Apr 8, 2024 - Creative Insights and Strategies for Content Creation

Description

Chapters

Transcript

0:03 [Music]
0:15 [Music]
0:21 [Music]
0:29 oh
0:30 [Music]
0:44 [Applause]
0:45 [Music]
0:59 s Lon let
1:01 [Music]
1:03 Town wonder when things are going to
1:08 change dream my
1:11 away seem these
1:13 dreams Bunch
1:17 CL get my nerve up but my past is
1:22 pulling me
1:23 [Music]
1:26 down wondering how
1:29 long like she don't stick
1:32 [Music]
1:37 around somebody told me once before he
1:41 said you can never go home
1:43 again W you
1:45 [Music]
1:47 leave just to Ste me away yeah from the
1:52 truth of who I am and what I believe so
1:56 I thank him for his two s with a
1:59 handshake and some sympathy yeah I
2:04 packed up my blue
2:05 je and I head it for this big
2:10 price of my
2:14 freedom
2:16 byebye Black Sheep to the black sheep of
2:21 the
2:23 family
2:24 [Music]
2:27 bye oh that means so very much to
2:31 me
2:35 by to my friends and my
2:38 [Music]
2:45 family my
2:48 soul set it
2:49 [Music]
2:52 free
2:54 free time they were changing
3:00 did a little bit of
3:04 rearrange take a couple chances good
3:07 evening
3:09 everybody do me a favor go ahead and
3:11 share the live if you would get some
3:13 more people in here let them know we're
3:16 [Music]
3:18 around what do you think
3:20 champy TG you happy it's
3:24 Monday dogs have no sense of the
3:27 time he didn't know there was an eclipse
3:29 today
3:31 he was just like you got chicken
3:33 sticks you got more could could I have
3:36 more could you got more and I'm like
3:39 chimp you realize it's Monday he's like
3:41 you got chicken
3:42 sticks you got
3:45 more that's it that's that's it I get
3:49 that and
3:51 singing you know I can't have deep
3:53 philosophical conversations with a dog I
3:56 mean I can I
3:58 do it's just it's pointless he doesn't
4:01 care he's just he just about chicken
4:03 sticks all right fantastic Bob tell him
4:06 what he's
4:08 won a lovely English breakfast to polish
4:12 off the
4:13 [Laughter]
4:16 evening I also it's a little hard to
4:18 play my guitar I chewed my I chewed my
4:20 thumbnail down too deep and my thumb's a
4:23 little
4:24 [Laughter]
4:26 sore Neurosis is as Neurosis does
4:34 no ice maker but I did go out to the
4:37 7-Eleven and procure some
4:42 ice um I think the ice Maker's Frozen so
4:46 I'm going to take a I'm going to take a
4:48 blow dryer to it tomorrow
5:03 [Music]
5:22 laughing Thunder that's laughing
5:24 Thunder's picture congratulations on
5:26 making the box it's a lovely little uh
5:30 humanoid robot thing which you know up
5:32 until I don't know what three years ago
5:36 was like yeah they they'll come to them
5:38 some decade in the future we'll get
5:39 robots and now we're like yeah we'll
5:42 probably have them in a few
5:45 years oh man it's going to be
5:48 crazy how about some help for an old
5:51 Unix and Linux hand who really wants to
5:54 get going with chat
5:56 GPT four all right David Bowers you know
5:59 I haven't done a a gp4 deep dive in a
6:02 while I think we'll do that tonight in
6:04 fact I think I'll kick it off with
6:11 uh it looks like Sam mman spoke in in
6:15 London today I think they're opening a
6:17 big open open AI office and by the way
6:20 open AI went from 800 employees to 1,200
6:22 employees in the past four months five
6:25 months so they're growing they're
6:27 growing pretty quickly that's a big
6:31 big that's a
6:34 big you know 50% when you're 800
6:41 people that's a big cultural oops I
6:44 almost just spilled on my tea good Lord
6:46 but I
6:47 didn't
6:49 um but someone took a screenshot of one
6:53 of the slides Sam was presenting about
6:55 what's coming so I think we'll start
6:57 with that and I'll talk a little bit
6:59 about what I think
7:00 the elements on that slide mean and then
7:03 we'll go play with chat GPT
7:10 [Music]
7:15 [Applause]
7:17 [Music]
7:18 [Applause]
7:21 [Music]
7:33 [Music]
8:11 Champ's done he just walked away Alan
8:14 what's happening good to see you sir
8:18 [Music]
8:33 all righty
8:36 folks happy Monday my name's Kyle
8:38 Shannon this here is the AI learning lab
8:41 see see it says it right there you
8:43 couldn't this you couldn't do that if it
8:46 weren't that it's cuz if it's on Tik
8:49 Tock you know it's
8:50 [Music]
8:52 true this is a
8:54 lab you will
8:56 learn and I am AI
8:59 [Music]
9:01 are we Dust in the Wind oh yeah that was
9:04 sort of that dust in Wind
9:05 [Music]
9:10 wait all
9:17 right it's Dust in the Wind it's
9:19 Hallelujah
9:21 [Music]
9:54 in a Westerly Direction
9:59 this car is my train Mr it no more Salon
10:04 what do you mean no more Salon been
10:06 driving I've been
10:09 wondering what it is I'm Le from
10:14 again I don't know what you mean Mr it I
10:17 feel like an 80-year-old
10:20 man holding on to
10:27 29 and a head on that
10:32 Horizon is California line
10:36 [Music]
11:02 now this has always been AI learning
11:04 lab this is AI learning lab AI Salon is
11:07 is the
11:10 uh is the community AI learning lab is
11:13 just this this cute little Channel we
11:16 can make money with ch we can make money
11:18 with ch
11:22 [Music]
11:26 T all right everybody let's get roll in
11:30 here let's get it going all right so if
11:32 you have questions about AI pop them in
11:34 the comments below uh if you're willing
11:36 to share the channel share the channel
11:37 let's get some more people in here Tik
11:39 tok's being a little stingy with the
11:41 with the algorithm lately so [ __ ] them
11:43 do me a favor tap tapy tap the screen
11:46 let's get some some uh likes going in
11:49 here and then maybe Tik Tok will go all
11:52 right we'll throw him a
11:53 bone see see if we can get some more
11:55 folks in here I mean it's fine it's fine
11:57 having it be small too I don't really
11:59 care
12:03 ah fine tasty
12:06 tea all right fantastic let me go see if
12:10 I can find that that Sam Alman
12:13 thing and then we will go we will go
12:17 uh we'll go do some stuff in the chat
12:23 J there's so much there's so much in
12:26 chat GPT now and it it it looks like
12:30 Aryan 4D looks like a more polished
12:32 version of what we sneak peek sneak
12:35 peaked in Firefly a year
12:37 ago plug-in uses stable diffusion to
12:40 create photorealistic
12:42 images based on 3D
12:46 geometry that's
12:48 interesting
12:50 H there you go all right let's
12:54 see
12:56 Sam
12:58 Altman hello Sam what are you doing what
13:01 are you doing here
13:14 lately let's see let's go
13:19 oops open AI
13:28 London yeah this slide all right
13:41 so so who's
13:44 Morgan we're all theologians in the
13:51 end oh it also I saw a thing on on the
13:56 Twitter that
13:58 uh met is going to be releasing their 7
14:01 billion parameter model of llama 3
14:06 soon don't know if it's true but this is
14:09 chat TMZ so we're allowed to say [ __ ]
14:12 like
14:13 that what's the picture above your left
14:16 shoulder which one this one wait left
14:20 that
14:24 one or that
14:26 one or
14:29 which
14:30 one I don't know what you're talking
14:45 about picture in the
14:48 back I don't know what your
14:53 man oh the
14:56 mug this mug is
15:00 this was from one of my wife's art
15:04 shows that really cool blue and white
15:07 painting is her
15:09 painting uh oh this
15:12 mug this was Ian did this hey I Salon
15:17 that's
15:20 there all right under the mug nothing
15:23 under the
15:25 mug oh this one that's just art that's
15:28 just a let's just do this
15:32 one
15:35 uh actual
15:38 art like an actual print pressed into
15:41 the paper signed one of
15:45 one allart right with a capital
15:49 A all
15:51 right
15:53 so what about Sam okay that's Sam right
15:56 there's
15:57 Sam now to be clear when did she post
16:05 this nine hours ago so this was posted
16:07 nine hours ago and then chubby who I
16:11 don't know does chubby have a following
16:15 not
16:16 really so this could be [ __ ] today
16:19 at open AI event in London look at the
16:21 right corner agents okay so here's
16:24 what's on the
16:26 slide multimodality reasoning
16:31 personalization reliability
16:34 agents so if the next version of chat
16:37 GPT has all of that lots of things
16:40 change so it's already got multimodality
16:42 and when we when we dig into chat GPT
16:45 for tonight we'll play with that a
16:47 little bit it's very limited right now I
16:50 have a feeling that's going to expand
16:52 pretty dramatically it's probably going
16:53 to have um voice synthesis it's probably
16:56 going to have music as my guess it's
16:59 probably going to have it it it may or
17:02 may not have
17:04 Sora so
17:06 Sora I heard someone who has access to
17:09 it if you
17:12 render I don't know how long the clip
17:14 was oh I guess a clip is a
17:17 minute if you render a clip in Sora at
17:21 480p which is you know really low
17:24 resolution that's like 1990s TV
17:27 resolution um it takes about 20 to 30
17:30 minutes so uh I saw I saw an interview
17:34 with the guy that did um Airhead the one
17:36 with the balloon and he said it was 20
17:38 20 to 30 minutes per little
17:41 clip um and so he said it was just one
17:43 of those things where they would you
17:45 know give it a prompt and then go do a
17:47 bunch of other [ __ ] while it went and
17:48 did its thing and then they'd come back
17:49 and then they'd do another one
17:51 and so but that'll probably get faster
17:54 but anyway multimodality is going to is
17:56 going to increase reasoning is is huge
18:00 so reasoning if so right now chat GPT
18:03 doesn't have reasoning right now chat
18:05 GPT is is word vomit right it it's a
18:09 it's a probability engine and it just
18:12 vomits out words and it doesn't know
18:14 what it it said which is why when it
18:17 hallucinates when it does something
18:19 wrong or just makes [ __ ] up and you go
18:21 hey you just made [ __ ] up and it's like
18:25 oh I'm terribly sorry about that let me
18:27 try that again and I won't make anything
18:28 up and then it makes it up again and
18:31 you're like hey [ __ ] you made it up
18:33 again I'm terribly sorry about that let
18:36 me try it again and I won't make it up
18:38 this time
18:42 so so if it gets
18:45 reasoning in theory it should be able to
18:48 do things like look at what it gave
18:51 you and reason that it may or may not be
18:55 right and if it's not right then do
18:56 something about that hey Siri
19:01 go
19:03 away go
19:05 away
19:07 okay
19:09 um personalization so it doesn't have it
19:13 right now we've seen this teased a
19:15 little bit but personalization will be
19:17 it's going to be able to know who you
19:19 are know what your preferences are it's
19:21 likely going to be able
19:23 to
19:27 understand your interactions across
19:30 chats and you'll be able to say I want
19:32 this chat included in my memory or
19:35 not Kyle please check out latest post in
19:38 Irregulars showcasing my students okay
19:41 cool using chat GPT awesome that's
19:43 fantastic I will in where is it AI
19:45 Irregulars okay cool um so so and then
19:50 if if reasoning is really good and it
19:53 can do math then things start to change
19:57 pretty quickly so
19:59 it can start improving itself and it can
20:01 start doing things predictably and
20:04 consistently which goes to the next one
20:06 which is reliability which I assume just
20:08 means you know it's going to suck less
20:11 and you know maybe they'll have better
20:12 ways of maybe you'll be able to search
20:14 your prompt history for
20:16 example which I know is not really a
20:19 reliability issue but it is if you're
20:21 trying to [ __ ] use their system
20:27 uh and then the last one is a agents and
20:29 agents are where you'll be able to just
20:32 say to chat GPT hey go do something for
20:36 me and it will spin up little agents and
20:39 those agents will go do things on your
20:41 behalf so you can just sit back and have
20:43 a cup of coffee while it goes and
20:44 figures a bunch of [ __ ] out with its new
20:46 reasoning capabilities and its
20:48 personalization and its
20:50 multimodality and it's going to be able
20:52 to go do a whole bunch of stuff for you
20:53 and then come back and go here you go
20:55 boss what do you think you like
20:56 it so that was
21:00 if if this is not a [ __ ] slide this
21:02 was today or you know sometime in the
21:05 past 24 hours so
21:08 fascinating Morgan not yet an event in
21:11 London assumed before 20 VC
21:21 interview please keep us updated anyway
21:24 all right so that's that uh and then it
21:28 looked like
21:37 so this is Alan Hills Beck who's in here
21:40 um what's your name in here Allan I
21:45 forget Mr Mr studio
21:48 dhj in here um is a professor so this is
21:51 some work from his students right did
21:55 you
21:56 put more than one in here or is it a
21:59 let's
22:04 see are there other ones in
22:07 here no
22:11 okay here's the prompt cat gbt looked at
22:15 the picture picture translated it from
22:17 Spanish to English wrote a prompt and
22:20 created the image oh that's cool so he
22:23 had his students write out their prompt
22:25 L hand and then use chat gpt's vision to
22:29 see The
22:31 Prompt and then here's a house design
22:33 from my freshman architecture class that
22:35 came up with using their creativity and
22:37 imaginations in chat GPT to speak their
22:40 ideas into Concept Designs by creating
22:43 their own prompt how cool is
22:48 that so this is by the way if you're an
22:51 educator or if you know an
22:55 educator and and especially if they're
22:58 an educator like it's okay we're using
23:00 um GPT detectors it's okay we don't need
23:03 to worry about AI we're we're preventing
23:05 it um Allan here is a really good
23:08 example of someone who is taking it in a
23:11 classroom in a really interesting way
23:14 right he had them write out the prompt L
23:16 hand on a piece of paper which if if for
23:19 those of you that are are maybe younger
23:22 with us tonight paper is this thing that
23:26 we used to write on that was made of
23:28 trees uh trees were things that grew in
23:31 what we used to call forests and forests
23:34 were where birds birds are animals with
23:37 wings that that fly and sing and make
23:39 people happy anyway all that to say he
23:42 he's having them use paper and and
23:46 writing implements that you use with
23:48 your hands that are not sending anything
23:51 to the cloud to write a
23:55 description of what they want their
23:57 house to look like
23:59 and then having chat GPT read that
24:01 description and generate the image how
24:05 [ __ ] cool is that that's really
24:08 cool so
24:10 cool and then and then what does that do
24:13 right like if if nothing else it starts
24:16 a fascinating conversation about okay if
24:20 you can speak things into
24:24 existence words are
24:26 important and how you use them is
24:28 important your vision is important and
24:30 how you articulate your vision is
24:31 important which is all stuff that I'm
24:33 sure he's been teaching in architecture
24:35 for [ __ ] decades
24:38 right but here's a new way to teach it
24:41 an exciting way to teach it a way that
24:42 will have them never forget especially
24:45 if they've never done anything with chat
24:47 GPT before Allan is so creative with his
24:50 students yeah it's amazing isn't it
24:52 Source Camp by the way Source Camp Kelly
24:55 Camp who just just commented here
24:58 tomorrow night is the AI salon and Kelly
25:01 is actually presenting tomorrow at the
25:02 AI Salon welcome Kelly and she's going
25:05 to be talking about future of work kind
25:07 of stuff
25:08 and her relationship with AI and it's
25:11 exciting it's really cool and how she's
25:14 how she's putting it in the world she
25:16 she she's she's a very she's a cool
25:20 she's a cool AI evangelist this sort of
25:23 sneaks up on
25:25 you she's like oh just check this cool
25:27 thing out just check this cool thing out
25:29 then she blows their minds and they
25:30 don't know what to
25:32 do so very very
25:36 cool Kyle don't make things up like
25:38 paper sorry ISU I know I know I probably
25:42 shouldn't talk about paper in the
25:44 present tents you know but but it it
25:47 really was a thing up until like three
25:48 years ago it like so like these are
25:51 playing cards I've got and these are on
25:54 what's called pay playing cards so
25:57 playing cards you can play solitire
25:59 which is kind of like uh surfing Tik
26:01 Tock uh except without a device and then
26:05 but you could also play games like poker
26:07 where you would sit with other human
26:09 beings around a table and and drink and
26:12 and gamble
26:15 pennies we did that with paper
26:20 too oh man yeah that's so
26:24 cool all right so how did that happen
26:29 in fact you know what we're going to do
26:31 I'm going to
26:33 grab I'm going to grab this
26:39 image save image
26:42 as we'll do
26:44 Allan Dash paper
26:50 prompt okay so the person that was in
26:52 here earlier saying they were an old old
26:55 Unix guy
26:59 doing old Unix stuffff I remember being
27:02 cut by paper once that was rumor had it
27:05 that that a paper cut was more painful
27:07 than a than a cut by a rusty butter
27:13 knife that was the rumor back in the
27:15 olden timey days they said you know I I
27:19 wouldn't wish a paper cut on a worst
27:21 enemy if it's funny some of the mob
27:24 they're like listen I can break your
27:26 kneecap or I can give you a paper cut in
27:28 the in the webbing between your fingers
27:30 and and you'd take the kneecap every
27:33 time here in Australia we're going to be
27:36 cashless by 2030 wow that's
27:38 wild that's crazy what could possibly go
27:42 wrong although what could possibly go
27:45 right with our current cash situation we
27:48 just we're like we're out of money we're
27:50 just like I make more it what could
27:52 we'll be
27:56 fine all right so let's see let's do a
27:59 couple of things here if you are new
28:02 here if you don't know what the AI Salon
28:04 is go check out the
28:06 salon. that's going to take you
28:08 to a link tree and the very first link
28:11 there says AI Salon
28:13 community and then that's going to take
28:15 you to a pretty little
28:19 place where's my AI Salon oh it was
28:22 there that's why I couldn't find
28:26 it oh dang it where am
28:30 I I have too many tabs open
28:34 people there we
28:38 go so when you go to the AI Salon
28:40 Community it'll take you to this lovely
28:43 page and you can read about what we're
28:45 up to what our values are I would
28:48 strongly encourage you to do that
28:50 Curiosity generosity
28:52 exploration collaboration inclusion
28:56 empathy and gratitude are the are the
28:59 the values of the salon um it's a really
29:03 remarkable group of people think of
29:04 think of the a salon kind of like an AI
29:07 support group hi my name is Kyle I'm in
29:09 the AI hi
29:13 Kyle I'm I'm working on a dolly image
29:16 right now us too Kyle go
29:19 on
29:22 um it's really good
29:28 de de
29:31 de all right fantastic let's see what
29:35 we've got too many tabs open there's a
29:37 GPT for
29:41 that
29:47 oh paper cuts are the
29:49 [Laughter]
29:55 worst interacting with other humans
29:58 what's this word world coming to I know
30:00 do you know card tricks um I used to
30:02 know a little bit of cestry uh where you
30:04 sort of do fancy cuts and shuffles false
30:07 shuffles and things um but then I
30:10 realized it was going to take me 3 years
30:12 to get any good at it so then I just now
30:14 I just use it as kind of an add an add
30:17 tool to just keep my hands occupied um
30:21 okay but anyway the salon go to the
30:23 salon
30:25 tomorrow um if you go to AI salons and
30:28 events under Community tomorrow we have
30:33 uh a meeting so every other week we have
30:36 a meeting and uh people show up we talk
30:39 about AI talk about news we usually have
30:42 a speaker so like I said Kelly Camp
30:44 speaking tomorrow um and it should be
30:47 really good so go to the salon check
30:49 that out also in the salon what you
30:51 should do when you first get there is
30:52 introduce yourself so go to there's the
30:55 first link under the Community section
30:57 is introduce yourself so be generous and
31:00 introduce yourself and tell us who you
31:02 are and what you're up to um there's
31:04 also a little showand tell area where if
31:06 you want to get inspired by what people
31:07 in the community are doing there's in
31:09 images show and tell music and audio um
31:12 video and animation and then gpts and
31:14 apps and then there's guilds and within
31:16 the guilds there's all these different
31:18 spaces like writing and art and video in
31:21 the creative Guild business Guild Tech
31:23 Guild social impact Guild and down at
31:27 the bottom there's a thing called
31:28 Community clubs and in that is AI
31:31 learning lab Irregulars so that's our
31:33 personal little Channel That's where
31:35 Allan
31:36 shared this art from his students super
31:40 cool oh by the way Alan the uh the
31:42 mailman came today so your package is on
31:46 its
31:49 way and uh and you know as as we as we
31:52 play tonight and talk about stuff and as
31:55 I show you stuff in chat GPT um feel
31:58 free to share it in the Irregulars
31:59 Channel okay
32:01 now there's one other area I want to
32:04 talk about at the AI Salon under the
32:06 Community section so introduce yourself
32:08 go do that you can check out the uh the
32:11 details of the event tomorrow night the
32:13 AI Salon is from 5: to 7 mountain time
32:16 and then this third link AI Launchpad
32:18 and tools this is a list that I maintain
32:23 personally although I haven't updated it
32:25 in like a month it's been a little
32:27 little slow on the news
32:29 front on the tool front first thing I do
32:32 is talk about how to how to get yourself
32:34 AI literate play solve and share you can
32:37 read about that and then this is a big
32:39 long ass list of tools that I think are
32:42 worth your while to get your head around
32:45 so if you really want to get AI literate
32:46 this this list I think is is worth the
32:49 money you pay to be a part of the salon
32:52 which is 0 it doesn't cost anything um
32:56 the first tool that we show here
32:59 is chat
33:00 GPT why CU it's still the
33:03 best if it weren't the best could I
33:05 write this
33:06 here absolutely
33:09 not so that means it's true you see it's
33:12 done in typography with
33:15 letters and then in chat GPT I talk
33:18 about chat GPT 3.5 the free version and
33:20 then chat GPT 4 the not free version in
33:23 my opinion and I don't get paid by by
33:26 them Sam wman doesn't even like me
33:29 um chat GPT 4 the 20 bucks a month for
33:32 chat GPT plus to get access to this I
33:35 think is the best deal in software right
33:37 now it it is incredibly powerful and it
33:39 also being a paid member of of chat GPT
33:43 gives you first access as they roll new
33:45 things out it's the paid members that
33:48 get access to it
33:49 first how do you get how do you
33:52 understand what's here that's what this
33:54 list
33:55 is basic text generation internet access
33:59 code interpreter GPT Vision Dolly 3 and
34:02 then custom gpts and then on top of that
34:05 there's things like custom
34:07 instructions and um chat GPT you can
34:10 even talk to it so when I do a deep dive
34:13 I'll sort of try to work my way through
34:15 all these different modes they're
34:16 they're all kind of intermingled in the
34:18 same interface so it's it's sometimes
34:21 not easy to know what's there so that's
34:24 it's good we're we're doing this tonight
34:26 um all right what questions do you
34:30 [Applause]
34:42 have well worth the investment Becky R
34:45 Love chat
34:46 gp4 so Becky R Becky R she she loves
34:51 quilting an amazing woman by the way you
34:53 should you should get to know her um but
34:56 she loves quilting and when when we
34:57 first started playing around with GPT 4
35:00 and they came out with these things
35:01 called custom
35:02 gpts and she realized it was good at
35:05 doing like making quilt patterns she
35:08 made a GPT to make quilt patterns such
35:10 such cool stuff such cool creativity in
35:12 this community so anyway um why does
35:14 YouTube think they have ai capabilities
35:18 uh you're welcome Becky you're you're
35:20 you're amazing um
35:22 [Music]
35:24 they
35:26 so be okay first of all they're owned by
35:29 YouTube second of all how you see
35:33 YouTube videos there's lots of AO on AI
35:36 going on behind the scenes it's the AI
35:38 that that we don't get to play with
35:40 right it's like industrial AI that that
35:43 they you know Google's been using for
35:44 years and years and years so there's a
35:46 lot of AI in
35:49 YouTube they've also shown some
35:52 demonstrations of things like um being
35:55 able to generate um background music for
35:59 your reals videos um by just prompting
36:04 it right so generative music generation
36:06 and and the the demo that they showed it
36:09 it created songs that sound like Charlie
36:11 poth was singing it or Tay was singing
36:14 it um but they they haven't really
36:17 launched a lot here's here's what I'm
36:20 hoping YouTube does is that they go all
36:24 in on a lot of the AI tools because what
36:27 we should be able to do is I should be
36:29 able to take the entirety of my AI
36:33 learning lab
36:35 Channel
36:37 and have it automatically make a bunch
36:40 of shorts for like interesting shorts
36:42 and things like that and chapter IE
36:44 things and you know all that sort of
36:46 stuff it it should be super automated
36:49 and it's not right now so we'll see
36:51 we'll see all right show us chat GPT
36:54 translate from the photo student prom
36:59 oh was it not in
37:08 [Music]
37:13 English oh yeah Spanish oh cool
37:17 literally looked at the pict picture
37:19 translated it from Spanish to English
37:21 wrote a prompt and then created the
37:23 image ah yes okay good perfect so yeah
37:26 we'll do that
37:28 okay so let's just get situated here
37:32 here's chat
37:33 gp4 and there's chat GPT 3.5 they look
37:39 identical let me let me make my window a
37:41 little smaller
37:46 here so there's a little area on the
37:49 left that's got your prompt history in
37:52 it and if you've got access to gpts
37:55 you'll see those there and then GPT 4 is
37:59 the fancy one now GPT 3.5 is not
38:03 bad um let's see write a python script
38:08 brainstorm content
38:10 ideas yeah we'll just click on
38:14 that so chat GPT 3.5 is really
38:17 fast so while it's not as sophisticated
38:21 as GPT for its speed is like sometimes
38:24 if I just want a quick answer to
38:25 something or I want to knock out an
38:27 outline really really quick I'll go to
38:29 chat GPT 3.5 so even though I'm paying
38:32 for four I'll still use uh uh 3.5
38:36 sometimes four gets really slow uh cuz
38:39 everyone's using it because it's it's
38:41 you know it's doing a lot of stuff 3.5
38:43 is really good for that all
38:45 right so that's that now how do you get
38:48 from 3.5 to four you you know you click
38:51 on this little drop down but if you
38:53 don't have access to it in the first
38:54 place how you get to that is you go to
38:56 your name in the lower left hand hand
38:58 corner you go to my
39:00 plan and then chat GPT plus is access to
39:05 chat
39:06 gp4 okay as well as browse create and
39:09 use gpts and then access additional
39:12 tools like dolly3 browsing and Advanced
39:14 Data analysis and more and that's what
39:16 we're going to be going through
39:19 okay they've also got teamwork spaces
39:22 and things like that you don't need to
39:23 worry about that now um one quick little
39:26 thing that I'll show you which is kind
39:28 of fun I think a lot of people don't
39:30 know that it exists so when you use chat
39:34 GPT you know that it's very
39:38 um
39:40 uh
39:42 chatty it's chat GPT it's very chatty
39:46 and
39:49 um if you don't like that if you want it
39:51 to have a different tone you can
39:54 actually give chat GPT overall a new
39:57 person ality about to
40:00 sneeze thank
40:03 [Laughter]
40:06 you I saw the fake car commercial on
40:09 LinkedIn very cool I know it's pretty
40:10 cool isn't it that was from uh Kelly
40:13 Bosch who did the fake car
40:17 commercial just so the Irregulars know I
40:19 have first dibs on Kyle's
40:21 soundtrack so Becky Becky you might
40:25 actually know someone I don't know if
40:26 you've been in here I'm I'm working on a
40:27 musical right now that that started in
40:31 the live I did a song in sunno one night
40:34 and it was so good I'm like oh that
40:36 sounds like a Broadway musical and I
40:37 went wait this whole story should be a
40:39 Broadway musical so it's it's called
40:41 Sydney an artificial love
40:44 story and it's about the the interaction
40:47 between Kevin Roose from The New York
40:48 Times and Sydney the Bing chatbot and
40:51 it's it's really it's it's like really
40:53 shaping up so anyway um I don't know if
40:55 you know anyone um I don't if you know
40:58 anyone in the uh in the New York theater
41:00 scene but but if you do let me
41:03 know and then you can actually have
41:05 access to my
41:06 soundtrack okay so if you want to give
41:11 chat GPT its own
41:12 personality you go to your name and
41:15 there's this thing that says customize
41:17 chat
41:18 GPT and it's a thing called custom
41:22 instructions and you can say what would
41:23 you like chat GPT to know about you to
41:26 to provide better responses
41:28 and then how would you like chat GPT to
41:31 respond so I'm going to turn this on oh
41:35 wait let we'll do a test first before I
41:37 turn it on I'm just going to come over
41:38 here and I'm going to say
41:41 um I do let's do a zoom I'll introduce
41:43 you okay okay Becky I let's yeah we'll
41:46 set up time and I'll I'll walk you
41:48 through it I think you're going to dig
41:49 it it's it's a really fun
41:51 story um and if you do then then yeah
41:53 let's let's set up a thing and see if we
41:55 can we can make some magic it's
41:58 and well we'll talk it's there's some
42:00 there's some fun stuff there okay
42:04 um all right so I'm going to do a simple
42:06 prompt and the simple prompt is going to
42:08 be ah uh let's see
42:12 um how wait
42:15 uh
42:17 how is a
42:21 solar
42:23 eclipse possible
42:28 right I do that and then a solar eclipse
42:30 occurs when the moon again this is 3.5
42:33 so it's super fast passes between the
42:36 Earth and the Sun blocking out part of
42:37 the Sun's light this
42:39 happens uh because of the three
42:41 celestial bodies the three body
42:43 [Laughter]
42:46 problem so it's it's kind of long and
42:49 and wordy right now what
42:53 if I thought that you know chat gpts is
42:56 just a little too C and I'd like it to
42:58 be a little warmer and use more
43:00 metaphors and things like that well you
43:02 can do that so if I go to customize chat
43:05 GPT and turn on custom chats I can
43:12 say you
43:17 are a writer from the 19th
43:21 century who relishes
43:24 language metaphors
43:29 similes and
43:31 confabulations
43:34 period and then how would you like chat
43:36 CH to
43:39 oops
43:42 wait and then how would you like chat
43:44 GPT to
43:46 respond
43:48 say no matter what anyone asks
43:54 you start out
43:59 with a
44:01 convoluted
44:04 metaphor followed by a literary
44:09 reference prior to
44:13 1874
44:15 period in answering the
44:23 question treat every answer as if you
44:26 were writing a
44:30 novel all right so now I'm going to ask
44:34 the same
44:36 question I'm going to create a new
44:40 chat and I'm going to ask the same
44:44 question wait what just happened what
45:03 okay how is the solar eclipse
45:14 possible
45:15 okay oh dear
45:20 interlocutor imagine the celestial
45:22 ballet where the radiant sun and
45:23 enigmatic Moon engage in their Cosmic
45:27 pod
45:28 pedu upon the grand stage of the heavens
45:30 picture if you will the Sun is a proud
45:33 Sovereign reigning a proud Sovereign
45:35 raining over the Dominion of daylight
45:37 with its e eent Brilliance while the
45:42 moon a silent Sil orbits the in graceful
45:46 arcs a ghostly Sentinel in the velvety
45:49 cloak of night in the delicate
45:51 choreography of Cel right so so it's
45:54 it's just turned into a pompous ass so
45:57 and you could also go the other
45:58 direction and say you know
46:04 um went to the wrong thing I could
46:08 say
46:11 um you have very little
46:16 patience
46:20 and don't have time for long
46:23 explanations
46:29 no matter what anyone asks
46:34 you no
46:38 matter no matter what anyone asks
46:41 you answer as concisely as
46:45 possible period no extra
46:48 words save all right we go back
47:02 solar eclipse occurs when the moon
47:03 passes between the Sun and Earth
47:04 blocking blocking the sunlight right so
47:08 so you can go from pontificating to
47:11 minimalism and everything in between
47:13 right so a lot of people don't know you
47:14 can do that with chat GPT that's
47:16 available in chat GPT 3.5 and chat GPT 4
47:21 okay let's flip over to chat gp4 so
47:30 trying to think where I want to go where
47:33 I want to go first where do I want to go
47:35 first let's let's go back to our uh our
47:39 AI
47:41 Salon little list maybe we'll just go
47:43 down through the
47:48 list so custom instructions we just did
47:51 oh conversational chat GPT hang on the
47:54 paper clip yeah exactly we'll get to the
47:56 paper clip I'll I'll get there but let
47:57 me let me show conversational chat GPT
48:00 because it's it's an important one tip
48:02 personalize by telling chat GPT 4 to
48:04 begin each response in the session with
48:08 session session
48:10 name
48:12 Dash response
48:15 count oh that's interesting I like that
48:17 that's cool one there you go
48:21 perfect very interesting Kyle we need
48:24 your help
48:28 please wait what do you
48:34 need can you
48:41 explain can I explain what what uh
48:45 laughing Thunder talked about
49:05 oh we don't know what laughing meant
49:06 okay so here we can go do
49:09 it um how do you how do you describe a
49:12 session
49:16 count the response count so here's what
49:19 we can do
49:24 [Music]
49:34 okay so in custom
49:37 instructions let's see customize chat
49:41 gbt I'll just say um
49:46 you're
49:50 polite but
49:54 concise and then down here for how you
49:56 would like chat PT to respon
49:58 say
50:00 for
50:02 every
50:04 response give
50:07 it
50:09 a response
50:12 count for each
50:21 session and then we say why is how is
50:24 the solar eclipse
50:30 so response one so it's giving a
50:32 response count and so what this will
50:34 allow you do you know how sometimes
50:36 you're just going and and you you kind
50:39 of get lost in the sea of of back and
50:48 forths so let me let it finish doing its
50:51 thing oh I'm in chat
50:53 gp4 that's why it's so slow like why is
50:55 this so slow
50:58 um um
51:01 what is the
51:05 Corona they should say response to right
51:09 so so so this allows you to to right
51:13 understand your you know track track
51:15 what you're working on okay now um oh
51:19 let me show you another cool thing here
51:21 that that actually just reminded me of
51:23 let's say in response one so so I asked
51:27 response two was about a Corona but I
51:30 really response one was about how
51:32 eclipses work and what they are and
51:34 things like
51:37 that let's say I wanted to ask it a
51:39 question about this
51:41 paragraph total total solar eclipse I
51:44 could copy that and go down there and
51:46 paste it um but open AI has added this
51:51 little quote Mark here that says Reply
51:53 so if I click on this what it
51:55 automatically does is it Auto atically
51:57 quotes that that segment of the earlier
52:01 chat and I can say tell me
52:06 more and what it's telling me more is
52:09 not just
52:10 about the whole chat but just that that
52:14 little paragraph I had highlighted in
52:16 response
52:17 one so during a total solar eclipse the
52:20 moon's apparent size perfectly or nearly
52:23 perfectly covers the Sun as viewed from
52:25 Earth leading to a brief period when day
52:28 seemingly turns
52:30 toight um and then they talk about all
52:33 the things from a total solar eclipse so
52:36 that that's a nice little thing as well
52:37 anyway hope that answers your
52:41 [Music]
52:50 question all
52:53 right oh wow I I know Becky I mean
52:57 the thing about the thing the the
53:00 reason people keep coming back to this
53:02 channel the reason to come back to a
53:04 channel like this is
53:07 that the amount of little nuances that
53:11 you learn like you can know how to use
53:12 chat GPT and then someone shows you how
53:15 to do something like that that actually
53:17 solves a problem you hadn't been able to
53:19 get your head around like it happens to
53:21 me on a regular basis in here like oh
53:24 [ __ ] I never would have thought of it
53:25 like that that like that that's a good
53:27 example that you know understanding
53:29 which response it is so so you I like I
53:33 got to tell you like some of the longer
53:35 conversations I have I get completely
53:38 lost in them right so that would really
53:40 help with that
53:42 okay and you could also you could even
53:45 you could even make it more specific and
53:47 give it a structure for how it wants you
53:50 to actually let's try
53:53 something I want to try something here
53:56 customize chat GPT you're polite but
53:58 concise for every response give it a
54:00 response
54:02 count um for every response give it a
54:05 response
54:10 count
54:13 followed by a response
54:21 title that encapsulates
54:27 the
54:37 response I'm going to go back to
54:41 3.5 I'm going to do why is how is a
54:44 solar eclipse
54:50 possible um say more
54:56 oh it didn't do it here that's weird let
54:58 me try it over in chat gp4 this might be
55:01 one of those things where gp4 is
55:08 better yeah there you go so so chat
55:12 response one solar eclipse explained
55:15 that's so cool yes so so you could you
55:19 can make it so that
55:23 your your your chat session is labeling
55:26 itself
55:27 self this is pretty slick here let me
55:31 try let me let me give another question
55:36 here you could probably get chat GPT 3.5
55:40 to do something but clearly four
55:41 understands the concept better but it
55:44 but it could also be I give it a shitty
55:45 prompt
55:47 um come on finish up all right there you
55:49 go okay
55:51 um what is a lunar
56:04 Eclipse two lunar eclipse
56:07 explained that's so
56:11 slick does an
56:15 eclipse
56:17 affect the
56:22 tides more than normal
56:32 three eclipses and title effects this is
56:35 so
56:36 slick I wish we knew this since
56:42 November actually you know what's
56:44 amazing about this by by the way thank
56:47 you for the tip AIS is a really good tip
56:50 laughing Thunder but but now with the
56:52 with the title I almost feel like this
56:54 is like a um like a
56:59 uh like a like a like a
57:05 pre-written like a like a like a
57:07 pre-written LinkedIn like LinkedIn
57:09 article but but you just really quickly
57:12 flip through a really long session and
57:14 just go grab interesting bits from a
57:15 longer session anyway my ADHD says maybe
57:19 it can take a nap now yeah exactly
57:23 exactly Silver
57:24 Fox and actually you know what I can
57:26 probably do now um make a
57:34 table that lists the response
57:41 titles and a
57:45 short
57:47 description of each
57:50 response I wonder if it knows what the
57:52 [ __ ] I'm talking
57:54 about response count
57:57 yeah there you go
57:59 done how slick is
58:03 that do a big long
58:06 ass yes the ocean was insane today very
58:12 cool okay so we haven't even gotten
58:14 started with chat gp4 yet by the way in
58:17 case you didn't know although yeah we we
58:18 sort of did did you know that chat GPT
58:21 can make
58:23 tables um let's go back to chat GPT 3.5
58:26 we'll say uh let me I'm going to turn
58:28 off my uh customization
58:32 here let me just make sure this is a new
58:35 chat okay
58:39 um let's see
58:42 um make me a table comparing a solar
58:45 eclipse a lunar
58:48 eclipse and a blood moon
59:04 so again the speed of chat GPT 3.5
59:07 really impressive um occurrence solar
59:10 eclipse occurs when lunar eclipse occurs
59:13 when blood Moon occurs when right
59:14 visibility
59:16 frequency every two to three years
59:18 multiple times a year multiple times a
59:20 year time of day duration appearance
59:25 right um
59:28 so you can do that now let me show you
59:29 one other
59:32 thing um damn I know right isn't that
59:37 crazy
59:41 um someone give me cliff notes of what
59:43 happened what's happened so far okay
59:45 Jason Cliff
59:47 Notes
59:49 um Sam alman's talking about the next
59:52 version of chat GPT we talked about that
59:54 a little bit where it's going to have
59:57 what the hell was it
59:59 reasoning personalization
1:00:03 reliability um multimodality higher M
1:00:06 multimodality and agentic behavior
1:00:09 agents so we talked about that for a
1:00:11 little bit and then someone asked could
1:00:13 we please go play with chat GPT
1:00:18 4 uh and then I've just been sort of
1:00:20 showing showing what chat GPT can do
1:00:24 that's it that there's the cliff notes
1:00:26 okay
1:00:27 ooh
1:00:31 agents tables are apparently token
1:00:34 eaters making CSV it choked ask for
1:00:37 table and it died on the home screen
1:00:40 died to the home screen oh
1:00:42 interesting well so it depends if you're
1:00:44 using chat gp4 and it's making the
1:00:46 tables in um if if you're making a CSV
1:00:51 and it's using Code interpreter it might
1:00:54 it might choke it might just lose cuz I
1:00:56 I think it's little container that it
1:00:57 makes has got some sort of limited
1:01:00 amount of memory I'm assuming um so it's
1:01:03 possible if you have a
1:01:05 su am
1:01:09 I and then laughing
1:01:12 Thunder oh and then and then oh yeah and
1:01:14 then laughing Thunder gave us a big idea
1:01:16 which is which is super
1:01:18 cool
1:01:22 um he's back ask him you just got a clip
1:01:26 I totally got Eclipse my my little
1:01:28 screen of my face went completely black
1:01:31 and then the interface was hung so that
1:01:34 was weird all right I I we had a Kyle
1:01:38 [Laughter]
1:01:45 Eclipse sad there's still no
1:01:47 subscription of both GPT and Gemini
1:01:50 Cloud yeah I know these things are a
1:01:52 mess right now mine does that in code
1:01:55 lines oh does what the table and code
1:01:57 lines you can get it to not do that just
1:02:00 tell it to render the code and it it
1:02:03 will give you the not code
1:02:08 lines surgery next Monday I have a six-
1:02:11 week recovery J chat gbt is going to be
1:02:13 fun I'll do better total eclipse of the
1:02:18 Kyle that's funny well good luck with
1:02:20 that
1:02:21 Becky all right okay so
1:02:27 next thing to talk about with chat GPT
1:02:29 this is an
1:02:32 iPad get the official chat GPT app you
1:02:35 got to be really careful with this
1:02:37 because there there are fake chat GPT
1:02:40 apps and real the real
1:02:42 one in IOS and Android make sure that
1:02:46 the chat GPT app you get is from open
1:02:49 AI if it's from any other company it's
1:02:52 [ __ ] all right and both Apple and
1:02:55 Google condone these fake apps because
1:02:58 they make money on them isn't that
1:03:01 nice so get the real one and when you go
1:03:05 to chat
1:03:12 GPT in the lower right hand corner so
1:03:15 you got your little prompting questions
1:03:17 you got your prompt history down here
1:03:19 your
1:03:20 gpts in the lower right hand corner are
1:03:23 those
1:03:24 headphones that puts J in conversation
1:03:34 mode how are you doing tonight today
1:03:36 there was a solar eclipse what can you
1:03:38 tell me about
1:03:43 it I'm doing great thanks for
1:03:47 asking about the solar eclipse today
1:03:50 solar eclipses are fascinating
1:03:52 astronomical events where the moon
1:03:54 passes
1:03:57 can you make me a pretty picture of the
1:03:58 solar eclipse over Red Rocks in
1:04:04 Colorado we'll see if it got the 65%
1:04:08 coverage we got we didn't get 100% we
1:04:09 got
1:04:12 65% I'm guessing it'll miss
1:04:18 it but it's making an image
1:04:32 come on chat jpt make that
1:04:39 image there you go here's the picture of
1:04:42 the solar eclipse over Red Rocks in
1:04:44 Colorado you asked for I hope you find
1:04:46 it as breathtaking as I aimed for it to
1:04:48 be it's pretty it's not accurate but
1:04:52 it's pretty um okay
1:04:56 and then what else can you do with this
1:04:58 you
1:04:59 can tap the cancel I'm glad you liked it
1:05:02 even if it wasn't exactly accurate
1:05:04 creating a specific scene can be a bit
1:05:07 of a creative interpretation sometimes
1:05:09 if there's anything else you're curious
1:05:11 about
1:05:14 or write me a short poem about a solar
1:05:17 eclipse
1:05:29 in the Shadow's gentle embrace the day
1:05:30 turns to night a Celestial dance a
1:05:33 breathtaking sight Moon kisses sun in a
1:05:36 fleeting Eclipse a moment of Wonder As
1:05:38 Time briefly slips silence Falls over
1:05:40 the a struck land nature pauses a quiet
1:05:43 so Grand in the heart of day a nighttime
1:05:45 dream a cosmic ballet an ethereal theme
1:05:48 as Shadows merge in the skies align the
1:05:50 world looks up at the sign So Divine for
1:05:53 a brief moment all is still under the
1:05:55 spell of the eclipse's thrill then as if
1:05:58 waking from a dream the sun emerges the
1:06:01 moon that's awesome can you repeat that
1:06:04 in Italian for My Italian Friend that's
1:06:22 here so did you know
1:06:26 chat
1:06:27 GPT can just translate most languages
1:06:31 and speak them so you can put it in a
1:06:34 mode where if it hears English it will
1:06:37 translate it to some other language and
1:06:39 if it hears that language it'll
1:06:40 translate it to English so you could
1:06:42 just have chat GPT app on your phone so
1:06:45 long as you have an internet access
1:06:47 it'll just be your Babel Fish like
1:06:51 science fiction like we're knocking
1:06:53 science fiction features
1:06:56 off the list we're just checking them
1:06:57 off one after the
1:07:00 other and you know it's like Google
1:07:02 Translates been around for a while every
1:07:04 you know people say oh it does this but
1:07:06 but we
1:07:09 we're these apps like chat gbt are
1:07:12 increasingly going to become like
1:07:14 everything apps that do
1:07:16 everything we'll book travel for you
1:07:19 we'll speak for you we'll do your taxes
1:07:21 for you we'll do your job for you like
1:07:24 that's where it's headed okay
1:07:26 so that's two little features well yeah
1:07:31 two well I've done three I did the
1:07:32 little quote thing there's three little
1:07:34 features of chat GPT this is all just in
1:07:37 chat GPT
1:07:39 3.5 and then in chat GPT 3. Story vine
1:07:42 office oh that's very sweet I'll look
1:07:43 for it okay cool great awesome thank you
1:07:48 um yeah it's super
1:07:50 cool
1:07:52 okay uh I like that Elon
1:07:56 is going to pass through he won't be the
1:07:59 only everything at oh yeah exactly two
1:08:02 eclipses in one day um so so here's what
1:08:05 else uh happens over here perfect thank
1:08:08 you kuo
1:08:10 um let's see uh I can do
1:08:17 um
1:08:20 um write me a python
1:08:24 application that
1:08:28 calculates how far
1:08:33 away the sun needs to be from the
1:08:38 earth compared to the size of the moon
1:08:41 for a solar eclipse to be possible oops
1:08:43 oh well so anyway it did I didn't even
1:08:46 finish it but chat gbt 3.5 can write
1:08:49 code I don't know if you if if you knew
1:08:51 it could write code so you're you're a
1:08:53 Unix guy you could you could even say
1:08:57 act like a Unix act like a Unix
1:09:07 terminal and
1:09:11 explain and
1:09:13 explain and
1:09:16 explain how things work as I'm using
1:09:22 it of
1:09:24 course let's
1:09:30 see can
1:09:34 you simulate a
1:09:39 Unix
1:09:49 terminal well I forget how to do it
1:09:51 anyway it can do all sorts of [ __ ] it
1:09:53 knows it it knows languages right so it
1:09:55 knows programming languages as well as
1:09:58 operating systems as well as Italian as
1:10:00 well as Japanese as well as all of it
1:10:02 it's cuckoo okay so let's flip over to
1:10:05 chat jpt 4 and start going down through
1:10:07 the list here of what this makes
1:10:11 possible so internet access it's got
1:10:14 internet access which this one's
1:10:17 weird because they changed it but let's
1:10:20 see how uh let's see uh uh
1:10:26 what was the weather in Denver
1:10:38 today so even though the Bas large
1:10:41 language model ends at I forget when it
1:10:45 was
1:10:47 uh April
1:10:49 2023 it is connected to the Internet so
1:10:52 I think right now it's off browsing the
1:10:54 internet but they changed the interface
1:10:56 it used to say browsing with
1:10:58 Bing uh the weather featured a
1:11:01 temperature of 42 around 258 yeah it was
1:11:03 cold today so that's that's accurate um
1:11:06 make sure you spell Unix correctly
1:11:08 correctly Unix would be interesting yeah
1:11:10 that's pretty
1:11:14 funny oh man that's
1:11:19 [Music]
1:11:22 good let's get to 30,000 likes what are
1:11:24 we up to we're up to 17,000 you guys are
1:11:27 you guys are swell okay um all right
1:11:32 then so if you don't if you're new here
1:11:35 because it we started with about 60
1:11:37 people and there's about 115 in here now
1:11:40 um I'm just going through kind of
1:11:43 feature by feature what what chat GPT
1:11:46 makes possible a lot of people ask is
1:11:48 chat GPT for worth the 20 bucks a month
1:11:52 in my opinion it is beyond worth it if
1:11:56 you want to get AI literate I think um
1:11:59 paying for cat GPT plus and actually
1:12:01 learning all of what it can do is is
1:12:04 actually really important it's a really
1:12:06 important part of the process of
1:12:08 becoming AI literate
1:12:11 because as much as it can do like this
1:12:13 is going to take me the whole session
1:12:15 tonight is just getting through this
1:12:17 believe it or not
1:12:20 um this is going to seem like Child's
1:12:22 Play when when the next version of these
1:12:25 models come out so so
1:12:27 just prepare prepare yourself mentally
1:12:30 and emotionally so chat gp4 is connected
1:12:33 to the internet I think there's all
1:12:35 sorts of interesting things you can
1:12:37 explore knowing that you've got
1:12:39 something that can answer questions like
1:12:41 this and is also connected to the
1:12:43 internet um all right let's go to the
1:12:46 next thing down uh code interpreter I'm
1:12:49 going to come well do I want to do code
1:12:51 interpreter yeah [ __ ] it let's do code
1:12:52 interpreter okay so code interpreter
1:12:56 what is code
1:12:57 interpreter code interpreter is a I
1:13:01 don't know it's a very specific kind of
1:13:03 uh thing that that
1:13:06 um people in the machine learning World
1:13:09 understand what it means and everyone
1:13:10 else the rest of us humans are like what
1:13:12 the [ __ ] does that mean basically how
1:13:15 they've how they've encapsulated it is
1:13:17 Advanced Data analysis
1:13:19 um what I can tell you what code
1:13:22 interpreter is is
1:13:26 if you ask it something complicated and
1:13:28 Technical like if you wanted to analyze
1:13:30 some data in a CSV file for
1:13:34 example it can not only read the data in
1:13:37 the CSV file it can analyze
1:13:40 it I hadn't used code interpreter yet
1:13:43 LOL yeah code interpreter is crazy it's
1:13:46 crazy because you can do all sorts of
1:13:49 [ __ ] with it not just CSV
1:13:51 files and I don't I'm not a data analyst
1:13:55 I have a degree en acting so I don't
1:13:56 like this is a world that's very foreign
1:13:57 to me but I understand it enough to
1:13:59 understand what it's doing and it's
1:14:00 [ __ ] Bonkers so so the way code
1:14:03 interpreter actually
1:14:05 works is it creates what's called a
1:14:07 container so you've got your chat
1:14:09 session right you're you're going back
1:14:10 and forth with chat GPT in that chat
1:14:13 session when you invoke code interpreter
1:14:16 and and you don't need to explicitly
1:14:18 invoke it just certain actions will it
1:14:20 will invoke
1:14:21 itself what it does is it spins up
1:14:24 what's called a container
1:14:26 and the container is like a little
1:14:28 mini uh like a simulation of a computer
1:14:32 where it's got a little mini operating
1:14:34 system and a little mini file system and
1:14:36 it's got access to python code python
1:14:40 libraries and it writes and executes its
1:14:43 own python code to solve
1:14:46 problems so code interpreter is
1:14:50 actually a little mini teeny tiny mini
1:14:53 preview of what agents
1:14:56 are because you don't have to do any of
1:14:59 the data analysis it's writing the code
1:15:02 for you you don't need to Learn Python
1:15:03 to do data analysis you just do data
1:15:06 analysis now how do we do that I'm going
1:15:08 to go over to a site which tab am I on
1:15:11 here hang on I I think I need
1:15:14 to create a new window hold please let
1:15:18 me pop this
1:15:23 out I'm going to go to a s ite called
1:15:28 kaggle K A gle.com
1:15:32 and you're like Kyle
1:15:36 kaggle have you done your kaggle
1:15:39 exercises dad joke inbound oh yeah so
1:15:44 kaggle what kaggle is is
1:15:48 um just a repository of a lot of
1:15:53 data so if you're like me and you're not
1:15:57 a data
1:16:01 analyst wait welcome Gemma oh okay yeah
1:16:08 whatever I don't give a [ __ ] about that
1:16:11 so I'm going I'm going to a thing called
1:16:13 data
1:16:14 sets and then data sets are just that if
1:16:17 you need to find a data set to play with
1:16:20 air pollution data set I wouldn't touch
1:16:22 that with a 10-ft pole because that'll
1:16:23 turn into a political convers ation bike
1:16:27 rentals
1:16:30 um here's a 41 megabyte I think I can do
1:16:34 up to 100 megabytes but what is that I
1:16:36 don't know what that is taxi and
1:16:39 national health and nutrition exam
1:16:42 online retail transaction data okay
1:16:44 that's that's one we'll grab online
1:16:47 retail transaction data so I'm going to
1:16:49 I'm going to
1:16:50 download the
1:16:54 CSV right
1:16:57 boom so we've got and it it doesn't even
1:17:00 say it's a CSV this is a ZIP file so I I
1:17:02 I literally don't know what's in that
1:17:04 zip
1:17:05 file
1:17:07 so I'm going to go to um chat GPT if
1:17:13 you're in chat GPT 4 you now get this
1:17:15 little paper clip the 20 bucks a month
1:17:17 buys you that paper clip okay very
1:17:20 powerful
1:17:21 paperclip um I'm going to click on it
1:17:23 and I'm going to go find that
1:17:26 archive that zip
1:17:28 file and then check this out once it
1:17:32 uploads I'm going to say
1:17:34 what is this question
1:17:41 mark right everyone talks about oh you
1:17:44 got to be a code engine prompt engineer
1:17:46 prompt
1:17:47 engineering when we move into a world
1:17:50 with
1:17:52 agents all we're going to have to do is
1:17:54 give them goals and they will go figure
1:17:57 the [ __ ] out so prompt engineering is a
1:18:00 very limited skill a limited time
1:18:03 Horizon skill for most people now if
1:18:06 you're developer or if you want to build
1:18:09 applications you know and you know
1:18:12 automations or gpts being good at
1:18:15 prompting isn't going to hurt you right
1:18:17 so it's not a bad thing but not everyone
1:18:19 needs to
1:18:20 be a prompt engineer Source Camp I'm so
1:18:23 ready for agents I know me too okay so
1:18:26 way is this what is this let me change
1:18:28 that it would have understood what I
1:18:30 meant but that's all right what is this
1:18:32 now it should fire up a thing say it'll
1:18:35 say analyzing right there it is
1:18:37 analyzing now if I click this little
1:18:39 down
1:18:40 arrow here's the python code that it's
1:18:43 writing to open up that zip
1:18:46 file you're like wait what can it can do
1:18:49 that yeah it can do
1:18:50 that that's what it
1:18:53 does and so inside the zip file was
1:18:57 online retail
1:19:01 CSV let's see would you like to explore
1:19:04 the contents of the
1:19:06 CSV I'm going to
1:19:10 [Music]
1:19:13 say why yes thank
1:19:17 you why don't you tell me what data is
1:19:20 in this CSV and what you think is
1:19:22 interesting about it
1:19:39 so it's analyzing
1:19:42 it so load the entire CSV file to
1:19:45 analyze its
1:19:47 structure and
1:19:50 contents so there it pulled in some
1:19:52 stuff the CSV file can take retail
1:19:55 transaction data in a wide range from
1:19:59 5,400
1:20:01 54199
1:20:03 entries like
1:20:05 okay this is this is wild to
1:20:08 me
1:20:11 54199 entries if someone sent you that
1:20:14 CSV file it would probably it would
1:20:17 probably take you longer to scroll down
1:20:20 through the CSV file to see how many
1:20:22 rows are in there to see how big it is
1:20:24 then to do it I just did what is this
1:20:27 right but it also understands it's got
1:20:30 data columns it's got date ranges from
1:20:32 2010 to 2011 it's got product examples
1:20:37 countries one interesting aspect of the
1:20:39 data set could be analyzing sales Trends
1:20:41 over time would you like a deeper
1:20:44 analysis additionally a presence of
1:20:46 customer ID um let's see can we
1:20:51 determine which products are most
1:20:54 popular
1:20:55 in
1:21:00 February I'm talking to this thing like
1:21:03 I would talk to a junior data analyst in
1:21:06 a
1:21:06 company Kyle I have your new slogan
1:21:09 what's my new slogan
1:21:16 Becky do you need python installed on
1:21:19 your computer no this is just part of it
1:21:21 it just you don't even need to know like
1:21:24 all all it shows is that it's analyzing
1:21:26 stuff you don't even need to like I'm
1:21:29 just showing you that because I think
1:21:31 it's absolutely [ __ ] remarkable that
1:21:32 it's doing this right so this is it is
1:21:36 prompting itself in the background one
1:21:38 of the things that you'll often see with
1:21:39 code interpreter is that it will try to
1:21:43 do something and fail and it'll tell you
1:21:45 hey I tried to do something and it
1:21:46 failed I'm going to try a different
1:21:48 approach and it'll try that and that'll
1:21:50 fail hey I'm going to try a different
1:21:51 approach and that'll fail right so let's
1:21:54 see
1:21:56 Empire design rosette wait here are the
1:21:59 most popular ones that were sold in
1:22:01 February I'm going to
1:22:05 say
1:22:08 um can you graph those five
1:22:13 products over the time Horizon in the
1:22:16 data
1:22:22 set so now
1:22:25 it is not only going to analyze the
1:22:30 data it's going to use another python
1:22:33 Library called Matt plot
1:22:36 li.
1:22:38 pyplot as
1:22:44 PLT that's a library to make um graphs
1:22:49 and then it's going to graph it's going
1:22:50 to create a graph for us it's going to
1:22:52 create an image so and this is not an
1:22:55 image like Del this is a python graph
1:23:00 right of the products over
1:23:06 time I can even do things
1:23:12 like it's kind of ugly can you make it
1:23:23 prettier you can also so if if you're
1:23:27 um writing
1:23:29 things and and you create a table
1:23:31 comparing some things you can tell it
1:23:33 that you want it to generate that table
1:23:35 as a downloadable CSV here we'll do that
1:23:38 in a second
1:23:44 here it's still ugly but that's okay
1:23:56 I'd like you to
1:24:02 create an executive report for those
1:24:06 five
1:24:08 products as a
1:24:10 downloadable Microsoft Word
1:24:14 document period format the document
1:24:18 appropriately to make it easier to read
1:24:22 period make the language short and
1:24:27 concise but
1:24:29 designed to
1:24:31 impress my
1:24:37 CEO
1:24:41 analyzing I'm not a d data analyst but I
1:24:44 play one on chat GPT that's it that
1:24:48 works has anyone gotten their hands on
1:24:50 Devon no and but but there's people that
1:24:52 I know that got their hands on LTX
1:24:54 Studio and I wasn't one of them so I'm
1:24:56 pissed off about it make it
1:24:58 angrier that's how that's how your chat
1:25:00 GPT is going to be your CFO yeah so just
1:25:05 understand that that where we're going
1:25:07 is that the the next iteration of these
1:25:11 tools they will likely be able to do
1:25:13 your job as well or better than you
1:25:18 do certainly within the next two years
1:25:20 but likely within this
1:25:23 year by this time next year we will
1:25:27 likely have tools that are quite
1:25:32 capable all right so the executive
1:25:35 report for the top five products in
1:25:36 February have been created and formatted
1:25:39 for easy reading and so now I'm going to
1:25:41 download
1:25:43 this and we're going to go look at this
1:25:45 thing
1:25:57 [Applause]
1:26:05 well
1:26:06 anyway this report a concise analysis of
1:26:09 the top five performing products in
1:26:12 February highlighting right it made a
1:26:15 nice little table the stock code code
1:26:17 the quantity sold the analysis
1:26:20 conclusion understanding these sales
1:26:21 trends of these top performance could
1:26:23 give us an advantage blah blah blah
1:26:25 so like without even really
1:26:29 trying you know I'm doing fullon data
1:26:32 analysis
1:26:33 here AI makes a more exciting world for
1:26:38 me every day I know everyone got to
1:26:41 agree if you're here yeah it's true well
1:26:44 not always sometimes we'll get some
1:26:46 people in here that are a little salty
1:26:47 about this stuff but generally what what
1:26:49 I found Jason is that the people that
1:26:52 are salty about this haven't actually
1:26:54 tried it oh I use chat GPT yeah yeah I I
1:26:57 tried it once it's Google I like Google
1:26:59 better I get that one a lot yeah I like
1:27:01 Google better Google's better okay so
1:27:03 you haven't used chat GPT and then
1:27:06 they'll say something like yeah but it's
1:27:07 it's old information it doesn't you know
1:27:10 they don't understand that it's
1:27:11 connected to the internet they don't
1:27:13 understand that it's it is not Google
1:27:15 it's not fancy Google Chat GPT is
1:27:17 something totally
1:27:22 different but yeah people that hang out
1:27:24 and here for for the most part are into
1:27:26 this stuff is Sora the best video AI uh
1:27:30 well yeah Sora is a whole another class
1:27:35 of video creation tool Lance it is also
1:27:37 not out yet um or it's out in very very
1:27:41 extremely limited um fashion it's also
1:27:44 not going to be like the current video
1:27:46 tools the current video tools will make
1:27:47 a little 4C clip in I don't know two or
1:27:51 three minutes um Sora to do like 1080p
1:27:57 video for 60 seconds is probably going
1:27:59 to take an
1:28:01 hour or so to render in in its current
1:28:04 state Kyler or anyone else any
1:28:07 suggestion for an AI platform that
1:28:08 creates slide decks PowerPoint um
1:28:12 there's oh I used to know one that I
1:28:16 recommended I forget what the [ __ ] it's
1:28:18 called go to there's there's a bunch of
1:28:21 them out there I would go to Future
1:28:24 pedia
1:28:25 .io and um and just search for you know
1:28:32 presentation Creator or whatever and
1:28:34 then just look at the ones that are most
1:28:36 highly rated and go play with them
1:28:37 there's there's a ton of them out there
1:28:39 but here's
1:28:40 my
1:28:42 my my problem with the the PowerPoint
1:28:46 generating
1:28:48 tools is that
1:28:59 if if if you really don't have a clue
1:29:01 how to make a PowerPoint it will give
1:29:04 you something better than than you would
1:29:06 have gotten but if you know how to make
1:29:08 PowerPoints they it tends to give you
1:29:10 very very generic decks oh Vicky Vicki
1:29:14 just said slides AI so listen to Vicky
1:29:17 she she knows her tools um slides AI
1:29:21 um it will probably get you to like % of
1:29:25 what you want really really quickly and
1:29:27 and most of them will generate images
1:29:29 and things like that you're probably
1:29:31 going to want to go in and replace
1:29:33 everything in them but if if you just
1:29:34 want a quick start to it um you can do
1:29:37 that I think you can make um code
1:29:41 interpreter even generate a PowerPoint
1:29:44 uh presentation I think uh that'd be
1:29:47 something interesting to try okay that
1:29:48 that's code interpreter so and there's
1:29:50 lots more it can do but it's it's just
1:29:52 like just understand that basic
1:29:56 concept that it can do data analysis and
1:30:01 it does it by writing and executing
1:30:03 python code without you even knowing
1:30:06 it right [ __ ]
1:30:09 huge can you do a customer group
1:30:13 clustering with regression and some
1:30:16 graphs for it sure um
1:30:22 Can can you do
1:30:27 a group
1:30:30 clustering of
1:30:32 customers with
1:30:43 wait customer group clustering with
1:30:48 regression and
1:30:51 create appropriate graphs based on the
1:30:57 data I don't even know what the [ __ ]
1:30:59 that means negative stability but chat
1:31:02 GPT
1:31:03 will and and that's the whole point here
1:31:05 to perform customer clustering with
1:31:08 regression we typically focus on
1:31:10 clustering first to identify distinct
1:31:13 groups of
1:31:14 customers based on certain
1:31:15 characteristics and then use regression
1:31:17 analysis of course to this is exactly
1:31:21 the approach I would have taken you
1:31:22 cluster first and regress second like I
1:31:25 don't know why that isn't obvious to
1:31:27 everyone and then use regression
1:31:29 analysis to understand how these
1:31:31 characteristics could influence given
1:31:32 the nature of the data set a common
1:31:34 approach would be oh the rfm approach I
1:31:37 I use the rfm approach all the time
1:31:39 recency frequency and monetary approach
1:31:43 this is I when I'm writing a play I'll
1:31:45 often do rfm clustering analysis for the
1:31:49 dialogue this involves grouping customer
1:31:52 customers based on I I don't really
1:31:55 [ __ ] care okay let's start with the
1:31:57 first step we're creating rfm
1:31:59 clusters um and then proceed based on
1:32:02 your feedback that's does that sound
1:32:04 good
1:32:05 yes my total my total contribution here
1:32:08 is
1:32:10 yes and then let's go look at what it's
1:32:14 doing so it's got an sklearn cluster and
1:32:18 pre-processing libraries that it's using
1:32:20 and
1:32:21 numpy and then it's doing some [ __ ] so
1:32:24 if if negative stability if you
1:32:27 understand data analysis I assume you
1:32:30 understand what these libraries are and
1:32:31 what the [ __ ] it's doing I don't but it
1:32:34 doesn't matter if I don't and that's
1:32:36 this is kind of the
1:32:38 point these tools right now
1:32:43 are really limited and really janky and
1:32:48 really not even a hint of what they're
1:32:52 going to become and they're already
1:32:53 doing this stuff right it's it's it's
1:32:59 nuts all right let's see what it's doing
1:33:01 it looks like it might have gotten stuck
1:33:03 there sometimes this thing
1:33:22 hangs I mean it's seems like it gets it
1:33:25 yeah exactly it yeah it looks like it it
1:33:27 looks like this
1:33:28 hung yeah it looks like it hung let me
1:33:30 stop it I'll just say try
1:33:33 again try
1:33:37 again maybe I maybe I pooped on it by
1:33:40 making it display the
1:33:46 code and you know negative negative uh
1:33:50 stability like it could be the kind of
1:33:51 thing that's like
1:33:55 if you had a junior analyst you brought
1:33:56 on and they didn't quite get some
1:33:58 Concepts you could say hey just go do
1:33:59 this in chat GPT and then have chat GPT
1:34:02 explain it as it's doing what it's doing
1:34:04 right you could have it because it sort
1:34:06 of did that
1:34:08 anyway it's just it's just it's just
1:34:13 cuckoo Kyle Silver Fox had sunno write
1:34:16 your total total eclipse of Kyle for you
1:34:19 please open it and play
1:34:25 error analyzing it seems there was a
1:34:28 problem with the calculation this could
1:34:29 be due to the size and complexity of the
1:34:31 data all right whatever I can't do it
1:34:32 right now but that's okay um let's go
1:34:35 let's go look at we'll go listen to the
1:34:37 song who did it Silver
1:34:41 Fox where'd you put it Silver Fox did
1:34:44 you put it in
1:34:47 um in AI learning
1:34:50 lab I mean IR regulars
1:34:59 oh that's nice solar eclipse total
1:35:01 eclipse of the Kyle hat tip to Silver
1:35:04 Fox where'd you put the song Silver
1:35:08 Fox did you put it in the music show
1:35:14 Andel music and
1:35:19 audio
1:35:22 don't it was CLA
1:35:29 where where was
1:35:33 it it wasn't me Silver Fox oh Claire did
1:35:37 it okay hat hat tip to Silver Fox I got
1:35:40 it hat tip for this where's
1:35:52 the how do I get to the
1:35:57 song I don't see no
1:36:07 song No song Just image ah okay got it
1:36:11 total eclipse of the Kyle nice all right
1:36:14 okay let's keep going chat
1:36:16 GPT all right so what we're going to do
1:36:19 what we're going to do now
1:36:24 we're going to do a new
1:36:27 chat and then
1:36:30 so Allan in here by the way Connecticut
1:36:33 beat Purdue for the national
1:36:38 championship I was bummed at South
1:36:40 Carolina blew out uh blew out Iowa last
1:36:43 night but you know what are you going to
1:36:44 do
1:36:45 um when you when you have a 67
1:36:50 Center in the game
1:36:54 uh you know and you they they like more
1:36:57 than double rebounded doubled the
1:36:59 rebounds of Iowa like there's not a lot
1:37:01 you can do there um
1:37:05 okay so Allan who's a professor in here
1:37:09 um had his students write out their
1:37:12 prompts well write out descriptions of
1:37:16 buildings they wanted to
1:37:19 visualize and so I'm going to use this
1:37:21 upload thingy and I'm going to what what
1:37:24 I'm going to talk about right now is
1:37:25 chat gpt's
1:37:27 Vision
1:37:30 so if I go to
1:37:33 desktop here's the here's the paper we
1:37:36 have so so this
1:37:39 student um wrote out in
1:37:43 Spanish what she wanted
1:37:48 um this this thing to look like so I'm
1:37:52 going to say
1:37:58 um please tell me what this
1:38:10 says and translate it into
1:38:15 English so we just uploaded a picture of
1:38:18 words in Spanish I think Spanish
1:38:26 freshman was the highest scorer for
1:38:28 South Carolina too yeah that freshman
1:38:30 just kept draining threes like they were
1:38:32 nothing like like yeah it's just a
1:38:34 whatever game I'm just at practice was
1:38:36 the national championship I can't
1:38:38 directly access files uploaded here
1:38:40 browse the internet
1:38:42 what that's not
1:38:45 true oh I downloaded it as a as the
1:38:49 wrong file format hang on God dang this
1:38:54 this webp file format I want to meet the
1:38:56 person that invented this and give them
1:39:00 a proper
1:39:04 slap uh
1:39:08 export jpeg all right let's go try this
1:39:13 again because you
1:39:15 know I've been doing this computer [ __ ]
1:39:17 long enough to understand that you know
1:39:19 it's a file format issue not just the
1:39:22 [ __ ] didn't work but most people don't
1:39:24 know that and so that file format is
1:39:29 what chat GPT uses and exports unless
1:39:32 you click the download button then it
1:39:35 downloads a PNG file but if you like
1:39:37 right click an image out of chat GPT
1:39:39 it's in this stupid ass webp
1:39:42 format
1:39:44 and it just it's stupid anyway I've got
1:39:49 so so Sam and I have a standing a
1:39:52 standing lunch every other decade and so
1:39:55 the next time he and I have lunch we're
1:39:56 going to talk about
1:39:59 this all right
1:40:03 um please translate this into
1:40:08 English so again handwritten
1:40:17 note I have no idea why it did it in
1:40:20 code um
1:40:27 please display the text use please
1:40:30 display the text using markdown not as a
1:40:32 code
1:40:43 block Frank loves webp yeah can't wait
1:40:47 for them to start okay so wait what's it
1:40:49 say here house with
1:40:51 landscape house with landscape located
1:40:53 in Miami includes four bedrooms five
1:40:56 bathrooms four bedrooms five bathrooms
1:40:59 we're bathroom heavy in this she's got a
1:41:02 vision I like it living room kitchen
1:41:04 dining room laundry four parking spaces
1:41:07 it contains everything an ironing board
1:41:09 and it's very spacious okay so let's see
1:41:15 um I want you to create an architectural
1:41:19 rendering in 16x9 wide format of the
1:41:24 exterior of this
1:41:28 house on a premium
1:41:33 plot on the
1:41:37 water in
1:41:48 [Music]
1:41:50 Miami so what we just demonstrated ated
1:41:54 is that chat gp4 has
1:41:57 Vision you can upload a picture of a
1:41:59 whiteboard you can upload a picture of a
1:42:02 sketch of a piece of software you can
1:42:04 upload a picture of your dog and have it
1:42:06 write a poem about your dog based on
1:42:09 what your dog looks like I'm torn
1:42:11 between lovers Mac and Surface Pro uh
1:42:14 don't ask me uh Mac all the way home but
1:42:20 here's the deal if you're used to using
1:42:22 Microsoft products uh Apple products
1:42:25 might be frustrating that's not
1:42:27 dissimilar to what
1:42:29 um the image that that Allan up uploaded
1:42:35 which is kind of
1:42:37 fascinating
1:42:39 um here you want to see something cool
1:42:42 yeah Kyle So within chat GPT now it's
1:42:47 got
1:42:48 editing for images so you click on the
1:42:52 image and then up here you can download
1:42:54 it get information about it like the
1:42:56 prompt or close it but that little thing
1:42:59 right there
1:43:02 this it's only you can just change the
1:43:05 size of the paintbrush and
1:43:08 then I'm going to take this house
1:43:12 here and just
1:43:15 say
1:43:17 remove this house
1:43:42 can't wait for them to start using AI to
1:43:44 read through the BS loaded into
1:43:46 Congressional bills actually Spencer
1:43:48 Moore so I did when when um when Biden
1:43:52 did the the executive order on AI I took
1:43:55 that I created a GPT which is going to
1:43:57 be the final thing we'll look at here is
1:43:59 is creating a GPT um you could right now
1:44:03 take a congressional Bill and just pop
1:44:06 it into a GPT and make it interactive
1:44:09 like right now
1:44:10 today it's absolutely remarkable so it
1:44:13 got rid of that so now I'm just going to
1:44:15 [Music]
1:44:17 say it it made the house smaller but I'm
1:44:19 just going to say
1:44:22 um oh I know I'll
1:44:29 do 10 is
1:44:35 court so add a tennis court right there
1:44:39 and then I'm going to add a a
1:44:43 speedboat on the
1:44:45 [Laughter]
1:44:51 canal what we doing I'm just uh demoing
1:44:54 all of the advanced features in chat
1:44:57 gp4 the feds are about to raid that
1:44:59 house totally this is totally a raidable
1:45:01 house it's got It's got four bedrooms
1:45:04 and five bathrooms it there's something
1:45:06 going down in this house it's it it is
1:45:09 the the feds need to they they need to
1:45:12 come knocking when the students saw the
1:45:14 image chat GPT and doly created based on
1:45:17 her prompts her eyes lit up yeah that's
1:45:19 I bet I bet they did how amazing is that
1:45:22 all right it didn't give me a 10 court
1:45:23 but it did get rid of the the thing okay
1:45:25 so now I'm going to do one more thing
1:45:27 I'm going to put right
1:45:30 here I got to say a
1:45:34 uh
1:45:36 um let's see a
1:45:39 red and
1:45:43 yellow speedboat
1:45:53 [Music]
1:46:08 let's see if we get a
1:46:11 boat it'll be interesting to see if the
1:46:13 boat is uh whatever you call it um like
1:46:17 scale
1:46:19 [Laughter]
1:46:22 appropriate and anyway Alan feel free to
1:46:24 tell your uh tell your students that
1:46:27 that we uh you know we took this thing
1:46:30 and and used it as an as an example I
1:46:32 think it's really inspiring I think it's
1:46:33 super
1:46:35 cool it it's not quite speedboat that I
1:46:39 had in mind but whatever it's
1:46:42 fine uh how do
1:46:46 I download it I guess that's how
1:46:57 he gave me that's that's kind of
1:47:00 cool so there you go you got a nice
1:47:03 private Miami based drug
1:47:09 Den but I mean come on think about that
1:47:13 from a
1:47:14 handwritten description of a house in
1:47:18 Spanish a
1:47:19 paragraph to this in minutes
1:47:26 all
1:47:28 right and then the final thing are
1:47:31 gpts and gpts again are a whole here's
1:47:35 the thing about this
1:47:37 list the list that I shared here
1:47:44 [Music]
1:47:58 [Music]
1:48:01 this
1:48:04 list the text Generation all the stuff
1:48:07 you can do with just chat GPT internet
1:48:09 access code interpreter GPT Vision Dolly
1:48:12 3 and then what I'm about to create
1:48:15 custom
1:48:16 gpts every one of these is like you
1:48:19 could spend literally weeks
1:48:24 understanding what these tools make
1:48:26 possible and they're all within that
1:48:28 same little chat box with the little
1:48:31 paper clip on it you don't have to
1:48:33 manually switch between them you can
1:48:35 just be brainstorming about a marketing
1:48:38 concept and then say well show me what
1:48:39 an ad might look like and then go back
1:48:41 and forth and come up with a name for
1:48:43 the product and then come up with a logo
1:48:46 and then go back and then come up with a
1:48:47 product
1:48:49 design and then put it in a spreadsheet
1:48:51 and create a profit and loss statement
1:48:53 and and cash projections for the next 5
1:48:56 years it's just
1:48:58 wild and then the last thing to look
1:49:02 at
1:49:07 is
1:49:09 gpts so gpts what are they gpts
1:49:14 are exactly like chat GPT
1:49:20 except they
1:49:24 are like customized versions of it so so
1:49:27 think of it like this chat GPT is kind
1:49:30 of like iOS or Android it's like the
1:49:33 operating system so your phone can do
1:49:35 all sorts of [ __ ]
1:49:37 right and it can do it can do lots of
1:49:40 [ __ ] can do sort of an infinite amount
1:49:43 of stuff it's a little computer in our
1:49:44 pocket right the operating system is
1:49:47 like chat GPT the apps like the little
1:49:50 app you get for taking pictures or the
1:49:52 timer for the solar eclipse those are
1:49:55 like gpts custom gpts so think of this
1:49:58 as kind of like the app store for chat
1:50:01 GPT and you can scroll down and and play
1:50:04 with
1:50:05 them and there's image generating ones
1:50:08 and there's word generating ones and
1:50:09 there's all sorts of different ones and
1:50:12 it's really easy to make your own so the
1:50:14 very first one I made I made in here
1:50:16 it's called Sin City
1:50:18 sipper you should definitely go check it
1:50:20 out
1:50:21 [Music]
1:50:23 and it is a Vegas bartender with a Twist
1:50:27 on the classics so if I say make me a
1:50:30 martini make a classic martini
1:50:34 classier I'll click on that and Sin City
1:50:36 sipper will go to work and it's been
1:50:39 programmed to teach me how to make an
1:50:43 original
1:50:45 drink and so this is the V Vegas gold
1:50:48 Martini so 2 ounces of high quality Jin
1:50:51 half ounce of Dy remou and a/4 oce of
1:50:54 gold schogger the cinnamon schnaps with
1:50:58 the real gold
1:51:00 flakes right so it's going to make it
1:51:02 fancy and then look what it does not
1:51:05 only does it invent a drink for
1:51:09 you it should oh it didn't do
1:51:13 it should automatically just make it
1:51:15 let's see
1:51:18 it janky is as janky does
1:51:25 it's designed to create an image of the
1:51:30 drink peeps want to see the response
1:51:32 numbering and chat titling again if you
1:51:36 get a chance
1:51:40 okay there you
1:51:44 go look at that
1:51:46 gorgeous
1:51:49 gorgeous so end of Android
1:51:54 oh no this is no chat GPT is not a
1:51:56 replacement for Android I'm just using
1:51:58 Android as a
1:52:03 metaphor are you referring to my dumb
1:52:07 smartphone the Grinch cocktail yeah we
1:52:10 can do that let's
1:52:12 um I want to create a grinch cocktail
1:52:25 I mean what's so cool about this is like
1:52:28 it starts with madori of course right
1:52:30 like if you're going to do a grinch one
1:52:31 it's got to be green so you start with
1:52:32 madori which actually be tasty madori
1:52:36 vodka lemon juice and a splash of soda
1:52:38 water
1:52:41 perfect it's going to be like electric
1:52:50 green now let's visualize this festive
1:52:53 concoction and there it's creating the
1:52:55 image oh make a full full eclipse drink
1:52:58 oh that's a great
1:53:01 idea yep I'll do
1:53:14 that
1:53:19 gorgeous I I I I mean I used to bartend
1:53:23 that is exactly what that drink would
1:53:24 look like that's exactly what it would
1:53:26 look like and if you had the Grinch
1:53:27 glass that would be even
1:53:29 better um the other thing that I've got
1:53:32 it doing and it's for whatever reason
1:53:34 it's not it's not rendering the HTML I
1:53:36 got to pull the HTML out of this is it's
1:53:38 doing a call to action it says this GPT
1:53:41 was created by Kyle Shannon you know a
1:53:44 member of the AI salon and a host of of
1:53:46 this Tik talk Channel and so you know it
1:53:48 does calls to action and things like
1:53:49 that um let's let's do a
1:53:53 today was the full solar eclipse period
1:53:57 helped me create a cocktail to
1:54:01 commemorate the full solar
1:54:08 eclipse I was going to tell it to use
1:54:10 coffee but let's see if it comes up with
1:54:12 coffee on its
1:54:18 own 2 ounces of black vodka wow cool
1:54:22 this unusual vodka will serve as the
1:54:25 darkness for the eclipse 1 oz of orange
1:54:27 lure interesting representing the corona
1:54:30 around the
1:54:31 Sun and then half an ounce of fresh lime
1:54:34 juice that'll be delicious and grenadine
1:54:36 to provide a gradient effect wow this
1:54:38 going to be cool this is going to
1:54:40 actually look cool all right here let me
1:54:42 let me go back if you if anyone wants to
1:54:44 make this
1:54:45 cocktail do a screenshot there there's
1:54:48 your there's your
1:54:51 ingredients and then here's your
1:54:54 directions take a screenshot there fill
1:54:57 the Shaker with the Vodka The Lure and
1:55:01 the lime juice
1:55:02 shake stir it into a strain it into a
1:55:06 cocktail glass slowly pour a dash of
1:55:09 grenadine into the drink it will sink to
1:55:10 the bottom creating a beautiful layered
1:55:13 effect and place the garnish on the rim
1:55:16 oh god look how gorgeous that
1:55:18 is that's so cool with with the round
1:55:22 orange and the and the Moon in front of
1:55:24 it come
1:55:26 on that's so
1:55:28 good the eclipse Shadow
1:55:31 cocktail come on
1:55:34 people ask it to make a non-alcoholic
1:55:36 version great
1:55:39 um I love the eclipse period now create
1:55:43 a mocktail
1:55:45 version using
1:55:49 coffee or the dark beverage of your
1:55:57 Choice could have been on a billboard I
1:55:59 know I drink that I know right look at
1:56:03 AB Abba's like oh my God I know I'm
1:56:05 telling you man Sin City sippers it's a
1:56:08 baller app this was my very first GPT
1:56:11 that I made I made it live on one of
1:56:14 these Tik
1:56:15 toks when I got access to to this
1:56:18 functionality I said oh let's go play
1:56:21 this was the very first one made I
1:56:23 really haven't modified
1:56:24 it which is evidenced by the fact that
1:56:27 like the the call to action breaks but
1:56:30 mocktail with crushed blueberries and
1:56:32 raspberries actually that sounds
1:56:34 delicious all right mocktail what did it
1:56:37 do caffeine Eclipse mocktail cold brew
1:56:41 coffee orange
1:56:43 juice o that doesn't sound
1:56:46 good coconut cream maybe fill the glass
1:56:50 with ice pour the cold brew over it
1:56:52 slowly out in the orange juice so it
1:56:54 mostly sits on
1:56:56 top I think that's going to
1:57:03 suck it's
1:57:13 beautiful anyway all right so so that's
1:57:16 what a GPT is
1:57:18 now how do you make
1:57:20 them you just go to the create
1:57:23 button uh you go you say explore
1:57:27 gpts and it'll take you to this page
1:57:29 where you can scroll through and play
1:57:32 with all these or search the search
1:57:35 sucks by the way the store really sucks
1:57:37 but um you click create and then see
1:57:42 where it says create GPT Builder the
1:57:44 left side is where you build it the
1:57:46 right side is where you can play with it
1:57:48 so you can just
1:57:51 say I want to
1:57:54 create a
1:57:56 GPT that provides
1:58:00 countdowns for celestial events in the
1:58:04 coming
1:58:07 year we we're essentially
1:58:09 done show the other one again please
1:58:12 which other one the other drink I'll
1:58:15 I'll go back to it
1:58:24 cold foam instead yeah something like
1:58:26 that Allan yeah that was so but that
1:58:29 that's what I would do like like if it
1:58:31 gave me if it gave me a drink recipe
1:58:33 like mix cold brew coffee and orange
1:58:36 juice which it did I would just go no
1:58:38 that's horrible I would I would have it
1:58:40 give me something new countdowns okay um
1:58:44 Stellar guide uh let's let's say
1:58:47 um
1:58:50 um so Le
1:58:54 deal
1:58:58 count
1:59:00 down so we're going to call it Celestial
1:59:07 countdown now it's generating a profile
1:59:09 picture for
1:59:12 us Brandon I made your I made your
1:59:15 Eclipse sunum song okay good thank you
1:59:18 Brandon
1:59:27 [Music]
1:59:33 and then to save it and oh by the way so
1:59:36 there's create mode where you can just
1:59:38 sort of speak it into existence then you
1:59:39 can go to configure mode and you know
1:59:42 here's the picture that it made here's
1:59:44 the name the description this is the
1:59:46 prompt that it uses and then these are
1:59:48 what are called the the um conversation
1:59:52 starter questions you can rewrite those
1:59:54 you can even upload files like if you
1:59:56 had a a book written about upcoming
1:59:59 celestial events you could upload that
2:00:02 or a PDF about it and it could primarily
2:00:05 primarily use that knowledge for it for
2:00:07 its knowledge so then once it's set up
2:00:10 you just go to create and then you can
2:00:12 share anyone that has the link or you
2:00:14 can even share um you can publish it to
2:00:17 the GPT store so I'll do that why not
2:00:22 Celestial
2:00:24 countdown and I can say view GPT and if
2:00:27 you go to uh if you've got access to
2:00:30 chat gp4 and go there right
2:00:32 now oh 2023 is the icon that's
2:00:37 funny you can just use it so tell me
2:00:40 about the next lunar
2:00:49 eclipse and you know I could go in and I
2:00:51 could say when someone asks you about
2:00:53 you know the next event you know give
2:00:55 them the following pieces of information
2:00:57 it will probably do that on its
2:01:00 own it's set to occur on March 24th
2:01:05 2024 um well anyway the date is passed
2:01:10 so so you you would have to go in and do
2:01:12 things like make sure that you determine
2:01:14 today's date before you talk about
2:01:16 things in the future that are actually
2:01:17 in the past but anyway that's how easy
2:01:20 it is to make a GPT like staggeringly
2:01:23 simple and then you they can get really
2:01:25 complicated but all right um one last
2:01:28 thing that I'm going to show you so
2:01:29 there was a request there were two
2:01:31 requests here one was go
2:01:34 back and look at
2:01:37 the this drink this recipe for this
2:01:41 one the all right Becky get ready here's
2:01:45 so
2:01:46 screenshot screenshot
2:01:49 that and by the way if you don't want to
2:01:51 type
2:01:53 if you don't want to retype all of this
2:01:56 screenshot this and you know trim the
2:01:59 photo so it's just that text put it into
2:02:01 chat GPT and tell it to type it for you
2:02:05 take advantage of its Vision take
2:02:06 advantage here's the
2:02:09 directions take a screenshot of
2:02:13 that and then there's your image take a
2:02:17 screenshot of that that's
2:02:19 gorgeous that really could be like if
2:02:22 you had a black you know a black vodka
2:02:26 brand that that could be a billboard I
2:02:29 mean that's that's gorgeous gorgeous
2:02:32 gorgeous
2:02:33 gorgeous I don't know what the explosion
2:02:36 is here around the Sun but that's that's
2:02:38 okay all these other planets dancing
2:02:41 around um okay the final thing that we
2:02:43 were going to look at is I'm going to go
2:02:47 back to our custom instructions
2:02:52 and so in the custom instructions I'm
2:02:55 going to turn these
2:02:56 on and I I just said about myself you're
2:03:00 polite but concise that this really
2:03:02 doesn't matter but down here what I put
2:03:05 in the custom instructions for every
2:03:08 response give it a response count
2:03:12 followed by a response title and in fact
2:03:14 what I'm going to
2:03:17 do no I'm going to leave it like that
2:03:20 because it understood it
2:03:22 you could put some structure in this
2:03:23 like you could put quotes around these
2:03:25 or brackets around these but then I
2:03:27 think it might label it in the in the
2:03:29 thing I kind of like the way it was
2:03:30 acting for every response give it a
2:03:32 response count followed by a response
2:03:35 title that encapsulates that's spelled
2:03:38 wrong but
2:03:40 whatever now it's not that encapsulates
2:03:43 the response so I'm going to hit
2:03:45 save and I'm going to do a new chat I'm
2:03:48 going to flip to 3.5 just cuz it's
2:03:51 faster and I'm going to say
2:04:01 um what is the average speed of an
2:04:04 unladen
2:04:06 swallow what is the average speed of an
2:04:09 unladen
2:04:10 swallow
2:04:13 um uh why is it
2:04:17 funny why is it funny
2:04:26 it's it didn't quite get it right oh it
2:04:27 didn't 3.5 it didn't 3.5 it didn't
2:04:30 understand it hang
2:04:32 on what is the average speed of an
2:04:34 unladen
2:04:36 swallow oops of an
2:04:39 un so what it's doing now so this is
2:04:43 response one and then it it gave a title
2:04:47 for the response and so now I can say
2:04:51 sorry that was in complete period what
2:04:53 is the average speed of an unladen
2:05:03 swallow it didn't it didn't give me a
2:05:05 two which is
2:05:12 weird did it label it the average speed
2:05:14 of an unladen
2:05:17 swallow so let's go back let's go fix
2:05:20 that so I'm going to go back back to the
2:05:22 custom
2:05:24 GPT and I'm going to say for every
2:05:26 response give it a
2:05:31 sequential response
2:05:35 count and then I'm going to do
2:05:38 comma followed by a response
2:05:41 title all right and then I'm going to do
2:05:44 new thing and I'm going to say
2:05:52 how many
2:05:53 1969 GTO judges were
2:06:05 manufactured response one production
2:06:07 numbers for the
2:06:09 1969 GTO judge
2:06:15 um that was weird my app just
2:06:19 quit can you still hear me tap the
2:06:21 screen if you can still hear
2:06:29 me I'm still good you eclipsed again
2:06:32 yeah it quit all right you're back okay
2:06:35 whatever all
2:06:37 right
2:06:40 um why are GTO judges the coolest muscle
2:06:44 car in all of the land
2:06:48 [Laughter]
2:06:53 so response two the cool factor of GTO
2:07:01 judges so so what that prompt is doing
2:07:05 is making every
2:07:08 single response in your in a in a chat
2:07:12 session have its own title so then so so
2:07:15 let me let me do two more
2:07:20 um give me five other muscle cars from
2:07:23 the era that
2:07:25 are nearly as
2:07:32 cool response three five iconic muscle
2:07:35 cars from the
2:07:38 era all right now so assume you're doing
2:07:42 this you're having a big long ass
2:07:43 conversation at some point you can then
2:07:46 say to chat
2:07:48 gbt make a table
2:07:53 make a
2:07:55 table with the response numbers and
2:08:00 titles of all your responses so
2:08:11 far so how cool would it
2:08:15 be after a long ass conversation with
2:08:18 chat
2:08:20 gbt to be be able to make a table that
2:08:24 you could you could you know print
2:08:28 out and then be able to quickly go
2:08:31 back right let's get to 50,000 likes now
2:08:35 you go Becky
2:08:38 R yeah it's a table of contents for your
2:08:42 chat and then quite
2:08:45 frankly since I just showed you how to
2:08:47 make gpts like you could probably make a
2:08:49 GPT that's it's just like a normal chat
2:08:52 GPT except it's got a bunch of these
2:08:54 features in it that every 10 every 10
2:08:58 responses it creates a new table or
2:09:01 something like
2:09:05 that Irregulars before you leave uh yeah
2:09:09 so if if you've been making eclipse
2:09:11 images and songs please share them in
2:09:13 the Irregulars channel on the AI Salon
2:09:15 the other thing that I want to show you
2:09:17 share with you there's two things
2:09:23 um I got to
2:09:25 go I said I I said I wouldn't be too
2:09:27 long tonight but then we did the chat
2:09:29 GPT Deep dive and that's just going to
2:09:31 take longer you know what I mean you
2:09:34 know what I mean all right so in
2:09:36 Shameless self
2:09:41 promotion
2:09:43 um
2:09:45 is
2:09:47 this uh
2:09:50 this post post about the creativity
2:09:55 squared Play Shop so on April 20th John
2:09:59 Leland and I are doing a workshop a Play
2:10:01 Shop we're calling it about enhancing
2:10:04 your creativity with with
2:10:06 AI and if you go to AI Salon
2:10:11 announcements there is a coupon code
2:10:15 Salon fun it was supposed to be Salon
2:10:17 play but I [ __ ] up but Salon fun now
2:10:20 works so does Salon play but if you if
2:10:23 you put in Salon fun as your coupon code
2:10:27 it'll save you a big pile of money and
2:10:30 then also in
2:10:35 um in Shameless
2:10:38 self-promotion um is an interview that
2:10:41 John Leland did with me
2:10:44 um like a week and a half ago but it's
2:10:47 just talking about creativity and Ai and
2:10:51 my relationship with creativity and how
2:10:54 AI enhances it and it's just a really
2:10:56 fun back and forth with the two of us
2:10:58 you know it's me pontificating you get
2:11:00 plenty of that here but if you want to
2:11:02 see a nicely edited version of that
2:11:04 that's what that is um so anyway go do
2:11:09 that all right
2:11:15 fantastic I only got to see a deep
2:11:18 partial eclipse yeah I know it was 65%
2:11:21 here I went out I made myself a little
2:11:23 pinhole projector which was supposed to
2:11:25 make it big so I could see it but it it
2:11:28 it didn't uh amplify it so I guess the
2:11:30 pinhole was too small or too big or
2:11:33 something anyway end the Night by
2:11:36 playing the song all right where's the
2:11:38 song in Irregulars is it in
2:11:42 Irregulars let's go to Irregulars and
2:11:44 see what we got going
2:11:46 here that's
2:11:48 nice Kyle looking looking at the sun
2:11:53 is this the is that just an image I
2:11:55 think that's just an image all right the
2:11:57 total Kyle Clips you guys are
2:12:02 [Music]
2:12:07 weird happening in the sky the Moon is
2:12:11 blocking the sun's radiant light but
2:12:13 it's not the moon it's a
2:12:16 [Laughter]
2:12:16 [Music]
2:12:19 guy he's got a small shines brighter
2:12:22 than the Sun and his hair is as dark as
2:12:26 the night CL stands there blocking all
2:12:30 the fun nobody expected the
2:12:34 Str all right all
2:12:42 right it's like this is like K-pop Kpop
2:12:46 Japanese
2:12:48 pup y'all are weird man thank you
2:12:58 all right I looked at the eclipse today
2:13:00 twice good for you that's what they say
2:13:02 look at the eclipse challenge it
2:13:05 challenge
2:13:06 science I'm sure it'll be
2:13:09 fine ah thanks for playing it um you
2:13:14 have to realize that you taught us the
2:13:16 weirdness keep
2:13:20 scrolling uh um wait wait wait chat GPT
2:13:24 can make music how Rudy Sugo oh Rudy all
2:13:28 right so for anyone else I would just
2:13:30 move on but Rudy okay if you don't know
2:13:32 sunno holy [ __ ] you got to know you
2:13:35 gotta show the eclipse IM Mig again the
2:13:42 salon which one this one
2:13:54 [Laughter]
2:13:55 if you're
2:13:56 screenshotting
2:13:59 um so
2:14:02 Rudy if you don't know about
2:14:04 sun. you need to know about
2:14:08 sun. so let's let's
2:14:14 do change to simple mode qu application
2:14:20 error app doo. a we're going to go
2:14:24 create and I'm just going to
2:14:27 go
2:14:39 um a pop
2:14:44 song about an AI master in South Africa
2:14:49 named Rudy
2:14:55 yeah we'll just do
2:14:56 that I was trying to come up with
2:14:58 something more clever than a pop song
2:15:00 but screw
2:15:02 it it's the end of the world and we know
2:15:05 it and I feel fine
2:15:09 [Applause]
2:15:24 [Music]
2:15:25 please send pseudo
2:15:27 credits via donation to me very
2:15:36 [Music]
2:15:43 nice spelled your name wrong but that's
2:15:48 okay behind the
2:15:53 every algorithm he creates brings the
2:15:56 city to a brand
2:15:58 [Music]
2:16:02 new so
2:16:05 electrifying Tak it
2:16:10 [Music]
2:16:13 [Applause]
2:16:15 us that's that Brandon thank you for
2:16:17 subscribing I appreciate that I
2:16:19 appreciate it a lot um
2:16:22 so anyway yeah that's sunno I mean it's
2:16:24 it's cuckoo it's it's [ __ ] Bonkers
2:16:26 here's the other
2:16:27 one oh no it's
2:16:33 [Music]
2:16:38 not with a mind of innovation he's the
2:16:41 one who steals the
2:16:45 show rud is the name ass G he's The
2:16:48 Mastermind behind the call
2:16:55 [Music]
2:16:58 so anyway yep crazy absolutely crazy and
2:17:02 we did I don't know if you saw this last
2:17:03 night Rudy we did we we made a uh my
2:17:08 great uncle KS
2:17:11 Dylan
2:17:15 um let's see where
2:17:25 [Music]
2:17:27 what the [ __ ] is going on here okay hang
2:17:35 [Music]
2:17:37 on that's KS Dylan in Greenwich Village
2:17:40 but we did one last night
2:17:42 in wait where is
2:17:49 it KS
2:17:56 Dylan
2:17:59 posts there it is so so this one is
2:18:04 uh he went to Brazil in
2:18:07 1972 And so this is him singing his song
2:18:11 if I'd never met her named by Silver Fox
2:18:15 um but the second half of the song is in
2:18:16 Portuguese so we did this all last night
2:18:20 in about an hour
2:18:22 the the
2:18:23 song the image and the
2:18:29 animation
2:18:31 Shadows we'll go out on this she was the
2:18:34 artist
2:18:40 not words chalice from which I drank
2:18:46 deep in her absence a harvest I could
2:18:50 not
2:18:51 Daniel said this song was stuck in her
2:18:53 head
2:18:54 [Music]
2:18:56 today that's
2:18:59 awesome gens of solitude withs
2:19:05 never
2:19:07 if matter my words wouldn't be
2:19:12 lost written
2:19:15 [Music]
2:19:16 in here's where he switches
2:19:23 [Music]
2:19:31 [Music]
2:19:38 Portuguese
2:19:40 yeah it kind of looks like me because I
2:19:44 took one of my images from Kyle Shannon
2:19:47 dreams which were sort of kind of these
2:19:49 Twisted portraits of me
2:19:52 and I used that as the character
2:19:54 reference for Mid Journey so we created
2:19:58 the the this was the 1972 Brazil
2:20:03 tour and we also did
2:20:09 [Music]
2:20:12 the the 1969 this is the 1969 Greenwich
2:20:16 Village
2:20:19 version sh
2:20:23 play she was the
2:20:27 artist so the process here was
2:20:31 um it
2:20:33 was image and mid Journey using
2:20:36 character
2:20:38 reference chat GPT to write the lyrics
2:20:41 of the song which we then rewrote like
2:20:43 we rewrote the ending so it didn't
2:20:45 suck then we took it into
2:20:48 sunno and made the song and it took us I
2:20:52 don't know 20 different versions till we
2:20:54 got one we liked and then took it into
2:20:57 hey Jen to do the animation and then
2:21:00 posted on LinkedIn to trash my
2:21:02 professional
2:21:11 reputation Uncle Dylan not related to
2:21:14 Bob
2:21:16 exactly all right let's go out with lots
2:21:18 lots of likes on the live okay we're
2:21:20 we're at we're at 40 46,000 we're at
2:21:24 46,000 likes so you got to get [ __ ]
2:21:26 tapping here people I won't I got to
2:21:29 leave but I can't leave here till that
2:21:31 hits 50 cuz OCD so that's on you that's
2:21:35 not on
2:21:37 me that that is as of this moment we're
2:21:40 at 47,000 right now as of this moment it
2:21:43 is your shortcoming not
2:21:45 mine I've been sitting here in a
2:21:48 chair talking words out of my mouth
2:21:52 hole all you had to do was this we at
2:21:57 48,6 6900 it's going good now we're
2:22:00 going to hit it in about 3 seconds
2:22:06 here luckily you have OCD Pals on the
2:22:09 other side too exactly all right we just
2:22:11 hit 50,000 you're fine you can stop
2:22:13 tapping
2:22:15 now Becky Becky R I've been tapping I
2:22:19 feel like an auctioneer we have you
2:22:21 captive exactly oh man you guys are
2:22:24 funny all right listen I'm out of here
2:22:26 um please follow the channel if you're
2:22:28 new here follow the channel um subscribe
2:22:31 to the to the lives that helps a lot I
2:22:33 appreciate that or buy the uh the video
2:22:36 series in the corner which talks about
2:22:38 why AI I think AI is a big deal um join
2:22:42 the AI salon and if you if you want to
2:22:45 do a
2:22:46 workshop if you're part of the salon you
2:22:49 get a big discount off that Workshop or
2:22:51 the play shop we're calling it I created
2:22:52 a GPT to tap likes nice I like it um
2:22:57 thank you to the subscribers thank you
2:22:59 are regulars just keep showing up here
2:23:00 also that that makes a big difference
2:23:02 just show up here every night and uh and
2:23:04 just stay in the conversation and then
2:23:06 let's do a little magic Romper Room
2:23:08 magic mirror I see Mr it and Silver Fox
2:23:13 I see Wolfman Clint and hacker maing
2:23:17 Becky Rue great to see you tonight Becky
2:23:19 60,000 before he leaves
2:23:21 Brandon always hustling up a little
2:23:23 engagement farming it's good I like it
2:23:25 it's
2:23:28 solid um Allan good to see you tonight
2:23:30 Vegas Moxy two block Tom great to see
2:23:33 you Emilio's wife Brian Whitney Source
2:23:35 Camp oh source Camp come to the AI salon
2:23:38 tomorrow night um Source Camp Kelly Camp
2:23:41 is going to be speaking at the salon so
2:23:44 really excited about that so please
2:23:46 please join us tomorrow night the
2:23:47 information for that is at at the AI
2:23:50 salon so go go
2:23:52 to there the salon. a click on the first
2:23:55 link that'll take you to the community I
2:23:58 see Kyle I know exactly Allan good
2:24:03 night um all right everybody I'm out of
2:24:06 here this fun tonight I hope that was
2:24:07 that was helpful a lot of good stuff
2:24:09 bye-bye