Mar 8, 2024 - How AI is Transforming Tarot and Astrology!

Description

Chapters

Transcript

0:01 [Music]
0:29 oh
0:32 [Music]
0:35 it's not simple to
0:38 say most
0:40 days I don't recognize me these shoes
0:45 and that place and its
0:48 Pat have taken more than I gave
0:53 them it's not easy to
0:56 know I'm not
0:59 anything like I used to be although it's
1:02 true I was never attention sweet
1:07 Center I still remember that
1:10 girl she's
1:12 imperfect but she
1:14 tries she is
1:17 good but she
1:19 lies she is hard on
1:24 herself she is broken and won't ask for
1:28 help she is m
1:31 but she's
1:32 kind she's
1:35 lonely most of the time she is all of
1:39 this mixed up and baked beautiful
1:43 pie she's gone but she used to be
1:48 [Music]
1:51 mine it's not what I ask
1:55 for
1:57 sometimes just through back door
2:00 and carves a person makes you believe
2:04 it's all true right
2:06 chy now I've got
2:11 [Music]
2:29 you
2:32 [Music]
2:40 happy Friday night
2:41 everybody Friday night date
2:44 [Music]
2:49 night what is
2:53 happening it's official Sam Alman is
2:56 back on the board of open
2:59 AI
3:01 he and Greg Brockman
3:03 were
3:06 found to have acted appropriately or not
3:10 inappropriately enough to be
3:12 [Laughter]
3:16 fired not a big surprise but some new
3:20 board members over there that's
3:29 interesting
3:31 [Music]
3:41 [Music]
3:59 Satan was suit and high does it work at
4:01 the D
4:03 Queen does it listen to rock and roll
4:07 does he feed the mean sing
4:11 Hallelujah
4:13 Hallelujah
4:16 Hallelujah
4:20 hujjah what about
4:22 Jesus didn't or two yeah hang out with
4:26 the prostitutes have a drink or two
4:38 [Music]
4:54 [Music]
5:08 [Applause]
5:08 [Music]
5:35 all right let's get this show on the bo
5:38 let's get it going let's get it going
5:39 let do this let do this the puppy vocals
5:42 the puppy vocals is where it's at I
5:45 singing
5:46 backup the dog leads the vocals I'm just
5:51 harmonizing not
5:54 well off key you can make money
5:58 with good evening good evening good
6:01 evening we've got our project for the
6:05 night I think it's going to be a fun
6:07 project this was a cool thing that
6:08 happened last night if you didn't see it
6:11 hello fellow AI en Enthusiast gritty
6:14 wisdom turn it into a regular irregular
6:17 gritty WIS wisdom showing up here a lot
6:19 I like it I like to see it they're
6:22 gritty and wise that's how we like it
6:26 around here gritty and wise buckle up
6:29 buttercup AI ain't for the faint of
6:31 heart I think we're we are we are
6:36 all we are all discovering that that
6:39 this
6:41 is this is a [ __ ] grind and I don't
6:44 mean these lives I mean this is a grind
6:46 too you
6:49 know Friday night honey are we gonna go
6:51 to the movies tonight we're GNA watch
6:53 Tik Tok
6:57 again I'm sorry for doing that to you
7:03 what are you going to do what are you
7:04 going to do huh there's no sleeping in
7:10 AI still crying at
7:15 baseball oh
7:17 man are you
7:21 crying AI should translate dog to text
7:24 wouldn't that be great all right it
7:25 looks like we got some some tapy tap uh
7:28 competition going on so if if you want
7:30 to give yourself carpal tunnel feel free
7:31 to
7:32 tap uh the ambitious of people in this
7:35 channel go bu themselves autot tappers
7:38 on
7:43 Amazon oh
7:46 man gritty wisdom signed up for the AI
7:48 Salon but needs to do my intro gritty
7:51 wisdom you have two hours starting right
7:53 now go write your go write your
7:56 intro it's not like you're going to
7:58 learn anything from me
8:00 [Laughter]
8:04 I feel like this is like you know two
8:07 hours of me rambling and 15 minutes of
8:09 Education every
8:12 night lately I've been rambling I you
8:15 know what's going on is I just I just
8:17 think that there's I think we're in a
8:18 weird [ __ ]
8:21 lull and it just who's I talking to
8:25 today oh Jim Ross Jim Ross from Three
8:28 Mile Storage cuz I we we had AI office
8:32 office hours on LinkedIn today and Jim
8:34 Ross was there I like Jim you know what
8:36 do you what are you innovating an AI he
8:38 was like like I haven't been doing [ __ ]
8:41 for like the past month I don't think he
8:43 said [ __ ] but you know he's a he's a
8:45 polite man he's a professional as
8:47 opposed to my
8:49 myself
8:51 um and I said yeah I'm I'm I'm the same
8:54 way like I there hasn't been much in
8:56 fact the the the tarot card the tarot
9:00 deck reinvention from last night was the
9:03 first thing that I did where I was that
9:06 that I've done anything that has sort of
9:07 lasted longer than the moment I'm doing
9:09 it
9:10 in and that was pretty cool I actually
9:12 had a good interaction with pi this
9:14 morning it well it was actually it was
9:17 good until it wasn't I'll I'll talk
9:18 about that
9:20 too but but what we're going to do
9:22 tonight is I'm going to build a
9:25 GPT and I'm going to build so this is
9:27 going to be we're going to do kind of a
9:30 deep dive and this is going to get
9:32 boring so take a nap um you
9:36 know put on Judge Judy I whatever is
9:40 Judge Judy still
9:43 on how [ __ ] old am I I'm really
9:45 [ __ ]
9:46 old uh things got ex exciting for me I
9:49 demoed a custom co-pilot to the
9:51 marketing team at my work today gritty
9:55 wisdom that's [ __ ]
9:58 huge okay so so so follow gritty
10:01 wisdom's lead here here's here's my uh
10:04 my tried andrue recipe for for
10:12 AI I was I was going to say AI literacy
10:15 but none of us are AI literate right now
10:18 cuz it's just shit's changing too fast
10:21 we we can be more AI
10:28 literate uh
10:32 um
10:34 the the the thing to do the thing to do
10:37 right now is just play with the tools
10:41 just like show up to [ __ ] like this join
10:43 the AI Salon find people that are
10:45 playing with AI and just be in the
10:47 conversation build [ __ ] solve problems
10:49 do all that but the most important thing
10:51 of all this is talk about what you're
10:53 doing talk about it out loud like post
10:57 [ __ ] I I think I went a little far with
11:01 posting [ __ ] about AI on LinkedIn
11:04 because someone that's known me for um I
11:08 don't know 20
11:09 years asked my co-founder the other day
11:13 does Kyle still work for storyvine I'm
11:15 the
11:17 CEO so so so I've maybe been going a
11:22 little overboard with the with the AI
11:24 stuff but that's okay um I would rather
11:26 air on the side of that because I think
11:28 it's actually important and it's funny
11:30 one of our uh one of our board members
11:32 called Monique today my
11:34 co-founder and said I just said sat in
11:37 on this this uh big innovation Summit
11:39 thing and and someone demoed their video
11:42 AI stuff and he goes your stuff's way
11:45 better thanks yeah we know like we know
11:48 it's it doesn't suck we're we've been
11:50 doing this for a while but
11:54 um so so it's good but that that thing
12:00 you're doing like demoing it for people
12:02 at work demoing it you know just talking
12:05 about it at
12:06 work cuz what what'll happen is if
12:09 you're at one of those places where
12:10 you're like I can't even talk about it
12:11 at work cuz it's banned we're not
12:13 allowed to use
12:15 AI talk about what you did on the
12:17 weekend talk about the kids book you
12:19 made talk about the peanut butter cookie
12:21 recipe you invented talk about you use
12:23 Sin City sipper
12:25 and came up with 50 drink recipes for
12:28 your Super Bowl party based on the booze
12:30 you had in your cupboard someone reached
12:32 out to me and told me he did that right
12:34 cool fun
12:35 stuff I keep getting messages from Tik
12:38 Tok to tell me to say hi to you it's
12:40 been a while hi Becky
12:42 Ru well good for Tik Tok look at Tik
12:45 Tock just connecting us like old Pals it
12:49 has been a while but good to see
12:51 you I see you on Tik Tok every now and
12:53 again so I don't feel like and I
12:54 certainly see you in here um but thank
12:57 you for saying hi um
13:00 I don't I don't feel disconnected from
13:02 you um I'm talking with everyone Uber
13:04 drivers this is good this woman works
13:06 I'm talking with everyone Uber drivers
13:08 co-workers Home Depot people my
13:11 doctor and others yes that's that's it
13:15 and and why is that why
13:18 because
13:21 um right now very few people are
13:25 actually using these tools very few it's
13:28 somewhere in the neighborhood if I'm I'm
13:30 going to say at this point it's probably
13:31 like
13:32 75% or
13:35 more somewhere between 75 and 82% it was
13:38 82% back in uh
13:41 November have not used chat
13:44 GPT not only haven't made images and
13:47 haven't done this haven't done that I
13:48 haven't even used chat GPT like write me
13:50 a poem about a wet
13:55 dog
13:58 and
14:03 and these tools are getting way more
14:05 powerful very very quickly did you
14:07 remember your laptop yes I did I am I am
14:10 all set up as if I were a
14:12 professional
14:14 um these tools are getting way more
14:18 powerful way more quickly and and you
14:20 know there's evidence
14:24 that things are starting to happen and
14:27 and one of the big ones for for me was
14:30 when open AI re released the or
14:34 announced Sora their video
14:37 model I think it was three or 4 days
14:39 after that there was a big story in
14:42 variety the the Hollywood
14:45 rag that Tyler Perry cancelled an $800
14:48 million Studio
14:50 extension after seeing Sora after seeing
14:53 that video model I did an Oscar pool
14:56 with AI for work today oh that's that's
14:58 a that's a perfect that's you know what
15:01 Jay Jay Winston Welsh that's a perfect
15:04 use of AI right use it to do the the
15:07 Oscar ballots right perfect use because
15:10 everyone's like oh that's kind of cool
15:11 how did you do that so quickly did you
15:13 download it off the internet no I just
15:14 wrote it up for you I use chat GPT oh
15:18 you can do that oh I didn't know it
15:20 could do I thought it I thought it it
15:22 its latest information was
15:27 1927 um I told my doctor uploaded blood
15:30 work into AI he's fascinated over its
15:32 responses oh that's fascinating I I did
15:35 he did he laugh at it or was he like oh
15:39 that knows as much as I do that's that's
15:41 pretty
15:42 curious um that's super cool so talk
15:47 about this stuff because here's what's
15:48 going to
15:51 happen there's going to
15:53 be I don't
15:56 know there's going to be a dozen more
15:59 high-profile stories or 20 or 50 it
16:03 doesn't really
16:05 matter where
16:08 AI changes something like changes it
16:12 dramatically so for me Tyler Perry was
16:15 the first one that was like the first
16:16 shoe to
16:18 drop and there's like I'm seeing like
16:20 little slivers of like of the 42,000
16:23 Tech workers laid off some high
16:25 percentage is attributed to
16:28 AI you starting to see little things
16:30 like that but what's going to happen is
16:33 it's going to become apparent to a lot
16:35 of
16:36 people oh [ __ ] I got to get my [ __ ]
16:40 together oh [ __ ] ai's coming right we're
16:42 not there
16:44 yet but there's like there's
16:46 Rumblings I can feel I can feel the
16:49 Earth rumbling a bit and I think it's
16:52 going to be this
16:53 year there's there's going to be either
16:56 a GPT 5 drops or like a sexier version
17:00 of Gemini or Sora or just something like
17:05 we don't know Apple you know June is not
17:07 that far away and Apple's going to drop
17:11 a 10 to 100 time improved version of
17:15 Siri the I I saw a tweet yesterday I
17:19 think from the Siri product manager at
17:22 Apple who said shortterm Siri's going to
17:26 get way better longterm Siri's going to
17:29 get a amazing some version of that hell
17:32 I was laying out laying on an ER gurnie
17:35 talking with the nurses about AI a few
17:37 weeks ago Jim Ross had a great
17:41 chat I hope you're okay I saw you today
17:44 so so hey
17:47 Siri go
17:52 away go okay I found this
17:57 okay you can't talk about her she gets
18:00 very jealous you probably already talked
18:02 about Nvidia no what the [ __ ] happened
18:04 with Nvidia which you mean them hitting
18:07 their fourth quarter
18:09 earnings or or their or their CEO saying
18:13 that you don't need to learn coding
18:15 anymore that that telling your kids to
18:18 learn to code is no longer what you
18:20 should be doing which which Nvidia
18:25 story yeah Three Mile Storage is Jim
18:28 Ross I just doubted you Jim
18:30 in in in crypto land I just doxed you
18:33 buddy you're doxed you're out of
18:37 there I'll tell you where he lives tell
18:40 I'll tell you I'll tell you the address
18:41 of everyone is one of his self storage
18:45 units don't look in unit 427 just don't
18:49 just don't go just don't look in unit
18:50 427 I think everybody knows that Jim
18:54 told us today he goes to Vegas once a
18:55 month what do you think that's about
18:57 it's unit
18:58 427
18:59 but we don't talk about that we
19:01 certainly don't talk about that in the
19:02 public forum it's coffee by the
19:06 way yeah it's 10 o'clock at night and
19:08 I'm drinking
19:09 coffee um I chat GPT a marketing plan
19:12 for a Beachside sunglasses vendor today
19:15 that's awesome he couldn't get
19:21 away he's like uh lady I got other
19:24 customers you're like I don't see any
19:25 other customers here look what else it
19:28 does
19:32 oh my goodness goodness goodness wait
19:35 what was that will uh why is AI not been
19:38 shut down due to copyright infringement
19:40 it clearly has trained on others rights
19:44 so so what's this what what you're
19:49 asking Merlin is a great
19:53 question with a not great
19:57 answer um
20:03 when
20:05 AI models are trained when when the the
20:08 stuff goes through the training set it
20:10 ceases to exist in its current form one
20:14 of the misconceptions about chat GPT and
20:17 the image diffusion models and things
20:20 like that is that They're copying and
20:23 pting other people's work and it's it's
20:25 simply not the case um it's more like
20:30 the equivalent of the like the G and GPT
20:32 stands for generative this stuff's being
20:34 generated and the nature of how it
20:37 works when you see something that you
20:41 know resembles something someone else
20:43 did it may resemble it but it resembles
20:46 it kind of like if I went to art school
20:49 for four years and studied nothing but
20:51 pop art and Keith Herring and Andy
20:53 Warhol and all that sort of [ __ ] and I
20:57 came out of art school and all of my
20:58 work was influenced by that it some of
21:01 it would probably look like that it's
21:03 it's more akin to
21:05 that and it there's a another metaphor
21:08 might be you know these things turn into
21:10 what are called vectors they're like
21:12 mathematical points in
21:14 space and then the diffusion models are
21:17 a bunch of Highly noised images that
21:19 were tagged with with words about what
21:22 was on those images when they originally
21:24 created and they exist as sort of these
21:27 noise images and when you a search it
21:30 filters out all the words that are not
21:33 relevant finds images that are tagged
21:35 and it averages them all together in a
21:38 single noised image and just denoises
21:40 that with the same algorithm that noised
21:43 all the original images it literally
21:44 doesn't know what it's
21:47 creating that
21:49 said a lot of these things have been
21:51 trained on training on on stuff that's
21:54 been copywritten at the core of the New
21:57 York Times law lawsuit is you know you
22:00 train up on all of our stuff if you
22:02 listen to Sam alman's response at Davos
22:04 it's really interesting he says the
22:08 reason we were cutting a deal with the
22:09 New York Times wasn't about the training
22:11 data we don't need the New York Times
22:13 for training
22:14 data it was for when people do a search
22:18 we wanted to be able to search the New
22:21 York Times database and pull it up as
22:22 citations and give them a branded thing
22:25 and we wanted to pay for
22:27 that but he said the training day he
22:29 goes there's so much data in these
22:32 models that
22:35 that how did he he the metaphor he used
22:37 was if you're trying to learn High
22:39 School level
22:41 biology after the third
22:44 book the books get less and less
22:46 relevant right if if you've if you've
22:49 learned biology from 2,000 books the
22:53 2001st book isn't all that interesting
22:57 he said the other problem with training
22:59 on the New York Times
23:02 is even if they trained on the New York
23:04 Times data at New York times.com which I
23:07 don't know if they did but even if they
23:09 did how many times have New York Times
23:12 articles been copy and pasted out on the
23:14 Internet or sections of them out into
23:19 articles like
23:22 millions they have no way of stopping
23:24 that right
23:27 so think of think of of the training
23:29 sets of these large language models and
23:32 of the image generation models and
23:34 things like that are more akin to a
23:37 library or a
23:40 museum where we're taking the collective
23:42 output of humanity and putting it in a
23:44 model that we all get to collaborate
23:47 with
23:49 now is that fair to someone it might not
23:54 be um is there a way to easily track if
23:59 if I take this little grain of sugar and
24:03 I toss it in a bowl and put a bunch of
24:05 other sugar in it and flour in it and
24:07 butter and eggs and stuff and I Stir It
24:09 Up and I stick it in the oven and it
24:11 bakes it into a cake and that cake comes
24:14 out can I find that grain of sugar again
24:17 and then how much is that grain of sugar
24:18 worth like that's the kind of minutia
24:21 that we're talking about here so there's
24:23 not a good answer for it
24:26 um the
24:30 the
24:31 um the companies haven't been shut down
24:34 I don't think because it's a very gray
24:36 area and there is there is
24:40 a there is a massive benefit to humanity
24:44 argument here and and there's a really
24:48 difficult technical argument here for
24:51 those you know claiming copyright like
24:54 the the Sarah Silverman all the all the
24:56 people that were suing uh uh mid Journey
25:00 or not mid Journey was it mid
25:03 Journey no it was open Ai
25:06 and Google whoever the hell they were
25:09 suing I think four out of five of their
25:12 cases got thrown out of court already
25:14 because they they can't prove that you
25:16 know you you can you can write a clever
25:19 prompt to get it to give you something
25:21 very close to what Sarah Silverman wrote
25:23 but it's it's not copying and pasting
25:26 from her work and that's relatively easy
25:28 to prove so so I don't know it's listen
25:31 I understand the anger I understand the
25:34 rage what I would strongly encourage you
25:36 to do is get educated on how these
25:39 models
25:40 work and I would you know I would
25:45 get I would get AI literate I'd start
25:47 using
25:48 them because here's the thing that
25:50 you're going to you're going to be able
25:52 to do I'm this I'm fairly confident on
25:54 Adobe is already building this
25:57 system
26:01 what adobe's done which I think is a a
26:04 more responsible way to do
26:07 this is they trained their image models
26:10 on public domain images and images that
26:12 they owned outright for Adobe
26:15 stock and then they're allowing artists
26:18 to opt out or opt
26:21 into being a part of the Adobe training
26:25 set and then if you opt into it they
26:28 figuring out ways to do attribution and
26:31 compensation and I think what that's
26:33 going to look like is rather than trying
26:35 to
26:36 track was your model used on the output
26:40 side which is the way copyright law is
26:42 right now you produce all this work and
26:44 people go you stole my [ __ ] I'm going to
26:46 sue
26:48 you I think what ad adobe's going to do
26:51 is say if you're if you've got an
26:53 official model in the Adobe system you
26:56 get paid on the generation side so if
26:59 someone wants to use a Kyle Shannon
27:01 branded photography
27:03 model for photographs that I took they
27:07 can select my name and say I want to
27:09 create an image based on this model and
27:11 it's it's my official model and and
27:13 Adobe will allow me to have attribution
27:16 and get compensated for that it's not
27:18 there yet but they're talking about that
27:21 and I see that being a really big
27:24 business so rather than fighting on the
27:27 scraps on the output side
27:30 as we're moving into a world of infinite
27:32 content
27:35 generation get educated about how the
27:37 models actually work and get involved in
27:39 projects where you get attribution on
27:41 the input side on the generation side
27:44 isn't the analogy like a doctor they
27:46 read books and they train they make a
27:47 diagnosis chat GPT it's very much like
27:50 that it's
27:52 the these things are modeled after the
27:55 way our brains work they're a much more
27:57 simplified version of the way our brains
27:59 work but they're not dissimilar go
28:02 listen to Jeffrey Hinton who's called
28:04 the grandfather of
28:07 AI his most recent talk he talks about
28:10 this he says he says large language
28:12 models are intelligent in a very similar
28:15 way to the way humans are
28:17 intelligent how do you think AI will
28:19 impact healthc care thoughts I think
28:20 it's going to be I think it's going to
28:23 be
28:24 profound um and and really dramatic but
28:28 I think Healthcare healthare anywhere
28:30 where you've got human lives at stake
28:33 are going to move way slower healthc
28:34 care moves incredibly
28:36 slow so if you're talking about things
28:39 like treatment and virtual nursing and
28:42 virtual therapy and virtual
28:46 um support in that kind of way I think
28:49 that's going to go slower I think you're
28:51 always going to have humans in the loop
28:53 there where where I think healthc care
28:55 is going to accelerate dramatically is
28:57 on the research side
28:59 right so you've got Alpha fold from Deep
29:02 Mind predicted 400,000 protein
29:06 folding things but I don't
29:09 know apparently they take a lot of
29:12 compute power to figure out how the [ __ ]
29:14 proteins are going to fold it's really
29:15 complicated our bodies suck they're
29:19 really complicated and Deep Mind from
29:22 Google they created this thing called
29:24 Alpha fold and it just predicted 400,000
29:26 of these things and just they're proving
29:29 to be right or very very close to it
29:32 very very
29:34 high uh predictability
29:38 score Deep Mind followed that up
29:42 with creating this isn't in healthare
29:45 but it this some of it will will affect
29:47 health
29:49 care about two months ago they put a
29:51 they put out a paper in nature
29:53 magazine they
29:56 invented 2.2 million new
29:59 materials of which 380,000
30:03 were stable crystals which I don't know
30:08 [ __ ] about Material Science but I what
30:10 what I do know is when you say things
30:12 like stable crystals you start thinking
30:14 of things like Battery
30:16 Technology and you know energy
30:19 transmission and you know uh room
30:22 temperature superconductors which by the
30:24 way looks like there were some
30:25 breakthroughs in
30:27 that to put the 380,000
30:30 new stable crystals into
30:34 perspective human
30:36 beings up to that point had discovered
30:41 22,000 and then human beings plus
30:44 computer models computer simulations or
30:47 computer whatever
30:50 analysis discovered another
30:53 44,000 so you had 66,000 known stable
30:57 crystals
30:59 for all of
31:01 humanity and deep mind just dumped
31:04 380,000 onto that World
31:11 overnight so I think I think I think the
31:14 impacts on Health Care like
31:18 once once we can use AI to do things
31:21 like completely decode the genome and
31:25 how how all this [ __ ] works and how
31:27 disease works and how it's vulnerable
31:29 and susceptible once you start to crack
31:32 the nut of single diseases just dominoes
31:36 are going to start to fall is it true
31:38 that you can get thinking images from
31:39 brain scans
31:41 yes there this made the news a lot more
31:44 about 6 months ago but I I assume things
31:46 are still going on so uh down in the
31:49 University of Texas in
31:51 Austin um they were doing fmri scans of
31:56 people that were watching video and then
31:59 they took those scans and recreated the
32:02 video they took scans of brain
32:05 patterns of people watching video and
32:08 they recreated them after training them
32:11 on the tra data set they also did that
32:13 with images I think they did it with
32:15 images in real time they also did um
32:20 there's a guy in
32:21 Italy paraplegic motorcycle accident you
32:25 know lost the use of his
32:27 legs they did an fmri
32:30 scan this this [ __ ]
32:34 insane told him to think about moving
32:36 his legs so they were fmri scanning him
32:39 he' think about moving his legs so they
32:42 trained the model on
32:44 that and then sent the signals from
32:48 electrodes in his brain Bluetooth them
32:51 around his injury to electrodes that
32:54 were in his lower spine connected to the
32:56 nerves in his legs and now he's
33:01 walking so
33:04 yeah I think everything
33:07 from Vision to hearing to brain injury
33:12 to um congenital diseases all that sort
33:16 of stuff is going to be impacted really
33:19 really dramatically um I do think that
33:22 anywhere where you've got human lives at
33:25 stake it's going to move slower right so
33:27 airplane
33:29 you know Air Traffic Control uh we're
33:32 going to want humans in the room doctors
33:34 you know surgeries I mean they've had
33:36 robotic surgery now for since I was
33:39 young right 20 30 40 years um now with
33:44 AI assists it's getting really really
33:46 really good you know if you look at what
33:48 they're doing with drones drones flying
33:51 around are getting really good so a AI
33:53 assisted robotic surgery is going to be
33:55 you know going to be big massive so
33:59 yeah so I think it's going to be big
34:00 here's here's another thing where I
34:01 think healthcare is going to be
34:03 dramatically
34:04 impacted so Google when they when they
34:08 release Gemini they release three models
34:10 Gemini Ultra which is the Big Daddy we
34:12 haven't really we we can play with it
34:14 now but it's not the it's not the fancy
34:16 one Google pro and Google Nano Google
34:19 Nano was is designed to
34:22 live on to to function on mobile devices
34:29 without an internet
34:31 connection um
34:34 anthropic Claude they just released
34:37 three different models of Claude the
34:39 smallest one is called
34:41 Haiku that's designed to run locally on
34:44 on not super powerful devices can't wait
34:48 for AI to help to defeat cancer no
34:51 [ __ ] I I'm I'm in the uh the uh lime
34:55 disease Camp I'm ready for it to cure
34:57 [ __ ] lime disease I
34:59 have two two sons upstairs that are
35:02 [ __ ] just devastated by that [ __ ]
35:04 disease cuz it's not one disease it's
35:07 like
35:09 12 so yeah
35:13 yeah um oh but anyway so and and then
35:17 Apple all of their new silicon the M1 M2
35:21 and M3
35:23 chips all have pretty significant
35:26 machine learning capabil abilities and
35:28 they're developing the software to be
35:30 able to run [ __ ] on it so I think what
35:33 you're going to see is Siri is likely
35:36 going to get really really good my guess
35:39 is a lot of her processing is going to
35:41 happen locally on your
35:43 device so imagine having something that
35:46 as capable as gpt3 maybe even gp4 that
35:50 can just run locally on a phone without
35:53 connection to the rest of the world and
35:56 go anywhere in the world
35:59 that doesn't have high quality medical
36:02 Medical Care high quality medical
36:05 education and just all of a sudden on
36:08 that device is all the diagnosis stuff
36:11 the stuff that that Becky was talking
36:12 about earlier I uploaded my blood chart
36:14 to this thing and had AI do it imagine
36:17 being able to do that anywhere in the
36:19 [ __ ] world even without internet
36:20 connection and
36:22 obesity yes
36:25 exactly so so yeah so I think it's going
36:28 to be massive but
36:32 it's why I'm so passionate about this
36:35 why I keep coming to these things it's
36:37 not just going to be the medical field
36:39 it's going to be every field that
36:40 Material Science thing I talked
36:42 about it's going to be every field it's
36:44 going to be mathematics it's going to be
36:46 science it's going to be Consulting it's
36:48 going to be art it's going to be film
36:51 making it's going to be music
36:52 composition it's going to be music
36:56 production it's going to be live
36:58 performance is going to be dramatically
37:00 impacted by this
37:03 stuff all of it everything anything
37:07 that's at all computer
37:09 mediated is going to be transformed in
37:11 the next 5 years like transformed like
37:14 we won't do it the same way we used to
37:17 and and a lot of the bad press right now
37:21 with AI is the fear of that change right
37:24 the going to take our jobs the AI the
37:28 robots are going to do our jobs what are
37:29 we going to do it's going to take our
37:30 jobs it's going to take some and it's
37:33 going to generate a lot it's going to
37:36 generate a lot of
37:39 jobs what are they I have no [ __ ]
37:44 idea but I know I I I know cuz I I
37:48 started one of the early digital ageny
37:51 like I know the world of advertising
37:52 agencies what I can promise
37:55 you is that the way advert iing agencies
37:59 do what they do is going to transform
38:03 dramatically P I'm jazzed about our
38:05 dystopian future it's so true he
38:10 is I think any
38:14 agency I think any agency that
38:21 doesn't get real serious about
38:25 Reinventing their business model
38:29 is going to be in
38:30 trouble any
38:33 agency and like and there and I'll tell
38:36 you specifically what I
38:37 mean in the ad
38:41 business how you make your money is the
38:44 billable
38:46 hour so the longer people work the more
38:49 money you make from the holding
38:52 companies down that's the business model
38:56 that's how when when holding companies
38:58 buy agencies I sold my agency to Omnicom
39:01 so I know I know this from
39:04 firsthand they do what's called an
39:06 earnout and the earnout is is based on
39:09 if you make a dollar of Revenue we'll
39:11 pay you
39:12 four and we'll do that for the next five
39:15 years so go make as much money as you
39:17 can go acquire companies go grow your
39:20 revenue and so it is built into the
39:25 structure like like at the at the DNA
39:27 level
39:28 of holding companies and ad agencies is
39:31 sell more billable
39:34 hours and along comes a technology that
39:36 says oh that thing that used to take you
39:39 two weeks I can now do that in an hour
39:43 with with a human sitting there you know
39:47 editing it and making it good I can do
39:49 it in five minutes and have it suck but
39:51 I can do it in an hour and have it be
39:54 good what happens to the rest of those
39:56 billable hours
40:58 far behind on this stuff why haven't
40:59 they innovated they haven't innovated
41:01 cuz they didn't have to they had an
41:03 effective
41:04 Monopoly search was search and that was
41:07 it and you did it the Google way or you
41:09 just go [ __ ]
41:13 yourself and they did it for more than a
41:16 decade like in the in the 90 plus
41:19 percent of
41:24 search so they have they they have no
41:28 incentive to
41:30 innovate and they are
41:32 structurally
41:34 incapacitated because they optimized for
41:38 printing money having an effective
41:40 Monopoly and then Along Comes shitty
41:43 little Sam
41:45 mman he's
41:47 like oh Microsoft likes us they gave us
41:49 a billion dollars here's this thing
41:50 called chat gbt
41:53 yay and a 100 million people use chat
41:57 GPT in the first six
42:00 weeks and goo it was called I think it
42:02 was called a red alert Google went into
42:04 Red Alert they brought back Sergey Bren
42:06 and Larry pige uh guys come back you're
42:08 head of AI uh
42:12 help Google [ __ ] invented the
42:15 Transformer the T in GPT not sure if you
42:18 knew
42:19 this stands for
42:21 Transformer that Google
42:26 invented this is absolutely one of those
42:30 I can't believe we're losing to these
42:32 guys
42:34 situations that is my exact Kevin
42:36 McAllister moment super excited I don't
42:39 know which one I said but that's super
42:45 cool so Google's kind of screwed because
42:49 their whole business
42:51 model the entire business
42:54 model is based on making making you
42:58 click as many times as
43:01 possible after you
43:04 search search here's all the results
43:06 click click click click
43:08 click comparing all of Google to one
43:11 model no it's so that yeah that's not
43:13 fair pay it's it's not all it's not a
43:15 single model within Google but it but
43:16 it's the large majority of it is
43:18 advertising Revenue unlike Microsoft
43:21 Google can't fall back
43:23 on massive software sales or well they
43:27 probably have some decent cloud services
43:29 but it's dwarfed by advertising
43:32 right so the bulk of the
43:35 revenue comes from getting you to click
43:38 more on websites right click click click
43:42 every time you click money goes in their
43:46 pocket having a
43:48 tool that just gives you the
43:51 answer [ __ ] that up [ __ ] it up
43:55 bad uh am I in am I jaded thinking we
43:58 will never get all of these amazing
44:00 things because capitalism says
44:03 no
44:10 um I was super optimistic about the
44:13 internet in
44:14 1994 so much so that I did I got this
44:17 obsessive with you know worldwide web
44:20 development back then I started one of
44:21 the first digital agencies so I like I
44:24 know it and you know the early days the
44:26 early days of the internet were really
44:30 um really
44:33 amazing because it was open and anyone
44:36 could do it and there it was not
44:39 Consolidated and then over the years
44:42 that changed right it became here's how
44:45 we do banners here's how we do websites
44:48 everything got templatized all the
44:49 creativity got bled out of it it you
44:52 know companies like Google and Facebook
44:55 and Twitter emerge as the sort of big
44:57 three logos that represent most of the
44:59 internet
45:01 traffic
45:08 um I actually think that meta right now
45:11 this is this is kind of weird and ironic
45:13 and it it's going to be interesting to
45:15 see what they do with llama
45:17 3 I think meta op sourcing llama 2
45:26 commercially makes makes it much more
45:30 likely
45:32 that we're all going to get to play with
45:35 with very powerful AI for a long long
45:38 time that it that it won't get that it
45:40 won't consolidate down
45:43 into um three
45:46 companies what capitalism says is it
45:49 will right what capitalism says
45:53 is it it will consolidate into three
45:56 companies 20 30 years from
45:59 now
46:02 um
46:04 but I think this AI stuff's going to
46:06 [ __ ] up the H like I think it's G to I
46:10 think it's going to fundamentally alter
46:13 where the power is in the
46:15 economic drivers of
46:19 society and so I think there's going to
46:21 be a massive
46:24 battle where the centers of Power are do
46:27 everything they [ __ ] can to
46:30 resist losing
46:32 control not to AI but to people using AI
46:37 to [ __ ] with like like what I what I
46:40 said about the the advertising holding
46:42 companies billions and billions of
46:44 dollars right billions and billions of
46:50 dollars if they resist what's coming too
46:53 long they're going to get upended that
46:56 that whole industry is just going to get
46:58 turn into total chaos and I think that
47:00 happens in every industry on probably
47:03 different timelines but this is not like
47:05 slight little changes it's like your
47:07 fundamental core business model gets
47:10 flipped upside
47:13 down so they're going to resist this
47:16 [ __ ]
47:18 hard they're going to try to shut some
47:20 [ __ ]
47:23 down
47:26 but if you've got opsource models that
47:30 are gp4 strong GPT 5
47:35 strong
47:37 um you can't put that Genie back in the
47:39 bottle but didn't Microsoft jump the AI
47:42 train to reinvigorate their software
47:44 model that was stagnated yeah they
47:46 absolutely did and and what it what it
47:49 allows Mike what it allows Microsoft to
47:52 do is and and the the CEO of Microsoft
47:56 said this
47:59 what allows Microsoft to do is lose
48:02 money
48:03 on
48:05 co-pilot they don't they don't have to
48:07 charge for it they can put it out there
48:09 get people to use that try to take I
48:11 think I forget what it was
48:13 the every
48:15 point of share that they gain from
48:19 Microsoft and search is worth like $2
48:21 billion do to them a year something like
48:23 that Google is the fifth iteration of AI
48:26 specific Hardware remember that they're
48:28 invested i p I don't I don't disagree
48:31 that that Google
48:33 is is probably technically Superior on a
48:37 lot of this I'm just saying from a
48:40 business model
48:47 perspective
48:49 unraveling the structures that have been
48:51 put in place over the last 10 or 15
48:53 years having an effective Monopoly isn't
48:55 going to happen overnight and a company
48:58 like open AI That's got 800 employees
49:00 can turn on a [ __ ] dime like that's
49:03 that's where Google's kind of [ __ ]
49:05 right
49:06 now sans's capitalism we'd still have
49:09 horses instead of
49:10 cars yeah so that this is so big boy
49:14 pants good name for this discussion this
49:17 is this is this is big boy pants kind
49:19 kind of [ __ ] going on um um this is the
49:25 whole point that Sam Alman said that
49:27 that he now has emails that says Elon
49:29 Musk clearly
49:31 understands
49:34 that they couldn't raise enough money as
49:36 a
49:38 nonprofit to build the AI that they need
49:41 to build if they want to get to AGI they
49:44 need buckets of money Sam hman right now
49:47 is trying to raise 5 to7 trillion dollar
49:52 in capital like in like Hey brother can
49:55 you spare a trillion
49:58 5 to7
50:00 trillion for software no to
50:04 build a chip
50:07 manufacturing industry that would dwarf
50:11 Nvidia because we need that much compute
50:14 power and if you watch the the recent
50:15 interview with the Nvidia CEO he
50:20 agrees we need as much compute power as
50:22 we can possibly like we need more than
50:25 we can
50:26 imagine
50:28 so yeah we need capitalism to do
50:31 this I think that the meta thing will
50:34 balance it out to keep it from going to
50:36 just a few like the the concern that
50:38 Elon Musk has which I think is
50:40 valid
50:42 is if the most powerful AI are trapped
50:45 inside two or three companies they can
50:48 just
50:49 effectively take over the world and
50:52 that's the fear but by meta open
50:56 sourcing their foundational
50:59 models and apparently what they're
51:02 training llama 3 on is [ __ ] insane
51:04 like
51:05 600,000 gpus which probably doesn't mean
51:09 much but put it in perspective for you
51:11 inflection so Pi the thing that we're
51:14 going to play with tonight if I don't
51:16 talk for [ __ ] four
51:19 hours um they
51:23 raised
51:25 1.2 Trill or 1 2
51:30 billion and are spending
51:33 995 million of that on 22,000 h100
51:42 gpus 22,000 meta is training llama 3 on
51:47 the equivalent of
51:49 600,000 it's like a 22 billion doll
51:52 investment just in compute to train
51:55 llama 3 that they claim they're going to
51:58 open
51:59 source so that's going to be a GPT
52:03 5 or 4.5 or five or six who knows level
52:11 model that they're going to open
52:15 source can we have some Frank as a
52:18 pallet cleanser from the serious
52:20 grown-up talk and I think that's a
52:22 lovely idea if you don't know who Frank
52:25 is Frank
52:31 there's
52:32 Frank so Frank Frank is
52:36 a is a brand new media channel you can
52:41 find on Elon musk's cute cute little
52:44 chat service there called X so if you go
52:47 to the Daily Jank on Twitter formerly
52:51 known as X or X formerly known as
52:55 Twitter you see Frank Frank and
52:59 Frank he's a
53:02 newsman he's a good newsman I like
53:06 him add an existing account daily
53:12 Jank let's see do I have it in here
53:16 daily Jank yes I do
53:26 boom
53:27 so here's the daily Jank with
53:30 Frank Frank is an
53:34 awesome news
53:36 wiener and so he does daily so here's
53:39 here's
53:42 Frank's yeah welcome to the Daily Jank
53:44 with Frank it's March 8th 2024 it's a
53:46 Friday I'm Frank well what went on in a
53:50 week of AI that was quite a bit I'll
53:52 tell you that right now so so uh so
53:55 here's one here's one I didn't talk
53:56 about before mid Journey you know it's
53:58 this image generation thing makes all
54:00 the pretty pictures anyway last Saturday
54:02 night a bunch of people from competitor
54:05 stability AI sneak into their servers
54:07 and get [ __ ] busted they caught them
54:09 all so now they're all banned and they
54:11 can't use it and everybody's going
54:12 boohoo so so they're kind of fighting
54:15 then you got Elon Musk fighting with Sam
54:16 Alman we talked about that a lot and and
54:18 then you got the the the different
54:20 models going you got your you got your
54:22 open AI right the jet gp4 then you got
54:25 this anthropic would you come up with
54:27 clae 3 clae 3 is apparently all live
54:29 over here this thing's really exciting
54:31 and then yesterday Pi comes out and
54:33 they're like well we got one that's as
54:34 good as chat gp42 so nobody knows what's
54:37 going on everybody's going where's chat
54:39 GPT 5 where's gp5 we want to see gp5 and
54:42 Sam alman's like yeah whatever anyway
54:45 that's what's going on in AI for the
54:47 week hope you're well this has been a
54:49 daily Jank with Frank see you soon all
54:52 right so go follow the daily Jank on the
54:55 Twitter and get your news from [ __ ]
54:58 hot
54:59 [Laughter]
55:05 dog you see you see what's happening
55:07 here people you see what's happening
55:10 here thank you Silver
55:14 Fox oh so the daily Jank the daily Jank
55:18 if you don't know the history of Frank
55:20 so one one night I came into the live
55:23 and someone said have you played with
55:25 Adobe expresses new animate from audio
55:28 tool and I'm like I don't know what that
55:30 is and so we went and played with it and
55:32 the first dude we did was this little
55:34 pencil we we've called him Farber
55:36 because I thought that's what the
55:37 pencils were called is Farber Castell
55:40 and then Vicky told me it's Faber
55:41 Castell but whatever his name's Farber
55:44 now but but he's Frank's number two get
55:46 it and then so he's really
55:50 dull so we did him first and then I did
55:54 the original Frank video where where do
55:56 I have the original Frank video um do I
56:00 have it here in his channel hang on I
56:03 might have it in his
56:07 channel Elon goes after
56:11 Sammy hey how's it going no that's not
56:13 the original one I should probably put
56:15 the original one up here let me see let
56:18 me see if I can go find it where do I
56:20 have it
56:25 um
56:33 cu the original one's really [ __ ]
56:37 stupid yeah this is a good pallet
56:39 cleanser from from all that is is AI G
56:43 to save us or kill us
56:45 [Laughter]
56:55 talk d That's not
56:59 him there's that there's
57:04 [Music]
57:07 that Frank where are
57:14 you was he this long ago I don't think
57:17 he was I must not have him on
57:24 LinkedIn oh add
57:27 what are we going to do with
57:31 you let's see do I have him oh I know
57:35 where I have
57:37 him I've got him
57:43 in express. adobe.com
57:49 and then we're going to go to animate
57:52 with
57:55 audio
57:57 and
58:00 then where's your
58:06 stuff come
58:09 on this is
58:14 it
58:15 rutro Kyle YouTube oh right I put it up
58:18 on YouTube but I think this is it yeah
58:21 this is
58:24 it all right here here's the here's the
58:27 original Frank this was the thing that
58:29 set it all off so so the way this works
58:31 the way you do this Adobe Express thing
58:33 you just record audio you pick a
58:36 character you pick a background and then
58:37 you record audio and it it syncs the
58:40 lips up and it does all the animation so
58:42 you don't have to animate you just talk
58:44 so I just did one of my stupid [ __ ]
58:46 voices and then this is this was this
58:49 was the thing that turned into we should
58:50 turn this into a news show so I come
58:53 down to the field to play the game and
58:54 they're like hey Frank Frank what are
58:56 you doing doing I'm like I'm here for
58:58 the game they're like Frank it's
59:00 Thursday there's no game today I'm like
59:02 you got to be [ __ ] kidding me there's
59:04 no game today I'm ready for the game
59:06 they're like Frank Frank you look
59:08 [ __ ] delicious I'm like hey hey hey
59:11 back up hey whoa whoa whoa whoa wly
59:13 fella get your hands off my
59:15 butt oh my goodness so so I leave I go
59:19 down to the the Piggly Wiggly there I
59:20 get myself a soda I come back they're
59:22 like they're like they're like Frank uh
59:23 we we got an issue like what's that it's
59:26 like Tommy's here what do you mean
59:29 Tommy's here yeah Tommy's here what's
59:32 going on with Tommy Tommy's hungry oh no
59:35 no no no no Frank it's going to be okay
59:40 it's not going to be okay you know T
59:42 doesn't stop a [ __ ] eaten oh my oh
59:44 Tommy Tommy no oh God oh
59:48 God that's that's Frank that this
59:50 Frank's final
59:55 day
1:00:03 this is always a good laugh thank you
1:00:05 Kyle you're welcome digital
1:00:08 Danny so anyway so Frank has now turned
1:00:11 into a media Empire wait are we making a
1:00:14 GPT of tarot cards today yes we are Jane
1:00:16 do I just this was the this was the
1:00:19 hourong hello all right let's go let's
1:00:21 go build some some [ __ ] okay so here's
1:00:24 what
1:00:25 happened
1:00:30 yesterday inflection the company
1:00:34 behind
1:00:35 Pi launched a new version of their large
1:00:39 language model called
1:00:42 2.5 inflection
1:00:46 2.5 and claimed it was better than gp4
1:00:52 um so I tested it yesterday and we did
1:00:56 two tests and one of the now is here I'm
1:01:00 going to the the salon. A and if you go
1:01:04 there that's going to take you to a link
1:01:05 tree and you click click on the top
1:01:08 link and that will take you to this here
1:01:15 site which is the AI Salon oh damn
1:01:22 it Tik Tok just made me prove I was
1:01:25 human so so this is the AI Salon site uh
1:01:29 and if you're new here if you want to
1:01:30 join it it doesn't cost anything um
1:01:33 please do me a favor read about who we
1:01:36 are why we do what we do and what our
1:01:39 values are really important
1:01:44 um I love how Kyle's brain goes pivot
1:01:48 that's it it's like we're going we're
1:01:49 going we're going [ __ ] it let's move um
1:01:52 so go so go check that stuff out um and
1:01:55 then join and my request would be my my
1:01:58 the first thing I'd like you to do if
1:02:00 you decide to join the salon is in the
1:02:03 community section the first link is
1:02:05 introduce yourself so go to introduce
1:02:07 yourself and introduce
1:02:09 yourself um and then in was it in water
1:02:13 cooler yeah so I created this article of
1:02:17 this thing that happened last night
1:02:19 using pi so so the first thing I asked
1:02:21 it when it when I found out it was
1:02:23 upgraded to
1:02:25 2.5 is I thought what is the most
1:02:29 nonsensical thing I can ask it that you
1:02:31 know it shouldn't be able
1:02:34 to solve and and so I said how would
1:02:37 Alexander Hamilton reimagine government
1:02:40 using the what did I say the concepts of
1:02:44 of quantum
1:02:46 mechanics what did I say incorporate the
1:02:49 theory of Quant if Alexander Hamilton
1:02:52 were to incorporate the theory of
1:02:53 quantum mechanics into a governance
1:02:56 system
1:02:57 how would it be different than what he
1:02:59 created for America and then it told me
1:03:01 and it was pretty good and I said okay
1:03:03 that's pretty smart so why don't you
1:03:04 design it and so then it came up with
1:03:07 Quantum legislature Quantum Judiciary
1:03:10 Quantum executive and Quantum
1:03:11 communication all of these are really
1:03:13 good it was just really good so that was
1:03:15 the first thing I did with it so I
1:03:17 posted that
1:03:18 there and then the next thing I did with
1:03:21 it I just pivoted back back to that
1:03:23 comment I just pivoted and said Okay I
1:03:24 want to do something completely
1:03:26 different
1:03:28 and what what popped in my head was a
1:03:30 Tarot deck and I I was like well what if
1:03:32 you know what could could we somehow
1:03:34 like do something different with that
1:03:37 and I thought what if you asked
1:03:40 Pi how she I use Voice number five it's
1:03:44 a she she's an Aussie here here I'll let
1:03:46 you I'll let you hear
1:03:50 her this is voice number
1:03:54 five hey there I'm with my friends again
1:03:56 and we're talking about the um tery
1:04:00 project that we talked about yesterday
1:04:02 and
1:04:04 [Music]
1:04:06 today hello Kyle great to chat with you
1:04:08 again and I'm glad you're discussing the
1:04:10 tury project with your friends as a
1:04:13 reminder tarology is a hypothetical
1:04:15 Tarot deck that combines traditional
1:04:18 tarot symbolism with astrological
1:04:20 elements so so that's what it is so I I
1:04:23 asked her I
1:04:25 said
1:04:27 if you were to add 12 cards to the Tarot
1:04:29 deck what would they be in
1:04:31 why and like not only did she answer it
1:04:34 like she answered it really [ __ ]
1:04:36 brilliantly she goes well I would create
1:04:38 a new minor Arcana called the the suit
1:04:43 of
1:04:44 orbs and there'd be a card for each of
1:04:47 the as astrological symbols the ter the
1:04:51 the astrology symbols and I was like oh
1:04:53 that's [ __ ] genius and so and so we
1:04:56 went went back and forth last night and
1:04:57 asked it some more questions and it and
1:04:59 it it spit out a bunch of stuff that was
1:05:02 really
1:05:03 good and then someone on the live I
1:05:05 forget who it was began with a p and it
1:05:07 had a v in it Pavlov or pav some someone
1:05:10 I I said we' got to name it what's
1:05:12 what's some combination of tarot and
1:05:14 astrology and they came up with
1:05:16 tarology t a r o l o g y and so I said
1:05:20 I'm going to go make a a GPT of that and
1:05:23 then I'll I'll post it somewhere and and
1:05:25 someone last night goes no make it live
1:05:27 here so that's so the project tonight is
1:05:29 we're going to make that
1:05:32 GPT
1:05:33 um and so fair warning this could get
1:05:37 boring and you're like but Kyle this
1:05:40 Channel's already boring all right then
1:05:42 this could get more
1:05:47 boring you can make
1:05:49 money because one of the things I've
1:05:51 learned in here is if I'm just [ __ ]
1:05:55 around it's it's fun and kind of
1:05:57 engaging when I actually try to build
1:05:59 something it can get boring because
1:06:01 building [ __ ] is sometimes boring so I'm
1:06:03 GNA try to keep it entertaining and not
1:06:05 get too precious about
1:06:08 stuff but I will share something cool
1:06:10 with you so this afternoon let me go let
1:06:13 me go over to Pi so we have it in front
1:06:15 of us let's get rid of
1:06:18 Frank um pi.
1:06:23 a oh P guy got a fancy new landing page
1:06:29 [Laughter]
1:06:37 woo where's how why can I not sign
1:06:42 in oh log
1:06:51 in all right
1:06:53 so so this morning I woke up thinking
1:06:56 about
1:06:57 theology can't be as bad as zappier yeah
1:07:01 that if you want to learn zappier you're
1:07:03 not going to learn it here I tried that
1:07:05 one night and I I could I could actually
1:07:08 hear here in Denver faces hitting their
1:07:11 phone screen falling asleep as I was as
1:07:14 I was doing that that was
1:07:16 bad um so I woke up this morning and
1:07:21 and I wanted to explore the T ology
1:07:26 thing and dig a little bit deeper and
1:07:28 and look at look at what it it had
1:07:30 created
1:07:31 and things like that and so I had I had
1:07:36 her what's her name her name's Z I
1:07:38 forget what it stands for Z she told me
1:07:40 her name I well she made up a name and
1:07:43 we nicknamed her Z so z z
1:07:49 um I told her to give me a list of
1:07:52 the the orb cards for the different
1:07:58 um tarot
1:08:00 signs and it only went up to
1:08:03 Jack and I realized that um a Tarot suit
1:08:07 has 14 cards not not 13 Zephyr that was
1:08:11 it Vicki oh you got I people that have
1:08:14 good brains just always blow me away
1:08:16 like I my brain is interesting but I
1:08:19 wouldn't call it I wouldn't call it a
1:08:20 solid retrieval
1:08:25 mechanism
1:08:27 um and I re so I realized that there are
1:08:30 only 12 Zodiac signs and we need 14
1:08:34 cards in a major Arcana
1:08:37 boring while building is entertaining if
1:08:40 you're in the game yeah that's actually
1:08:41 a good point Papa JJ um and so I said to
1:08:45 Pi today I said
1:08:48 um we've got a problem
1:08:53 here yeah
1:08:57 so let me make this
1:09:00 bigger so this was from this morning so
1:09:04 so I'm I'm walking around like pre
1:09:05 coffee like I haven't showered or
1:09:07 anything and I just have my earphone in
1:09:09 and I'm talking to pie I'm having a a a
1:09:12 p conversation today inventing a whole
1:09:15 new Tarot
1:09:19 deck and so so she came up with this the
1:09:22 Ace of orbs is Aries Taurus Gemini
1:09:25 cancer right she she does all this and
1:09:27 we get up to the Knight of orbs and I'm
1:09:29 like oh [ __ ] we're missing the king and
1:09:31 queen and so I
1:09:35 said um how many cards are in the minor
1:09:37 Arcana there's
1:09:40 14 and then I said um you recommended
1:09:44 the Zodiac as an appropriate addition
1:09:46 but there are only 12 of them how do you
1:09:48 reconcile that or how do we reconcile
1:09:51 that and just the immediate answer is to
1:09:55 reconcile reconcile the 12 Zodiac signs
1:09:57 with the 14 card structure of the minor
1:10:00 Arcana we would include two additional
1:10:02 cards represent representing the lunar
1:10:07 nodes which are significant points in
1:10:09 astrology where the moon's orbit
1:10:11 intersects the Earth's orbit around the
1:10:13 Sun the North Node and the South
1:10:17 node how [ __ ] cool is
1:10:20 that like Bang by adding these two cards
1:10:23 we maintain the 14 card structure for
1:10:26 the orb suit and brought in the
1:10:28 astrological scope of the Tarot
1:10:31 deck
1:10:34 um and so the North Node represents
1:10:37 Mission and growth and the South node
1:10:39 represents past and Karma and and then I
1:10:41 said well you know this is the Knight of
1:10:45 orbs this needs to be the queen of orbs
1:10:47 and this needs to be the the king of
1:10:50 orbs excellent point Kyle right so then
1:10:53 it gives me a new list
1:10:57 and so here's our new
1:11:00 list South let's oh we can't okay so so
1:11:05 now here's
1:11:10 okay
1:11:14 what here's the challenge at
1:11:17 hand in order to create a
1:11:21 GPT for this new
1:11:24 deck we need to know what the rules are
1:11:27 we need to do know what the look and
1:11:28 feel of these cards are we there's a lot
1:11:31 we need to
1:11:33 know so I got some of that work done so
1:11:36 so part of what I'm about to
1:11:38 do is embrace the Jank and there's parts
1:11:43 of this conversation that are really
1:11:45 really good and we want to keep there's
1:11:46 parts of this conversation that are
1:11:48 trash and we want to throw out and so
1:11:50 how do you do that well how I do that is
1:11:53 is actually just really
1:11:54 shitty
1:11:57 I just use
1:11:58 notes so here's here here's
1:12:03 notes so all I'm going to do is I'm
1:12:05 going to
1:12:07 go back and forth
1:12:16 between between this God damn
1:12:22 it embrace the Jank embrace the Jank
1:12:25 embrace the J Jank these [ __ ] tools
1:12:31 suck is this it I think
1:12:34 so okay how do
1:12:39 I want I go
1:12:42 up okay copy okay so we're going to copy
1:12:45 there we're going to come over here
1:12:46 we're going to
1:12:48 paste and we're going to paste with
1:12:51 retain
1:12:52 style uh wait select uh select all
1:12:58 paste and we're going to change those
1:13:01 all
1:13:04 to H we'll change those to body [ __ ] it
1:13:08 okay oh I see I've got it really big
1:13:11 okay I'm just I've got it big so you can
1:13:13 see it because I because I care about
1:13:16 you I care about you
1:13:19 all so okay so so here's the plan plan
1:13:23 is I'm going to go get what's good from
1:13:27 PI and put it in a relatively good order
1:13:32 here and then I'm going to either save
1:13:35 it as a PDF or just copy and paste it
1:13:38 and I'm going to go to
1:13:39 Claud now why I'm going to go to Claude
1:13:42 is because Claude just got a new version
1:13:44 that they're claiming is better than
1:13:46 gp4 but what claude's got that gp4
1:13:49 doesn't is CLA uh the the the new big
1:13:52 fat
1:13:54 Cloud opus
1:13:56 has a 200,000 token context window so we
1:13:59 can take a big ass long thing out of Pi
1:14:04 paste it over there and we can fill it
1:14:05 in those are all minor Arcana you're
1:14:08 missing the major Arcana which starts
1:14:10 with a fool yeah no this is just so the
1:14:13 the list that I just copied this is a
1:14:15 new minor Arcana that are going to be
1:14:19 augmented to the existing Tarot deck so
1:14:21 we're going from a 78 card Tarot deck to
1:14:24 a 92 card Tarot deck that's got this new
1:14:27 minor Arcana in
1:14:29 it and so so what we've got to go do is
1:14:33 we've got to go get all of the good
1:14:35 stuff from this conversation and move it
1:14:38 over there then we're going to take that
1:14:39 over to Claude and we're going to have
1:14:41 Claude fill in any gaps and give us sort
1:14:44 of a complete rule set and then we're
1:14:47 going to take that and that's what we're
1:14:48 going to build the GPT out of so that's
1:14:50 that's what I'm doing right
1:14:54 now
1:15:10 so I'm going to start at the beginning
1:15:11 and just kind of okay so here's Quantum
1:15:13 legislature so so all right let's start
1:15:18 fresh I said our government was a
1:15:20 steaming pile of dog poo I think that's
1:15:22 being
1:15:24 kind and then I said okay let's change
1:15:27 gears why are there the number of tarot
1:15:29 cards there are it explains that I says
1:15:32 if you were to add a dozen more cards to
1:15:34 the Tarot deck what would be your
1:15:36 justification how would you organize
1:15:38 them and what would they
1:15:40 be if if I were to add 12 more cards to
1:15:43 the Tarot deck I'd call them orbs each
1:15:45 suit would depict a different archa
1:15:48 astrological sign
1:15:51 archaeological organization wise the
1:15:53 orbs follow suit would be structured as
1:15:56 the other minor Arcana suits with cards
1:15:59 numbered from 1 to zero and four Court
1:16:01 cards page Knight queen and king they
1:16:03 would maintain the consistency of the
1:16:06 deck while adding unique things okay so
1:16:07 we're going to grab this description we
1:16:09 don't need the stupid [ __ ]
1:16:12 emojis and we're going to go to oh this
1:16:16 I can't the fact I can't hold down the
1:16:17 shift key and select this
1:16:22 stuff a new suit called orbs
1:16:29 okay copy so we're going to go over
1:16:35 here we're going to paste that in so
1:16:38 we're going to say
1:16:40 um I'm
1:16:42 creating a
1:16:45 new
1:16:50 expanded
1:16:53 um oh she came up with a really good so
1:16:56 I've got a really good description of
1:16:59 it hang on let me go let me I got I got
1:17:01 to scroll to the bottom this is okay so
1:17:05 seriously what I'm doing right now if
1:17:06 you haven't played with
1:17:10 AI so one of the brilliant things about
1:17:13 Pi is it's conversational so I I can
1:17:15 just literally be talking to it and it's
1:17:17 taking notes of of what I said and what
1:17:19 it said and
1:17:22 right but conversation is messy and
1:17:26 and so shit's all over the place so part
1:17:30 of the art of
1:17:31 AI is sifting through the [ __ ] you
1:17:34 generated right cuz some of it's just
1:17:36 awful and then some of it's really good
1:17:39 and
1:17:40 so one of the misconceptions about AI is
1:17:44 you push a button and and stuff comes
1:17:46 out and then people like oh but it's not
1:17:48 good that's not the way it works in
1:17:50 general what you're doing is you're kind
1:17:51 of going back and forth with these tools
1:17:53 doing some sort of collaborative stuff
1:17:55 getting your idea or your head around
1:17:57 the idea that it came up with I would
1:18:00 love to see this deck come into
1:18:01 existence so Mary Mary you'll see so one
1:18:04 of the things that I did was was I
1:18:06 talked about um launching this as a
1:18:08 Kickstarter and I I talked about this in
1:18:11 my office hours today that this is one
1:18:13 of the challenges of AI is because it's
1:18:16 so good at taking an idea from a brain
1:18:19 fart to an
1:18:20 execution is that you can in about an
1:18:24 hour go from hey here's a stupid idea to
1:18:27 holy [ __ ] this is now a business I got
1:18:29 to deal with and it's like how far do I
1:18:32 want to take it so what I'm going to
1:18:34 start with is is with the GPT but what
1:18:37 I'm going to if I can get to it and
1:18:39 maybe this will sort of expand into
1:18:41 tomorrow is I want to create the artwork
1:18:44 for these new cards and and then that
1:18:46 becomes a question do you just create an
1:18:50 augmentation to existing tarot cards and
1:18:53 you just add a new
1:18:55 um a new 12 cards or do you do a whole
1:18:58 new deck called
1:18:59 theology deck in fact while I'm at
1:19:03 it why don't we go to
1:19:24 um
1:19:25 [Music]
1:19:37 he's really annoying when he sings like
1:19:39 that doesn't anyone complain oh people
1:19:43 complain a lot what's going on
1:19:50 here sign in why is this not signing in
1:19:54 oh it did
1:19:57 fine search
1:20:13 domains search
1:20:15 domains search the
1:20:24 domains
1:20:44 ter tery deck.com
1:20:51 what what is going
1:20:54 on
1:21:14 continue one item
1:21:18 oh looks good keep going wait what did I
1:21:21 just do
1:21:23 [ __ ] I don't want domain protection
1:21:26 [ __ ] go daddy everything is a goddamn
1:21:29 marketing up cell just give me a
1:21:36 domain okay okay here's here's part of
1:21:39 my job here is I'm here to voice the
1:21:41 frustration we all flip and feel there's
1:21:44 a movie in a book called tery yeah the
1:21:46 book called tarology I I asked Pi about
1:21:50 that so I think we're
1:21:54 okay
1:21:56 [Music]
1:22:06 we'll just grab it it may be
1:22:09 irresponsible the book the book is not
1:22:12 is has nothing to do with this I think
1:22:13 we're okay I talked to P and I had a
1:22:15 conversation about this
1:22:17 today
1:22:23 um is more ways to
1:22:41 pay why can I not pay with my PayPal
1:22:43 balance
1:22:53 whatever order summary ordered all right
1:22:56 so we got a website
1:22:58 okay welcome to chat
1:23:01 add now where was I where is my I
1:23:14 need where's
1:23:17 Pi did I lose my sight oh did I just do
1:23:21 it here
1:23:22 probably pi. a
1:23:27 okay we're
1:23:30 back oh so I wanted to go
1:23:45 okay this got really weird she was doing
1:23:48 really good for a while and then she
1:23:49 just decided to not be able to talk to
1:23:52 me about it anymore that was weird I'll
1:23:54 talk about that in a sec
1:23:56 second I'm Bound by my programming yeah
1:23:59 this was this was very frustrating she
1:24:01 went full Jank on
1:24:09 me here's the kickstarter I asked her to
1:24:11 help me do Kickstarter we'll come back
1:24:13 to
1:24:24 that
1:24:27 okay that's the
1:24:29 design what I'm trying to find the book
1:24:32 theology by en Enrique Enriquez was
1:24:35 published in 2018 despite the shared
1:24:38 name the book focuses on the poetic and
1:24:40 philosophical exploration of the tarot
1:24:43 de Mar rather than introducing new cards
1:24:46 so all
1:24:48 right um she says she doesn't think
1:24:50 there's a a trademark
1:24:53 issue um the domain
1:24:55 appears to be taken but it wasn't I just
1:24:58 got
1:25:01 it
1:25:21 um tarology intertwin intertwining tarot
1:25:24 and
1:25:25 astrology tarology interweaving tarot
1:25:29 divination and astrological wisdom here
1:25:32 we
1:25:34 go okay so that's
1:25:36 the that's the so that's the title and
1:25:41 whatever
1:25:42 subhead of this here
1:25:48 thing title elegantly communicates
1:25:50 blending the tarot's Divination
1:25:52 tradition with astrology's Cosmic
1:25:54 knowledge created a rich tapestry of
1:25:57 insight Rich
1:25:58 tapestry it's so [ __ ]
1:26:01 AI a rich
1:26:04 tabry the subhead elegantly communicates
1:26:07 the blending of tarot's divination
1:26:09 tradition with astrology Cosmic
1:26:12 knowledge the rich tapestry of insights
1:26:15 and guidance the Transcendence of the
1:26:17 boundaries of Convention of
1:26:21 decks all right that's good enough all
1:26:23 right let's go over let's go back to
1:26:24 notes
1:26:26 I'm creating a new
1:26:28 expanded
1:26:32 um I'll just call it a Tarot deck for
1:26:35 now
1:26:39 deck
1:26:41 that
1:26:44 adds a new
1:26:47 minor Arcana based on
1:26:53 astrology the deck
1:26:58 is
1:27:02 called and then I'll pop that thing
1:27:07 in
1:27:12 tery
1:27:15 the
1:27:17 subtitle
1:27:20 is all right interweaving tarot devation
1:27:24 with us astrological wisdom might not be
1:27:26 perfect but that's okay we we we at
1:27:29 least
1:27:31 get we can at least get
1:27:34 um claw up to speed on what this is cuz
1:27:37 what I need to do is I need to capture
1:27:39 enough of what we generated to get
1:27:42 Claude to understand it so it can help
1:27:43 us fill in the gaps um anyone have any
1:27:46 question this is fun to watch I've been
1:27:48 part of the medical metaphysical
1:27:50 community over 20 years folks will love
1:27:52 this oh thanks DK yeah I'm I'm really
1:27:55 excited about this my my um my
1:27:58 co-founder Monique and I are are super
1:28:00 into it too we and my my wife always
1:28:03 talks to the angels and we we do all
1:28:05 this stuff and and so even as we're
1:28:08 running our business we'll get we'll get
1:28:09 tarot readings and and talk about how
1:28:12 are things going so we're totally into
1:28:14 this
1:28:15 um and and what what's funny about this
1:28:18 idea
1:28:21 is again it was like I did this wasn't
1:28:26 an idea I had right this was I was just
1:28:29 trying to do something
1:28:31 nonsensical to see how good the new
1:28:34 model was and it actually came up with
1:28:38 some cool thing that's like hey
1:28:40 this like I would be interested in
1:28:44 playing with this right in understanding
1:28:46 this
1:28:47 in and I think that's the way that a lot
1:28:50 of innovation is going to happen
1:28:52 because when you're interacting with a
1:28:54 large language model you're essentially
1:28:57 collaborating with all of humanity or as
1:29:00 much of humanity as you know got their
1:29:03 [ __ ] put on the internet they either put
1:29:05 it there themselves or you know we
1:29:07 uploaded scanned books or whatever it
1:29:09 was right just all the stuff on the
1:29:12 internet is what you're interacting with
1:29:15 and so when you're doing a mashup like
1:29:17 this it's you're kind of
1:29:19 collaborating with Humanity which is
1:29:21 just it's just a
1:29:23 wild new kind of thing okay so we're at
1:29:30 uh uh let's see
1:29:33 the
1:29:35 addition is a new suit called orbs to
1:29:39 represent a new minor a new
1:29:45 suit and minor
1:29:48 Arcana called
1:29:52 orbs okay so there they are so okay so
1:29:55 now let's go back over here and let's
1:29:57 figure out what else we need to pull
1:29:58 from here we'll go we'll go down through
1:30:00 the through our list and I know this is
1:30:01 kind of painful and shitty and boring
1:30:03 right now but I'll try to keep busy and
1:30:06 and entertaining and listen we can make
1:30:09 a dog
1:30:13 [Music]
1:30:19 sing No
1:30:21 champy guess we can't oh I heard heard
1:30:27 [Music]
1:30:33 him well I heard heard there was a
1:30:38 secret
1:30:39 C David played and it pleased a
1:30:44 [Music]
1:30:46 lot you don't really care for music do
1:30:53 you goes like this the
1:30:57 fourth wait the fourth the
1:31:00 fifth might have fall a major
1:31:03 lift B King
1:31:06 composing
1:31:12 Hallelujah he's
1:31:14 sleeping I woke him
1:31:17 up all right so anyway back to the
1:31:19 boring [ __ ] work when you actually have
1:31:21 to do work it gets
1:31:23 boring
1:31:26 [Laughter]
1:31:33 there's the avocado
1:31:41 joke oh that was that was something
1:31:43 different hang
1:31:48 on there's Alexander Hamilton okay let's
1:31:52 start
1:31:53 fresh
1:31:57 okay here's where we
1:32:02 were what would be the most powerful
1:32:04 cards and
1:32:06 why I got those completely
1:32:10 wrong my apologies the orb suit the page
1:32:14 of orbs okay that it got
1:32:18 wrong page of orbs Gemini Sagittarius
1:32:21 this it got wrong so like that's
1:32:23 different now
1:32:25 shuffles the deck we can have the GPT do
1:32:28 that so we don't need to do
1:32:32 that here's a new nine card layout
1:32:35 called the celestial
1:32:38 Journey the present the past the future
1:32:40 the spiritual self the physical self the
1:32:43 challenge the
1:32:45 strength I think we can come up with new
1:32:48 layouts over in Claude or in or in GPT
1:32:53 but all right here are the cards I've
1:32:56 pulled for each
1:33:09 position fantastic I'm I'm thrilled that
1:33:12 your friends are impressed Kyle this was
1:33:14 last night when you guys are like holy
1:33:15 [ __ ] this is really good challenge
1:33:23 accepted
1:33:32 in the new Tarot
1:33:35 deck okay so this was this was this
1:33:38 morning I
1:33:51 think Ace of orbs Aries okay this might
1:33:54 be good please list them in order
1:33:57 include their rank the sign and a brief
1:33:59 Des description of the symbols on the
1:34:03 card ace of orbs area single okay so
1:34:07 let's grab this cuz show
1:34:11 more five of orbs 6 S 8 n show more
1:34:20 10 see it [ __ ] up the
1:34:23 order
1:34:26 yeah th this is and it repeated
1:34:30 things
1:34:32 but okay so let's how we going to do
1:34:35 this let's just do it one at a time so
1:34:36 we'll go Aries there's a a gentle little
1:34:40 description so there's
1:34:45 Aries and then we'll do this as
1:34:50 bold just so we can read
1:34:53 it
1:34:57 two orbs
1:35:02 nestled and so these descriptions are
1:35:05 really important because these
1:35:08 become how we
1:35:10 make
1:35:12 the the images right so we need the
1:35:15 description of the cards then we need
1:35:17 the meaning of the
1:35:19 cards and then we need things like rules
1:35:22 for how this shit's going to work when
1:35:24 you incorporate these and then I was
1:35:27 thinking we also need wait is this still
1:35:30 right Gemini cancer yeah this is
1:35:33 right then we need things like
1:35:36 um can you do unique layouts with just
1:35:40 the new minor Arcana with just the orbs
1:35:43 right can you just do an astrology
1:35:45 layouts with the minor Arcana that could
1:35:48 be kind of
1:35:50 interesting so it's kind of like
1:35:52 inventing a whole new thing well it is
1:35:56 that's what we're doing here people
1:35:58 we're innovating Virgo is Virgo next yes
1:36:01 this is good
1:36:07 copy Libra
1:36:12 next oh and by the way anyone tells you
1:36:16 that any of this AI generated [ __ ] isn't
1:36:20 copyrightable [ __ ] that like this is a
1:36:23 lot of work right this is like the whole
1:36:24 point of what the copyright office said
1:36:28 is it's got to have human contribution
1:36:31 for it to be good you could absolutely
1:36:33 argue that this was just spit out of
1:36:36 Pi's brain right like I had a
1:36:37 conversation with it it was spit out of
1:36:38 the brain but like I had the
1:36:41 conversation I'm the one that asked pi
1:36:43 to do this and then I'm the one that
1:36:44 steered the conversation and I'm the one
1:36:46 that's choosing what's shitty and what's
1:36:48 not
1:36:49 shitty and then I'm the one copying and
1:36:52 pasting this shitty ass
1:36:56 [ __ ] out of the janky ass tools that
1:37:00 exist right
1:37:02 now and then I'm going to stick them in
1:37:04 another thing and then we're going to
1:37:06 create some other mess and then we're
1:37:08 going to stick that in the GPT you know
1:37:10 what I mean it's a lot of [ __ ]
1:37:12 work tell me this was AI generated I
1:37:15 I'll I I will have words with your
1:37:19 Senior
1:37:20 Management all right Scorpio is the
1:37:24 eight is that right yes we're still good
1:37:28 here and there's probably some work to
1:37:30 do we can actually do this when we go
1:37:32 over to Claude there's probably some
1:37:34 work here to do here to
1:37:37 say are these are these in the right
1:37:41 order right like should should they be
1:37:43 in a different order for some
1:37:45 reason but we'll we'll we'll figure that
1:37:47 when we get to
1:37:53 it
1:37:54 [Music]
1:37:56 [Applause]
1:37:58 Capricorn
1:38:06 copy
1:38:11 Aquarius
1:38:15 Aquarius and Pisces and then it see and
1:38:19 then it started over oh that's what it
1:38:22 did cuz it it needed a queen and a king
1:38:24 so it's it started over that's why it
1:38:26 repeated [ __ ] okay so
1:38:30 now uh-oh did I just quit this
1:38:33 no now I need to go
1:38:35 find my apologies for the O
1:38:44 oversight okay so here's okay the North
1:38:50 Node would symbolize the souls this card
1:38:54 would
1:38:55 symbolize the Soul's Mission potential
1:38:58 growth and life's
1:39:00 purpose so that's not
1:39:03 really what's on the card but that's
1:39:06 okay let's just at least
1:39:11 capture
1:39:18 oops cuz there cuz these are new the
1:39:20 these are new Concepts right we we know
1:39:23 what the the zodia things are these
1:39:25 North and South nodes are weird the
1:39:28 expanded orb now
1:39:32 includes would these
1:39:36 nodes be the king and queen
1:39:39 respectively yes
1:39:43 okay upon further reflection Kyle I
1:39:46 believe that placing lunar nodes as the
1:39:48 queen and king of orbs is indeed the
1:39:51 most suitable Arrangement here's why in
1:39:54 traditional tarot the
1:39:57 courts hierarchy typically progresses
1:40:00 from page youngest to King to to Knight
1:40:04 queen king and most mature assigning the
1:40:07 lunar nodes to Queen and King maintains
1:40:10 this structure placing them in more
1:40:12 mature and influential roles balance the
1:40:14 king and queen often represent balance I
1:40:18 mean this is
1:40:19 so astrological significance the lunar
1:40:22 nodes are important
1:40:24 okay so let's capture all of
1:40:28 this considering these factors I'm
1:40:30 confident these are the right ones okay
1:40:32 so I got to do this shitty ass thing
1:40:35 again don't scroll too
1:40:38 fast this is so
1:40:42 frustrating hang
1:40:44 on watching an old man trying to use a
1:40:46 computer oh
1:40:50 yeah um so here's I'm going to go note
1:40:56 um
1:40:59 the queen and
1:41:03 king oh wait this should be the north
1:41:07 lunar
1:41:09 node
1:41:11 oops and the South
1:41:15 lunar
1:41:18 node the king and
1:41:21 queen of
1:41:24 orbs
1:41:29 um extend beyond the
1:41:34 12
1:41:38 signs of the
1:41:43 zodiac um and
1:41:50 here is the
1:41:52 context for why we added
1:41:58 them and by the way let me let me take a
1:42:03 little break here for a
1:42:09 second
1:42:17 okay
1:42:20 why what makes AI different
1:42:26 if you think about how you would do this
1:42:29 Pro let's say you had this
1:42:31 idea in the olden way in the olden timey
1:42:35 days in the olden timey days you know
1:42:43 2022 if if we were sitting around and we
1:42:46 were like you know what would be cool
1:42:48 you know we did a little smoked a little
1:42:50 doobie maybe you know had a
1:42:54 uh micro
1:42:56 do like you know it' be
1:42:59 cool what if you expanded the Tarot deck
1:43:02 to include astrology
1:43:07 cool and
1:43:09 then what does it take for that idea to
1:43:12 survive well it takes the high
1:43:15 people one of the high people
1:43:18 remembering it enough to remember it
1:43:20 Beyond everyone being a high when they
1:43:22 thought it was a brilliant idea
1:43:23 [Music]
1:43:24 and then motivated enough to actually
1:43:27 try
1:43:30 to understand how they would do that how
1:43:33 you would do it right and so if you're
1:43:36 super into astrology and into tarot and
1:43:39 know all this stuff you could maybe just
1:43:41 pull it out of your head and say Here's
1:43:43 how we might do it but if you didn't
1:43:45 what would you have to do you'd go to
1:43:47 [ __ ]
1:43:49 Google and you'd type in astrology and
1:43:52 this and that it would take you to 8,000
1:43:54 sites that were trying to sell you their
1:43:56 version of a Tarot deck or their version
1:43:58 of a coaching series on how to give
1:44:01 tarot readings right and you'd spend 45
1:44:04 minutes or 2 hours or 10 hours or two
1:44:07 weeks searching and researching just
1:44:10 pulling [ __ ] out of sites and just
1:44:12 pasting it in
1:44:14 places all to get
1:44:18 to this thing that like we're just
1:44:21 copying and pasting right
1:44:24 so so what does what does AI do is it
1:44:28 takes it gives you the ability to have
1:44:31 these really Vapor thin ideas that like
1:44:35 in this case it was just kind of a like
1:44:36 I was just trying to stump it I wasn't
1:44:39 expecting it to give a good answer so it
1:44:41 wasn't even really an idea it just like
1:44:43 this Vapor
1:44:45 thin
1:44:48 notion and then
1:44:50 within whatever it was 10 minutes of
1:44:52 going not even 5 minutes on the live
1:44:54 here people were like holy [ __ ] this is
1:44:57 good I'm like holy [ __ ] this is good and
1:44:59 as I'm going through this thing I'm like
1:45:01 like this is like the amount of time it
1:45:04 would have taken just to block this out
1:45:06 to get it to this point would be
1:45:09 weeks it's [ __ ] insane
1:45:12 okay enough comedy jokes let's keep it
1:45:16 moving people people are getting bored
1:45:17 they're going to sleep I can hear the
1:45:19 phones hitting their faces do you all do
1:45:22 that where you you're like
1:45:24 watching Tik Tok and you're holding your
1:45:26 phone above your face in bed and you
1:45:28 just drop it on your
1:45:30 face hey man you've been saving text
1:45:34 wait what was that have I been saving
1:45:36 text yeah because wouldn't this suck if
1:45:38 I just lost all
1:45:42 this okay so here's our updated list we
1:45:46 have that the final arrangement horb
1:45:52 harmonizes
1:45:54 tery Expanding
1:46:00 Horizons this is not bad let's pull this
1:46:04 over
1:46:12 here maybe the book is
1:46:20 called subtitle um let's see
1:46:24 marketing
1:46:42 blurb and then I think this is
1:46:46 possible book title because the other
1:46:49 thing I was thinking is like this is
1:46:51 absolutely a [ __ ] book right it's
1:46:55 absolutely a
1:46:57 GPT it's absolutely a set of
1:47:03 cards I mean there's something here I
1:47:06 think we should I I don't know there's
1:47:08 something to play with
1:47:11 here tery Expanding Horizons that might
1:47:14 be the book title delve deeper into the
1:47:17 cosmic connections between the tarot
1:47:20 between tarot and astrology with this
1:47:22 Innovative 92 card deck designed to
1:47:29 enrich
1:47:31 designed
1:47:37 by the
1:47:40 collective output of
1:47:45 humanity
1:47:48 harnessed let's see that was earned
1:47:51 let's see collec
1:48:03 harnessed
1:48:05 harnessed
1:48:07 with I'm just going to put generative AI
1:48:09 for now generative
1:48:11 AI just to piss people off because
1:48:14 actually you know what's you know what's
1:48:16 really interesting about this this is
1:48:18 kind of the most
1:48:21 human of Endeavors right right this is
1:48:24 this is like the most the most human of
1:48:26 Endeavors is like the tarot and
1:48:28 astrology is like this just this deeply
1:48:31 human you know connecting with the
1:48:33 universe kind of
1:48:35 thing and I think that
1:48:40 the the bad rap that AI is getting is
1:48:43 that these are cold machines that
1:48:45 they're robots that the robots are out
1:48:46 to get us and if you if you understand
1:48:49 how they work how they work is like a
1:48:52 calculator which is a cold non- feeling
1:48:55 machine but the reason that working with
1:48:58 these tools is not like working with a
1:49:00 cold unfeeling machine is because what
1:49:03 they were trained on was the collective
1:49:05 output of
1:49:07 humanity right so you're collaborating
1:49:09 with Humanity so I think we put that
1:49:12 right in the description okay let's go
1:49:14 back over here and see what else we
1:49:16 got maybe make it tighter tarology
1:49:19 explore Beyond I liked what we had
1:49:21 before discover the universe with tarot
1:49:22 through this groundbreaking that was
1:49:25 crap what's the subhead unveiling the
1:49:28 cosmic connection that's bad how about
1:49:32 intertwining intertwining this is okay
1:49:34 that we got oh interweaving okay that
1:49:37 was good take a look on the internet and
1:49:39 see if there's something called theology
1:49:41 so so uh TMI um this part of the
1:49:45 conversation I hopped in the shower and
1:49:47 I just put on pie so I was continuing my
1:49:50 my brainstorming my brainstorming
1:49:54 conversation uh while I was while I was
1:49:56 in the
1:49:57 shower that was when I was finding out
1:49:59 if if I was uh uh if someone had already
1:50:02 done this cuz what I would have done is
1:50:04 if someone had already done this I just
1:50:06 would have stopped or we' have just made
1:50:07 a GPT or something like that but you
1:50:09 know I don't want to I don't want to
1:50:11 step on anyone's toes but that's all
1:50:13 right all right from the research it
1:50:15 seems that gold golden thread tarot is
1:50:20 modern
1:50:22 um no one else done this it seems like I
1:50:25 misunderstood your question is there a
1:50:27 website called
1:50:34 tery
1:50:40 if uh the likelihood of a trademark okay
1:50:44 that was this was trademark [ __ ] when
1:50:45 was the book
1:50:49 written oh this was cool can you find me
1:50:52 3 different printers to print custom
1:50:55 tarot
1:50:56 decks and it it gave it to me so so part
1:51:01 of this I guess uh I won't do it now
1:51:04 I'll come back to this and pull it out
1:51:06 but
1:51:07 um so like there's business stuff in
1:51:10 here too right because part of this is a
1:51:11 creative exercise but then part of it's
1:51:13 like well wait a minute if I wanted to
1:51:14 turn these into actual tarot cards how
1:51:16 would you [ __ ] do that well you'd go
1:51:19 to some of the prince tarot cards well
1:51:21 you just ask Pi for it and there they
1:51:22 are
1:51:23 so that was pretty
1:51:26 cool okay here's a challenge I want to
1:51:28 describe how I might create a design
1:51:30 style for the entire deck that's
1:51:32 familiar to tarot readers but also
1:51:34 familiar to astrology people okay this
1:51:36 is good great idea to create a design
1:51:39 style for
1:51:41 tery we must consider the following
1:51:43 approach classic tarot
1:51:46 structure astrology inspired imagery
1:51:49 color
1:51:51 palette borders and typography y art
1:51:54 style okay so let's go grab
1:52:03 this because this will help feed when we
1:52:06 when we want Claude to make us a bunch
1:52:08 of prompts we
1:52:11 can use that as a design okay let's
1:52:15 scroll to the bottom here
1:52:18 okay
1:52:22 oops
1:52:25 I don't want that to be a number how do
1:52:26 I make it not a
1:52:28 number there we
1:52:31 go
1:52:33 um
1:52:38 design of the
1:52:48 deck uh how about
1:52:52 design
1:52:53 imagery and
1:52:56 structure of the
1:52:59 deck okay so we got
1:53:05 that okay I'd like to create a design
1:53:07 brief design brief for the tery deck
1:53:12 overview so I had it write a design
1:53:15 brief let's show all
1:53:17 that I can't shift select that
1:53:21 huh that is just shitty
1:53:27 programming janky is as janky does
1:53:30 people technical specifications this
1:53:33 gave 2.75 by 4.75 that feels small for
1:53:37 tarot but
1:53:39 whatever I won't put
1:53:42 technical specifications in
1:53:50 there yeah I will and I'll just update
1:53:52 them oh [ __ ] okay I gotta go all the way
1:53:55 down damn
1:53:57 it all
1:53:58 [Music]
1:54:05 right okay
1:54:08 copy so then we'll
1:54:19 do so there's the design
1:54:22 brief
1:54:27 overview design
1:54:29 objectives beautiful okay we have a
1:54:32 design brief
1:54:34 now if I were to write
1:54:38 a would you recommend creating an
1:54:41 exercise or two that uses only the
1:54:43 astrology cards I'm going to come back
1:54:45 to that I didn't like what she came up
1:54:48 with although at some point I
1:54:51 said
1:54:57 let's Al let's also come up with a five
1:54:59 card spread called the
1:55:01 star with one card in the center and
1:55:04 four around that I think that's already
1:55:05 an existing thing this deck is coming
1:55:09 into existence wow isn't it amazing I
1:55:11 know it's it's just
1:55:13 amazing
1:55:17 um here's the star
1:55:21 spread
1:55:26 the star spread invites the querent to
1:55:29 examine their issue from multiple angles
1:55:32 let let's grab
1:55:34 this here's the star
1:55:38 spread we can we can figure out if
1:55:40 that's if there's something that already
1:55:42 exists but let's just grab
1:55:46 it
1:55:49 um so this would go
1:55:53 design brief art
1:55:56 style okay so we go up here
1:56:02 um
1:56:04 uh
1:56:06 unique
1:56:10 spreads that
1:56:12 only use the
1:56:15 orb
1:56:21 cards
1:56:25 only
1:56:27 uses the orb
1:56:33 cards all
1:56:36 right and then I think what we got to do
1:56:38 is I've got to go in here and put in
1:56:44 some
1:56:48 subheading cuz this is getting hard to
1:56:50 read and none of this [ __ ] matters Ms
1:56:53 Right for for another um when we toss it
1:56:55 into CLA but it it matters for us as
1:57:00 we're trying to as as I'm trying to take
1:57:02 this add vomit of ideas out
1:57:09 of
1:57:12 um out of Pi and put them into here it's
1:57:15 really easy to get lost
1:57:21 so
1:57:27 and oh by the way if the if if you're
1:57:29 sitting out there going absolutely
1:57:31 [ __ ] crazy you know there's an easier
1:57:33 way to do this I don't care I know this
1:57:35 is like watching an old man use a
1:57:37 computer because it is watching an old
1:57:39 man use a
1:57:41 computer this is how we did it in 1995
1:57:44 and that's how I'm going to do it
1:57:50 today I don't care about your stupid
1:57:53 [ __ ] brilliant
1:58:02 ideas you know there's a tool that would
1:58:05 make this much more trivial to I
1:58:08 know but I'm an old man get off my
1:58:15 lawn okay there we areology the movie I
1:58:21 but we're not that far from that
1:58:23 so then I asked it about crowdfunding
1:58:25 and it came up with a whole Kickstarter
1:58:27 crowdfunding
1:58:30 Campaign which was good did a cost
1:58:35 analysis um oh this is one that I wanted
1:58:39 to get if we create a symbol for the
1:58:42 back of the cards what would it be it
1:58:43 came up with three good ones so the
1:58:46 celestial mandala create a mandala like
1:58:49 design featuring the 12 Zodiac signs
1:58:51 around a circular pattern
1:58:53 symbolizing unity and Harmony in the
1:58:55 center plac the classic tarot symbol
1:58:57 such as a pentagram or Fool's Journey
1:58:59 path Cosmic wheel design a circular
1:59:03 symbol inspired by the astrological
1:59:05 birth
1:59:06 chart with a section representing the
1:59:10 zodiac sign incorporate tarot symbolism
1:59:14 like the four suits of the major Arcana
1:59:16 around the wheel circumference and then
1:59:18 this was the one I really
1:59:20 liked Duality symbol combine the symbols
1:59:23 of the Sun and the Moon representing The
1:59:25 Duality found in both tarot and
1:59:28 astrology add subtle elements of the
1:59:30 zodiac signs or planets within the
1:59:33 design to reflect the tarology
1:59:36 theme I think I think I like that let's
1:59:39 see so that will go in the little
1:59:44 oops that'll go in the little design
1:59:51 area design objectives card illustration
1:59:55 guidelines borders and typography let's
1:59:57 add it here we We'll add a
2:00:02 five and we'll say um just
2:00:07 paste
2:00:14 oops Duality symbol so let's say um back
2:00:19 of the
2:00:21 card
2:00:24 symbol Duality symbol combine this
2:00:28 okay all right there's that that symbol
2:00:31 would do that give me a detailed
2:00:32 description of the first three cards oh
2:00:34 okay so this is where it got really
2:00:38 interesting
2:00:42 um I wanted it to write me prompts for
2:00:45 the designs of the cards like
2:00:48 detailed image
2:00:50 prompts and it did it all the way up
2:00:52 until the ninth card and when I asked it
2:00:55 to do the 10th card it safety guard
2:00:57 rails kicked in and it refused to
2:00:59 [ __ ] do it anymore so I got most of
2:01:01 the way there so part of what we need to
2:01:03 actually do is go is go
2:01:10 uh go fix this
2:01:21 [ __ ] okay gotcha let's focus on the orb
2:01:24 so
2:01:25 Aries
2:01:28 Aries depicts a
2:01:30 ram
2:01:32 okay so how many do we have we have
2:01:34 three what about the next three
2:01:43 okay so there's the first three should I
2:01:47 just add these to that existing list no
2:01:50 I think what we'll
2:01:51 do is we'll go we'll go to the bottom
2:01:55 here and we'll go new
2:02:00 subhead are we just watching an old man
2:02:03 copy and paste yeah that's pretty much
2:02:05 what we do around
2:02:07 here
2:02:11 um uh let's see
2:02:15 card face
2:02:18 image
2:02:21 prompts
2:02:30 oops oh rut row
2:02:34 paste shift return shift
2:02:43 return
2:02:46 okay and then it did
2:02:51 cancer Leo and Virgo so it was good it
2:02:55 was good through through
2:02:58 Virgo ah can't copy and paste right ah
2:03:03 Virgo okay copy go back over
2:03:06 here go here
2:03:11 paste this is annoying this is annoying
2:03:16 this is really
2:03:20 unnecessary this is this is one of the
2:03:23 one of the problems with ADHD is that
2:03:28 you've got this weird combination of
2:03:31 complete [ __ ] chaos and then OCD on
2:03:34 certain stupid details like line
2:03:40 breaks
2:03:42 Scorpio
2:03:45 Sagittarius okay
2:03:50 Sagittarius see I think some of these
2:03:52 are shitty like like nine it didn't do
2:03:55 good the Sagittarius card features a
2:03:58 centur the symbol of this fire
2:04:04 sign like I feel like it was good up to
2:04:09 here so let me grab Libra and
2:04:21 Scorpio
2:04:32 and then I
2:04:32 [Music]
2:04:42 need from Sagittarius
2:04:51 on
2:05:15 oops
2:05:16 okay so those those I need
2:05:19 still see here it said
2:05:23 ah I'm sorry Kyle since this is a Tarot
2:05:25 deck I can't provide detailed
2:05:27 interpretations of the cards as they are
2:05:29 often used for personal spiritual or div
2:05:32 divination purposes this requires the
2:05:35 reader's own
2:05:36 interpretation and I'm like just
2:05:38 continue with this description of the
2:05:40 cards you were doing before I'm sorry
2:05:42 Kyle I can't do that I'm sorry Dave I'm
2:05:44 afraid I can't do that like we are we're
2:05:47 [ __ ] living in science fiction where
2:05:49 the the [ __ ] tool was working just
2:05:51 fine and then goes I'm sorry Dave I'm
2:05:53 afraid I can't do that now granted I'm
2:05:56 not out in space trying to get back in
2:05:58 the spaceship but but still it's it's
2:06:02 the same [ __ ] experience
2:06:06 people I can understand your frustration
2:06:09 I can I can understand you feeling
2:06:11 disappointed or confused but my
2:06:13 programming doesn't let me do this then
2:06:16 I had it list out the things and then I
2:06:20 had it describe
2:06:29 Sagittarius and it sort of did that and
2:06:31 then it just it it just then it was like
2:06:34 I'm not even familiar with what you're
2:06:35 talking about what is this theology
2:06:40 project
2:06:41 unfortunately my apologies
2:06:47 okay oh okay so wait
2:06:51 um
2:06:54 interesting and the 13th and 14th cards
2:06:56 are lunar nodes what might those look
2:06:58 like you're right the last two cards are
2:07:01 the are the nodes the North
2:07:04 Node deps a compass with a thing facing
2:07:09 north representing the Soul's purpose
2:07:11 the background might include a path
2:07:13 through the forest or a Starry
2:07:16 Sky so let's
2:07:20 grab this
2:07:28 and we'll put it
2:07:31 here and then let's grab
2:07:40 this and we'll put it here but we need
2:07:43 to we need to
2:07:46 um ah we need to
2:07:49 um because this is the queen of War it
2:07:52 shouldn't just be a compass it should
2:07:54 right it should be the
2:07:56 queen
2:07:59 so yeah this is this is frustrating like
2:08:02 the these the four card the four face
2:08:03 cards are are a little frustrating so
2:08:05 we're going to have to fix that we we
2:08:06 can go fix that and probably chat GPT I
2:08:09 think I'll use chat GPT to fix some of
2:08:11 this
2:08:12 [ __ ] because I think it's got a context
2:08:14 W window big enough to just take the the
2:08:16 stuff that I I just created okay it's a
2:08:18 bit tricky okay and then this was
2:08:20 something else someone was talking about
2:08:22 someone was talking about deacons and
2:08:24 lots and that I should probably
2:08:26 incorporate those and then I was like
2:08:29 now this is starting to feel like
2:08:30 [ __ ]
2:08:31 work
2:08:34 okay it's [ __ ] it's [ __ ] it's
2:08:37 after midnight in uh wherever the hell
2:08:40 you
2:08:40 are we we've drained we've drained from
2:08:44 about 120 people to 68 so so you're
2:08:47 you're the hardcore
2:08:50 ones I told you this was going to be
2:08:52 boring I I promised you
2:08:55 um do you know how GPS writes IC Code I
2:09:00 know a signed copy of his
2:09:03 book um okay uh so so mods I don't I
2:09:07 don't know if anyone's still
2:09:08 modding but if anyone has any questions
2:09:10 let me scroll to the bottom does anyone
2:09:12 have any
2:09:13 questions um about what I'm doing and
2:09:15 then what I might do is I might go just
2:09:18 drop this all I might drop all of that
2:09:22 in a
2:09:23 GPT and then we'll come we'll come back
2:09:25 I don't know okay Silver Fox is here I
2:09:28 loved every minute of this great R
2:09:30 Rachel Sparkles thank
2:09:34 you um so basically what I've spent the
2:09:37 last hour doing was just copying and
2:09:41 pasting just trying to figure out what I
2:09:43 had there that was that that is the
2:09:46 foundation that we can now go fill in
2:09:49 right because what we have to fill in
2:09:54 is we have to fill
2:09:55 in the rest of the image stuff for the
2:09:59 cards and get those to a place we're
2:10:00 happy with them then we need to create a
2:10:03 meaning for each card that'll be fairly
2:10:05 straightforward cuz it's going to know
2:10:07 all that but we want it to do it in the
2:10:09 context of a of a of a re you know a
2:10:11 Tarot read a Tarot like reading a
2:10:14 tarology
2:10:20 reading
2:10:25 and then I think once we have that then
2:10:27 I think we're mostly
2:10:31 done because I think at that point oh oh
2:10:34 the other thing is I think once we
2:10:39 have once we have the meanings of the
2:10:43 cards then I think then I think we have
2:10:46 chat GPT come up with some rules of
2:10:48 engagement right how do you use this
2:10:50 deck just please post into salon my eyes
2:10:54 are closing and I'm losing
2:11:00 Consciousness that's
2:11:05 awesome thank you for being honest
2:11:07 that's terrific
2:11:10 um all right what else let's see well
2:11:12 done will you give subscribers personal
2:11:15 readings
2:11:17 yeah I'm gonna it's going to be my new
2:11:19 line of business is giving readings
2:11:22 meaning will you give subscribers
2:11:23 personal readings yeah I will uh meaning
2:11:27 an in Reverse reading for oh meaning and
2:11:30 in Reverse meaning that's right well
2:11:32 done okay actually yeah so so okay let's
2:11:35 let's go do a little bit of
2:11:37 fun
2:11:40 okay we we'll come back to
2:11:42 images so I'm going to say I'm creating
2:11:45 a new expanded
2:11:50 deck
2:11:52 I think we got to do this
2:11:55 in I think we got to do this in
2:11:59 Claude okay so here's what we're going
2:12:02 to
2:12:05 do
2:12:08 copy all right now I'm going over to
2:12:19 Cloud cloud.
2:12:25 and I'm going to paste in that Big
2:12:31 Blob notice I'm using Claude 3 Opus so
2:12:35 Opus is their fanciest Model 92 card
2:12:38 pickup yeah
2:12:40 exactly
2:12:41 [Music]
2:12:46 um put symbols on the card other than
2:12:49 the name of the card pneumonic
2:12:50 interpretive yeah that that will
2:12:52 definitely do laughing
2:12:55 Thunder use my creative lens and our oh
2:12:58 okay oh yeah that might be cool Corey
2:13:01 Sandler that would be a cool
2:13:04 mashup so Corey DM me because maybe
2:13:07 maybe we I think I think you doing the
2:13:12 artwork here could be really fun DM
2:13:16 me let's um cuz I cuz I think there's
2:13:19 the GPT and then I think there's I think
2:13:21 there is bringing these things to life
2:13:24 could be really fun um put symbols on
2:13:27 cards other than a card name yeah yeah
2:13:29 yeah U perhaps a screenplay will you
2:13:32 show the rest of your process updates
2:13:33 please absolutely I I'll so what we'll
2:13:36 do is we'll do this here we'll continue
2:13:38 this tomorrow just set up an AI to do
2:13:40 the reading that's what we're going to
2:13:42 do we're going to build a
2:13:43 GPT fingers just fell off but I'm okay
2:13:46 really okay hell yeah we'll do for sure
2:13:49 I'd left to yeah this will this will be
2:13:51 a fun one Cory cuz your your vision for
2:13:56 we can we can talk about this we can
2:13:57 come up with a style for all of the
2:14:00 things and cuz here's my vision for it I
2:14:04 think that
2:14:06 the I I watched some tarot guys on
2:14:09 YouTube and there's this one guy I watch
2:14:11 that I kind I kind of like he's really
2:14:13 nerdy but I kind of like him but the
2:14:15 deck that he uses it's like it's it's
2:14:18 all um like completely non-traditional
2:14:23 like like 70s acid art and so he knows
2:14:27 what they mean but like it doesn't have
2:14:29 any meaning so I want I want the cards
2:14:32 the I want the existing tarot cards to
2:14:34 feel familiar to people but have a
2:14:36 really unique feel to them inspired by
2:14:38 the you know the orb side but then I
2:14:41 want the orb ones to be really cool and
2:14:43 unique so right ABS [ __ ] lutely yeah
2:14:46 exactly
2:14:47 okay so so I pasted in all of our [ __ ]
2:14:51 there
2:14:54 so that's our document and then I'm
2:14:56 going to
2:15:01 say based
2:15:04 on based on
2:15:08 this
2:15:11 description of the
2:15:13 [Music]
2:15:15 tery
2:15:20 deck um
2:15:24 please give
2:15:33 me so I need
2:15:35 the um image descriptions
2:15:43 for I'm going to lose these
2:15:50 things I'm going to select all this and
2:15:53 copy it and I'll tell you why so I'm I'm
2:15:56 what I'm going to ask it to do is I'm
2:15:57 I'm going to ask it to give me image
2:15:59 descriptions that basically follow the
2:16:02 lead of these and so I I didn't because
2:16:05 those two were kind of written
2:16:06 differently so so I'm going to basically
2:16:09 say um at the bottom of the document you
2:16:12 know I need image descriptions for
2:16:16 everything from the nine of orbs to the
2:16:18 king so let me get rid of this
2:16:22 and make sure that I got all this copy
2:16:25 go over
2:16:28 here
2:16:30 paste based on this description
2:16:34 of a
2:16:40 new
2:16:42 deck please give
2:16:45 me
2:16:48 image
2:16:50 description
2:16:52 for the last cards at the
2:16:57 bottom
2:17:02 from the nine to the
2:17:07 king that match the
2:17:12 descriptions or that sort of match
2:17:15 the tone and
2:17:18 style of the image
2:17:23 descriptions uh for the first eight I
2:17:28 don't know if that's English
2:17:40 but
2:17:42 perfect Centaur the fire sign vast open
2:17:46 something Capricorn showcases a mountain
2:17:48 goat symbol of Earth climbing a steep
2:17:51 thing
2:17:52 page of orbs
2:17:54 Aquarius
2:17:56 good planet
2:17:58 Uranus Knight of orbs showcases the
2:18:01 knight's armor queen of orbs lth North
2:18:03 lunar node Mission and growth the queen
2:18:06 of orbs card features a real figure
2:18:09 dressed in flowing robes adorned with
2:18:12 the symbol of North lunar node the queen
2:18:15 stands at the base of winding
2:18:18 path beautiful the king of orbs
2:18:22 showcases wise introspective figure okay
2:18:25 these are [ __ ]
2:18:27 perfect you see people you see how this
2:18:30 pays
2:18:32 off and now so here's what's really cool
2:18:35 about this now that we've done all this
2:18:37 work and now that we're in Cloud that's
2:18:38 got a big context
2:18:40 window
2:18:43 it will just give us the [ __ ] we
2:18:50 need
2:18:53 I think it's g i think we're going be
2:18:54 able to knock this [ __ ] out I think I'm
2:18:55 going to knock it out so I know it's
2:18:58 late I know I'm a little long but
2:19:01 whatever no one here is happy most of
2:19:04 you are
2:19:05 asleep but this is this is what you have
2:19:08 to [ __ ] deal with with add all right
2:19:12 it's just how it
2:19:14 goes oh wait can I shift select here in
2:19:18 Claude yes there's some good program in
2:19:21 some actual [ __ ]
2:19:24 engineering
2:19:27 yay someone's actually used a piece of
2:19:29 software
2:19:35 before and bitter you
2:19:44 [Music]
2:19:46 betcha copy all
2:19:50 right okay so
2:19:59 now whatever they're not numbered but
2:20:01 that's
2:20:08 okay they're a little long it didn't it
2:20:11 made them a little long didn't it that's
2:20:13 all right doesn't
2:20:15 matter okay so now let's take all of
2:20:19 this copy
2:20:22 because that symbolism could become
2:20:25 important we'll go back to Claude we'll
2:20:27 do a brand new start we'll paste this
2:20:30 whole [ __ ] thing in here say based on
2:20:35 this new
2:20:37 card
2:20:45 set
2:20:48 that adds a minor
2:20:52 ARA
2:20:53 to the
2:20:57 traditional Tarot deck that adds a minor
2:21:03 Arcana based on
2:21:08 astrology to the traditional Tarot
2:21:14 deck
2:21:16 please
2:21:19 create a
2:21:22 description uh uh let's let's start
2:21:25 instead of a description for each card
2:21:27 please
2:21:33 um
2:21:42 please write an
2:21:46 overview of how
2:21:50 the
2:21:52 orb
2:21:57 suit should be used and
2:22:03 thought
2:22:05 about in a way
2:22:10 that is
2:22:12 consistent with
2:22:14 both
2:22:17 tarot
2:22:19 divination and
2:22:23 astrology
2:22:24 wisdom that's
2:22:27 good and I know I don't need to fix
2:22:29 typos but I did
2:22:36 okay the orb suit in the tarology de
2:22:39 deck serves as a bridge between a
2:22:41 traditional tarot divination and
2:22:43 astrological wisdom the numbered
2:22:46 cards Ace through 10 represent the
2:22:49 progression of energy and an expression
2:22:52 associated with the astrological
2:22:55 sign for example the for orbs might
2:22:58 indicate a need for emotional security
2:23:00 and nurturing okay
2:23:02 great yeah this is
2:23:05 perfect ah the orb suit can also be used
2:23:07 in astrologically focused spreads such
2:23:10 as the star spread mentioned in the
2:23:11 document oh this is good okay great
2:23:14 great
2:23:15 now
2:23:17 terrific now you now using that
2:23:24 description as your
2:23:27 guide
2:23:29 create
2:23:31 a meaning for each
2:23:35 card in
2:23:37 both the normal and
2:23:42 inverted
2:23:44 positions let's just see what it
2:23:47 does all of these please and thank yous
2:23:51 or going to make AI
2:23:53 sensient I I don't know how to not
2:23:57 thank these
2:23:59 things so wait upright okay Ace of orbs
2:24:03 Aries New Beginnings initial action
2:24:06 pioneering Spirit courage delayed starts
2:24:09 lack of Direction impulsiveness
2:24:11 self-doubt all right so this is a good
2:24:13 outline
2:24:18 perfect embracing personal growth resist
2:24:21 resistance to change this is [ __ ]
2:24:25 awesome okay great um I'm going to say
2:24:31 that's
2:24:35 a terrific outline or short reference
2:24:43 guide now write a
2:24:48 detailed guide for for each
2:24:55 card
2:24:58 position of about a
2:25:01 paragraph or
2:25:05 so each
2:25:08 card let's do in both
2:25:14 positions of about a paragraph or so
2:25:17 each all right now it should
2:25:22 because Claude can do long [ __ ] I like
2:25:24 to be polite I don't know how to not be
2:25:26 polite upright the Ace of orbs and the
2:25:28 upright position represents Dynamic
2:25:31 initiating energy of Aries the card this
2:25:34 is [ __ ] genius you know what's so
2:25:36 amazing about this is there has there is
2:25:40 so much that's been written on tarot and
2:25:44 astrology that that's the training set
2:25:47 is full of
2:25:52 wisdom right so it's going to write this
2:25:55 [ __ ] really good it's it's not going to
2:25:58 it's going to do this stuff good because
2:26:00 it's got so much in the training
2:26:02 set wonder about adding grock as well to
2:26:06 write details and see the difference
2:26:08 possibly yeah custom spreads yeah custom
2:26:11 spreads we'll do
2:26:12 next so I think what we'll do next is
2:26:15 we'll say
2:26:17 um give us custom spread concept
2:26:23 and then and then we'll have him write
2:26:25 write them up and then we're then I
2:26:27 think we're done and then what we'll do
2:26:29 tomorrow is we'll take all this stuff
2:26:30 we'll consolidate it we'll we'll we'll
2:26:32 make a GPT this makes my heart Happy a
2:26:35 that's really sweet I know this is
2:26:36 really cool isn't
2:26:40 it and and notice I mean with a 200,000
2:26:44 context or 200,000 token context window
2:26:47 it's about 150,000
2:26:49 words notice how long like this isn't
2:26:51 super fast but this is going pretty good
2:26:55 and I think I probably only have two
2:26:57 more I think you're allowed five prompts
2:27:00 per day for their Opus model so I
2:27:02 probably have one or two
2:27:05 left so so we're going to let it do this
2:27:08 and then and then I'll
2:27:12 do then I'll do the concepts for the
2:27:15 custom
2:27:17 spreads thank you Kyle Vegas Moxy
2:27:20 absolutely always Great Kyle thank you
2:27:22 Auto yeah we can we can start winding
2:27:24 down here so if you want to support the
2:27:26 channel follow me um you can subscribe
2:27:30 to the channel you can buy one of the
2:27:32 video series in the corners um you can
2:27:35 join the AI Salon which you should do if
2:27:38 if you're at at all trying to figure out
2:27:42 how to navigate this [ __ ] you should be
2:27:43 in community and I'm biased but I think
2:27:46 the AI Salon is the best [ __ ]
2:27:48 Community out there so that's the salon
2:27:50 on. a okay that's
2:27:53 great
2:27:54 okay okay this is
2:27:59 perfect now to make things
2:28:07 complete I want you
2:28:12 to
2:28:19 create
2:28:21 five
2:28:24 unique
2:28:32 spreads
2:28:36 that are a nod to
2:28:42 traditional
2:28:44 tarot but take the new
2:28:49 Arcana into
2:28:53 account the
2:28:56 smallest should be a three card
2:29:04 spread and there should
2:29:17 be a spread
2:29:23 for only the orb
2:29:29 cards
2:29:32 beyond the
2:29:34 star
2:29:37 spread in the
2:29:39 document the the fact that this can all
2:29:42 right so we're going to just see if this
2:29:43 thing can [ __ ] just
2:29:45 write
2:29:49 five
2:29:51 all
2:29:52 right the celestial
2:29:55 Trinity this simple spread provides
2:29:58 quick
2:30:02 overview card one the sun card two the
2:30:06 moon card three the rising sign the
2:30:08 elemental cross spread five
2:30:13 cards the Zodiac wheel 12 cards oh
2:30:17 that's so cool cards 1 through 12 lay
2:30:21 them in a circular
2:30:22 formation starting with Aries ascendant
2:30:25 and moving
2:30:30 counterclockwise each card represents
2:30:32 the energy and themes associated with
2:30:34 that astrological
2:30:36 house that's cool the lunar nodes spread
2:30:42 path the cosmic
2:30:45 Insight
2:30:47 six cards using only the orb cards
2:30:52 the core energy the supporting signs the
2:30:54 challenge the potential the integration
2:30:57 all right [ __ ] this this is [ __ ]
2:31:02 awesome okay so here's the
2:31:04 spreads so we're going to copy
2:31:08 this we're going to go over here so
2:31:10 let's
2:31:14 see
2:31:16 um that's art style okay so here's the
2:31:19 spread unique spread
2:31:22 so we're going to call
2:31:24 this the
2:31:28 tery
2:31:37 spreads all right there's
2:31:41 that and then we're going to go grab our
2:31:45 what the [ __ ] cards mean so we're
2:31:48 going to do the
2:31:51 oh wait we also need the okay here's how
2:31:54 it
2:31:54 works oh this is so
2:31:57 good see see all that work all that
2:31:59 boring work pays off
2:32:04 because
2:32:06 um you know now now we know what we need
2:32:09 okay so possible book
2:32:14 tarot organiz okay so there's
2:32:19 those
2:32:21 the oh tril tarology
2:32:25 ter ter God damn
2:32:27 it
2:32:31 terog spread so what we're going to do
2:32:33 here is we're going to go
2:32:35 um
2:32:37 how
2:32:40 tery
2:32:43 Works
2:32:45 boom that's how it [ __ ] works this is
2:32:49 so good
2:32:53 hey all right
2:32:56 boom the court
2:33:05 cards okay and then the next one is
2:33:11 uh
2:33:14 the oh you know what we'll do
2:33:18 um we'll do it down here me meaning of
2:33:23 the
2:33:26 cards and then we're going to grab all
2:33:29 of
2:33:34 these copy we're going to go the meaning
2:33:40 of the cards
2:33:42 um
2:33:47 reference
2:33:49 guide
2:33:52 of
2:33:55 the the meaning of
2:33:58 the orb
2:34:02 cards we're going to make that a
2:34:05 subhead and then we're going to go down
2:34:08 here we're go
2:34:09 here oops what did I just
2:34:17 do what did I
2:34:19 do
2:34:32 copy paste okay the meaning of the orb
2:34:37 cards
2:34:40 detailed
2:34:49 guide okay okay um go back over here now
2:34:53 we'll go get our
2:34:57 detailed you can see how this can get
2:35:00 really [ __ ] confusing right because
2:35:04 if you don't keep track of all this
2:35:05 stuff it's the it's like the the
2:35:08 greatest strength of this stuff is the
2:35:09 greatest
2:35:10 weakness
2:35:14 where there we go um the fact that it
2:35:18 makes so much stuff you can just get
2:35:20 totally lost in
2:35:23 it okay so
2:35:29 now now well I'll go you know what I'll
2:35:32 do I I'll do this all tomorrow okay I
2:35:35 got I got to get out of here you got to
2:35:37 get out of here we all got to get
2:35:38 [ __ ] out of
2:35:40 here I'm in
2:35:47 trouble um just downloaded this and
2:35:49 asked Cloud to give you a synopsis yeah
2:35:51 that's a good idea we'll we'll do that
2:35:53 tomorrow we'll sort of tighten this up
2:35:54 but now we've got okay so now this
2:35:56 thing's basically done right cuz from
2:36:00 here we can do artwork we can do cards
2:36:03 we can do a book we can do um a
2:36:07 GPT we'll start with the GPT we'll sort
2:36:09 of we'll take all of this we'll pull out
2:36:11 the [ __ ] that's not relevant for the
2:36:17 GPT but what I'm even thinking here
2:36:35 is there's probably get a the I know
2:36:38 there was a way in early
2:36:39 GPT someone had created a prompt that
2:36:43 will create a tarot
2:36:46 reading that showed the
2:36:49 cards and it was using
2:36:52 markdown to display the a JPEG from a
2:36:58 URL so I'm thinking if we make the if we
2:37:01 make the cards we can upload them to the
2:37:03 internet we could probably get a GPT to
2:37:06 actually do the readings and display the
2:37:09 cards with the
2:37:12 descriptions there should be a small
2:37:13 Hedgehog hidden in every card that's a
2:37:16 great idea just make a yeah a little
2:37:18 teeny Hedgehog and just Photoshop it in
2:37:20 [Laughter]
2:37:24 that's really cute um all right tery
2:37:31 exists far out
2:37:35 um I'm whipped you're whipped peace out
2:37:38 happy Friday Friday
2:37:40 night um you can make money
2:37:44 with see doing [ __ ] like this is work
2:37:47 right this is what I was saying earlier
2:37:48 that you come up with an idea like this
2:37:52 AI lets you materialize your idea really
2:37:55 really quickly which is super exciting
2:37:57 but to actually execute something it
2:38:01 always [ __ ] takes
2:38:03 work so there's no shortcuts but that's
2:38:06 okay because the level of work that we
2:38:08 get to do for the amount of effort that
2:38:11 we take is just insane this is insane
2:38:14 this is insane this is really cool all
2:38:15 right
2:38:17 peace have a good night