,
1:16:07 like, you know, we were at a place for a
1:16:10 while where you could say, "Make me a
1:16:13 presentation." And and it would just use
1:16:15 chat GPT to make a bunch of slides. Now
1:16:18 they're doing things that are more
1:16:19 sophisticated like going off and doing
1:16:22 actual research, compiling that
1:16:24 research, finding images to use as the
1:16:27 backgrounds of slides, actually coding
1:16:31 the slides so that they are designed and
1:16:35 visual and
1:16:36 beautiful. All right, so we're done. So
1:16:38 we've got this thing. So
1:16:43 now save. All right, what's it doing?
1:16:49 Oh, seven of seven.
1:16:51 Okay, let's finish up our presentation
1:16:54 with the final slide that brings us to
1:16:55 the present day covering digital
1:16:57 disruption and streaming war era. Oh, I
1:17:01 I know what I'm going to do. We're going
1:17:02 to we're going to have we're going to
1:17:04 have Jen Spark add one more
1:17:07 slide. We're going to have it add a
1:17:09 speculation slide. Claude's off coding
1:17:11 the more interactive version of its
1:17:13 thing. Um, Manis is still
1:17:17 thinking. So, if you haven't played with
1:17:19 these tools, it's fascinating to watch
1:17:21 them. Again, the interfaces on these
1:17:23 things are pretty shitty right now. So,
1:17:25 you got to click around to understand
1:17:27 what's going on, but once you kind of
1:17:29 get your head around it, it's it's
1:17:31 pretty slick. Genpark has file upload
1:17:35 limit of two megabytes. Yeah, I know
1:17:36 that's a
1:17:37 drag. Create a visually appealing modern
1:17:40 web-based training guide that's easy to
1:17:42 share and navigate. Yeah, that's pretty
1:17:44 good. That's good. Good little prompt.
1:17:46 Uh oh. What just happened
1:17:47 here? Oh, it's I don't know. It's doing
1:17:50 its thing. Okay. Um Gen Spark. Let's let
1:17:54 that finish that last
1:17:56 slide. I want to I want to do a
1:17:59 speculative slide at the end of this,
1:18:01 which is how is generative AI going to
1:18:04 impact Hollywood?
1:18:19 Oh, this looks
1:18:27 better. Oh, this is
1:18:35 cool. Journey through the evolution of
1:18:37 Hollywood movie money.
1:18:44 Oh, this is super cool. Okay, so 1950
1:18:47 the format was all
1:18:49 theatrical. MGM was the number one
1:18:53 studio. Number two was
1:18:55 Paramount. 20th Century Fox, Warner
1:18:57 Brothers,
1:18:59 RKO. A movie cost 46 cents. Test your
1:19:03 knowledge. What year did VHS defeat
1:19:06 Betamax? 1990.
1:19:09 Did I get it right? Oh, it didn't tell
1:19:12 me. Theater count 44 thou 45,000. Ticket
1:19:16 price 46
1:19:19 cents. All right. So, now let's go
1:19:22 to
1:19:25 1972. No, let's go to 19. Here we go.
1:19:30 1979. Betamax
1:19:33 VHS watch format
1:19:36 war. This is cool.
1:19:46 Tik Tok pin AI is going to be the age of
1:19:50 creative renaissance. Yeah, I'm calling
1:19:52 the the era that we're moving into the
1:19:56 um the great renaissance. And I don't
1:19:58 know if you know this, but in the word
1:20:00 renaissance are the letters AI right
1:20:02 next to each other.
1:20:10 All right. 1992, Betamax is gone. So,
1:20:15 when did Betamax
1:20:17 die? VHS wins.
1:20:20 1988 Betamax. This is good. Look at
1:20:22 this. This little bouncing thing. All
1:20:24 right, let me publish this. This is
1:20:25 pretty slick. Publish and copy link. So,
1:20:29 if you're on YouTube,
1:20:36 You can go you can go there and and for
1:20:38 those of you on the Tik Tok go over to
1:20:41 the AI
1:20:47 salon and let's
1:20:54 see let me put it
1:20:57 in water cooler for
1:20:59 now. We're going to talk about the AI
1:21:02 salon mastermind in a bit.
1:21:05 But in water cooler, I'm just going to
1:21:07 drop
1:21:11 in. We'll call this um
1:21:16 Hollywood
1:21:19 Hollywood
1:21:22 money
1:21:24 history
1:21:28 dashboard. All right. So you can go into
1:21:32 uh AI salon and go to water cooler and
1:21:36 you can get access to this cloud
1:21:41 application. All right. So copy link and
1:21:44 then let's go to it. Let's go look at it
1:21:46 as a standalone
1:21:52 app. Yeah, this is
1:21:55 slick. You can jump 10 years ahead.
1:22:02 Psycho, $800,000 budget, 60 million
1:22:13 gross. I feel like it should anchor the
1:22:16 the year at the
1:22:18 top. Oh, that's cool. It's got a little
1:22:21 play
1:22:24 button. Whoa. Look at the look at the
1:22:27 streaming. The the the bottom one is
1:22:30 streaming. Look at talk about
1:22:33 existential crisis. Okay. Okay.
1:22:36 Actually, this is this is really good.
1:22:38 This is actually very illustrative.
1:22:41 So if you go back to
1:22:44 1950s up
1:22:49 until
1:22:53 70
1:22:55 five when Betamax comes
1:22:58 out. So from 1950 to 1975, 25 years it
1:23:03 was just theatrical and then this thing
1:23:06 called called Betamax comes out and then
1:23:08 VHS comes out and notice how they're
1:23:11 eating into the share of movies. And
1:23:14 remember when all the movie people were
1:23:15 like they're copying our movies onto
1:23:17 these VHS tapes, they're stealing our
1:23:21 data. And then what the hell
1:23:25 happened? Oh yeah, the home video, the
1:23:28 format war 1980. Look at look at how
1:23:31 theatrical dropped from 80% to
1:23:36 40%. And then theatrical and home video
1:23:41 are now
1:23:42 even,
1:23:44 right? If you're in the movie business,
1:23:46 this is existential. This is the kind of
1:23:48 [ __ ] every business is going to be
1:23:50 confronting with all this AI [ __ ] So
1:23:53 their DVDs came
1:23:55 out. Theatrical drops to
1:24:00 25%. DVDs go to
1:24:03 35%. Now look, streaming comes in
1:24:08 in
1:24:11 2007, right? It's just this little blip.
1:24:15 Everyone's making fun of
1:24:17 it. 25% 35% 5%.
1:24:22 And now watch. B, right, DVDs go
1:24:27 away, right? They disappear
1:24:31 completely. No, they're still there.
1:24:33 They're 5%. Theatrical is at 20%.
1:24:37 Streaming is at
1:24:40 60%. That's That's radical. And And you
1:24:44 know what's going to happen with
1:24:46 uh with what you call it with you know
1:24:49 what I'm
1:24:50 saying? All
1:24:53 right. Disruption and streaming wars.
1:24:56 Okay. Okay. I want you
1:24:59 to add a
1:25:03 speculative
1:25:05 slide. Do
1:25:07 research on the
1:25:10 current state of AI
1:25:14 video
1:25:19 tools.
1:25:22 including
1:25:26 VO3. Um, current. What's up, champ? You
1:25:30 want cheese? Including V3.
1:25:33 And what? Create a
1:25:38 speculative slide. Oh, wait. and create
1:25:44 a prediction of how AI will
1:25:49 impact
1:25:51 [Music]
1:25:52 Holly would
1:25:59 um
1:26:01 predict the
1:26:05 impact
1:26:08 of democratizing
1:26:15 and
1:26:18 hyper personalizing
1:26:22 um movie
1:26:25 making. All
1:26:29 right. What are we doing today? We're
1:26:32 playing. We're playing. Uh uh uh AI
1:26:38 media should be the newcomer
1:26:41 category.
1:26:45 Yeah. Oh, go to 2020 and see
1:26:49 if if it changed theatrical
1:26:54 20 post pandemic Hollywood. So
1:26:58 2019 25% 2020
1:27:02 20% 2023
1:27:05 20%. So it basically knocked it down 5%.
1:27:09 I mean, there was that big dip, but it
1:27:10 kind of came back, I
1:27:13 guess. But yeah,
1:27:15 20% studio power rankings, Disney,
1:27:18 Warner Brothers, Universal, Sony,
1:27:20 Netflix number
1:27:28 five, huh? All right. Very cool. That's
1:27:31 that one. Uh, Gen Spark still doing its
1:27:34 thing, right?
1:27:39 What do you want? Winer. What do you
1:27:41 want?
1:27:52 [Music]
1:27:58 [Music]
1:28:09 [Music]
1:28:11 All right. What has Manis done? Where
1:28:14 are we with
1:28:16 Manis? Updating the plan.
1:28:25 diff
1:28:31 original. I don't know if you knew this
1:28:33 in Manis, but you can this they call it
1:28:37 Manis's computer. You can scroll back
1:28:40 through
1:28:41 it. So, as it's doing its thing, you can
1:28:45 kind of look at what it did. And then
1:28:47 you can actually save this as a movie
1:28:48 and play it as a movie. The algorithm
1:28:51 will s will suggest show ideas and then
1:28:54 generate them as you're wa
1:28:57 watching starf fighter. That's
1:29:01 cool. Yeah. Yeah. We're going to get to
1:29:04 the point where this is going to be
1:29:07 um movies will be completely dynamic and
1:29:12 hyperpersonalized. Tony
1:29:16 Casantino. There's also something weird
1:29:18 is it doesn't add up to 100%. Other
1:29:20 formats not in the graph. Yeah, it's
1:29:22 like that that thing's not perfect. The
1:29:24 CL the claud thing's not perfect. Again,
1:29:26 I'm just kind of doing these as one shot
1:29:29 oneshot things. I'm not doing any
1:29:30 refinement to them. If you were going to
1:29:32 actually publish that, well, I know I
1:29:35 published it, but if you were going to
1:29:36 publish it and, you know, mean it, you'd
1:29:39 go in and and fix all that stuff and it
1:29:42 like it could fix it all. It's just it
1:29:43 got lazy and and [ __ ] up. That's these
1:29:46 again, the tools right
1:29:49 now are very easy to demonize, right?
1:29:53 It's very easy to go, well, it didn't
1:29:55 get this right, didn't get that right,
1:29:56 didn't get that right. And people are
1:29:59 using AI's imperfections as a reason to
1:30:03 stay on the sidelines. And I think
1:30:05 that's the biggest [ __ ] mistake you
1:30:06 could do. I I think like right
1:30:10 now take whatever energy you have and
1:30:13 put it into learning everything you can
1:30:14 about AI. It doesn't matter anymore what
1:30:17 tool you pick. Just pick something and
1:30:19 just go play with it. Um and just get
1:30:22 curious about this stuff and understand
1:30:24 what these things make possible because
1:30:26 they're just going to get better and
1:30:27 better and better. I mean, they're
1:30:29 they're these things are getting to the
1:30:31 point now that we're, you know, we're in
1:30:34 the 85 90% range of like as good as it
1:30:38 should be. You know, it used to be 75 80
1:30:43 and it was like 80 85. Now we're like 85
1:30:45 90. That last 10%'s a [ __ ] but it's
1:30:48 not going to be that long until these
1:30:50 things just generate [ __ ]
1:30:53 perfection. And then what do you do,
1:30:55 right?
1:30:57 Okay, so the AI revolution. Okay,
1:31:04 good. Why did it make that slide so
1:31:11 tall? Wow, Kyle, speculative AI slide
1:31:14 came out looking fantastic. As someone
1:31:16 deeply involved at the intersection of
1:31:18 AI and creative storytelling with your
1:31:20 work at StoryVine in the AI learning
1:31:22 lab. Wait, how did it know
1:31:28 that? I I guess I've
1:31:33 done
1:31:34 research. I guess GenSpark has memory
1:31:39 now. That's kind of
1:31:42 surprising. Speculative. As someone
1:31:44 deeply involved at the intersection of
1:31:46 AI and creative storytelling with your
1:31:48 work at StoryVine in the AI learning
1:31:49 lab, I think you'll appreciate how
1:31:51 comprehensive it is. The slide explores
1:31:54 how tools like V3 and Google Flow are
1:31:56 poised to transform Hollywood's economic
1:31:59 landscape through democratization of
1:32:03 production. The cost comparison chart
1:32:05 really tells the story. Traditional film
1:32:07 making at $10,000 plus AI assisted at
1:32:10 500 and fully AI generated at 50 to 250
1:32:14 for a five minute short. The economic
1:32:17 barriers to entry are
1:32:19 collapsing. Hyperpersonalization re uh
1:32:22 revolution. This is cool. Okay. So now
1:32:26 view and
1:32:45 export. The Hollywood economy export
1:32:48 beta PDF
1:32:51 PPX. Publish, play
1:32:55 slides, edit in
1:32:58 Canva. So if
1:33:01 I play
1:33:11 slides, there you go. All right. So this
1:33:16 is this is what GenSpark created.
1:33:20 I don't think you can do any click on
1:33:21 each era to explore.
1:33:28 Oh, nothing
1:33:31 happens. All right, whatever. Streaming
1:33:34 wars. How do we go on? How do we move
1:33:38 on? Oh, it's
1:33:40 broken. Damn
1:33:43 it. Well, whatever. Slide two.
1:33:48 the studio system
1:33:54 collapse. I think that slide I think
1:33:57 this slide was supposed to jump
1:33:59 to each of those other slides and then
1:34:01 it's not doing it. Let's export this as
1:34:03 a as a
1:34:05 PowerPoint. Generating P PowerPoint. Oh,
1:34:08 god damn that thing.
1:34:18 1 to 10 minutes. Uh, okay. It's
1:34:22 exhausting. Everything like Yeah, I
1:34:25 know. It it it did research and designed
1:34:28 and made a whole presentation for me and
1:34:30 I'm annoyed that it's going to take 10
1:34:32 minutes to export
1:34:34 it. Click play slides for a nicer
1:34:36 presentation. Yeah, I did that and it it
1:34:39 they didn't work. Play slides.
1:34:43 And then there's Apple moving my windows
1:34:45 around for me in ways that I don't want.
1:34:47 Thank you,
1:34:49 Apple. So, if I roll over
1:34:53 these, they just they don't
1:35:01 work. Oh, it also didn't rewrite. It
1:35:04 didn't rewrite the first
1:35:07 slide. Okay, this is done now.
1:35:11 So, let's open this in
1:35:19 [Music]
1:35:30 [Music]
1:35:33 Keynote. Open with
1:35:37 Keynote.
1:35:39 Ah, why is it taking so
1:35:45 long?
1:35:51 Okay.
1:35:54 Play. All right, that sort of worked.
1:35:57 Oh, no it
1:36:01 didn't. Oh, Apple. Damn it. Sorry.
1:36:16 Studio system
1:36:26 collapse. This is pretty pretty amazing.
1:36:29 You gota you gotta you got to give it
1:36:32 that. It's a little hard to read,
1:36:39 but the DVD age DVD revenue peaked in
1:36:44 2004. Just been
1:36:47 sliding. Box office staying steady. VHS
1:36:54 revenue. Digital disruption. digital
1:36:58 sales
1:37:00 eclipse DVD sales in
1:37:06 2013 streaming
1:37:10 wars.
1:37:13 Huh, and then the AI revolution. And
1:37:16 again, these slides
1:37:19 are they're bigger than they can be
1:37:21 displayed.
1:37:29 Huh? All
1:37:31 right. Crazy crazy ass stuff,
1:37:35 man. All right. Manis Manis is still
1:37:38 off. Okay. Test picture book
1:37:41 functionality. Finalize and deliver
1:37:43 completed
1:37:45 projects. I'm continuing to develop the
1:37:47 picture book with all the interactive
1:37:49 stuff. Let's see what it did. Did it do
1:37:52 anything fun? Show us pictures that it
1:38:01 made. No. Jump to
1:38:07 live. All right. I've completed both
1:38:10 parts. Interactive
1:38:14 dashboard. Interactive picture book.
1:38:33 Share Manis' creation as an interactive
1:38:36 website. View all files in this
1:38:39 task. Interactive picture book. Here we
1:38:45 go. There's no pictures.
1:38:50 Well, I don't know. I don't know what to
1:38:52 say. I don't know what to say, people.
1:38:55 Manis, I thought Manis now had images
1:38:58 and videos, and I don't see it having
1:39:00 done that, but I could be clueless.
1:39:03 Again, the interface on these things is
1:39:06 kind of a mess right now. So, all right.
1:39:10 So, here's the thing. I'm going to wrap
1:39:12 it up now. Um, what I want you to do is
1:39:16 go on over
1:39:18 to the AI salon. So, go to the
1:39:22 salon.ai and then click on join our
1:39:24 community. If you're not a part of it,
1:39:26 if you are a part of it, just head over
1:39:27 there. But when you go to the A salon,
1:39:32 um we have a new area called the
1:39:34 mastermind, which is a new uh
1:39:39 subscription area within the salon that
1:39:42 includes clubs and hubs and things like
1:39:45 that. Um as well as tool talk, uh the
1:39:50 the showand tell area, things like that.
1:39:52 So this is a new
1:39:56 area. It went
1:39:58 live technically on Sunday, but we had
1:40:01 we had an issue with Stripe not
1:40:03 cooperating. So, we fixed that today.
1:40:05 So, it is now live. So, if you're not
1:40:06 part of the mastermind, please go join.
1:40:08 Um, and uh it's an area it's basically
1:40:11 think of it like a smaller, more focused
1:40:13 area for people to really get to know
1:40:15 one another, collaborate. Um, and uh the
1:40:18 AI salon in general remains free, but
1:40:21 this is a more focused, tightened area.
1:40:24 Um, so go do that.
1:40:26 And
1:40:28 next day, tomorrow, um, is AI salon
1:40:33 presents. So, if you go to the salon, go
1:40:34 to
1:40:35 events
1:40:37 and RSVP for this top
1:40:41 one. And we've got Danny Kravitz, who's
1:40:44 a creative, he's a professor of
1:40:47 screenwriting and does creativity
1:40:50 workshops and things like that. So, he's
1:40:52 speaking tomorrow night at AI Salon
1:40:54 Presents. All right. So go RSVP for
1:40:57 that. Come visit us. Come hang out with
1:40:59 us and uh and we'll do that. All right.
1:41:03 And then yeah, go play with tools like
1:41:06 GenSpark and Manis and and Deep Research
1:41:10 within all like all the tools now have
1:41:12 deep research. They all do some version
1:41:14 of this um of the stuff we did tonight.
1:41:17 I know it doesn't demo super well
1:41:20 because it's just like these things are
1:41:21 just kind of cranking along, but when
1:41:24 it's your project, watching them crank
1:41:26 along is actually quite fascinating.
1:41:28 Like watching it go surf and read and
1:41:31 understand websites and modify its plan
1:41:34 based on what it's learning is is pretty
1:41:37 amazing. All right, so go do that.
1:41:40 tomorrow night. Um, we'll probably be a
1:41:44 little late because we've got the AI
1:41:45 salon until 7 and then I got to come
1:41:47 home and eat and, you know, just do
1:41:50 life. So, it'll probably be 8:30 or 9.
1:41:52 Pro. I'll shoot for 8:30, but it it'll
1:41:54 it'll be one of those two. 8:30 or 9
1:41:56 Mountain time tomorrow night here. All
1:41:59 right. Beautiful. Fantastic. All right.
1:42:03 Well, we got through some stuff. We saw
1:42:05 some Perplexity. We saw some Madness. We
1:42:07 saw some Gen Spark. We saw some
1:42:11 cloud and we had good questions like
1:42:14 well when would you use those tools over
1:42:16 chat GPT and I don't have a good answer
1:42:18 for you when you need something more.
1:42:21 It's a crappy
1:42:26 answer. Oh man. All right. Same project,
1:42:30 three different models. Yeah, exactly.
1:42:32 Are you doing two projects at once? We
1:42:34 did three projects at once. I say Claude
1:42:36 wins. I would say that. Yeah, I I would
1:42:38 say Claude wins, Jen Spark comes in
1:42:41 second, and Manis was kind of
1:42:43 disappointing quite quite honestly. So,
1:42:46 oh man, thanks Rick. I mean, Kyle U for
1:42:51 if we include Perplexity. That's right.
1:42:53 Yeah, we did this started in Perplexity.
1:42:55 So, we did we did four different four
1:42:58 different projects. So, groovy. Um, all
1:43:01 right, everybody. Have yourself a
1:43:02 fantastic night. I will talk to you
1:43:04 later. Bye.